The DENX U-Boot and Linux Guide (DULG) for canyonlands _________________________________________________________________ Table of contents: * [1]1. Abstract * [2]2. Introduction + [3]2.1. Copyright + [4]2.2. Disclaimer + [5]2.3. Availability + [6]2.4. Credits + [7]2.5. Translations + [8]2.6. Feedback + [9]2.7. Conventions * [10]3. Embedded Linux Development Kit + [11]3.1. ELDK Availability + [12]3.2. ELDK Getting Help + [13]3.3. Supported Host Systems + [14]3.4. Supported Target Architectures + [15]3.5. Installation o [16]3.5.1. Product Packaging o [17]3.5.2. Downloading the ELDK o [18]3.5.3. Initial Installation o [19]3.5.4. Installation and Removal of Individual Packages o [20]3.5.5. Removal of the Entire Installation + [21]3.6. Working with ELDK o [22]3.6.1. Switching Between Multiple Installations + [23]3.7. Mounting Target Components via NFS + [24]3.8. Rebuilding ELDK Components o [25]3.8.1. ELDK Source Distribution o [26]3.8.2. Rebuilding Target Packages o [27]3.8.3. Rebuilding ELDT Packages + [28]3.9. ELDK Packages o [29]3.9.1. List of ELDT Packages o [30]3.9.2. List of Target Packages + [31]3.10. Rebuilding the ELDK from Scratch o [32]3.10.1. ELDK Build Process Overview o [33]3.10.2. Setting Up ELDK Build Environment o [34]3.10.3. build.sh Usage o [35]3.10.4. Format of the cpkgs.lst and tpkgs.lst Files + [36]3.11. Notes for Solaris 2.x Host Environment * [37]4. System Setup + [38]4.1. Serial Console Access + [39]4.2. Configuring the "cu" command + [40]4.3. Configuring the "kermit" command + [41]4.4. Using the "minicom" program + [42]4.5. Permission Denied Problems + [43]4.6. Configuration of a TFTP Server + [44]4.7. Configuration of a BOOTP / DHCP Server + [45]4.8. Configuring a NFS Server * [46]5. Das U-Boot + [47]5.1. Current Versions + [48]5.2. Unpacking the Source Code + [49]5.3. Configuration + [50]5.4. Installation o [51]5.4.1. Before You Begin # [52]5.4.1.1. Installation Requirements # [53]5.4.1.2. Board Identification Data o [54]5.4.2. Installation Using a BDM/JTAG Debugger o [55]5.4.3. Installation using U-Boot + [56]5.5. Tool Installation + [57]5.6. Initialization + [58]5.7. Initial Steps + [59]5.8. The First Power-On + [60]5.9. U-Boot Command Line Interface o [61]5.9.1. Information Commands # [62]5.9.1.1. bdinfo - print Board Info structure # [63]5.9.1.2. coninfo - print console devices and informations # [64]5.9.1.3. flinfo - print FLASH memory information # [65]5.9.1.4. iminfo - print header information for application image # [66]5.9.1.5. help - print online help o [67]5.9.2. Memory Commands # [68]5.9.2.1. base - print or set address offset # [69]5.9.2.2. crc32 - checksum calculation # [70]5.9.2.3. cmp - memory compare # [71]5.9.2.4. cp - memory copy # [72]5.9.2.5. md - memory display # [73]5.9.2.6. mm - memory modify (auto-incrementing) # [74]5.9.2.7. mtest - simple RAM test # [75]5.9.2.8. mw - memory write (fill) # [76]5.9.2.9. nm - memory modify (constant address) # [77]5.9.2.10. loop - infinite loop on address range o [78]5.9.3. Flash Memory Commands # [79]5.9.3.1. cp - memory copy # [80]5.9.3.2. flinfo - print FLASH memory information # [81]5.9.3.3. erase - erase FLASH memory # [82]5.9.3.4. protect - enable or disable FLASH write protection # [83]5.9.3.5. mtdparts - define a Linux compatible MTD partition scheme o [84]5.9.4. Execution Control Commands # [85]5.9.4.1. source - run script from memory # [86]5.9.4.2. bootm - boot application image from memory # [87]5.9.4.3. go - start application at address 'addr' o [88]5.9.5. Download Commands # [89]5.9.5.1. bootp - boot image via network using BOOTP/TFTP protocol # [90]5.9.5.2. dhcp - invoke DHCP client to obtain IP/boot params # [91]5.9.5.3. loadb - load binary file over serial line (kermit mode) # [92]5.9.5.4. loads - load S-Record file over serial line # [93]5.9.5.5. rarpboot- boot image via network using RARP/TFTP protocol # [94]5.9.5.6. tftpboot- boot image via network using TFTP protocol o [95]5.9.6. Environment Variables Commands # [96]5.9.6.1. printenv- print environment variables # [97]5.9.6.2. saveenv - save environment variables to persistent storage # [98]5.9.6.3. setenv - set environment variables # [99]5.9.6.4. run - run commands in an environment variable # [100]5.9.6.5. bootd - boot default, i.e., run 'bootcmd' o [101]5.9.7. Flattened Device Tree support # [102]5.9.7.1. fdt addr - select FDT to work on # [103]5.9.7.2. fdt list - print one level # [104]5.9.7.3. fdt print - recursive print # [105]5.9.7.4. fdt mknode - create new nodes # [106]5.9.7.5. fdt set - set node properties # [107]5.9.7.6. fdt rm - remove nodes or properties # [108]5.9.7.7. fdt move - move FDT blob to new address # [109]5.9.7.8. fdt chosen - fixup dynamic info o [110]5.9.8. Special Commands # [111]5.9.8.1. i2c - I2C sub-system o [112]5.9.9. Storage devices o [113]5.9.10. Miscellaneous Commands # [114]5.9.10.1. echo - echo args to console # [115]5.9.10.2. reset - Perform RESET of the CPU # [116]5.9.10.3. sleep - delay execution for some time # [117]5.9.10.4. version - print monitor version # [118]5.9.10.5. ? - alias for 'help' + [119]5.10. U-Boot Environment Variables + [120]5.11. U-Boot Scripting Capabilities + [121]5.12. U-Boot Standalone Applications o [122]5.12.1. "Hello World" Demo o [123]5.12.2. Timer Demo + [124]5.13. U-Boot Image Formats + [125]5.14. U-Boot Advanced Features o [126]5.14.1. Boot Count Limit * [127]6. Embedded Linux Configuration + [128]6.1. Download and Unpack the Linux Kernel Sources + [129]6.2. Kernel Configuration and Compilation + [130]6.3. Installation * [131]7. Booting Embedded Linux + [132]7.1. Introduction + [133]7.2. Flattened Device Tree Blob + [134]7.3. Passing Kernel Arguments + [135]7.4. Boot Arguments Unleashed + [136]7.5. Networked Operation with Root Filesystem over NFS o [137]7.5.1. Bootlog of tftp'd Linux kernel with Root Filesystem over NFS + [138]7.6. Boot from Flash Memory + [139]7.7. Standalone Operation with Ramdisk Image * [140]8. Building and Using Modules * [141]9. Advanced Topics + [142]9.1. Flash Filesystems o [143]9.1.1. Memory Technology Devices o [144]9.1.2. Journalling Flash File System o [145]9.1.3. Second Version of JFFS o [146]9.1.4. Compressed ROM Filesystem + [147]9.2. The TMPFS Virtual Memory Filesystem o [148]9.2.1. Mount Parameters o [149]9.2.2. Kernel Support for tmpfs o [150]9.2.3. Usage of tmpfs in Embedded Systems + [151]9.3. Adding Swap Space + [152]9.4. Splash Screen Support in Linux + [153]9.5. Root File System: Design and Building o [154]9.5.1. Root File System on a Ramdisk o [155]9.5.2. Root File System on a JFFS2 File System o [156]9.5.3. Root File System on a cramfs File System o [157]9.5.4. Root File System on a Read-Only ext2 File System o [158]9.5.5. Root File System on a Flash Card o [159]9.5.6. Root File System in a Read-Only File in a FAT File System + [160]9.6. Root File System Selection + [161]9.7. Overlay File Systems + [162]9.8. The Persistent RAM File system (PRAMFS) o [163]9.8.1. Mount Parameters o [164]9.8.2. Example * [165]10. Debugging + [166]10.1. Debugging of U-Boot o [167]10.1.1. Debugging of U-Boot Before Relocation o [168]10.1.2. Debugging of U-Boot After Relocation + [169]10.2. Linux Kernel Debugging o [170]10.2.1. Linux Kernel and Statically Linked Device Drivers o [171]10.2.2. Dynamically Loaded Device Drivers (Modules) o [172]10.2.3. GDB Macros to Simplify Module Loading + [173]10.3. GDB Startup File and Utility Scripts + [174]10.4. Tips and Tricks + [175]10.5. Application Debugging o [176]10.5.1. Local Debugging o [177]10.5.2. Remote Debugging + [178]10.6. Debugging with Graphical User Interfaces * [179]11. Simple Embedded Linux Framework * [180]12. Books, Mailing Lists, Links, etc. + [181]12.1. Application Notes + [182]12.2. Further Reading o [183]12.2.1. License Issues o [184]12.2.2. Linux kernel o [185]12.2.3. General Linux / Unix programming o [186]12.2.4. Network Programming o [187]12.2.5. C++ programming o [188]12.2.6. Java programming o [189]12.2.7. Power Architecture® Programming o [190]12.2.8. Embedded Topics + [191]12.3. Mailing Lists + [192]12.4. Links + [193]12.5. Tools * [194]13. Appendix + [195]13.1. Flat Device Tree + [196]13.2. Flat Device Tree + [197]13.3. BDI2000 Configuration file * [198]14. FAQ - Frequently Asked Questions + [199]14.1. ELDK o [200]14.1.1. ELDK Installation under FreeBSD o [201]14.1.2. ELDK Installation Hangs o [202]14.1.3. .gvfs: Permission Denied o [203]14.1.4. Installation on Local Harddisk o [204]14.1.5. System Include Files Missing o [205]14.1.6. patch: command not found o [206]14.1.7. ELDK Include Files Missing o [207]14.1.8. Using the ELDK on a 64 bit platform o [208]14.1.9. GDB Problems with BDI2000/BDI3000 on e500 Cores o [209]14.1.10. How can I check if Floating Point support is working? o [210]14.1.11. ELDK 2.x Installation Aborts o [211]14.1.12. Enable SSH Access + [212]14.2. U-Boot o [213]14.2.1. Can U-Boot be configured such that it can be started in RAM? o [214]14.2.2. Relocation cannot be done when using -mrelocatable o [215]14.2.3. Source object has EABI version 4, but target has EABI version 0 o [216]14.2.4. U-Boot crashes after relocation to RAM o [217]14.2.5. Warning - bad CRC, using default environment o [218]14.2.6. Net: No ethernet found o [219]14.2.7. Wrong debug symbols after relocation o [220]14.2.8. Decoding U-Boot Crash Dumps o [221]14.2.9. Porting Problem: cannot move location counter backwards o [222]14.2.10. U-Boot Doesn't Run after Upgrading my Compiler o [223]14.2.11. How Can I Reduce The Image Size? o [224]14.2.12. Erasing Flash Fails o [225]14.2.13. Ethernet Does Not Work o [226]14.2.14. Where Can I Get a Valid MAC Address from? o [227]14.2.15. Why do I get TFTP timeouts? o [228]14.2.16. Why is my Ethernet operation not reliable? o [229]14.2.17. How the Command Line Parsing Works # [230]14.2.17.1. Old, simple command line parser # [231]14.2.17.2. Hush shell # [232]14.2.17.3. Hush shell scripts # [233]14.2.17.4. General rules o [234]14.2.18. How can I load and uncompress a compressed image o [235]14.2.19. How can I create an uImage from a ELF file o [236]14.2.20. My standalone program does not work o [237]14.2.21. Linux hangs after uncompressing the kernel o [238]14.2.22. How can I implement automatic software updates? + [239]14.3. Linux o [240]14.3.1. Linux crashes randomly o [241]14.3.2. Linux crashes when uncompressing the kernel o [242]14.3.3. Linux Post Mortem Analysis o [243]14.3.4. Linux kernel register usage o [244]14.3.5. Linux Kernel Ignores my bootargs o [245]14.3.6. Cannot configure Root Filesystem over NFS o [246]14.3.7. Linux Kernel Panics because "init" process dies o [247]14.3.8. Unable to open an initial console o [248]14.3.9. System hangs when entering User Space (ARM) o [249]14.3.10. Mounting a Filesystem over NFS hangs forever o [250]14.3.11. Ethernet does not work in Linux o [251]14.3.12. Loopback interface does not work o [252]14.3.13. Linux kernel messages are not printed on the console o [253]14.3.14. Linux ignores input when using the framebuffer driver o [254]14.3.15. How to switch off the screen saver and the blinking cursor? o [255]14.3.16. BogoMIPS Value too low o [256]14.3.17. Linux Kernel crashes when using a ramdisk image o [257]14.3.18. Ramdisk Greater than 4 MB Causes Problems o [258]14.3.19. Combining a Kernel and a Ramdisk into a Multi-File Image o [259]14.3.20. Adding Files to Ramdisk is Non Persistent o [260]14.3.21. Kernel Configuration for PCMCIA o [261]14.3.22. Configure Linux for PCMCIA Cards using the Card Services package o [262]14.3.23. Configure Linux for PCMCIA Cards without the Card Services package # [263]14.3.23.1. Using a MacOS Partition Table # [264]14.3.23.2. Using a MS-DOS Partition Table o [265]14.3.24. Boot-Time Configuration of MTD Partitions o [266]14.3.25. Use NTP to synchronize system time against RTC o [267]14.3.26. Configure Linux for XIP (Execution In Place) # [268]14.3.26.1. XIP Kernel # [269]14.3.26.2. Cramfs Filesystem # [270]14.3.26.3. Hints and Notes # [271]14.3.26.4. Space requirements and RAM saving, an example o [272]14.3.27. Use SCC UART with Hardware Handshake o [273]14.3.28. How can I access U-Boot environment variables in Linux? o [274]14.3.29. The =appWeb= server hangs *OR* /dev/random hangs o [275]14.3.30. Swapping over NFS o [276]14.3.31. Using NFSv3 for NFS Root Filesystem o [277]14.3.32. Using and Configuring the SocketCAN Driver o [278]14.3.33. Telnet / SSH (dropbear) server not working + [279]14.4. Self o [280]14.4.1. How to Add Files to a SELF Ramdisk o [281]14.4.2. How to Increase the Size of the Ramdisk + [282]14.5. RTAI o [283]14.5.1. Conflicts with asm clobber list + [284]14.6. BDI2000 o [285]14.6.1. Where can I find BDI2000 Configuration Files? o [286]14.6.2. How to Debug Linux Exceptions o [287]14.6.3. How to single step through "RFI" instruction o [288]14.6.4. Setting a breakpoint doesn't work o [289]14.6.5. Remote 'g' packet reply is too long + [290]14.7. Motorola LITE5200 Board o [291]14.7.1. LITE5200 Installation Howto o [292]14.7.2. USB does not work on Lite5200 board * [293]15. Glossary 1. Abstract This is the DENX U-Boot and Linux Guide to Embedded [294]PowerPC, ARM and MIPS Systems. The document describes how to configure, build and use the firmware Das U-Boot (typically abbreviated as just "U-Boot") and the operating system Linux for Embedded [295]PowerPC, ARM and MIPS Systems. The focus of this version of the document is on canyonlands boards. This document was generated at 24 Jan 2012 - 20:15. * [296]2. Introduction + [297]2.1. Copyright + [298]2.2. Disclaimer + [299]2.3. Availability + [300]2.4. Credits + [301]2.5. Translations + [302]2.6. Feedback + [303]2.7. Conventions 2. Introduction This document describes how to use the firmware U-Boot and the operating system Linux in Embedded [304]Power Architecture®, ARM and MIPS Systems. There are many steps along the way, and it is nearly impossible to cover them all in depth, but we will try to provide all necessary information to get an embedded system running from scratch. This includes all the tools you will probably need to configure, build and run U-Boot and Linux. First, we describe how to install the Cross Development Tools [305]Embedded Linux Development Kit which you probably need - at least when you use a standard x86 PC running Linux or a Sun Solaris 2.6 system as build environment. Then we describe what needs to be done to connect to the serial console port of your target: you will have to configure a terminal emulation program like cu or kermit. In most cases you will want to load images into your target using ethernet; for this purpose you need [306]TFTP and [307]DHCP / [308]BOOTP servers. A short description of their configuration is given. A description follows of what needs to be done to configure and build the U-Boot for a specific board, and how to install it and get it working on that board. The configuration, building and installing of Linux in an embedded configuration is the next step. We use SELF, our Simple Embedded Linux Framework, to demonstrate how to set up both a development system (with the root filesystem mounted over NFS) and an embedded target configuration (running from a ramdisk image based on busybox). This document does not describe what needs to be done to port U-Boot or Linux to a new hardware platform. Instead, it is silently assumed that your board is already supported by U-Boot and Linux. The focus of this document is on canyonlands boards. 2.1. Copyright Copyright (C) 2001 - 2011 by Wolfgang Denk, DENX Software Engineering. Copyright (C) 2003 - 2011 by Detlev Zundel, DENX Software Engineering. Copyright (C) 2003 - 2011 by contributing authors You have the freedom to distribute copies of this document in any format or to create a derivative work of it and distribute it provided that you: * Distribute this document or the derivative work at no charge at all. It is not permitted to sell this document or the derivative work or to include it into any package or distribution that is not freely available to everybody. * Send your derivative work (in the most suitable format such as sgml) to the author. * License the derivative work with this same license or use [309]GPL. Include a copyright notice and at least a pointer to the license used. * Give due credit to previous authors and major contributors. It is requested that corrections and/or comments be forwarded to the author. If you are considering to create a derived work other than a translation, it is requested that you discuss your plans with the author. 2.2. Disclaimer Use the information in this document at your own risk. DENX disavows any potential liability for the contents of this document. Use of the concepts, examples, and/or other content of this document is entirely at your own risk. All copyrights are owned by their owners, unless specifically noted otherwise. Use of a term in this document should not be regarded as affecting the validity of any trademark or service mark. Naming of particular products or brands should not be seen as endorsements. 2.3. Availability The latest version of this document is available in a number of formats: * HTML [310]http://www.denx.de/wiki/publish/DULG/DULG-canyonlands.html * plain ASCII text [311]http://www.denx.de/wiki/publish/DULG/DULG-canyonlands.txt * PostScript European A4 format [312]http://www.denx.de/wiki/publish/DULG/DULG-canyonlands.ps * PDF European A4 format [313]http://www.denx.de/wiki/publish/DULG/DULG-canyonlands.pdf 2.4. Credits A lot of the information contained in this document was collected from several mailing lists. Thanks to anybody who contributed in one form or another. 2.5. Translations None yet. 2.6. Feedback Any comments or suggestions can be mailed to the author: Wolfgang Denk at [314]wd@denx.de. 2.7. Conventions [315]Descriptions [316]Appearance Warnings ALERT! Hint TIP Notes Note. Information requiring special attention Warning File Names file.extension Directory Names directory Commands to be typed a command Applications Names another application Prompt of users command under bash shell bash$ Prompt of root users command under bash shell bash# Prompt of users command under tcsh shell tcsh$ Environment Variables VARIABLE Emphasized word word Code Example ls -l * [317]3. Embedded Linux Development Kit + [318]3.1. ELDK Availability + [319]3.2. ELDK Getting Help + [320]3.3. Supported Host Systems + [321]3.4. Supported Target Architectures + [322]3.5. Installation o [323]3.5.1. Product Packaging o [324]3.5.2. Downloading the ELDK o [325]3.5.3. Initial Installation o [326]3.5.4. Installation and Removal of Individual Packages o [327]3.5.5. Removal of the Entire Installation + [328]3.6. Working with ELDK o [329]3.6.1. Switching Between Multiple Installations + [330]3.7. Mounting Target Components via NFS + [331]3.8. Rebuilding ELDK Components o [332]3.8.1. ELDK Source Distribution o [333]3.8.2. Rebuilding Target Packages o [334]3.8.3. Rebuilding ELDT Packages + [335]3.9. ELDK Packages o [336]3.9.1. List of ELDT Packages o [337]3.9.2. List of Target Packages + [338]3.10. Rebuilding the ELDK from Scratch o [339]3.10.1. ELDK Build Process Overview o [340]3.10.2. Setting Up ELDK Build Environment o [341]3.10.3. build.sh Usage o [342]3.10.4. Format of the cpkgs.lst and tpkgs.lst Files + [343]3.11. Notes for Solaris 2.x Host Environment 3. Embedded Linux Development Kit The Embedded Linux Development Kit (ELDK) includes the GNU cross development tools, such as the compilers, binutils, gdb, etc., and a number of pre-built target tools and libraries necessary to provide some functionality on the target system. It is provided for free with full source code, including all patches, extensions, programs and scripts used to build the tools. Some versions of [344]ELDK (4.1) are available in two versions, which use Glibc resp. uClibc as the main C library for the target packages. Packaging and installation is based on the RPM package manager. 3.1. [345]ELDK Availability The [346]ELDK is available * on DVD-ROM from [347]DENX Computer Systems * for download on the following server: [348]FTP [349]HTTP [350]ftp://ftp.denx.de/pub/eldk/ [351]http://www.denx.de/ftp/pub/eldk/ * for download on the following mirrors: [352]FTP [353]HTTP [354]ftp://ftp-stud.hs-esslingen.de/pub/Mirrors/eldk/ [355]http://ftp-stud.hs-esslingen.de/pub/Mirrors/eldk/ [356]ftp://mirror.switch.ch/mirror/eldk/ [357]http://mirror.switch.ch/ftp/mirror/eldk/ not available [358]http://mira.sunsite.utk.edu/eldk/ [359]ftp://ftp.sunet.se/pub/Linux/distributions/eldk/ [360]http://ftp.sunet.se/pub/Linux/distributions/eldk/ 3.2. [361]ELDK Getting Help Community support for the [362]ELDK is available through the [363]ELDK Mailing List. Previous postings to this mailing list are available from the [364]ELDK archives. Commercial support is also available; please feel free to contact [365]DENX Software Engineering GmbH. 3.3. Supported Host Systems The [366]ELDK can be installed onto and operate with the following operating systems: * [367]Fedora Core 4, 5, 6 [368]Fedora 7, 8, 9, 10, 11, 12 * [369]Red Hat Linux 7.3, 8.0, 9 * [370]SuSE Linux 8.x, 9.0, 9.1, 9.2, 9.3, 10.0 * [371]OpenSUSE 10.2, 10.3 (32 Bit); OpenSUSE 11.0 (32 and 64 Bit) * [372]Debian 3.0 (Woody), 3.1 (Sarge) and 4.0 (Etch) * [373]Ubuntu 4.10, 5.04, 6.10, 8.04, 9.04, 9.10, 10.04 * [374]FreeBSD 5.0 Users also reported successful installation and use of the [375]ELDK on the following host systems: * [376]Suse Linux 7.2, 7.3 * [377]Mandrake 8.2 * [378]Slackware 8.1beta2 * [379]Gentoo Linux 2006.1 NOTE: Note: It may be necessary, and is usually recommended, to install the latest available software updates on your host system. For example, on [380]Fedora systems, you can use yum or apt-get to keep your systems current. 3.4. Supported Target Architectures The [381]ELDK includes target components and supports code generation for the following [382]Power Architecture® types of processors: * ppc_4xx = AMCC 4xx processors without FPU * ppc_4xxFP = AMCC 4xx processors with FPU (440EP, 440EPx) * ppc_6xx = PowerPC processors based on 60x cores with FPU (This includes support for MPC5xxx, 7xx, 82xx, 83xx and P40xx processors). * ppc_74xx = 74xx processors (This includes support for MPC86xx processors). * ppc_8xx = MPC8xx processors without FPU * ppc_85xx = MPC85xx processors (without FPU, but with a SPE) * ppc_85xxDP = MPC8544 and similar processors using a e500v2 core (without FPU, but with a V.2 SPE); this includes the P10xx and P20xx QorIQ processors, too. * ppc64 = 64 bit Power Architecture processors with FPU (for example, PA6T) There is also an [383]ELDK for ARM and MIPS systems. 3.5. Installation 3.5.1. Product Packaging Stable versions of the [384]ELDK are distributed in the form of an ISO image, which can be either burned onto a DVD or mounted directly, using the loopback Linux device driver (Linux host only). For the Power Architecture® target, the [385]ELDK distribution was split into three independent ISO images: one targeting the 4xx family of processors (AMCC), one targeting the ppc64 family of processors and another one for the 8xx, 6xx, 74xx and 85xx families (Freescale). This makes the ISO images fit on standard DVDROM media. If you are not bound by the DVDROM size limitiation there is still a single image containing all 32-bit targets (AMCC and Freescale). Development versions of the [386]ELDK are available as directory trees so it is easy to update individual packages; instructions for download of these trees and creation of ISO images from it is described in section [387]3.5.2. Downloading the ELDK. The [388]ELDK contains an installation utility and a number of RPM packages, which are installed onto the hard disk of the cross development host by the installation procedure. The RPM packages can be logically divided into two parts: * Embedded Linux Development Tools (ELDT) * Target components The first part contains the cross development tools that are executed on the host system. Most notably, these are the GNU cross compiler, binutils, and gdb. For a full list of the provided ELDT packages, refer to section [389]3.9.1. List of ELDT Packages below. The target components are pre-built tools and libraries which are executed on the target system. The [390]ELDK includes necessary target components to provide a minimal working NFS-based environment for the target system. For a list of the target packages included in the [391]ELDK, refer to section [392]3.9.2. List of Target Packages below. The [393]ELDK contains several independent sets of the target packages, one for each supported target [394]architecture [395]CPU family. Each set has been built using compiler code generation and optimization options specific to the respective target [396]CPU family. 3.5.2. Downloading the [397]ELDK You can either download the ready-to-burn ISO-images from one of the mirror sites (see [398]3.1. ELDK Availability), or you can download the individual files of the [399]ELDK from the development directory tree and either use these directly for installation or create an ISO image that can be burned on DVD-ROM. Change to a directory with sufficient free disk space; for the Power Architecture® version of the [400]ELDK you need about 1.6 GiB, or twice as much (3.2 GiB) if you also want to create an ISO image in this directory. To download the ISO image from the ppc-linux-x86/iso directory of one of the mirror sites you can use standard tools like wget or ncftpget, for example: bash$ wget ftp://ftp.sunet.se/pub/Linux/distributions/eldk/4.2/ppc-linux-x86/is o/ppc-2008-04-01.iso If you want to download the whole [401]ELDK directory tree instead you can - for example - use the ncftp [402]FTP client: bash$ ncftp ftp.sunet.se ... ncftp / > cd /pub/Linux/distributions/eldk/4.2 ncftp /pub/Linux/distributions/eldk/4.2 > bin ncftp /pub/Linux/distributions/eldk/4.2 > get -R ppc-linux-x86/distribution ... ncftp /pub/Linux/distributions/eldk/4.2 > bye TIP If you don't find the ncftp tool on your system you can download the NcFTP client from [403]http://www.ncftp.com/download/ There are a few executable files (binaries and scripts) in the [404]ELDK tree. Make sure they have the execute permissions set in your local copy: bash$ for file in \ > tools/bin/rpm \ > tools/usr/lib/rpm/rpmd \ > install \ > ELDK_MAKEDEV \ > ELDK_FIXOWNER > do > chmod +x ppc-linux-x86/distribution/$file > done Now create an ISO image from the directory tree: bash$ mkisofs \ > -A "ELDK-4.2 -- Target: PowerPC -- Host: x86 Linux" \ > -publisher "(C) `date "+%Y"` DENX Software Engineering, www.denx.de" \ > -p "`id -nu`@`hostname` -- `date`" \ > -V ppc-linux-x86 \ > -l -J -R -o eldk-ppc-linux-x86.iso ppc-linux-x86/distribution This will create an ISO image eldk-ppc-linux-x86.iso in your local directory that can be burned on DVD or mounted using the loopback device and used for installation as described above. Of course you can use the local copy of the directory tree directly for the installation, too. Please refer to section [405]3.10.2. Setting Up ELDK Build Environment for instructions on obtaining the build environment needed to re-build the [406]ELDK from scratch. 3.5.3. Initial Installation The initial installation is performed using the install utility located in the root of the [407]ELDK ISO image directory tree. The install utility has the following syntax: $ ./install [-d ] [] [] ... -d Specifies the root directory of the [408]ELDK being installed. If omitted, the [409]ELDK goes into the current directory. Specifies the target [410]CPU family the user desires to install. If one or more parameters are specified, only the target components specific to the respective [411]CPU families are installed onto the host. If omitted, the target components for all supported target [412]architecture [413]CPU families are installed. Note: Make sure that the "exec" option to the mount command is in effect when mounting the [414]ELDK ISO image. Otherwise the install program cannot be executed. On some distributions, it may be necessary to modify the /etc/fstab file, adding the "exec" mount option to the cdrom entry - it may also be the case that other existing mount options, such as "user" prevent a particular configuration from mounting the [415]ELDK DVD with appropriate "exec" permission. In such cases, consult your distribution documentation or mount the DVD explicitly using a command such as "sudo mount -o exec /dev/cdrom /mnt/cdrom" (sudo allows regular users to run certain privileged commands but may not be configured - run the previous command as root without "sudo" in the case that "sudo" has not been setup for use on your particular GNU/Linux system). You can install the [416]ELDK to any empty directory you wish, the only requirement being that you have to have write and execute permissions on the directory. The installation process does not require superuser privileges. Depending on the parameters the install utility is invoked with, it installs one or more sets of target components. The ELDT packages are installed in any case. Refer to section [417]3.6. Working with ELDK for a sample usage of the [418]ELDK. ALERT! Note: If you intend to use the installation as a root filesystem exported over NFS, then you now have to finish the configuration of the [419]ELDK following the instructions in [420]3.7. Mounting Target Components via NFS. ALERT! Note: Installation of the Glibc- and uClibc-based [421]ELDK versions into one directory is not yet supported. ALERT! Note: Installation of the 32-bit and 64-bit [422]ELDK versions into one directory is not yet supported. 3.5.4. Installation and Removal of Individual Packages The [423]ELDK has an RPM-based structure. This means that on the ISO image, individual components of the [424]ELDK are in the form of RPM packages, and after installation, the [425]ELDK maintains its own database which contains information about installed packages. The RPM database is kept local to the specific [426]ELDK installation, which allows you to have multiple independent [427]ELDK installations on your host system. (That is, you can install several instances of [428]ELDK under different directories and work with them independently). Also, this provides for easy installation and management of individual [429]ELDK packages. To list the installed [430]ELDK RPM packages, use the following command: bash$ ${CROSS_COMPILE}rpm -qa To remove an [431]ELDK package, use the following command: bash$ ${CROSS_COMPILE}rpm -e To install a package, use the following command: bash$ ${CROSS_COMPILE}rpm -i To update a package, use the following command: bash$ ${CROSS_COMPILE}rpm -U For the above commands to work correctly, it is crucial that the correct rpm binary gets invoked. In case of multiple [432]ELDK installations and RedHat-based host system, there may well be several rpm tools installed on the host system. You must make sure, either by using an explicit path or by having set an appropriate PATH environment variable, that when you invoke rpm to install/remove components of a [433]ELDK installation, it is the [434]ELDK's rpm utility that gets actually invoked. The rpm utility is located in the bin subdirectory relative to the [435]ELDK root installation directory. To avoid confusion with the host OS (RedHat) rpm utility, the [436]ELDK creates symlinks to its rpm binary with the names such that it could be invoked using the ${CROSS_COMPILE}rpm notation, for all supported [437]$CROSS_COMPILE values. TIP The standard (host OS) rpm utility allows various macros and configuration parameters to specified in user-specific ~/.rpmrc and ~/.rpmmacros files. The [438]ELDK rpm tool also has this capability, but the names of the user-specific configuration files are ~/.eldk_rpmrc and ~/.eldk_rpmmacros, respectively. 3.5.5. Removal of the Entire Installation To remove the entire [439]ELDK installation, use the following command while in the [440]ELDK root directory: bash$ rm -rf where specifies the root directory of the [441]ELDK to be removed. 3.6. Working with [442]ELDK After the initial installation is complete, all you have to do to start working with the [443]ELDK is to set and export the CROSS_COMPILE environment variable. Optionally, you may wish to add the bin and usr/bin directories of your [444]ELDK installation to the value of your PATH environment variable. For instance, a sample [445]ELDK installation and usage scenario looks as follows: * Create a new directory where the [446]ELDK is to be installed, say: bash$ mkdir /opt/eldk * Mount a CD or an ISO image with the distribution: bash$ mount /dev/cdrom /mnt/cdrom * Run the installation utility included on the distribution to install into that specified directory: bash$ /mnt/cdrom/install -d /opt/eldk * After the installation utility completes, export the CROSS_COMPILE variable: bash$ export CROSS_COMPILE=ppc_4xxFP- * TIP The trailing '-' character in the CROSS_COMPILE variable value is optional and has no effect on the cross tools behavior. However, it is required when building Linux kernel and U-Boot images. * Add the directories /opt/eldk/usr/bin and /opt/eldk/bin to PATH: bash$ PATH=$PATH:/opt/eldk/usr/bin:/opt/eldk/bin * Compile a file: bash$ ${CROSS_COMPILE}gcc -o hello_world hello_world.c TIP You can also call the cross tools using the generic prefix ppc-linux- for example: bash$ ppc-linux-gcc -o hello_world hello_world.c * or, equivalently: bash$ /opt/eldk/usr/ppc-linux/bin/gcc -o hello_world hello_world.c The value of the CROSS_COMPILE variable must correspond to the target [447]CPU family you want the cross tools to work for. Refer to the table below for the supported CROSS_COMPILE variable values: 3.6.A Table of possible values for $CROSS_COMPILE CROSS_COMPILE Value Predefined Compiler Flag FPU present or not ppc_4xx- -mcpu=403 No ppc_4xxFP- -mcpu=405fp Yes ppc_6xx- -mcpu=603 Yes ppc_74xx- -mcpu=7400 Yes ppc_8xx- -mcpu=860 No ppc_85xx- -mcpu=8540 Yes ppc_85xxDP- -mcpu=8540 Yes ppc64-linux- -mcpu=powerpc64 Yes TIP For compatibility with older versions of the [448]ELDK and with other toolkits the following values for $CROSS_COMPILE can be used, too: ppc_7xx- and ppc_82xx-. These are synonyms for ppc_6xx. 3.6.1. Switching Between Multiple Installations No special actions are required from the user to switch between multiple [449]ELDK installations on the same host system. Which [450]ELDK installation is used is determined entirely by the filesystem location of the binary that is being invoked. This approach can be illustrated using the following example. Assume the directory /work/denx_tools/usr/bin, where the ppc-linux-gcc compiler binary has been installed, is a part of the PATH environment variable. The user types the command as follows: $ ppc_8xx-gcc -c myfile.c To load the correct include files, find the correct libraries, spec files, etc., the compiler needs to know the [451]ELDK root directory. The compiler determines this information by analyzing the shell command it was invoked with ( ppc_8xx-gcc - without specifying the explicit path in this example) and, if needed, the value of the PATH environment variable. Thus, the compiler knows that it has been executed from the /work/denx_tools/usr/bin directory. Then, it knows that the compiler is installed in the usr/bin subdirectory of the root installation directory, so the [452]ELDK, the compiler is a part of, has been installed in the subdirectories of the /work/denx_tools directory. This means that the target include files are in /work/denx_tools//usr/include, and so on. 3.7. Mounting Target Components via NFS The target components of the [453]ELDK can be mounted via NFS as the root file system for your target machine. For instance, for an 8xx-based target, and assuming the [454]ELDK has been installed into the /opt/eldk directory, you can use the following directory as the NFS-based root file system: /opt/eldk/ppc_8xx ALERT! Before the NFS-mounted root file system can work, you must create necessary device nodes in the //dev directory. This process requires superuser privileges and thus cannot be done by the installation procedure (which typically runs as non-root). To facilitate creation of the device nodes, the [455]ELDK provides a script named ELDK_MAKEDEV, which is located in the root of the [456]ELDK distribution ISO image. The script acccepts the following optional arguments: -d Specifies the root directory of the [457]ELDK being installed. If omitted, then the current directory is assumed. -a Specifies the target [458]CPU family directory. If omitted, all installed target [459]architecture directories will be populated with the device nodes. -h Prints usage. # /mnt/cdrom/ELDK_MAKEDEV -d /opt/eldk NOTE: Compared to older versions of the [460]ELDK, options and behaviour of this command have been changed significantly. Please read the documentation. ALERT! Some of the target utilities included in the [461]ELDK, such as mount and su, have the SUID bit set. This means that when run, they will have privileges of the file owner of these utilities. That is, normally, they will have the privileges of the user who installed the [462]ELDK on the host system. However, for these utilities to work properly, they must have superuser privileges. This means that if the [463]ELDK was not installed by the superuser, the file owner of the target [464]ELDK utilities that have the SUID bit set must be changed to root before a target component may be mounted as the root file system. The [465]ELDK distribution image contains an ELDK_FIXOWNER script, which you can use to change file owners of all the appropriate files of the [466]ELDK installation to root. The script accepts the same arguments as the ELDK_MAKEDEV script above. Please note that you must have superuser privileges to run this script. For instance, if you have installed the [467]ELDK in the /opt/eldk directory, you can use the following commands: # cd /opt/eldk # /mnt/cdrom/ELDK_FIXOWNER Please note, that in the case that the installation directory, where the new [468]ELDK distribution is being installed, is already populated with other [469]ELDK distributions, the execution of the ELDK_FIXOWNER script without arguments will make the script work with all installed [470]ELDK target [471]architecture directories. This could take some time. To save the time, please use the -a argument to specify the appropriate target [472]architecture. For instance: # cd /opt/eldk # /mnt/cdrom/ELDK_FIXOWNER -a ppc_8xx 3.8. Rebuilding [473]ELDK Components 3.8.1. [474]ELDK Source Distribution The [475]ELDK is distributed with the full sources of all the components, so you may rebuild any [476]ELDK package. The sources are provided in the form of SRPM packages, distributed as a separate ISO image. To rebuild a target or ELDT package, you must first install the appropriate source RPM package from the ISO image into the [477]ELDK environment. This can be done using the following command: $ ${CROSS_COMPILE}rpm -i /mnt/cdrom/SRPMS/.src.rpm After an [478]ELDK source RPM is installed using the above command, its spec file and sources can be found in the subdirectories of the /usr/src/denx subdirectory. The sections that follow provide detailed instructions on rebuilding ELDT and target components of the [479]ELDK. 3.8.2. Rebuilding Target Packages All the target packages can be rebuilt from the provided source RPM packages. At first you have to install the Source RPM itself: bash$ ${CROSS_COMPILE}rpm -iv .src.rpm Then you can rebuild the binary target RPM using the following command from the [480]ELDK environment: bash$ ${CROSS_COMPILE}rpmbuild -ba .spec In order for the rebuilding process to work correctly, the following conditions must be true: * The [481]$CROSS_COMPILE environment variable must be set as appropriate for the target [482]CPU family. * The /usr/ppc-linux/bin directory must be in PATH before the /usr/bin directory. This is to make sure that the command gcc results in the fact that the [483]ELDK cross compiler is invoked, rather than the host gcc. The newly built package can then be installed just as easily: bash$ ${CROSS_COMPILE}rpm -i .rpm 3.8.3. Rebuilding ELDT Packages All the ELDT packages allow for rebuilding from the provided source RPM packages using the following command from the [484]ELDK environment: $ unset CROSS_COMPILE $ /usr/bin/rpmbuild -ba In order for the rebuilding process to work correctly, make sure all of the following is true: * The [485]$CROSS_COMPILE environment variable must NOT be set. * Do NOT use the [486]$CROSS_COMPILE command prefix. * The /usr/ppc-linux/bin directory must NOT be in PATH. This is to make sure that the command gcc causes invokation of the host gcc, rather than the [487]ELDK cross compiler. Note that the newly built package should be installed with the "global" rpm, not with the arch specific one: bash$ /bin/rpm -i .rpm 3.9. [488]ELDK Packages 3.9.1. List of ELDT Packages [489]Package Name [490]Package Version autoconf 2.61-8 automake 1.10-5 bison 2.3-3 crosstool-devel 0.43-3 dtc 20070802-1 elocaledef 1-1 ftdump 20070802-1 gdb 6.7-2 genext2fs 1.4.1-1 info 4.8-15 ldd 0.1-1 libtool 1.5.22-11 make 3.81-6 mkcramfs 1.1-1 mkimage 1.3.1-1 mtd-utils 1.0.1-2 rpm 4.4.2-46_2 rpm-build 4.4.2-46_2 sed 4.1.4-1 texinfo 4.8-15 ALERT! Note: The crosstool 0.43 ELDT package provides the following packages: gcc 4.2.2, gcc-c++ 4.2.2, gcc-java 4.2.2, cpp 4.2.2 and binutils 2.17.90. For more information about the crosstool package please refer to [491]http://kegel.com/crosstool. 3.9.2. List of Target Packages [492]Package Name [493]Package Version acl 2.2.39-3.1 appweb 2.2.2-5 attr 2.4.32-2 autoconf 2.61-8 bash 3.2-9 bc 1.06-26 bind 9.4.1-8.P1 binutils 2.17.90-1 binutils-devel 2.17.90-1 boa 0.94.14-0.5.rc21 busybox 1.7.1-2 byacc 1.9.20050813-1 bzip2 1.0.4-10 bzip2-devel 1.0.4-10 bzip2-libs 1.0.4-10 ccid 1.2.1-10 chkconfig 1.3.34-1 coreutils 6.9-3 cpio 2.6-27 cpp 4.2.2-2 cracklib 2.8.9-11 cracklib-dicts 2.8.9-11 crosstool-targetcomponents 0.43-3 curl 7.16.2-1 cyrus-sasl 2.1.22-6 cyrus-sasl-devel 2.1.22-6 cyrus-sasl-lib 2.1.22-6 db4 4.5.20-5_2 db4-devel 4.5.20-5_2 db4-utils 4.5.20-5_2 device-mapper 1.02.17-7 device-mapper-devel 1.02.17-7 device-mapper-libs 1.02.17-7 dhclient 3.0.5-38 dhcp 3.0.5-38 diffutils 2.8.1-16 directfb 1.0.0-1 dosfstools 2.11-8 dropbear 0.50-1 dtc 20070802-1 duma 2.5.8-2 e2fsprogs 1.39-11 e2fsprogs-devel 1.39-11 e2fsprogs-libs 1.39-11 ethtool 5-1 expat 1.95.8-9 expat-devel 1.95.8-9 file 4.21-1 file-libs 4.21-1 findutils 4.2.29-2 flex 2.5.33-9 freetype 2.3.4-3 freetype-devel 2.3.4-3 ftdump 20070802-1 ftp 0.17-40 gawk 3.1.5-15 gcc 4.2.2-2 gcc-c++ 4.2.2-2 gcc-java 4.2.2-2 gdb 6.7-1 glib 1.2.10-26 glib2 2.12.13-1 glib2-devel 2.12.13-1 glib-devel 1.2.10-26 gmp 4.1.4-12.3 grep 2.5.1-57 groff 1.18.1.4-2 gzip 1.3.11-2 hdparm 6.9-3 httpd 2.2.4-4.1 httpd-devel 2.2.4-4.1 httpd-manual 2.2.4-4.1 initscripts 8.54.1-1 iproute 2.6.20-2 iptables 1.3.8-2 iputils 20070202-3 iscsitarget 0.4.15-1 kbd 1.12-22 kernel-headers 2.6.24-1 kernel-source 2.6.24-1 krb5-devel 1.6.1-2.1 krb5-libs 1.6.1-2.1 less 394-9 libattr 2.4.32-2 libattr-devel 2.4.32-2 libcap 1.10-29 libcap-devel 1.10-29 libpng 1.2.16-1 libpng-devel 1.2.16-1 libsysfs 2.1.0-1 libsysfs-devel 2.1.0-1 libtermcap 2.0.8-46.1 libtermcap-devel 2.0.8-46.1 libtirpc 0.1.7-7_2 libtirpc-devel 0.1.7-7_2 libtool 1.5.22-11 libtool-ltdl 1.5.22-11 libtool-ltdl-devel 1.5.22-11 libusb 0.1.12-7 libusb-devel 0.1.12-7 libuser 0.56.2-1 libuser-devel 0.56.2-1 libxml2 2.6.29-1 logrotate 3.7.5-3.1 lrzsz 0.12.20-22.1 lsof 4.78-5 ltp 20080131-eldk2 lvm2 2.02.24-1 m4 1.4.8-2 mailcap 2.1.23-1 make 3.81-6 MAKEDEV 3.23-1.2 man 1.6e-3 mdadm 2.6.2-4 microwindows 0.91-2 microwindows-fonts 0.91-1 mingetty 1.07-5.2.2 mktemp 1.5-25 module-init-tools 3.3-0.pre11.1.0 mtd-utils 1.0.1-2 ncompress 4.2.4-49 ncurses 5.6-17 ncurses-devel 5.6-17 net-snmp 5.4-14 net-snmp-devel 5.4-14 net-snmp-libs 5.4-14 net-snmp-utils 5.4-14 net-tools 1.60-82 newt 0.52.6-30 newt-devel 0.52.6-30 nfs-utils 1.1.0-1 ntp 4.2.4p2-1 open-iscsi 2.0-865.15 openldap 2.3.34-3 openldap-devel 2.3.34-3 openssl 0.9.8b-12_2 openssl-devel 0.9.8b-12_2 oprofile 0.9.2-8_2 pam 0.99.7.1-5.1 pam-devel 0.99.7.1-5.1 passwd 0.74-3 patch 2.5.4-29.2.2 pciutils 2.2.4-3_2 pciutils-devel 2.2.4-3_2 pcmciautils 014-9_2 pcre 7.0-2 pcsc-lite 1.3.3-1.0 pcsc-lite-devel 1.3.3-1.0 pcsc-lite-libs 1.3.3-1.0 perl 5.8.8-18_2 perl-libs 5.8.8-18_2 popt 1.12-1 portmap 4.0-65_2 postgresql 8.2.4-1_2 postgresql-devel 8.2.4-1_2 postgresql-libs 8.2.4-1_2 ppp 2.4.4-7 procps 3.2.7-14 psmisc 22.3-2 python 2.5.1-1 rdate 1.4-6 readline 5.2-4 readline-devel 5.2-4 routed 0.17-12_1 rpcbind 0.1.4-6 rpm 4.4.2-46_2 rpm-build 4.4.2-46_2 rpm-devel 4.4.2-46_2 rpm-libs 4.4.2-46_2 rsh 0.17-40 rsh-server 0.17-40 screen 4.0.3-50 sed 4.1.5-7 [494]SELF 1.0-13 setup 2.6.4-1_2 shadow-utils 4.0.18.1-15 slang 2.0.7-17 slang-devel 2.0.7-17 smartmontools 5.38-2 strace 4.5.15-1 sysfsutils 2.1.0-1 sysklogd 1.4.2-9 sysvinit 2.86-17 tar 1.15.1-26 tcp_wrappers 7.6-48 tcp_wrappers-devel 7.6-48 tcp_wrappers-libs 7.6-48 telnet 0.17-38 telnet-server 0.17-38 termcap 5.5-1.20060701.1 tftp 0.42-4 tftp-server 0.42-4 thttpd 2.25b-13 time 1.7-29 u-boot 1.3.1-1 udev 106-4.1 unixODBC 2.2.12-2 unzip 5.52-4 util-linux 2.13-0.52_2 vim-common 7.1.12-1 vim-minimal 7.1.12-1 vixie-cron 4.1-82 vsftpd 2.0.5-16_2 which 2.16-8 wireless-tools 28-4 wpa_supplicant 0.5.7-3 wu-ftpd 2.6.2-1 xdd 65.013007-1 xenomai 2.4.2-1 xinetd 2.3.14-12 zip 2.31-3 zlib 1.2.3-10 zlib-devel 1.2.3-10 ALERT! Note 1: Not all packages will be installed automatically; for example the boa and thttpd web servers are mutually exclusive - you will have to remove one package before you can (manually) install the other one. ALERT! Note 2: The crosstool 0.43 target package provides the following packages: glibc 2.6, glibc-common 2.6, glibc-devel 2.6, libstdc++ 4.2.2, libgcj 4.2.2, libgcj-devel 4.2.2 and libstdc++-devel 4.2.2. For more information about the crosstool package please refer to [495]http://kegel.com/crosstool ALERT! Note 3: The Xenomai and gcc-java packages are unavailable in ARM [496]ELDK version. 3.10. Rebuilding the [497]ELDK from Scratch In this section, you will find instructions on how to build the [498]ELDK from scratch, using the pristine package sources available on the Internet, and patches, spec files, and build scripts provided on the [499]ELDK source CD-ROM. 3.10.1. [500]ELDK Build Process Overview The [501]ELDK uses the Fedora 7 Linux distribution as source code reference. Any modifications to Fedora's sources the [502]ELDK has introduced are in the form of patches applied by the RPM tool while building the packages. Also, the [503]ELDK uses modified spec files for its RPM packages. So, the sources of almost every [504]ELDK package consist of the following parts: * Fedora pristine sources or * [505]ELDK source tarball, * [506]ELDK patches, * [507]ELDK spec file. The Fedora pristine sources may be obtained from the Internet, see [508]http://download.fedora.redhat.com/pub/fedora/linux/core. The [509]ELDK patches and spec files are available on the [510]ELDK source CD-ROM and from the DENX GIT repositories. Also, for convenience, the pristine Fedora sources are available here, too. Please use the following commands to check out a copy of one of the modules: git-clone git://www.denx.de/git/eldk/module The following [511]ELDK modules are available: [512]Module Name [513]Contents tarballs Source tarballs build Build tools, patches, and spec files SRPMS Fedora 7 sources Then you may switch to a specific release of the [514]ELDK using the "git-checkout" command; for example, to get the files for [515]ELDK release 4.1, please do the following from the module directory: git-checkout ELDK_4_2 It must be noted that some of the packages which are included in the [516]ELDK are not included in Fedora. Examples of such packages are appWeb, microwindows, and wu-ftpd. For these packages tarballs are provided in the DENX GIT repository. To facilitate building of the [517]ELDK, a build infrastructure has been developed. The infrastructure is composed of the following components: * ELDK_BUILD script * build.sh script * cpkgs.lst file * tpkgs.lst file * SRPMS.lst file * tarballs.lst file The ELDK_BUILD script is the main script of the [518]ELDK build procedure. It is the tool that you would normally use to build the [519]ELDK from scratch. In the simplest case, the script may be invoked without arguments, and it will perform all necessary steps to build the [520]ELDK in a fully automated way. You may pass the following optional arguments to the ELDK_BUILD script: -a target architecture: "arm", "ppc" or "ppc64", defaults to "ppc". -n an identification string for the build. Defaults to the value based on the build architecture and current date, and has the following format: -YYYY-MM-DD -v [521]ELDK version string -u build the uClibc-based [522]ELDK version (on the platforms and versions where this is available). -p Optional build directory. By default, build will place the work files and results in the current directory. ALERT! Warning: The [523]ELDK build scripts rely on standard behaviour of the RPM tool. Make sure you don't use non-standard settings in your personal ~/.rpmmacros file that might cause conflicts. build.sh is a supplementary script that is called by ELDK_BUILD to accomplish certain steps of the build. Refer to section [524]3.10.3. build.sh Usage below for more details. The cpkgs.lst and tpkgs.lst files are read by build.sh and must contain lines describing sub-steps of the eldt and trg build procedure steps. Essentially, the files contain the list of the ELDT and target packages to be included in the [525]ELDK. The SRPMS.lst file contains the list of the Fedora source RPM packages used during the [526]ELDK build. The tarballs.lst file contains the list of source tarballs of the packages that are included in the [527]ELDK but are not present in Fedora 7. For the ELDK_BUILD script to work correctly, it must be invoked from a certain build environment created on the host system. The build environment can be either checked out from the DENX GIT repository (see section [528]3.10.2. Setting Up [529]ELDK Build Environment below for details) or copied from the [530]ELDK build environment CD-ROM. To be more specific, the following diagram outlines the build environment needed for correct operation of the ELDK_BUILD script: / build/cross_rpms//SPECS/... SOURCES/... target_rpms//SPECS/... SOURCES/... install/install.c Makefile misc/ELDK_MAKEDEV ELDK_FIXOWNER README.html cpkgs.lst tpkgs.lst build.sh ELDK_BUILD SRPMS.lst tarballs.lst tarballs/.... SRPMS/.... SRPMS-updates/.... In subdirectories of the cross_rpms and target_rpms directories, the sources and RPM spec files of, respectively, the ELDT and target packages are stored. The install subdirectory contains the sources of the installation utility which will be built and placed in the root of the ISO image. tarballs directory contains the source tarballs of the packages that are included in the [531]ELDK but are not present in Fedora 7. The SRPMS and SRPMS-updates directories may contain the source RPM packages of Fedora 7. The ELDK_BUILD script looks for a package in the SRPMS directory and then, if the package is not found, in the SRPMS-updates directory. If some (or all) of the Fedora SRPMs needed for the build are missing in the directories, the ELDK_BUILD script will download the source RPMs automatically from the Internet. The [532]ELDK build environment CD-ROM provides a ready-to-use [533]ELDK build environment. Please refer to section [534]3.10.2. Setting Up [535]ELDK Build Environment below for detailed instructions on setting up the build environment. The ELDK_BUILD script examines the contents of the ELDK_PREFIX environment variable to determine the root directory of the [536]ELDK build environment. If the variable is not set when the script is invoked, it is assumed that the root directory of the [537]ELDK build environment is /opt/eldk. To build the [538]ELDK in the example directory layout given above, you must set and export the ELDK_PREFIX variable prior to invoking ELDK_BUILD. After all the build steps are complete, the following subdirectories are created in the [539]ELDK build environment: build//work/ - full ELDK environment build//logs/ - build procedure log files build//results/b_cdrom/ - binary cdrom tree, ready for mkisofs results/s_cdrom/ - source cdrom tree, ready for mkisofs results/d_cdrom/ - debuginfo cdrom tree, ready for mkisofs On Linux hosts, the binary and source ISO images are created automatically by the ELDK_BUILD script and placed in the results directory. On Solaris hosts, creating the ISO images is a manual step. Use the contents of the b_cdrom and s_cdrom directories for the contents of the ISO images. 3.10.2. Setting Up [540]ELDK Build Environment For your convenience, the [541]ELDK build environment CD-ROM provides full [542]ELDK build environment. All you need to do is copy the contents of the CD-ROM to an empty directory on your host system. Assuming the [543]ELDK build environment CD-ROM is mounted at /mnt/cdrom, and the empty directory where you want to create the build environment is named /opt/eldk, use the following commands to create the build environment: bash$ cd /opt/eldk bash$ cp -r /mnt/cdrom/* . These commands will create the directory structure as described in section [544]3.10.1. [545]ELDK Build Process Overview above. All necessary scripts and [546]ELDK specific source files will be placed in the build subdirectory, and the required tarballs can be found in the tarballs subdirectory. In the SRPMS subdirectory, you will find all the Fedora 7 SRPMS needed to build the [547]ELDK. Alternatively, you can obtain the [548]ELDK build environment from the DENX GIT repository. Two modules are provided for check out: build and tarballs. The first one contains the files for the build subdirectory in the build environment, and the second one contains source tarballs of the packages that are included in the [549]ELDK but are not present in Fedora 7. To create the [550]ELDK build environment from the DENX GIT repository, use the following commands (the example below assumes that the root directory of the build environment is /opt/eldk): bash$ cd /opt/eldk bash$ git-clone git://www.denx.de/git/eldk/build bash$ git-clone git://www.denx.de/git/eldk/tarballs bash$ git-clone git://www.denx.de/git/eldk/SRPMS ALERT! Note: To allow to install the [551]ELDK on as many as possible Linux distributions (including old systems), we use a Red Hat 7.3 host system for building. Also, Fedora Core 5 is known to work as a build environment. Other, especially more recent Linux distributions, will most likely have problems. We therefor provide a Red Hat 7.3 based root file system image than can run in some virtualization environment (like qemu etc.). Here is an application note with detailed instructions: [552]http://www.denx.de/wiki/DULG/AN2009_02_EldkReleaseBuildEnvironmen t 3.10.3. build.sh Usage If you wish to perform only a part of the [553]ELDK build procedure, for instance to re-build or update a certain package, it may sometimes be convenient to invoke the build.sh script manually, without the aid of the ELDK_BUILD script. Please note, however, that this approach is in general discouraged. The whole build procedure is logically divided into six steps, and the build.sh must be told which of the build steps to perform. The build steps are defined as follows: * rpm - build RPM * eldt - build ELDT packages * seldt - save ELDT SRPM packages to create a source ISO image later on * trg - build target packages * biso - prepare the file tree to create the binary ISO image * siso - prepare the file tree to create the source ISO image * diso - prepare the file tree to create the debuginfo ISO image Further, the eldt and trg build steps are devided into sub-steps, as defined in the cpkgs.lst and tpkgs.lst files (see below for details). You may specify which sub-steps of the build step are to be performed. The formal syntax for the usage of build.sh is as follows: bash$ ./build.sh [-a ] [-n ] [-p ] [-r ] \ [-w ] [] -a target architecture: "ppc", "ppc64", "arm" or "mips", defaults to "ppc". -n an identification string for the build. It is used as a name for some directories created during the build. You may use for example the current date as the build name. -p is the name of the directory that contains the build environment. Refer to [554]build overview above for description of the build environment. -r is the name of the directory where the resulting RPMs and SRPMs created on this step will be placed. -w is the name of the directory where the build is performed. is the name of the build step that is to be performed. Refer to the list of the build procedure steps above. is an optional parameter which identifies sub-steps of the step which are to be performed. This is useful when you want to re-build only some specific packages. The numbers are defined in the cpkgs.lst and tpkgs.lst files discussed below. You can specify a range of numbers here. For instance, "2 5" means do steps from 2 to 5, while simply "2" means do all steps starting at 2. ALERT! Please note that you must never use build.sh to build the [555]ELDK from scratch. For build.sh to work correctly, the script must be invoked from the build environment after a successful build using the ELDK_BUILD script. A possible scenario of build.sh usage is such that you have a build environment with results of a build performed using the ELDK_BUILD script and want to re-build certain ELDT and target packages, for instance, because you have updated sources of a package or added a new package to the build. When building the target packages (during the trg buildstep), build.sh examines the contents of the TARGET_CPU_FAMILY_LIST environment variable, which may contain a list indicating which target [556]CPU variants the packages must be built for. Possible [557]CPU variants are 4xx, 4xxFP, 6xx, 74xx, 8xx, 85xx and ppc64. For example, the command below rebuilds the target RPM listed in the tpckgs.lst file under the number of 47 (see section [558]3.10.4. Format of the cpkgs.lst and tpkgs.lst Files for description of the tpckgs.lst and cpkgs.lst files), for the 8xx and 85xx [559]CPUs: bash$ TARGET_CPU_FAMILY_LIST="8xx 85xx" \ > /opt/eldk/build.sh -a ppc \ > -n 2007-01-19 \ > -p /opt/eldk/build/ppc-2007-01-19 \ > -r /opt/eldk/build/ppc-2007-01-19/results \ > -w /opt/eldk/build/ppc-2007-01-19/work \ > trg 47 47 Note: If you are going to invoke build.sh to re-build a package that has already been built in the build environment by the ELDK_BUILD script, then you must first manually uninstall the package from [560]ELDK installation created by the build procedure under the work directory of the build environment. Note: It is recommended that you use the build.sh script only at the final stage of adding/updating a package to the [561]ELDK. For debugging purposes, it is much more convenient and efficient to build both ELDT and target packages using a working [562]ELDK installation, as described in the sections [563]3.8.2. Rebuilding Target Packages and [564]3.8.3. Rebuilding ELDT Packages above. 3.10.4. Format of the cpkgs.lst and tpkgs.lst Files Each line of these files has the following format: \ The [565]ELDK source CD-ROM contains the cpkgs.lst and tpkgs.lst files used to build this version of the [566]ELDK distribution. Use them as reference if you want to include any additional packages into the [567]ELDK, or remove unneeded packages. To add a package to the [568]ELDK you must add a line to either the cpkgs.lst file, if you are adding a ELDT package, or to the tpkgs.lst file, if it is a target package. Keep in mind that the relative positions of packages in the cpkgs.lst and tpkgs.lst files (the sub-step numbers) are very important. The build procedure builds the packages sequentially as defined in the *.lst files and installs the packages in the "work" environment as they are built. This implies that if a package depends on other packages, those packages must be specified earlier (with smaller sub-step numbers) in the *.lst files. Note: For cpkgs.lst, the package_version may be replaced by the special keyword "RHAUX". Such packages are used as auxiliary when building [569]ELDK 4.2 on non-Fedora hosts. These packages will be built and used during the build process, but will not be put into the [570]ELDK 4.2 distribution ISO images. 3.11. Notes for Solaris 2.x Host Environment If you use a Solaris 2.x host environment, you need additional freeware packages (mostly GNU tools) to install and especially to build the [571]ELDK packages. The following table lists all required packages that must be installed on the Solaris host system before attempting to build and/or install the [572]ELDK. All these files except those marked with (**) (and the RPM and zlib-1.1.2 packages, which are available at [573]ftp://rpmfind.net/linux/solaris are available for free download at [574]ftp://ftp.sunfreeware.com/pub/freeware/sparc/2.6/ Necessary Freeware Packages: [575]Package [576]Version [577]Instance [578]File Name autoconf(**) 2.13 SMCautoc autoconf-2.13-sol26-sparc-local.gz automake(**) 1.4 SMCautom automake-1.4-sol26-sparc-local.gz bash 2.05 SMCbash bash-2.05-sol26-sparc-local.gz binutils 2.11.2 SMCbinut binutils-2.11.2-sol26-sparc-local.gz bison 1.28 SMCbison bison-1.28-sol26-sparc-local.gz bzip2 1.0.1 SMCbzip2 bzip2-1.0.1-sol26-sparc-local.gz ddd(*) 3.0 TUBddd ddd-3.0-sol26-sparc-local.gz diffutils 2.7 GNUdiffut diffutils-2.7-sol26-sparc-local.gz expect(*) 5.25 NTexpect expect-5.25-sol26-sparc-local.gz fileutils 4.0 SMCfileu fileutils-4.0-sol26-sparc-local.gz flex 2.5.4a FSFflex flex-2.5.4a-sol26-sparc-local.gz gawk 3.1.0 SMCgawk gawk-3.1.0-sol26-sparc-local.gz gcc 2.95.3 SMCgcc gcc-2.95.3-sol26-sparc-local.gz gettext 0.10.37 SMCgtext gettext-0.10.37-sol26-sparc-local.gz gzip 1.3 SMCgzip gzip-1.3-sol26-sparc-local libiconv 1.6.1 SMClibi libiconv-1.6.1-sol26-sparc-local.gz libtool 1.4 SMClibt libtool-1.4-sol26-sparc-local.gz m4 1.4 SMCm4 m4-1.4-sol26-sparc-local.gz make(**) 3.79.1 SMCmake make-3.79.1-sol26-sparc-local.gz ncurses 5.2 SMCncurs ncurses-5.2-sol26-sparc-local.gz patch 2.5 FSFpatch patch-2.5-sol26-sparc-local.gz perl(**) 5.005_03 SMCperl perl-5.005_03-sol26-sparc-local.gz python 1.5.2 SMCpython python-1.5.2-sol26-sparc-local.gz rpm 2.5.2 RPM rpm-2.5.2.pkg sed 3.02 SMCsed sed-3.02-sol26-sparc-local.gz tar 1.13.19 SMCtar tar-1.13.19-sol26-sparc-local.gz tcl(*) 8.3.3 SMCtcl tcl-8.3.3-sol26-sparc-local.gz texinfo 4.0 SMCtexi texinfo-4.0-sol26-sparc-local.gz textutils 2.0 SMCtextu textutils-2.0-sol26-sparc-local.gz unzip 5.32 IZunzip unzip-5.32-sol26-sparc-local.gz wget 1.7 SMCwget wget-1.7-sol26-sparc-local.gz zlib(**) 1.0.4 SMCzlib zlib-1.0.4-sol26-sparc-local.gz zlib 1.1.2 - zlib-1.1.2.tar.gz The packages marked "(*)" are not absolutely required, but sooner or later you will need them anyway so we recommend to install them. The packages marked "(**)" are older versions of the ones currently available at [579]ftp://ftp.sunfreeware.com/pub/freeware/sparc/2.6/. You can obtain them from the DENX public [580]FTP server. The following symbolic links must be created in order to be able to build the [581]ELDK on a Solaris machine: /usr/local/bin/cc --> /usr/local/bin/gcc /usr/lib/libiconv.so.2 --> /usr/local/lib/libiconv.so.2 /usr/lib/libncurses.so.5 --> /usr/local/lib/libncurses.so.5 Additionally, to be able to build the [582]ELDK on Solaris, you must place newer GNU gettext macros to the /usr/local/share/aclocal directory. This can be accomplished as follows: * Download the [583]http://www.ibiblio.org/pub/packages/solaris/sparc/GNUgettext. 0.10.40.SPARC.32bit.Solaris.8.pkg.tgz package. * Untar the package to a temporary directory and copy the macros to the /usr/local/share/aclocal directory: $ cp GNUgettext/root/usr/local/share/aclocal/*.m4 /usr/local/share/aclocal * [584]4. System Setup + [585]4.1. Serial Console Access + [586]4.2. Configuring the "cu" command + [587]4.3. Configuring the "kermit" command + [588]4.4. Using the "minicom" program + [589]4.5. Permission Denied Problems + [590]4.6. Configuration of a [591]TFTP Server + [592]4.7. Configuration of a [593]BOOTP / [594]DHCP Server + [595]4.8. Configuring a NFS Server 4. System Setup Some tools are needed to install and configure U-Boot and Linux on the target system. Also, especially during development, you will want to be able to interact with the target system. This section describes how to configure your host system for this purpose. 4.1. Serial Console Access To use U-Boot and Linux as a development system and to make full use of all their capabilities you will need access to a serial console port on your target system. Later, U-Boot and Linux can be configured to allow for automatic execution without any user interaction. There are several ways to access the serial console port on your target system, such as using a terminal server, but the most common way is to attach it to a serial port on your host. Additionally, you will need a terminal emulation program on your host system, such as cu or kermit. 4.2. Configuring the "cu" command The cu command is part of the UUCP package and can be used to act as a dial-in terminal. It can also do simple file transfers, which can be used in U-Boot for image download. On [596]RedHat systems you can check if the UUCP package is installed as follows: $ rpm -q uucp If necessary, install the UUCP package from your distribution media. To configure cu for use with U-Boot and Linux please make sure that the following entries are present in the UUCP configuration files; depending on your target configuration the serial port and/or the console baudrate may be different from the values used in this example: (/dev/ttyS0, 115200 bps, 8N1): * /etc/uucp/sys: # # /dev/ttyS0 at 115200 bps: # system S0@115200 port serial0_115200 time any * /etc/uucp/port: # # /dev/ttyS0 at 115200 bps: # port serial0_115200 type direct device /dev/ttyS0 speed 115200 hardflow false You can then connect to the serial line using the command $ cu S0@115200 Connected. To disconnect, type the escape character '~' followed by '.' at the beginning of a line. See also: cu(1), info uucp. 4.3. Configuring the "kermit" command The name kermit stands for a whole family of communications software for serial and network connections. The fact that it is available for most computers and operating systems makes it especially well suited for our purposes. kermit executes the commands in its initialization file, .kermrc, in your home directory before it executes any other commands, so this can be easily used to customize its behaviour using appropriate initialization commands. The following settings are recommended for use with U-Boot and Linux: * ~/.kermrc: set line /dev/ttyS0 set speed 115200 set carrier-watch off set handshake none set flow-control none robust set file type bin set file name lit set rec pack 1000 set send pack 1000 set window 5 This example assumes that you use the first serial port of your host system (/dev/ttyS0) at a baudrate of 115200 to connect to the target's serial console port. You can then connect to the serial line: $ kermit -c Connecting to /dev/ttyS0, speed 115200. The escape character is Ctrl-\ (ASCII 28, FS) Type the escape character followed by C to get back, or followed by ? to see other options. ---------------------------------------------------- TIP Due to licensing conditions you will often find two kermit packages in your GNU/Linux distribution. In this case you will want to install the ckermit package. The gkermit package is only a command line tool implementing the kermit transfer protocol. TIP If you cannot find kermit on the distribution media for your Linux host system, you can download it from the kermit project home page: [597]http://www.columbia.edu/kermit/ 4.4. Using the "minicom" program minicom is another popular serial communication program. Unfortunately, many users have reported problems using it with U-Boot and Linux, especially when trying to use it for serial image download. It's use is therefore discouraged. 4.5. Permission Denied Problems The terminal emulation program must have write access to the serial port and to any locking files that are used to prevent concurrent access from other applications. Depending on the used Linux distribution you may have to make sure that: * the serial device belongs to the same group as the cu command, and that the permissions of cu have the setgid bit set * the kermit belongs to the same group as cu and has the setgid bit set * the /var/lock directory belongs to the same group as the cu command, and that the write permissions for the group are set 4.6. Configuration of a [598]TFTP Server The fastest way to use U-Boot to load a Linux kernel or an application image is file transfer over Ethernet. For this purpose, U-Boot implements the TFTP protocol (see the tftpboot command in U-Boot). To enable [599]TFTP support on your host system you must make sure that the [600]TFTP daemon program /usr/sbin/in.tftpd is installed. On [601]RedHat systems you can verify this by running: $ rpm -q tftp-server If necessary, install the [602]TFTP daemon program from your distribution media. Most Linux distributions disable the [603]TFTP service by default. To enable it for example on [604]RedHat systems, edit the file /etc/xinetd.d/tftp and remove the line disable = yes or change it into a comment line by putting a hash character in front of it: # default: off # description: The tftp server serves files using the trivial file transfer # protocol. The tftp protocol is often used to boot diskless # workstations, download configuration files to network-aware printers, # and to start the installation process for some operating systems. service tftp { socket_type = dgram protocol = udp wait = yes user = root server = /usr/sbin/in.tftpd server_args = -s /tftpboot # disable = yes per_source = 11 cps = 100 2 } Also, make sure that the /tftpboot directory exists and is world-readable (permissions at least "dr-xr-xr-x"). 4.7. Configuration of a [605]BOOTP / [606]DHCP Server BOOTP resp. DHCP can be used to automatically pass configuration information to the target. The only thing the target must "know" about itself is its own Ethernet hardware ([607]MAC) address. The following command can be used to check if [608]DHCP support is available on your host system: $ rpm -q dhcp If necessary, install the [609]DHCP package from your distribution media. Then you have to create the [610]DHCP configuration file /etc/dhcpd.conf that matches your network setup. The following example gives you an idea what to do: subnet 192.168.0.0 netmask 255.255.0.0 { option routers 192.168.1.1; option subnet-mask 255.255.0.0; option domain-name "local.net"; option domain-name-servers ns.local.net; host trgt { hardware ethernet 00:30:BF:01:02:D0; fixed-address 192.168.100.6; option root-path "/opt/eldk-4.2/ppc_4xx"; option host-name "canyonlands"; next-server 192.168.1.1; filename "/tftpboot/canyonlands/uImage"; } } With this configuration, the [611]DHCP server will reply to a request from the target with the ethernet address 00:30:BF:01:02:D0 with the following information: * The target is located in the subnet 192.168.0.0 which uses the netmask 255.255.0.0. * The target has the hostname canyonlands and the IP address 192.168.100.6. * The host with the IP address 192.168.1.1 will provide the boot image for the target and provide NFS server function in cases when the target mounts it's root filesystem over NFS. TIP The host listed with the next-server option can be different from the host that is running the [612]DHCP server. * The host provides the file /tftpboot/canyonlands/uImage as boot image for the target. * The target can mount the directory /opt/eldk-4.2/ppc_4xx on the NFS server as root filesystem. 4.8. Configuring a NFS Server For a development environment it is very convenient when the host and the target can share the same files over the network. The easiest way for such a setup is when the host provides NFS server functionality and exports a directory that can be mounted from the target as the root filesystem. Assuming NFS server functionality is already provided by your host, the only configuration that needs to be added is an entry for your target root directory to your /etc/exports file, for instance like this: /opt/eldk-4.2/ppc_4xx 192.168.0.0/255.255.0.0(rw,no_root_squash,sync) This line exports the /opt/eldk-4.2/ppc_4xx directory with read and write permissions to all hosts on the 192.168.0.0 subnet. After modifying the /etc/exports file you must make sure the NFS system is notified about the change, for instance by issuing the command: # /sbin/service nfs restart * [613]5. Das U-Boot + [614]5.1. Current Versions + [615]5.2. Unpacking the Source Code + [616]5.3. Configuration + [617]5.4. Installation o [618]5.4.1. Before You Begin # [619]5.4.1.1. Installation Requirements # [620]5.4.1.2. Board Identification Data o [621]5.4.2. Installation Using a BDM/JTAG Debugger o [622]5.4.3. Installation using U-Boot + [623]5.5. Tool Installation + [624]5.6. Initialization + [625]5.7. Initial Steps + [626]5.8. The First Power-On + [627]5.9. U-Boot Command Line Interface o [628]5.9.1. Information Commands # [629]5.9.1.1. bdinfo - print Board Info structure # [630]5.9.1.2. coninfo - print console devices and informations # [631]5.9.1.3. flinfo - print FLASH memory information # [632]5.9.1.4. iminfo - print header information for application image # [633]5.9.1.5. help - print online help o [634]5.9.2. Memory Commands # [635]5.9.2.1. base - print or set address offset # [636]5.9.2.2. crc32 - checksum calculation # [637]5.9.2.3. cmp - memory compare # [638]5.9.2.4. cp - memory copy # [639]5.9.2.5. md - memory display # [640]5.9.2.6. mm - memory modify (auto-incrementing) # [641]5.9.2.7. mtest - simple RAM test # [642]5.9.2.8. mw - memory write (fill) # [643]5.9.2.9. nm - memory modify (constant address) # [644]5.9.2.10. loop - infinite loop on address range o [645]5.9.3. Flash Memory Commands # [646]5.9.3.1. cp - memory copy # [647]5.9.3.2. flinfo - print FLASH memory information # [648]5.9.3.3. erase - erase FLASH memory # [649]5.9.3.4. protect - enable or disable FLASH write protection # [650]5.9.3.5. mtdparts - define a Linux compatible MTD partition scheme o [651]5.9.4. Execution Control Commands # [652]5.9.4.1. source - run script from memory # [653]5.9.4.2. bootm - boot application image from memory # [654]5.9.4.3. go - start application at address 'addr' o [655]5.9.5. Download Commands # [656]5.9.5.1. bootp - boot image via network using BOOTP/TFTP protocol # [657]5.9.5.2. dhcp - invoke DHCP client to obtain IP/boot params # [658]5.9.5.3. loadb - load binary file over serial line (kermit mode) # [659]5.9.5.4. loads - load S-Record file over serial line # [660]5.9.5.5. rarpboot- boot image via network using RARP/TFTP protocol # [661]5.9.5.6. tftpboot- boot image via network using TFTP protocol o [662]5.9.6. Environment Variables Commands # [663]5.9.6.1. printenv- print environment variables # [664]5.9.6.2. saveenv - save environment variables to persistent storage # [665]5.9.6.3. setenv - set environment variables # [666]5.9.6.4. run - run commands in an environment variable # [667]5.9.6.5. bootd - boot default, i.e., run 'bootcmd' o [668]5.9.7. Flattened Device Tree support # [669]5.9.7.1. fdt addr - select FDT to work on # [670]5.9.7.2. fdt list - print one level # [671]5.9.7.3. fdt print - recursive print # [672]5.9.7.4. fdt mknode - create new nodes # [673]5.9.7.5. fdt set - set node properties # [674]5.9.7.6. fdt rm - remove nodes or properties # [675]5.9.7.7. fdt move - move FDT blob to new address # [676]5.9.7.8. fdt chosen - fixup dynamic info o [677]5.9.8. Special Commands # [678]5.9.8.1. i2c - I2C sub-system o [679]5.9.9. Storage devices o [680]5.9.10. Miscellaneous Commands # [681]5.9.10.1. echo - echo args to console # [682]5.9.10.2. reset - Perform RESET of the CPU # [683]5.9.10.3. sleep - delay execution for some time # [684]5.9.10.4. version - print monitor version # [685]5.9.10.5. ? - alias for 'help' + [686]5.10. U-Boot Environment Variables + [687]5.11. U-Boot Scripting Capabilities + [688]5.12. U-Boot Standalone Applications o [689]5.12.1. "Hello World" Demo o [690]5.12.2. Timer Demo + [691]5.13. U-Boot Image Formats + [692]5.14. U-Boot Advanced Features o [693]5.14.1. Boot Count Limit 5. Das U-Boot 5.1. Current Versions Das U-Boot (or just "U-Boot" for short) is Open Source Firmware for Embedded [694]Power Architecture®, ARM, MIPS, x86 and other processors. The U-Boot project is hosted by DENX, where you can also find the project home page: [695]http://www.denx.de/wiki/U-Boot/ The current version of the U-Boot source code can be retrieved from the DENX [696]"git" repository. You can browse the "git" repositories at [697]http://git.denx.de/ The trees can be accessed through the git, HTTP, and rsync protocols. For example you can use one of the following commands to create a local clone of one of the source trees: git clone git://git.denx.de/u-boot.git u-boot/ git clone http://git.denx.de/u-boot.git u-boot/ git clone rsync://git.denx.de/u-boot.git u-boot/ For details please see [698]here. Official releases of U-Boot are also available through [699]FTP. Compressed tar archives can downloaded from the directory [700]ftp://ftp.denx.de/pub/u-boot/. 5.2. Unpacking the Source Code If you used [701]GIT to get a copy of the U-Boot sources, then you can skip this next step since you already have an unpacked directory tree. If you downloaded a compressed tarball from the DENX [702]FTP server, you can unpack it as follows: $ cd /opt/eldk/usr/src $ wget ftp://ftp.denx.de/pub/u-boot/u-boot-1.3.2.tar.bz2 $ rm -f u-boot $ bunzip2 < u-boot-1.3.2.tar.bz2 | tar xf - $ ln -s u-boot-1.3.2 u-boot $ cd u-boot 5.3. Configuration After changing to the directory with the U-Boot source code you should make sure that there are no build results from any previous configurations left: $ make distclean The following (model) command configures U-Boot for the canyonlands board: $ make canyonlands_config And finally we can compile the tools and U-Boot itself: $ make all By default the build is performed locally and the objects are saved in the source directory. One of the two methods can be used to change this behaviour and build U-Boot to some external directory: 1. Add O= to the make command line invocations: make O=/tmp/build distclean make O=/tmp/build canyonlands_config make O=/tmp/build all Note that if the 'O=output/dir' option is used then it must be used for all invocations of make. 2. Set environment variable BUILD_DIR to point to the desired location: export BUILD_DIR=/tmp/build make distclean make canyonlands_config make all Note that the command line "O=" setting overrides the BUILD_DIR environment variable. 5.4. Installation 5.4.1. Before You Begin 5.4.1.1. Installation Requirements The following section assumes that flash memory is used as the storage device for the firmware on your board. If this is not the case, the following instructions will not work - you will probably have to replace the storage device (probably ROM or EPROM) on such systems to install or update U-Boot. 5.4.1.2. Board Identification Data All canyonlands boards use a serial number for identification purposes. Also, all boards have at least one ethernet ([703]MAC) address assigned. You may lose your warranty on the board if this data gets lost. Before installing U-Boot or otherwise changing the software configuration of a board (like erasing some flash memory) you should make sure that you have all necessary information about such data. 5.4.2. Installation Using a BDM/JTAG Debugger A fast and simple way to write new data to flash memory is via the use of a debugger or flash programmer with a [704]BDM or [705]JTAG interface. In cases where there is no running firmware at all (for instance on new hardware), this is usually the only way to install any software at all. We use (and highly recommend) the BDI2000/BDI3000 by [706]Abatron . Other [707]BDM / [708]JTAG debuggers may work too, but how to use them is beyond the scope of this document. Please see the documentation for the tool you want to use. Before you can use the BDI2000 you have to configure it. A configuration file that can be used with canyonlands boards is included in section [709]13.3. BDI2000 Configuration file To install a new U-Boot image on your canyonlands board using a BDI2000, proceed as follows: Note: Included topic DULGData_canyonlands.InstallUBootUsingBDI2000 does not exist yet 5.4.3. Installation using U-Boot If U-Boot is already installed and running on your board, you can use these instructions to download another U-Boot image to replace the current one. ALERT! Warning: Before you can install the new image, you have to erase the current one. If anything goes wrong your board will be dead. It is strongly recommended that: * you have a backup of the old, working U-Boot image * you know how to install an image on a virgin system ALERT! Proceed as follows: => setenv u-boot /tftpboot/duts/canyonlands/u-boot.bin => run load update Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/u-boot.bin'. Load address: 0x200000 Loading: ########################### done Bytes transferred = 393216 (60000 hex) Un-Protected 3 sectors ... done Erased 3 sectors Copy to Flash... done Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => reset U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) CPU: AMCC PowerPC 460EX Rev. B at 1066.667 MHz (PLB=266 OPB=88 EBC=88) Security/Kasumi support Bootstrap Option H - Boot ROM Location I2C (Addr 0x52) Internal PCI arbiter enabled 32 kB I-Cache 32 kB D-Cache Board: Canyonlands - AMCC PPC460EX Evaluation Board, 1*PCIe/1*SATA, Rev. 16 I2C: ready DRAM: 512 MB (ECC not enabled, 533 MHz, CL4) FLASH: 64 MB NAND: 128 MiB PCI: Bus Dev VenId DevId Class Int PCIE1: link is not up. DTT: 1 is 32 C Net: ppc_4xx_eth0, ppc_4xx_eth1 Type run flash_nfs to mount root filesystem over NFS Hit any key to stop autoboot: 0 => version U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) => 5.5. Tool Installation U-Boot uses a special image format when loading the Linux kernel or ramdisk or other images. This image contains (among other things) information about the time of creation, operating system, compression type, image type, image name and CRC32 checksums. The tool mkimage is used to create such images or to display the information they contain. When using the [710]ELDK, the mkimage command is already included with the other [711]ELDK tools. If you don't use the [712]ELDK then you should install mkimage in some directory that is in your command search PATH, for instance: $ cp tools/mkimage /usr/local/bin/ 5.6. Initialization To initialize the U-Boot firmware running on your canyonlands board, you have to connect a terminal to the board's serial console port. The default configuration of the console port on the canyonlands board uses a baudrate of 115200/8N1 (115200 bps, 8 Bit per character, no parity, 1 stop bit, no handshake). If you are running Linux on your host system we recommend either kermit or cu as terminal emulation programs. Do not use minicom, since this has caused problems for many users, especially for software download over the serial port. For the configuration of your terminal program see section [713]4.1. Serial Console Access Make sure that both hardware and software flow control are disabled. 5.7. Initial Steps In the default configuration, U-Boot operates in an interactive mode which provides a simple command line-oriented user interface using a serial console on port UART1. In the simplest case, this means that U-Boot shows a prompt (default: =>) when it is ready to receive user input. You then type a command, and press enter. U-Boot will try to run the required action(s), and then prompt for another command. To see a list of the available U-Boot commands, you can type help (or simply ?). This will print a list of all commands that are available in your current configuration. [Please note that U-Boot provides a lot of configuration options; not all options are available for all processors and boards, and some options might be simply not selected for your configuration.] => => hel With the command help you can get additional information about most commands: => help tftpboot tftpboot - boot image via network using TFTP protocol Usage: tftpboot [loadAddress] [[hostIPaddr:]bootfilename] => help setenv printenv setenv - set environment variables Usage: setenv name value ... - set environment variable 'name' to 'value ...' setenv name - delete environment variable 'name' printenv - print environment variables Usage: printenv - print values of all environment variables printenv name ... - print value of environment variable 'name' => Most commands can be abbreviated as long as the string remains unambiguous: => help fli tftp flinfo - print FLASH memory information Usage: flinfo - print information for all FLASH memory banks flinfo N - print information for FLASH memory bank # N tftpboot - boot image via network using TFTP protocol Usage: tftpboot [loadAddress] [[hostIPaddr:]bootfilename] => 5.8. The First Power-On HELP Note: If you bought your canyonlands board with U-Boot already installed, you can skip this section since the manufacturer probably has already performed these steps. Connect the port labeled UART1 on your canyonlands board to the designated serial port of your host, start the terminal program, and connect the power supply of your canyonlands board. You should see messages like this: => => => reset U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) CPU: AMCC PowerPC 460EX Rev. B at 1066.667 MHz (PLB=266 OPB=88 EBC=88) Security/Kasumi support Bootstrap Option H - Boot ROM Location I2C (Addr 0x52) Internal PCI arbiter enabled 32 kB I-Cache 32 kB D-Cache Board: Canyonlands - AMCC PPC460EX Evaluation Board, 1*PCIe/1*SATA, Rev. 16 I2C: ready DRAM: 512 MB (ECC not enabled, 533 MHz, CL4) FLASH: 64 MB NAND: 128 MiB PCI: Bus Dev VenId DevId Class Int PCIE1: link is not up. DTT: 1 is 32 C Net: ppc_4xx_eth0, ppc_4xx_eth1 Type run flash_nfs to mount root filesystem over NFS Hit any key to stop autoboot: 0 => You can interrupt the "Count-Down" by pressing any key. If you don't you will probably see some (harmless) error messages because the system has not been initialized yet. ALERT! In some cases you may see a message *** Warning - bad CRC, using default environment This is harmless and will go away as soon as you have initialized and saved the environment variables. At first you have to enter the serial number and the ethernet address of your board. Pay special attention here since these parameters are write protected and cannot be changed once saved (usually this is done by the manufacturer of the board). To enter the data you have to use the U-Boot command setenv, followed by the variable name and the data, all separated by white space (blank and/or TAB characters). Use the variable name serial# for the board ID and/or serial number, and ethaddr for the ethernet address, for instance: => setenv ethaddr !!!!!!FILL_THIS!!!!!! => setenv serial# 86BA-5AA1-BD9 Use the printenv command to verify that you have entered the correct values: => printenv serial# ethaddr ## Error: "serial#" not defined ethaddr=00:10:ec:01:08:84 => Please double check that the printed values are correct! You will not be able to correct any errors later! If there is something wrong, reset the board and restart from the beginning; otherwise you can store the parameters permanently using the saveenv command: => saveenv Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => 5.9. U-Boot Command Line Interface The following section describes the most important commands available in U-Boot. Please note that U-Boot is highly configurable, so not all of these commands may be available in the configuration of U-Boot installed on your hardware, or additional commands may exist. You can use the help command to print a list of all available commands for your configuration. For most commands, you do not need to type in the full command name; instead it is sufficient to type a few characters. For instance, help can be abbreviated as h. ALERT! The behaviour of some commands depends on the configuration of U-Boot and on the definition of some variables in your U-Boot environment. ALERT! Almost all U-Boot commands expect numbers to be entered in hexadecimal input format. (Exception: for historical reasons, the sleep command takes it's argument in decimal input format.) ALERT! Be careful not to use edit keys besides 'Backspace', as hidden characters in things like environment variables can be very difficult to find. 5.9.1. Information Commands 5.9.1.1. bdinfo - print Board Info structure => help bdinfo bdinfo - print Board Info structure Usage: bdinfo => The bdinfo command (short: bdi) prints the information that U-Boot passes about the board such as memory addresses and sizes, clock frequencies, [714]MAC address, etc. This information is mainly needed to be passed to the Linux kernel. => bdi memstart = 0x00000000 memsize = 0x20000000 flashstart = 0xFC000000 flashsize = 0x04000000 flashoffset = 0x00000000 sramstart = 0x00000000 sramsize = 0x00000000 bootflags = 0xFFFE6530 intfreq = 1066.667 MHz busfreq = 266.667 MHz ethaddr = 00:10:ec:01:08:84 eth1addr = 00:10:ec:81:08:84 IP addr = 192.168.100.6 baudrate = 115200 bps => 5.9.1.2. coninfo - print console devices and informations => help conin coninfo - print console devices and information Usage: coninfo => The coninfo command (short: conin) displays information about the available console I/O devices. => conin List of available devices: serial 80000003 SIO stdin stdout stderr serial1 00000003 .IO serial0 00000003 .IO nc 80000003 SIO => The output contains the device name, flags, and the current usage. For example, the output serial 80000003 SIO stdin stdout stderr means that the serial device is a system device (flag 'S') which provides input (flag 'I') and output (flag 'O') functionality and is currently assigned to the 3 standard I/O streams stdin, stdout and stderr. 5.9.1.3. flinfo - print FLASH memory information => help flinfo flinfo - print FLASH memory information Usage: flinfo - print information for all FLASH memory banks flinfo N - print information for FLASH memory bank # N => The command flinfo (short: fli) can be used to get information about the available flash memory (see Flash Memory Commands below). => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 FC020000 FC040000 FC060000 FC080000 FC0A0000 FC0C0000 FC0E0000 FC100000 FC120000 FC140000 FC160000 FC180000 FC1A0000 FC1C0000 E FC1E0000 FC200000 FC220000 FC240000 FC260000 FC280000 FC2A0000 FC2C0000 FC2E0000 FC300000 FC320000 FC340000 FC360000 FC380000 FC3A0000 FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E FF920000 E FF940000 E FF960000 E FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => 5.9.1.4. iminfo - print header information for application image => help iminfo iminfo - print header information for application image Usage: iminfo addr [addr ...] - print header information for application image starting at address 'addr' in memory; this includes verification of the image contents (magic number, header and payload checksums) => iminfo (short: imi) is used to print the header information for images like Linux kernels or ramdisks. It prints (among other information) the image name, type and size and verifies that the CRC32 checksums stored within the image are OK. => tftp ${ram_ws} ${bootfile} Waiting for PHY auto negotiation to complete.... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0) Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/uImage'. Load address: 0x100000 Loading: T ################################################################# ################################################################# #### done Bytes transferred = 1958609 (1de2d1 hex) => imi ${ram_ws} ## Checking Image at 00100000 ... Legacy image found Image Name: Linux-2.6.32.7-00007-g08eba26 Created: 2010-02-04 17:54:22 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1958545 Bytes = 1.9 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => HELP Like with many other commands, the exact operation of this command can be controlled by the settings of some U-Boot environment variables (here: the verify variable). See below for details. 5.9.1.5. help - print online help => help help help - print online help Usage: help [command ...] - show help information (for 'command') 'help' prints online help for the monitor commands. Without arguments, it prints a short usage message for all commands. To get detailed help information for specific commands you can type 'help' with one or more command names as arguments. => The help command (short: h or ?) prints online help. Without any arguments, it prints a list of all U-Boot commands that are available in your configuration of U-Boot. You can get detailed information for a specific command by typing its name as argument to the help command: => help protect protect - enable or disable FLASH write protection Usage: protect on start end - protect FLASH from addr 'start' to addr 'end' protect on start +len - protect FLASH from addr 'start' to end of sect w/addr 'start'+'len'-1 protect on N:SF[-SL] - protect sectors SF-SL in FLASH bank # N protect on bank N - protect FLASH bank # N protect on all - protect all FLASH banks protect off start end - make FLASH from addr 'start' to addr 'end' writable protect off start +len - make FLASH from addr 'start' to end of sect w/addr 'start'+'len'-1 wrtabl e protect off N:SF[-SL] - make sectors SF-SL writable in FLASH bank # N protect off bank N - make FLASH bank # N writable protect off all - make all FLASH banks writable => 5.9.2. Memory Commands 5.9.2.1. base - print or set address offset => help base base - print or set address offset Usage: base - print address offset for memory commands base off - set address offset for memory commands to 'off' => You can use the base command (short: ba) to print or set a "base address" that is used as address offset for all memory commands; the default value of the base address is 0, so all addresses you enter are used unmodified. However, when you repeatedly have to access a certain memory region (like the internal memory of some embedded [715]Power Architecture® processors) it can be very convenient to set the base address to the start of this area and then use only the offsets: => base Base Address: 0x00000000 => md 0 0xc 00000000: 00ff43a6 00000000 ffffffff ffffffff ..C............. 00000010: 00ff43a6 00000000 ffffffff ffffffff ..C............. 00000020: 0c904d01 320b4481 1ea3d0a2 c498293a ..M.2.D.......): => base 0x100000 Base Address: 0x00100000 => md 0 0xc 00100000: 0e0a0e81 bd86200a 60a19054 2c12c402 ...... .`..T,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): => 5.9.2.2. crc32 - checksum calculation The crc32 command (short: crc) can be used to caculate a CRC32 checksum over a range of memory: => crc 0x100004 0x3FC CRC32 for 00100004 ... 001003ff ==> 8083764e => When used with 3 arguments, the command stores the calculated checksum at the given address: => crc 0x100004 0x3FC 0x100000 CRC32 for 00100004 ... 001003ff ==> 8083764e => md 0x100000 4 00100000: 8083764e bd86200a 60a19054 2c12c402 ..vN.. .`..T,... => As you can see, the CRC32 checksum was not only printed, but also stored at address 0x100000. 5.9.2.3. cmp - memory compare => help cmp cmp - memory compare Usage: cmp [.b, .w, .l] addr1 addr2 count => With the cmp command you can test of the contents of two memory areas is identical or not. The command will either test the whole area as specified by the 3rd (length) argument, or stop at the first difference. => cmp 0x100000 0x200000 0x400 word at 0x00100000 (0x8083764e) != word at 0x00200000 (0x27051956) Total of 0 words were the same => md 0x100000 0xc 00100000: 8083764e bd86200a 60a19054 2c12c402 ..vN.. .`..T,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): => md 0x200000 0xc 00200000: 27051956 552d426f 6f742032 3030392e '..VU-Boot 2009. 00200010: 31312e31 20284665 62203035 20323031 11.1 (Feb 05 201 00200020: 30202d20 30383a35 373a3132 29000000 0 - 08:57:12)... => Like most memory commands the cmp can access the memory in different sizes: as 32 bit (long word), 16 bit (word) or 8 bit (byte) data. If invoked just as cmp the default size (32 bit or long words) is used; the same can be selected explicitely by typing cmp.l instead. If you want to access memory as 16 bit or word data, you can use the variant cmp.w instead; and to access memory as 8 bit or byte data please use cmp.b. ALERT! Please note that the count argument specifies the number of data items to process, i. e. the number of long words or words or bytes to compare. => cmp.l 0x100000 0x200000 0x400 word at 0x00100000 (0x8083764e) != word at 0x00200000 (0x27051956) Total of 0 words were the same => cmp.w 0x100000 0x200000 0x800 halfword at 0x00100000 (0x8083) != halfword at 0x00200000 (0x2705) Total of 0 halfwords were the same => cmp.b 0x100000 0x200000 0x1000 byte at 0x00100000 (0x80) != byte at 0x00200000 (0x27) Total of 0 bytes were the same => 5.9.2.4. cp - memory copy => help cp cp - memory copy Usage: cp [.b, .w, .l] source target count => help cp cp - memory copy Usage: cp [.b, .w, .l] source target count => The cp is used to copy memory areas. => cp 0x100000 0x200000 0x10000 => The cp understands the type extensions .l, .w and .b : => cp.l 0x200000 0x100000 0x10000 => cp.w 0x200000 0x100000 0x20000 => cp.b 0x200000 0x100000 0x40000 => 5.9.2.5. md - memory display => help md md - memory display Usage: md [.b, .w, .l] address [# of objects] => The md can be used to display memory contents both as hexadecimal and ASCII data. => md 0x100000 00100000: 8083764e bd86200a 60a19054 2c12c402 ..vN.. .`..T,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): => 00100030: 58f5c828 6029e009 d0718131 154b105b X..(`)...q.1.K.[ 00100040: 9019a424 7423a001 e064013c 016a0070 ...$t#...d.<.j.p 00100050: d0809820 12437140 0064e018 424be2a9 ... .Cq@.d..BK.. => This command, too, can be used with the type extensions .l, .w and .b : => => md.w 0x100000 00100000: 8083 764e bd86 200a 60a1 9054 2c12 c402 ..vN.. .`..T,... 00100010: c101 d028 0043 8198 ...(.C.. => md.b 0x10000 The last displayed memory address and the value of the count argument are remembered, so when you enter md again without arguments it will automatically continue at the next address, and use the same count again. => md.b 0x100000 0x20 00100000: 2f 83 00 00 40 9e ff 38 38 60 00 00 4b ff ff 3c /...@..88`..K..< 00100010: 83 5e 00 0c 80 9e 00 08 2b 9a 00 ff 82 9e 00 10 .^......+....... => md.w 0x100000 00100000: 2f83 0000 409e ff38 3860 0000 4bff ff3c /...@..88`..K..< 00100010: 835e 000c 809e 0008 2b9a 00ff 829e 0010 .^......+....... 00100020: 82be 0014 7f45 d378 409d 000c 3b40 00ff .....E.x@...;@.. 00100030: 38a0 00ff 2b95 00ff 409d 0008 3aa0 00ff 8...+...@...:... => md 0x100000 00100000: 2f830000 409eff38 38600000 4bffff3c /...@..88`..K..< 00100010: 835e000c 809e0008 2b9a00ff 829e0010 .^......+....... 00100020: 82be0014 7f45d378 409d000c 3b4000ff .....E.x@...;@.. 00100030: 38a000ff 2b9500ff 409d0008 3aa000ff 8...+...@...:... 00100040: 8002021c 3bfb000a 7f9f0040 419d002c ....;......@A.., 00100050: 2f9a0000 419e0014 7c1f0050 3925ffff /...A...|..P9%.. 00100060: 7f890040 419d0014 7fe3fb78 4bf1401d ...@A......xK.@. 00100070: 7c651b78 48000014 3c00bfff 6000ffff |e.xH...<...`... => 5.9.2.6. mm - memory modify (auto-incrementing) => help mm mm - memory modify (auto-incrementing address) Usage: mm [.b, .w, .l] address => The mm is a method to interactively modify memory contents. It will display the address and current contents and then prompt for user input. If you enter a legal hexadecimal number, this new value will be written to the address. Then the next address will be prompted. If you don't enter any value and just press ENTER, then the contents of this address will remain unchanged. The command stops as soon as you enter any data that is not a hex number (like .): => => mm 0x100000 00100000: 8083764e ? 0 00100004: bd86200a ? 0xaabbccdd 00100008: 60a19054 ? 0x01234567 0010000c: 2c12c402 ? . => md 0x100000 0x10 00100000: 00000000 aabbccdd 01234567 2c12c402 .........#Eg,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): 00100030: 58f5c828 6029e009 d0718131 154b105b X..(`)...q.1.K.[ => Again this command can be used with the type extensions .l, .w and .b : => => mm.w 0x100000 00100000: 0000 ? 0x0101 00100002: 0000 ? 0x0202 00100004: aabb ? 0x4321 00100006: ccdd ? 0x8765 00100008: 0123 ? . => md 0x100000 0x10 00100000: 01010202 43218765 01234567 2c12c402 ....C!.e.#Eg,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): 00100030: 58f5c828 6029e009 d0718131 154b105b X..(`)...q.1.K.[ => => => mm.b 0x100000 00100000: 01 ? 0x48 00100001: 01 ? 0x65 00100002: 02 ? 0x6c 00100003: 02 ? 0x6c 00100004: 43 ? 0x6f 00100005: 21 ? 0x20 00100006: 87 ? 0x20 00100007: 65 ? 0x20 00100008: 01 ? . => md 0x100000 0x10 00100000: 48656c6c 6f202020 01234567 2c12c402 Hello .#Eg,... 00100010: c101d028 00438198 7ab01239 62406128 ...(.C..z..9b@a( 00100020: 0c900d05 320b4581 1ca3d0a2 c498293a ....2.E.......): 00100030: 58f5c828 6029e009 d0718131 154b105b X..(`)...q.1.K.[ => 5.9.2.7. mtest - simple RAM test => help mtest mtest - simple RAM read/write test Usage: mtest [start [end [pattern [iterations]]]] => The mtest provides a simple memory test. => => mtest 0x100000 0x200000 Pattern 00000000 Writing... Reading...Pattern FFFFFFFF Writing... Reading.. .Pattern 00000001 Writing... Reading...Pattern FFFFFFFE Writing... Reading. ..Pattern 00000002 Writing... Reading...Pattern FFFFFFFD Writing... Reading ...Pattern 00000003 Writing... Reading...Pattern FFFFFFFC Writing... Readin g...Pattern 00000004 Writing... Reading...Pattern FFFFFFFB Writing... Readi ng...Pattern 00000005 Writing... Reading...Pattern FFFFFFFA Writing... Read ing...Pattern 00000006 Writing... Reading...Pattern FFFFFFF9 Writing... Rea ding...Pattern 00000007 Writing... Reading...Pattern FFFFFFF8 Writing... Re ading...Pattern 00000008 Writing... Reading...Pattern FFFFFFF7 Writing... R eading...Pattern 00000009 Writing... Reading...Pattern FFFFFFF6 Writing... Reading...Pattern 0000000A Writing... Reading...Pattern FFFFFFF5 Writing... Reading...Pattern 0000000B Writing... Reading...Pattern FFFFFFF4 Writing... Reading...Pattern 0000000C Writing... Reading...Pattern FFFFFFF3 Writing.. . Reading...Pattern 0000000D Writing... Reading...Pattern FFFFFFF2 Writing. .. Reading...Pattern 0000000E Writing... Reading...Pattern FFFFFFF1 Writing ... Reading...Pattern 0000000F Writing... Reading... => ALERT! This tests writes to memory, thus modifying the memory contents. It will fail when applied to ROM or flash memory. ALERT! This command may crash the system when the tested memory range includes areas that are needed for the operation of the U-Boot firnware (like exception vector code, or U-Boot's internal program code, stack or heap memory areas). 5.9.2.8. mw - memory write (fill) => help mw mw - memory write (fill) Usage: mw [.b, .w, .l] address value [count] => The mw is a way to initialize (fill) memory with some value. When called without a count argument, the value will be written only to the specified address. When used with a count, then a whole memory areas will be initialized with this value: => md 0x100000 0x10 00100000: 0000000f 00000010 00000011 00000012 ................ 00100010: 00000013 00000014 00000015 00000016 ................ 00100020: 00000017 00000018 00000019 0000001a ................ 00100030: 0000001b 0000001c 0000001d 0000001e ................ => mw 0x100000 0xaabbccdd => md 0x100000 0x10 00100000: aabbccdd 00000010 00000011 00000012 ................ 00100010: 00000013 00000014 00000015 00000016 ................ 00100020: 00000017 00000018 00000019 0000001a ................ 00100030: 0000001b 0000001c 0000001d 0000001e ................ => mw 0x100000 0 6 => md 0x100000 0x10 00100000: 00000000 00000000 00000000 00000000 ................ 00100010: 00000000 00000000 00000015 00000016 ................ 00100020: 00000017 00000018 00000019 0000001a ................ 00100030: 0000001b 0000001c 0000001d 0000001e ................ => This is another command that accepts the type extensions .l, .w and .b : => mw.w 0x100004 0x1155 6 => md 0x100000 0x10 00100000: 00000000 11551155 11551155 11551155 .....U.U.U.U.U.U 00100010: 00000000 00000000 00000015 00000016 ................ 00100020: 00000017 00000018 00000019 0000001a ................ 00100030: 0000001b 0000001c 0000001d 0000001e ................ => mw.b 0x100007 0xff 7 => md 0x100000 0x10 00100000: 00000000 115511ff ffffffff ffff1155 .....U.........U 00100010: 00000000 00000000 00000015 00000016 ................ 00100020: 00000017 00000018 00000019 0000001a ................ 00100030: 0000001b 0000001c 0000001d 0000001e ................ => 5.9.2.9. nm - memory modify (constant address) => help nm nm - memory modify (constant address) Usage: nm [.b, .w, .l] address => The nm command (non-incrementing memory modify) can be used to interactively write different data several times to the same address. This can be useful for instance to access and modify device registers: => => nm.b 0x100000 00100000: 00 ? 0x48 00100000: 48 ? 0x65 00100000: 65 ? 0x6c 00100000: 6c ? 0x6c 00100000: 6c ? 0x6f 00100000: 6f ? . => md 0x100000 8 00100000: 6f000000 115511ff ffffffff ffff1155 o....U.........U 00100010: 00000000 00000000 00000015 00000016 ................ => The nm command too accepts the type extensions .l, .w and .b. 5.9.2.10. loop - infinite loop on address range => help loop loop - infinite loop on address range Usage: loop [.b, .w, .l] address number_of_objects => The loop command reads in a tight loop from a range of memory. This is intended as a special form of a memory test, since this command tries to read the memory as fast as possible. ALERT! This command will never terminate. There is no way to stop it but to reset the board! => loop 100000 8 5.9.3. Flash Memory Commands 5.9.3.1. cp - memory copy => help cp cp - memory copy Usage: cp [.b, .w, .l] source target count => help cp cp - memory copy Usage: cp [.b, .w, .l] source target count => The cp command "knows" about flash memory areas and will automatically invoke the necessary flash programming algorithm when the target area is in flash memory. => cp.b 0x100000 0xFF900000 0x40000 Copy to Flash... done => ALERT! Writing to flash memory may fail when the target area has not been erased (see erase below), or if it is write-protected (see protect below). => cp.b 0x100000 0xFF900000 0x40000 Copy to Flash... Can't write to protected Flash sectors => ALERT! Remember that the count argument specifies the number of items to copy. If you have a "length" instead (= byte count) you should use cp.b or you will have to calculate the correct number of items. 5.9.3.2. flinfo - print FLASH memory information The command flinfo (short: fli) can be used to get information about the available flash memory. The number of flash banks is printed with information about the size and organization into flash "sectors" or erase units. For all sectors the start addresses are printed; write-protected sectors are marked as read-only (RO). Some configurations of U-Boot also mark empty sectors with an (E). => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 FC020000 FC040000 FC060000 FC080000 FC0A0000 FC0C0000 FC0E0000 FC100000 FC120000 FC140000 FC160000 FC180000 FC1A0000 FC1C0000 E FC1E0000 FC200000 FC220000 FC240000 FC260000 FC280000 FC2A0000 FC2C0000 FC2E0000 FC300000 FC320000 FC340000 FC360000 FC380000 FC3A0000 FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E FF920000 E FF940000 E FF960000 E FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => 5.9.3.3. erase - erase FLASH memory => help era erase - erase FLASH memory Usage: erase start end - erase FLASH from addr 'start' to addr 'end' erase start +len - erase FLASH from addr 'start' to the end of sect w/addr 'start'+'len'-1 erase N:SF[-SL] - erase sectors SF-SL in FLASH bank # N erase bank N - erase FLASH bank # N erase all - erase all FLASH banks => The erase command (short: era) is used to erase the contents of one or more sectors of the flash memory. It is one of the more complex commands; the help output shows this. Probably the most frequent usage of this command is to pass the start and end addresses of the area to be erased: => era 0xFF900000 0xFF95FFFF ... done Erased 3 sectors => ALERT! Note that both the start and end addresses for this command must point exactly at the start resp. end addresses of flash sectors. Otherwise the command will not be executed. Another way to select certain areas of the flash memory for the erase command uses the notation of flash banks and sectors: Technically speaking, a bank is an area of memory implemented by one or more memory chips that are connected to the same chip select signal of the [716]CPU, and a flash sector or erase unit is the smallest area that can be erased in one operation. For practical purposes it is sufficient to remember that with flash memory a bank is something that eventually may be erased as a whole in a single operation. This may be more efficient (faster) than erasing the same area sector by sector. [It depends on the actual type of flash chips used on the board if such a fast bank erase algorithm exists, and on the implementation of the flash device driver if is actually used.] In U-Boot, flash banks are numbered starting with 1, while flash sectors start with 0. To erase the same flash area as specified using start and end addresses in the example above you could also type: => era 1:455-456 Erase Flash Sectors 455-456 in Bank # 1 .. done => To erase a whole bank of flash memory you can use a command like this one: => era bank 1 Erase Flash Bank # 1 - Warning: 5 protected sectors will not be erased! ............................................................................... ............................................................................... ............................................................................... ............................................................................... ............................................................................... ............................................................................... ................................. done => ALERT! Note that a warning message is printed because some write protected sectors exist in this flash bank which were not erased. With the command: => era all Erase Flash Bank # 1 - Warning: 5 protected sectors will not be erased! ............................................................................... ............................................................................... ............................................................................... ............................................................................... ............................................................................... ............................................................................... ................................. done => the whole flash memory (except for the write-protected sectors) can be erased. 5.9.3.4. protect - enable or disable FLASH write protection => help protect protect - enable or disable FLASH write protection Usage: protect on start end - protect FLASH from addr 'start' to addr 'end' protect on start +len - protect FLASH from addr 'start' to end of sect w/addr 'start'+'len'-1 protect on N:SF[-SL] - protect sectors SF-SL in FLASH bank # N protect on bank N - protect FLASH bank # N protect on all - protect all FLASH banks protect off start end - make FLASH from addr 'start' to addr 'end' writable protect off start +len - make FLASH from addr 'start' to end of sect w/addr 'start'+'len'-1 wrtabl e protect off N:SF[-SL] - make sectors SF-SL writable in FLASH bank # N protect off bank N - make FLASH bank # N writable protect off all - make all FLASH banks writable => The protect command is another complex one. It is used to set certain parts of the flash memory to read-only mode or to make them writable again. Flash memory that is "protected" (= read-only) cannot be written (with the cp command) or erased (with the erase command). Protected areas are marked as (RO) (for "read-only") in the output of the flinfo command: => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 E FC020000 E FC040000 E FC060000 E FC080000 E FC0A0000 E FC0C0000 E FC0E0000 E FC100000 E FC120000 E FC140000 E FC160000 E FC180000 E FC1A0000 E FC1C0000 E FC1E0000 E FC200000 E FC220000 E FC240000 E FC260000 E FC280000 E FC2A0000 E FC2C0000 E FC2E0000 E FC300000 E FC320000 E FC340000 E FC360000 E FC380000 E FC3A0000 E FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 E FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E FF920000 E FF940000 E FF960000 E FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => prot on 0xFF900000 0xFF97FFFF Protected 4 sectors => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 E FC020000 E FC040000 E FC060000 E FC080000 E FC0A0000 E FC0C0000 E FC0E0000 E FC100000 E FC120000 E FC140000 E FC160000 E FC180000 E FC1A0000 E FC1C0000 E FC1E0000 E FC200000 E FC220000 E FC240000 E FC260000 E FC280000 E FC2A0000 E FC2C0000 E FC2E0000 E FC300000 E FC320000 E FC340000 E FC360000 E FC380000 E FC3A0000 E FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 E FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E RO FF920000 E RO FF940000 E RO FF960000 E RO FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => era 0xFF900000 0xFF97FFFF - Warning: 4 protected sectors will not be erased! done Erased 4 sectors => prot off 1:455 Un-Protect Flash Sectors 455-455 in Bank # 1 => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 E FC020000 E FC040000 E FC060000 E FC080000 E FC0A0000 E FC0C0000 E FC0E0000 E FC100000 E FC120000 E FC140000 E FC160000 E FC180000 E FC1A0000 E FC1C0000 E FC1E0000 E FC200000 E FC220000 E FC240000 E FC260000 E FC280000 E FC2A0000 E FC2C0000 E FC2E0000 E FC300000 E FC320000 E FC340000 E FC360000 E FC380000 E FC3A0000 E FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 E FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E RO FF920000 E RO FF940000 E RO FF960000 E RO FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => era 1:455 Erase Flash Sectors 455-455 in Bank # 1 . done => ALERT! The actual level of protection depends on the flash chips used on your hardware, and on the implementation of the flash device driver for this board. In most cases U-Boot provides just a simple software-protection, i. e. it prevents you from erasing or overwriting important stuff by accident (like the U-Boot code itself or U-Boot's environment variables), but it cannot prevent you from circumventing these restrictions - a nasty user who is loading and running his own flash driver code cannot and will not be stopped by this mechanism. Also, in most cases this protection is only effective while running U-Boot, i. e. any operating system will not know about "protected" flash areas and will happily erase these if requested to do so. 5.9.3.5. mtdparts - define a Linux compatible [717]MTD partition scheme U-Boot implements two different approaches to define a [718]MTD partition scheme that can be shared easily with the linux kernel. The first one is to define a single, static partition in your board config file, for example: #undef CONFIG_JFFS2_CMDLINE #define CONFIG_JFFS2_DEV "nor0" #define CONFIG_JFFS2_PART_SIZE 0xFFFFFFFF /* use whole device */ #define CONFIG_JFFS2_PART_SIZE 0x00100000 /* use 1MB */ #define CONFIG_JFFS2_PART_OFFSET 0x00000000 The second method uses the Linux kernel's mtdparts command line option and dynamic partitioning: #define CONFIG_JFFS2_CMDLINE #define MTDIDS_DEFAULT "nor1=zuma-1,nor2=zuma-2" #define MTDPARTS_DEFAULT "mtdparts=zuma-1:-(jffs2),zuma-2:-(user)" Command line of course produces bigger images, and may be inappropriate for some targets, so by default it's off. The mtdparts command offers an easy to use and powerful interface to define the contents of the environment variable of the same name that can be passed as boot argument to the Linux kernel: => help mtdparts mtdparts - list partition table mtdparts delall - delete all partitions mtdparts del part-id - delete partition (e.g. part-id = nand0,1) mtdparts add [@] [] [ro] - add partition mtdparts default - reset partition table to defaults ----- this command uses three environment variables: 'partition' - keeps current partition identifier partition := := ,part_num 'mtdids' - linux kernel mtd device id <-> u-boot device id mapping mtdids=[,,...] := = := 'nand'|'nor' := mtd device number, 0... := unique device tag used by linux kernel to find mtd device (mtd->n ame) 'mtdparts' - partition list mtdparts=mtdparts=[;...] := :[,...] := unique device tag used by linux kernel to find mtd device (mtd->n ame) := [@][][] := standard linux memsize OR '-' to denote all remaining space := partition start offset within the device := '(' NAME ')' := when set to 'ro' makes partition read-only (not used, passed to k ernel) For example, on some target system the mtdparts command might display this information: => mtdparts device nor0 , # parts = 4 #: name size offset mask_flags 0: firmware 0x00100000 0x00000000 1 1: kernel 0x00180000 0x00100000 0 2: small-fs 0x00d80000 0x00280000 0 3: big-fs 0x01000000 0x01000000 0 active partition: nor0,0 - (firmware) 0x00100000 @ 0x00000000 defaults: mtdids : nor0=TQM5200-0 mtdparts: mtdparts=TQM5200-0:1m(firmware),1536k(kernel),3584k(small-fs),2m(init rd),8m(misc),16m(big-fs) The partition table printed here obviously differs from the default value for the mtdparts variable printed in the last line. To verify this, we can check the current content of this variable: => print mtdparts mtdparts=mtdparts=TQM5200-0:1024k(firmware)ro,1536k(kernel),13824k(small-fs),16 m(big-fs) and we can see that it exactly matches the partition table printed above. Now let's switch back to the default settings: => mtdparts default => mtdparts device nor0 , # parts = 6 #: name size offset mask_flags 0: firmware 0x00100000 0x00000000 0 1: kernel 0x00180000 0x00100000 0 2: small-fs 0x00380000 0x00280000 0 3: initrd 0x00200000 0x00600000 0 4: misc 0x00800000 0x00800000 0 5: big-fs 0x01000000 0x01000000 0 active partition: nor0,0 - (firmware) 0x00100000 @ 0x00000000 defaults: mtdids : nor0=TQM5200-0 mtdparts: mtdparts=TQM5200-0:1m(firmware),1536k(kernel),3584k(small-fs),2m(init rd),8m(misc),16m(big-fs) => print mtdparts mtdparts=mtdparts=TQM5200-0:1m(firmware),1536k(kernel),3584k(small-fs),2m(initr d),8m(misc),16m(big-fs) Then we delete the last 4 partitions ("small-fs", "initrd", "misc" and "big-fs") ... => mtdparts del small-fs => mtdparts del initrd => mtdparts del misc => mtdparts del big-fs => mtdparts device nor0 , # parts = 2 #: name size offset mask_flags 0: firmware 0x00100000 0x00000000 0 1: kernel 0x00180000 0x00100000 0 active partition: nor0,0 - (firmware) 0x00100000 @ 0x00000000 defaults: mtdids : nor0=TQM5200-0 mtdparts: mtdparts=TQM5200-0:1m(firmware),1536k(kernel),3584k(small-fs),2m(init rd),8m(misc),16m(big-fs) ... and combine the free space into a singe big partition: => mtdparts add nor0 - new-part => mtdparts device nor0 , # parts = 3 #: name size offset mask_flags 0: firmware 0x00100000 0x00000000 0 1: kernel 0x00180000 0x00100000 0 2: new-part 0x01d80000 0x00280000 0 active partition: nor0,0 - (firmware) 0x00100000 @ 0x00000000 defaults: mtdids : nor0=TQM5200-0 mtdparts: mtdparts=TQM5200-0:1m(firmware),1536k(kernel),3584k(small-fs),2m(init rd),8m(misc),16m(big-fs) => print mtdparts mtdparts=mtdparts=TQM5200-0:1m(firmware),1536k(kernel),30208k(new-part) 5.9.4. Execution Control Commands 5.9.4.1. source - run script from memory => help source source - run script from memory Usage: source [addr] - run script starting at addr - A valid image header must be present => With the source command you can run "shell" scripts under U-Boot: You create a U-Boot script image by simply writing the commands you want to run into a text file; then you will have to use the mkimage tool to convert this text file into a U-Boot image (using the image type script). This image can be loaded like any other image file, and with source you can run the commands in such an image. For instance, the following text file: echo echo Network Configuration: echo ---------------------- echo Target: printenv ipaddr hostname echo echo Server: printenv serverip rootpath echo can be converted into a U-Boot script image using the mkimage command like this: bash$ mkimage -A ppc -O linux -T script -C none -a 0 -e 0 \ > -n "autoscr example script" \ > -d ./testsystems/dulg/testcases/example.script /tftpboot/duts/canyonlands/exa mple.scr Image Name: autoscr example script Created: Mon Feb 8 16:36:04 2010 Image Type: PowerPC Linux Script (uncompressed) Data Size: 157 Bytes = 0.15 kB = 0.00 MB Load Address: 0x00000000 Entry Point: 0x00000000 Contents: Image 0: 149 Bytes = 0 kB = 0 MB Now you can load and execute this script image in U-Boot: => tftp 0x100000 /tftpboot/duts/canyonlands/example.scr Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/example.scr'. Load address: 0x100000 Loading: # done Bytes transferred = 221 (dd hex) => imi ## Checking Image at 00100000 ... Legacy image found Image Name: autoscr example script Created: 2010-02-08 15:36:04 UTC Image Type: PowerPC Linux Script (uncompressed) Data Size: 157 Bytes = 0.2 kB Load Address: 00000000 Entry Point: 00000000 Contents: Image 0: 149 Bytes = 0.1 kB Verifying Checksum ... OK => source 0x100000 ## Executing script at 00100000 Network Configuration: ---------------------- Target: ipaddr=192.168.100.6 hostname=canyonlands Server: serverip=192.168.1.1 rootpath=/opt/eldk/ppc_4xxFP => 5.9.4.2. bootm - boot application image from memory => help bootm bootm - boot application image from memory Usage: bootm [addr [arg ...]] - boot application image stored in memory passing arguments 'arg ...'; when booting a Linux kernel, 'arg' can be the address of an initrd image When booting a Linux kernel which requires a flat device-tree a third argument is required which is the address of the device-tree blob. To boot that kernel without an initrd image, use a '-' for the second argument. If you do not pass a third a bd_info struct will be passed instead Sub-commands to do part of the bootm sequence. The sub-commands must be issued in the order below (it's ok to not issue all sub-commands): start [addr [arg ...]] loados - load OS image ramdisk - relocate initrd, set env initrd_start/initrd_end fdt - relocate flat device tree cmdline - OS specific command line processing/setup bdt - OS specific bd_t processing prep - OS specific prep before relocation or go go - start OS => The bootm command is used to start operating system images. From the image header it gets information about the type of the operating system, the file compression method used (if any), the load and entry point addresses, etc. The command will then load the image to the required memory address, uncompressing it on the fly if necessary. Depending on the OS it will pass the required boot arguments and start the OS at it's entry point. The first argument to bootm is the memory address (in RAM, ROM or flash memory) where the image is stored, followed by optional arguments that depend on the OS. Linux requires the flattened device tree blob to be passed at boot time, and bootm expects its third argument to be the address of the blob in memory. Second argument to bootm depens on whether an initrd initial ramdisk image is to be used. If the kernel should be booted without the initial ramdisk, the second argument should be given as "-", otherwise it is interpreted as the start address of initrd (in RAM, ROM or flash memory). To boot a Linux kernel image without a initrd ramdisk image, the following command can be used: => bootm ${kernel_addr} - ${fdt_addr} If a ramdisk image shall be used, you can type: => bootm ${kernel_addr} ${ramdisk_addr} ${fdt_addr} Both examples of course imply that the variables used are set to correct addresses for a kernel, fdt blob and a initrd ramdisk image. ALERT! When booting images that have been loaded to RAM (for instance using [719]TFTP download) you have to be careful that the locations where the (compressed) images were stored do not overlap with the memory needed to load the uncompressed kernel. For instance, if you load a ramdisk image at a location in low memory, it may be overwritten when the Linux kernel gets loaded. This will cause undefined system crashes. 5.9.4.3. go - start application at address 'addr' => help go go - start application at address 'addr' Usage: go addr [arg ...] - start application at address 'addr' passing 'arg' as arguments => U-Boot has support for so-called standalone applications. These are programs that do not require the complex environment of an operating system to run. Instead they can be loaded and executed by U-Boot directly, utilizing U-Boot's service functions like console I/O or malloc() and free(). This can be used to dynamically load and run special extensions to U-Boot like special hardware test routines or bootstrap code to load an OS image from some filesystem. The go command is used to start such standalone applications. The optional arguments are passed to the application without modification. For more informatoin see [720]5.12. U-Boot Standalone Applications. 5.9.5. Download Commands 5.9.5.1. bootp - boot image via network using [721]BOOTP/TFTP protocol => help bootp bootp - boot image via network using BOOTP/TFTP protocol Usage: bootp [loadAddress] [[hostIPaddr:]bootfilename] => 5.9.5.2. dhcp - invoke [722]DHCP client to obtain IP/boot params => help dhcp dhcp - boot image via network using DHCP/TFTP protocol Usage: dhcp [loadAddress] [[hostIPaddr:]bootfilename] => 5.9.5.3. loadb - load binary file over serial line (kermit mode) => help loadb loadb - load binary file over serial line (kermit mode) Usage: loadb [ off ] [ baud ] - load binary file over serial line with offset 'off' and baudrate 'baud' => With kermit you can download binary data via the serial line. Here we show how to download uImage, the Linux kernel image. Please make sure, that you have set up kermit as described in section [723]4.3. Configuring the "kermit" command and then type: => loadb 100000 ## Ready for binary (kermit) download ... Ctrl-\c (Back at denx.denx.de) ---------------------------------------------------- C-Kermit 7.0.197, 8 Feb 2000, for Linux Copyright (C) 1985, 2000, Trustees of Columbia University in the City of New York. Type ? or HELP for help. Kermit> send /bin /tftpboot/pImage ... Kermit> connect Connecting to /dev/ttyS0, speed 115200. The escape character is Ctrl-\ (ASCII 28, FS) Type the escape character followed by C to get back, or followed by ? to see other options. ---------------------------------------------------- = 550260 Bytes ## Start Addr = 0x00100000 => iminfo 100000 ## Checking Image at 00100000 ... Image Name: Linux-2.4.4 Created: 2002-07-02 22:10:11 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 550196 Bytes = 537 kB = 0 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK 5.9.5.4. loads - load S-Record file over serial line => help loads loads - load S-Record file over serial line Usage: loads [ off ] [ baud ] - load S-Record file over serial line with offset 'off' and baudrate 'baud' => 5.9.5.5. rarpboot- boot image via network using RARP/TFTP protocol => help rarp rarpboot - boot image via network using RARP/TFTP protocol Usage: rarpboot [loadAddress] [[hostIPaddr:]bootfilename] => 5.9.5.6. tftpboot- boot image via network using [724]TFTP protocol => help tftp tftpboot - boot image via network using TFTP protocol Usage: tftpboot [loadAddress] [[hostIPaddr:]bootfilename] => 5.9.6. Environment Variables Commands 5.9.6.1. printenv- print environment variables => help printenv printenv - print environment variables Usage: printenv - print values of all environment variables printenv name ... - print value of environment variable 'name' => The printenv command prints one, several or all variables of the U-Boot environment. When arguments are given, these are interpreted as the names of environment variables which will be printed with their values: => printenv ipaddr hostname netmask ipaddr=192.168.100.6 hostname=canyonlands netmask=255.255.0.0 => Without arguments, printenv prints all a list with all variables in the environment and their values, plus some statistics about the current usage and the total size of the memory available for the environment. => printenv bootdelay=5 baudrate=115200 loads_echo= preboot=echo;echo Type "run flash_nfs" to mount root filesystem over NFS;echo netdev=eth0 nfsargs=setenv bootargs root=/dev/nfs rw nfsroot=${serverip}:${rootpath} ramargs=setenv bootargs root=/dev/ram rw addip=setenv bootargs ${bootargs} ip=${ipaddr}:${serverip}:${gatewayip}:${netma sk}:${hostname}:${netdev}:off panic=1 addtty=setenv bootargs ${bootargs} console=ttyS0,${baudrate} addmisc=setenv bootargs ${bootargs} initrd_high=30000000 kernel_addr_r=400000 ramdisk_addr_r=C00000 hostname=canyonlands ramdisk_file=canyonlands/uRamdisk rootpath=/opt/eldk/ppc_4xxFP flash_self=run ramargs addip addtty addmisc;bootm ${kernel_addr} ${ramdisk_addr } ${fdt_addr} flash_nfs=run nfsargs addip addtty addmisc;bootm ${kernel_addr} - ${fdt_addr} net_nfs=tftp ${kernel_addr_r} ${bootfile}; tftp ${fdt_addr_r} ${fdt_file}; run nfsargs addip addtty addmisc;bootm ${kernel_addr_r} - ${fdt_addr_r} net_self_load=tftp ${kernel_addr_r} ${bootfile};tftp ${fdt_addr_r} ${fdt_file}; tftp ${ramdisk_addr_r} ${ramdisk_file}; net_self=run net_self_load;run ramargs addip addtty addmisc;bootm ${kernel_addr _r} ${ramdisk_addr_r} ${fdt_addr_r} fdt_file=canyonlands/canyonlands.dtb update=protect off 0xFFFA0000 FFFFFFFF;era 0xFFFA0000 FFFFFFFF;cp.b ${fileaddr} 0xFFFA0000 ${filesize};setenv filesize;saveenv upd=run load update nload=tftp 200000 canyonlands/u-boot-nand.bin nupdate=nand erase 0 100000;nand write 200000 0 100000;setenv filesize;saveenv nupd=run nload nupdate pciconfighost=1 pcie_mode=RP:RP ethaddr=00:10:ec:01:08:84 eth1addr=00:10:ec:81:08:84 hostname=canyonlands sr=tftp 200000 canyonlands/u-boot.bin-sr;protect off 0xFFFA0000 FFFFFFFF;era 0x FFFA0000 FFFFFFFF;cp.b ${fileaddr} 0xFFFA0000 ${filesize};setenv filesize;savee nv srlinux=setenv bootfile canyonlands/uImage-sr;setenv fdt_file canyonlands/canyo nlands.dtb-sr;run net_nfs bootcmd=run srlinux fdtaddr=800000 uboot_file=canyonlands/u-boot.bin-duts load=tftp 200000 ${u-boot} dzu_net_nfs=setenv bootfile dzu/canyonlands/uImage;setenv fdt_file dzu/canyonla nds/canyonlands.dtb;run net_nfs bootargs=root=/dev/ram rw ip=192.168.100.6:192.168.1.1:192.168.1.254:255.255.0. 0:canyonlands:eth0:off panic=1 console=ttyS0,115200 ethact=ppc_4xx_eth0 bootfile=/tftpboot/duts/canyonlands/uImage bar=This is a new example. cons_opts=console=tty0 console=ttyS0,${baudrate} test=echo This is a test;printenv ipaddr;echo Done. test2=echo This is another Test;printenv hostname;echo Done. kernel_addr=0xFC000000 ramdisk_addr=0xFC200000 fdt_addr=0xFC1E0000 fdt_addr_r=0x00b00000 u-boot=/tftpboot/duts/canyonlands/u-boot.bin fileaddr=200000 gatewayip=192.168.1.254 netmask=255.255.0.0 ipaddr=192.168.100.6 serverip=192.168.1.1 stdin=serial stdout=serial stderr=serial ver=U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) Environment size: 2780/16379 bytes => 5.9.6.2. saveenv - save environment variables to persistent storage => help saveenv saveenv - save environment variables to persistent storage Usage: saveenv => All changes you make to the U-Boot environment are made in RAM only. They are lost as soon as you reboot the system. If you want to make your changes permanent you have to use the saveenv command to write a copy of the environment settings to persistent storage, from where they are automatically loaded during startup: => saveenv Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => 5.9.6.3. setenv - set environment variables => help setenv setenv - set environment variables Usage: setenv name value ... - set environment variable 'name' to 'value ...' setenv name - delete environment variable 'name' => To modify the U-Boot environment you have to use the setenv command. When called with exactly one argument, it will delete any variable of that name from U-Boot's environment, if such a variable exists. Any storage occupied for such a variable will be automatically reclaimed: => setenv foo This is an example value. => printenv foo foo=This is an example value. => setenv foo => printenv foo ## Error: "foo" not defined => When called with more arguments, the first one will again be the name of the variable, and all following arguments will (concatenated by single space characters) form the value that gets stored for this variable. New variables will be automatically created, existing ones overwritten. => printenv bar ## Error: "bar" not defined => setenv bar This is a new example. => printenv bar bar=This is a new example. => Remember standard shell quoting rules when the value of a variable shall contain characters that have a special meaning to the command line parser (like the $ character that is used for variable substitution or the semicolon which separates commands). Use the backslash (\) character to escape such special characters, or enclose the whole phrase in apstrophes ('). Use "${name}" for variable expansion (see [725]14.2.17. How the Command Line Parsing Works for details). => setenv cons_opts 'console=tty0 console=ttyS0,${baudrate}' => printenv cons_opts cons_opts=console=tty0 console=ttyS0,${baudrate} => TIP There is no restriction on the characters that can be used in a variable name except the restrictions imposed by the command line parser (like using backslash for quoting, space and tab characters to separate arguments, or semicolon and newline to separate commands). Even strange input like "=-/|()+=" is a perfectly legal variable name in U-Boot. ALERT! A common mistake is to write setenv name=value instead of setenv name value There will be no error message, which lets you believe everything went OK, but it didn't: instead of setting the variable name to the value value you tried to delete a variable with the name name=value - this is probably not what you intended! Always remember that name and value have to be separated by space and/or tab characters! 5.9.6.4. run - run commands in an environment variable => help run run - run commands in an environment variable Usage: run var [...] - run the commands in the environment variable(s) 'var' => You can use U-Boot environment variables to store commands and even sequences of commands. To execute such a command, you use the run command: => setenv test echo This is a test\;printenv ipaddr\;echo Done. => printenv test test=echo This is a test;printenv ipaddr;echo Done. => run test This is a test ipaddr=192.168.100.6 Done. => You can call run with several variables as arguments, in which case these commands will be executed in sequence: => setenv test2 echo This is another Test\;printenv hostname\;echo Done. => printenv test test2 test=echo This is a test;printenv ipaddr;echo Done. test2=echo This is another Test;printenv hostname;echo Done. => run test test2 This is a test ipaddr=192.168.100.6 Done. This is another Test hostname=canyonlands Done. => TIP If a U-Boot variable contains several commands (separated by semicolon), and one of these commands fails when you "run" this variable, the remaining commands will be executed anyway. TIP If you execute several variables with one call to run, any failing command will cause "run" to terminate, i. e. the remaining variables are not executed. 5.9.6.5. bootd - boot default, i.e., run 'bootcmd' => help boot boot - No help available. => The bootd (short: boot) executes the default boot command, i. e. what happens when you don't interrupt the initial countdown. This is a synonym for the run bootcmd command. 5.9.7. Flattened Device Tree support U-Boot is capable of quite comprehensive handling of the flattened device tree blob, implemented by the fdt family of commands: => => help fd 5.9.7.1. fdt addr - select FDT to work on First, the blob that is to be operated on should be stored in memory, and U-Boot has to be informed about its location by the fdt addr command. Once this command has been issued, all subsequent fdt handling commands will use the blob stored at the given address. This address can be changed later on by issuing fdt addr or fdt move command. Here's how to load the blob into memory and tell U-Boot its location: => print fdt_addr_r fdt_addr_r=0x00b00000 => print fdt_file fdt_file=/tftpboot/duts/canyonlands/canyonlands.dtb => tftp ${fdt_addr_r} ${fdt_file} Waiting for PHY auto negotiation to complete... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0) Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/canyonlands.dtb'. Load address: 0xb00000 Loading: T # done Bytes transferred = 10190 (27ce hex) => fdt addr ${fdt_addr_r} => 5.9.7.2. fdt list - print one level Having selected the device tree stored in the blob just loaded, we can inspect its contents. As an FDT usually is quite extensive, it is easier to get information about the structure by looking at selected levels rather than full hierarchies. fdt list allows us to do exactly this. Let's have a look at the hierarchy one level below the cpus node: => fdt list /cpus libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => 5.9.7.3. fdt print - recursive print To print a complete subtree we use fdt print. In comparison to the previous example it is obvious that the whole subtree is printed: => fdt print /cpus libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => 5.9.7.4. fdt mknode - create new nodes fdt mknode can be used to attach a new node to the tree. We will use the fdt list command to verify that the new node has been created and that it is empty: => fdt list / libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt mknode / testnode libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list / libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list /testnode libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => 5.9.7.5. fdt set - set node properties Now, let's create a property at the newly created node; again we'll use fdt list for verification: => fdt set /testnode testprop testvalue libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list /testnode libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => 5.9.7.6. fdt rm - remove nodes or properties The fdt rm command is used to remove nodes and properties. Let's delete the test property created in the previous paragraph and verify the results: => fdt rm /testnode testprop libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list /testnode libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt rm /testnode libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list / libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => 5.9.7.7. fdt move - move FDT blob to new address To move the blob from one memory location to another we will use the fdt move command. Besides moving the blob, it makes the new address the "active" one - similar to fdt addr: => fdt move ${fdt_addr_r} $CFG_RAM_WS_BASE fdt - flattened device tree utility commands Usage: fdt addr [] - Set the fdt location to fdt boardsetup - Do board-specific set up fdt move - Copy the fdt to and make it active fdt resize - Resize fdt to size + padding to 4k addr fdt print [] - Recursive print starting at fdt list [] - Print one level starting at fdt set [] - Set [to ] fdt mknode - Create a new node after fdt rm [] - Delete the node or fdt header - Display header info fdt bootcpu - Set boot cpuid fdt memory - Add/Update memory node fdt rsvmem print - Show current mem reserves fdt rsvmem add - Add a mem reserve fdt rsvmem delete - Delete a mem reserves fdt chosen [ ] - Add/update the /chosen branch in the tree / - initrd start/end addr NOTE: Dereference aliases by omiting the leading '/', e.g. fdt print ethernet0. => fdt list / libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt mknod / foobar libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt list / libfdt fdt_path_offset() returned FDT_ERR_BADMAGIC => fdt addr ${fdt_addr_r} => fdt list / / { #address-cells = <0x2>; #size-cells = <0x1>; model = "amcc,canyonlands"; compatible = "amcc,canyonlands"; dcr-parent = <0x1>; aliases { }; cpus { }; memory { }; interrupt-controller0 { }; interrupt-controller1 { }; interrupt-controller2 { }; interrupt-controller3 { }; sdr { }; cpr { }; l2c { }; plb { }; }; => 5.9.7.8. fdt chosen - fixup dynamic info One of the modifications made by U-Boot to the blob before passing it to the kernel is the addition of the /chosen node. Linux 2.6 Documentation/powerpc/booting-without-of.txt says that this node is used to store "some variable environment information, like the arguments, or the default input/output devices." To force U-Boot to add the /chosen node to the current blob, fdt chosen command can be used. Let's now verify its operation: => fdt list / / { #address-cells = <0x2>; #size-cells = <0x1>; model = "amcc,canyonlands"; compatible = "amcc,canyonlands"; dcr-parent = <0x1>; aliases { }; cpus { }; memory { }; interrupt-controller0 { }; interrupt-controller1 { }; interrupt-controller2 { }; interrupt-controller3 { }; sdr { }; cpr { }; l2c { }; plb { }; }; => fdt chosen => fdt list / / { #address-cells = <0x2>; #size-cells = <0x1>; model = "amcc,canyonlands"; compatible = "amcc,canyonlands"; dcr-parent = <0x1>; chosen { }; aliases { }; cpus { }; memory { }; interrupt-controller0 { }; interrupt-controller1 { }; interrupt-controller2 { }; interrupt-controller3 { }; sdr { }; cpr { }; l2c { }; plb { }; }; => fdt list /chosen chosen { bootargs = "root=/dev/ram rw ip=192.168.100.6:192.168.1.1:192.168.1.254:255. 255.0.0:canyonlands:eth0:off panic=1 console=ttyS0,115200"; }; => Note: fdt boardsetup performs board-specific blob updates, most commonly setting clock frequencies, etc. Discovering its operation is left as an excercise for the reader. 5.9.8. Special Commands 5.9.8.1. i2c - I2C sub-system => help i2c i2c - I2C sub-system Usage: i2c speed [speed] - show or set I2C bus speed i2c md chip address[.0, .1, .2] [# of objects] - read from I2C device i2c mm chip address[.0, .1, .2] - write to I2C device (auto-incrementing) i2c mw chip address[.0, .1, .2] value [count] - write to I2C device (fill) i2c nm chip address[.0, .1, .2] - write to I2C device (constant address) i2c crc32 chip address[.0, .1, .2] count - compute CRC32 checksum i2c probe - show devices on the I2C bus i2c reset - re-init the I2C Controller i2c loop chip address[.0, .1, .2] [# of objects] - looping read of device i2c sdram chip - print SDRAM configuration information => 5.9.9. Storage devices This chapter introduces commands to work with storage devices, i.e. ATA, CF, SATA, SCSI, USB, NAND, etc. connected to the board. 5.9.10. Miscellaneous Commands 5.9.10.1. echo - echo args to console => help echo echo - echo args to console Usage: echo [args..] - echo args to console; \c suppresses newline => The echo command echoes the arguments to the console: => echo The quick brown fox jumped over the lazy dog. The quick brown fox jumped over the lazy dog. => 5.9.10.2. reset - Perform RESET of the [726]CPU => help reset reset - Perform RESET of the CPU Usage: reset => The reset command reboots the system. => => reset U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) CPU: AMCC PowerPC 460EX Rev. B at 1066.667 MHz (PLB=266 OPB=88 EBC=88) Security/Kasumi support Bootstrap Option H - Boot ROM Location I2C (Addr 0x52) Internal PCI arbiter enabled 32 kB I-Cache 32 kB D-Cache Board: Canyonlands - AMCC PPC460EX Evaluation Board, 1*PCIe/1*SATA, Rev. 16 I2C: ready DRAM: 512 MB (ECC not enabled, 533 MHz, CL4) FLASH: 64 MB NAND: 128 MiB PCI: Bus Dev VenId DevId Class Int PCIE1: link is not up. DTT: 1 is 40 C Net: ppc_4xx_eth0, ppc_4xx_eth1 Type run flash_nfs to mount root filesystem over NFS Hit any key to stop autoboot: 0 => 5.9.10.3. sleep - delay execution for some time => help sleep sleep - delay execution for some time Usage: sleep N - delay execution for N seconds (N is _decimal_ !!!) => The sleep command pauses execution for the number of seconds given as the argument: => sleep 5 => 5.9.10.4. version - print monitor version => help version version - print monitor version Usage: version => You can print the version and build date of the U-Boot image running on your system using the version command (short: vers): => version U-Boot 2009.11.1 (Feb 05 2010 - 08:57:12) => 5.9.10.5. ? - alias for 'help' You can use ? as a short form for the help command (see description above). 5.10. U-Boot Environment Variables The U-Boot environment is a block of memory that is kept on persistent storage and copied to RAM when U-Boot starts. It is used to store environment variables which can be used to configure the system. The environment is protected by a CRC32 checksum. This section lists the most important environment variables, some of which have a special meaning to U-Boot. You can use these variables to configure the behaviour of U-Boot to your liking. * autoload: if set to "no" (or any string beginning with 'n'), the rarpb, bootp or dhcp commands will perform only a configuration lookup from the [727]BOOTP / [728]DHCP server, but not try to load any image using [729]TFTP. * autostart: if set to "yes", an image loaded using the rarpb, bootp, dhcp, tftp, disk, or docb commands will be automatically started (by internally calling the bootm command). * baudrate: a decimal number that selects the console baudrate (in bps). Only a predefined list of baudrate settings is available. When you change the baudrate (using the "setenv baudrate ..." command), U-Boot will switch the baudrate of the console terminal and wait for a newline which must be entered with the new speed setting. This is to make sure you can actually type at the new speed. If this fails, you have to reset the board (which will operate at the old speed since you were not able to saveenv the new settings.) If no "baudrate" variable is defined, the default baudrate of 115200 is used. * bootargs: The contents of this variable are passed to the Linux kernel as boot arguments (aka "command line"). * bootcmd: This variable defines a command string that is automatically executed when the initial countdown is not interrupted. This command is only executed when the variable bootdelay is also defined! * bootdelay: After reset, U-Boot will wait this number of seconds before it executes the contents of the bootcmd variable. During this time a countdown is printed, which can be interrupted by pressing any key. Set this variable to 0 boot without delay. Be careful: depending on the contents of your bootcmd variable, this can prevent you from entering interactive commands again forever! Set this variable to -1 to disable autoboot. * bootfile: name of the default image to load with [730]TFTP * cpuclk: (Only with MPC859 / MPC866 / MPC885 processors) On some processors, the [731]CPU clock frequency can be adjusted by the user (for example to optimize performance versus power dissipation). On such systems the cpuclk variable can be set to the desired [732]CPU clock value, in MHz. If the cpuclk variable exists and its value is within the compile-time defined limits (CFG_866_CPUCLK_MIN and CFG_866_CPUCLK_MAX = minimum resp. maximum allowed [733]CPU clock), then the specified value is used. Otherwise, the default [734]CPU clock value is set. * ethaddr: Ethernet [735]MAC address for first/only ethernet interface (= eth0 in Linux). This variable can be set only once (usually during manufacturing of the board). U-Boot refuses to delete or overwrite this variable once it has been set. * eth1addr: Ethernet [736]MAC address for second ethernet interface (= eth1 in Linux). * eth2addr: Ethernet [737]MAC address for third ethernet interface (= eth2 in Linux). ... * initrd_high: used to restrict positioning of initrd ramdisk images: If this variable is not set, initrd images will be copied to the highest possible address in RAM; this is usually what you want since it allows for maximum initrd size. If for some reason you want to make sure that the initrd image is loaded below the CFG_BOOTMAPSZ limit, you can set this environment variable to a value of "no" or "off" or "0". Alternatively, you can set it to a maximum upper address to use (U-Boot will still check that it does not overwrite the U-Boot stack and data). For instance, when you have a system with 16 MB RAM, and want to reserve 4 MB from use by Linux, you can do this by adding "mem=12M" to the value of the "bootargs" variable. However, now you must make sure that the initrd image is placed in the first 12 MB as well - this can be done with => setenv initrd_high 00c00000 Setting initrd_high to the highest possible address in your system (0xFFFFFFFF) prevents U-Boot from copying the image to RAM at all. This allows for faster boot times, but requires a Linux kernel with zero-copy ramdisk support. * ipaddr: IP address; needed for tftp command * loadaddr: Default load address for commands like tftp or loads. * loads_echo: If set to 1, all characters received during a serial download (using the loads command) are echoed back. This might be needed by some terminal emulations (like cu), but may as well just take time on others. * mtdparts: This variable (usually defined using the [738]mtdparts command) allows to share a common [739]MTD partition scheme between U-Boot and the Linux kernel. * pram: If the "Protected RAM" feature is enabled in your board's configuration, this variable can be defined to enable the reservation of such "protected RAM", i. e. RAM which is not overwritten by U-Boot. Define this variable to hold the number of kB you want to reserve for pRAM. Note that the board info structure will still show the full amount of RAM. If pRAM is reserved, a new environment variable "mem" will automatically be defined to hold the amount of remaining RAM in a form that can be passed as boot argument to Linux, for instance like that: => setenv bootargs ${bootargs} mem=\${mem} => saveenv This way you can tell Linux not to use this memory, either, which results in a memory region that will not be affected by reboots. * serverip: [740]TFTP server IP address; needed for tftp command. * serial#: contains hardware identification information such as type string and/or serial number. This variable can be set only once (usually during manufacturing of the board). U-Boot refuses to delete or overwrite this variable once it hass been set. * silent: If the configuration option CONFIG_SILENT_CONSOLE has been enabled for your board, setting this variable to any value will suppress all console messages. Please see doc/README.silent for details. * verify: If set to n or no disables the checksum calculation over the complete image in the bootm command to trade speed for safety in the boot process. Note that the header checksum is still verified. The following environment variables may be used and automatically updated by the network boot commands (bootp, dhcp, or tftp), depending the information provided by your boot server: * bootfile: see above * dnsip: IP address of your Domain Name Server * gatewayip: IP address of the Gateway (Router) to use * hostname: Target hostname * ipaddr: see above * netmask: Subnet Mask * rootpath: Pathname of the root filesystem on the NFS server * serverip: see above * filesize: Size (as hex number in bytes) of the file downloaded using the last bootp, dhcp, or tftp command. 5.11. U-Boot Scripting Capabilities U-Boot allows to store commands or command sequences in a plain text file. Using the mkimage tool you can then convert this file into a script image which can be executed using U-Boot's autoscr command. For example, assume that you will have to run the following sequence of commands on many boards, so you store them in a text file, say "setenv-commands": bash$ cat setenv-commands setenv loadaddr 00200000 echo ===== U-Boot settings ===== setenv u-boot /tftpboot/TQM860L/u-boot.bin setenv u-boot_addr 40000000 setenv load_u-boot 'tftp ${loadaddr} ${u-boot}' setenv install_u-boot 'protect off ${u-boot_addr} +${filesize};era ${u-boot_add r} +${filesize};cp.b ${loadaddr} ${u-boot_addr} ${filesize};saveenv' setenv update_u-boot run load_u-boot install_u-boot echo ===== Linux Kernel settings ===== setenv bootfile /tftpboot/TQM860L/uImage setenv kernel_addr 40040000 setenv load_kernel 'tftp ${loadaddr} ${bootfile};' setenv install_kernel 'era ${kernel_addr} +${filesize};cp.b ${loadaddr} ${kerne l_addr} ${filesize}' setenv update_kernel run load_kernel install_kernel echo ===== Ramdisk settings ===== setenv ramdisk /tftpboot/TQM860L/uRamdisk setenv ramdisk_addr 40100000 setenv load_ramdisk 'tftp ${loadaddr} ${ramdisk};' setenv install_ramdisk 'era ${ramdisk_addr} +${filesize};cp.b ${loadaddr} ${ram disk_addr} ${filesize}' setenv update_ramdisk run load_ramdisk install_ramdisk echo ===== Save new definitions ===== saveenv bash$ To convert the text file into a script image for U-Boot, you have to use the mkimage tool as follows: bash$ mkimage -T script -C none -n 'Demo Script File' -d setenv-commands setenv .img Image Name: Demo Script File Created: Mon Jun 6 13:33:14 2005 Image Type: PowerPC Linux Script (uncompressed) Data Size: 1147 Bytes = 1.12 kB = 0.00 MB Load Address: 0x00000000 Entry Point: 0x00000000 Contents: Image 0: 1139 Bytes = 1 kB = 0 MB bash$ On the target, you can download this image as usual (for example, using the "tftp" command). Use the "autoscr" command to execute it: => tftp 100000 /tftpboot/TQM860L/setenv.img Using FEC ETHERNET device TFTP from server 192.168.3.1; our IP address is 192.168.3.80 Filename '/tftpboot/TQM860L/setenv.img'. Load address: 0x100000 Loading: # done Bytes transferred = 1211 (4bb hex) => imi 100000 ## Checking Image at 00100000 ... Image Name: Demo Script File Created: 2005-06-06 11:33:14 UTC Image Type: PowerPC Linux Script (uncompressed) Data Size: 1147 Bytes = 1.1 kB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => autoscr 100000 ## Executing script at 00100000 ===== U-Boot settings ===== ===== Linux Kernel settings ===== ===== Ramdisk settings ===== ===== Save new definitions ===== Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => TIP Hint: maximum flexibility can be achieved if you are using the Hush shell as command interpreter in U-Boot; see section [741]14.2.17. How the Command Line Parsing Works 5.12. U-Boot Standalone Applications U-Boot supports "standalone" applications, which are loaded dynamically; these applications can have access to the U-Boot console I/O functions, memory allocation and interrupt services. A couple of simple examples are included with the U-Boot source code: 5.12.1. "Hello World" Demo examples/hello_world.c contains a small "Hello World" Demo application; it is automatically compiled when you build U-Boot. It's configured to run at address 0x00040004, so you can play with it like that: => loads ## Ready for S-Record download ... ~>examples/hello_world.srec 1 2 3 4 5 6 7 8 9 10 11 ... [file transfer complete] [connected] ## Start Addr = 0x00040004 => go 40004 Hello World! This is a test. ## Starting application at 0x00040004 ... Hello World argc = 7 argv[0] = "40004" argv[1] = "Hello" argv[2] = "World!" argv[3] = "This" argv[4] = "is" argv[5] = "a" argv[6] = "test." argv[7] = "" Hit any key to exit ... ## Application terminated, rc = 0x0 Alternatively, you can of course use TFTP to download the image over the network. In this case the binary image (hello_world.bin) is used. TIP Note that the entry point of the program is at offset 0x0004 from the start of file, i. e. the download address and the entry point address differ by four bytes. => tftp 40000 /tftpboot/hello_world.bin ... => go 40004 This is another test. ## Starting application at 0x00040004 ... Hello World argc = 5 argv[0] = "40004" argv[1] = "This" argv[2] = "is" argv[3] = "another" argv[4] = "test." argv[5] = "" Hit any key to exit ... ## Application terminated, rc = 0x0 5.12.2. Timer Demo ALERT! This example is only available on MPC8xx [742]CPUs. TIP This example, which demonstrates how to register a [743]CPM interrupt handler with the U-Boot code, can be found in examples/timer.c. Here, a [744]CPM timer is set up to generate an interrupt every second. The interrupt service routine is trivial, just printing a '.' character, but this is just a demo program. The application can be controlled by the following keys: ? - print current values of the CPM Timer registers b - enable interrupts and start timer e - stop timer and disable interrupts q - quit application => loads ## Ready for S-Record download ... ~>examples/timer.srec 1 2 3 4 5 6 7 8 9 10 11 ... [file transfer complete] [connected] ## Start Addr = 0x00040004 => go 40004 ## Starting application at 0x00040004 ... TIMERS=0xfff00980 Using timer 1 tgcr @ 0xfff00980, tmr @ 0xfff00990, trr @ 0xfff00994, tcr @ 0xfff00998, t cn @ 0xfff0099c, ter @ 0xfff009b0 Hit 'b': [q, b, e, ?] Set interval 1000000 us Enabling timer Hit '?': [q, b, e, ?] ........ tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0xef6, ter=0x0 Hit '?': [q, b, e, ?] . tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x2ad4, ter=0x0 Hit '?': [q, b, e, ?] . tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x1efc, ter=0x0 Hit '?': [q, b, e, ?] . tgcr=0x1, tmr=0xff1c, trr=0x3d09, tcr=0x0, tcn=0x169d, ter=0x0 Hit 'e': [q, b, e, ?] ...Stopping timer Hit 'q': [q, b, e, ?] ## Application terminated, rc = 0x0 5.13. U-Boot Image Formats U-Boot operates on "image" files which can be basically anything, preceeded by a special header; see the definitions in include/image.h for details; basically, the header defines the following image properties: * Target Operating System (Provisions for OpenBSD, NetBSD, FreeBSD, 4.4BSD, Linux, SVR4, Esix, Solaris, Irix, SCO, Dell, NCR, LynxOS, pSOS, QNX, RTEMS, U-Boot, ARTOS, Unity OS, Integrity; Currently supported: Linux, NetBSD, VxWorks, QNX, RTEMS, ARTOS, Unity OS, Integrity). * Target [745]CPU Architecture (Provisions for Alpha, ARM, AVR32, BlackFin, IA64, M68K, Microblaze, MIPS, MIPS64, NIOS, NIOS2, [746]Power Architecture®, IBM S390, SuperH, Sparc, Sparc 64 Bit, Intel x86; Currently supported: ARM, AVR32, BlackFin, M68K, Microblaze, MIPS, MIPS64, NIOS, NIOS2, [747]Power Architecture®, SuperH, Sparc, Sparc 64 Bit, Intel x86). * Compression Type (Provisions for uncompressed, gzip, bzip2, lzo; Currently supported: uncompressed, gzip, bzip2, lzo). * Load Address * Entry Point * Image Name * Image Timestamp The header is marked by a special Magic Number, and both the header and the data portions of the image are secured against corruption by CRC32 checksums. 5.14. U-Boot Advanced Features 5.14.1. Boot Count Limit The Open Source Development Labs Carrier Grade Linux Requirements Definition version 2.0 ([748]http://www.osdl.org/docs/carrier_grade_linux_requirements_defini tion___version_20_final_public_draft.pdf) contains the following requirement definition (ID PLT.4.0, p. 44): CGL shall provide support for detecting a repeating reboot cycle due to recurring failures and will go to an offline state if this occurs. This feature is available in U-Boot if you enable the CONFIG_BOOTCOUNT_LIMIT configuration option. The implementation uses the following environment variables: bootcount: This variable will be automatically created if it does not exist, and it will be updated at each reset of the processor. After a power-on reset, it will be initialized with 1, and each reboot will increment the value by 1. bootlimit: If this variable exists, its contents are taken as the maximum number of reboot cycles allowed. altbootcmd: If, after a reboot, the new value of bootcount exceeds the value of bootlimit, then instead of the standard boot action (executing the contents of bootcmd) an alternate boot action will be performed, and the contents of altbootcmd will be executed. If the variable bootlimit is not defined in the environment, the Boot Count Limit feature is disabled. If it is enabled, but altbootcmd is not defined, then U-Boot will drop into interactive mode and remain there. It is the responsibility of some application code (typically a Linux application) to reset the variable bootcount, thus allowing for more boot cycles. ALERT! At the moment, the Boot Count Limit feature is available only for MPC8xx, MPC82xx and MPC5200 [749]Power Architecture® processors. * [750]ubootBootcountAccess.c: C-source: bootcount access through /proc file system * [751]6. Embedded Linux Configuration + [752]6.1. Download and Unpack the Linux Kernel Sources + [753]6.2. Kernel Configuration and Compilation + [754]6.3. Installation 6. Embedded Linux Configuration 6.1. Download and Unpack the Linux Kernel Sources You can download the Linux Kernel Sources from our anonymous [755]git server at [756]http://git.denx.de/. To checkout the module for the first time, proceed as follows: bash$ cd /opt/eldk/usr/src bash$ git clone git://git.denx.de/linux-2.6-denx.git linux-2.6-denx bash$ cd linux-2.6-denx bash$ git checkout -b duts remotes/origin/44x/canyonlands Branch duts set up to track remote branch 44x/canyonlands from origin. Switched to a new branch 'duts' bash$ 6.2. Kernel Configuration and Compilation The canyonlands board is fully supported by DENX Software Engineering. This means that you will always be able to build a working default configuration with just minimal interaction. Please be aware that you will need the "powerpc" cross development tools for the following steps. Make sure that the directory which contains the binaries of your [757]ELDK are in your PATH. To be sure that no intermediate results of previous builds are left in your Linux kernel source tree you can clean it up as follows: bash$ make mrproper The following command selects a standard configuration for the canyonlands board that has been extensively tested. It is recommended to use this as a starting point for other, customized configurations: bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- 44x/canyonlands_defconfig TIP Note: The name of this default configuration file is arch/powerpc/configs/44x/canyonlands_defconfig . By (recursively) listing the contents of the arch/powerpc/configs/ directory you can easily find out which other default configurations are available. If you don't want to change the default configuration you can now continue to use it to build a kernel image: bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- uImage Otherwise you can modify the kernel configuration as follows: bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- config or bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- menuconfig ALERT! Note: Because of problems (especially with some older Linux kernel versions) the use of "make xconfig" is not recommended. bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- uImage The make target uImage uses the tool mkimage (from the U-Boot package) to create a Linux kernel image in arch/powerpc/boot/uImage which is immediately usable for download and booting with U-Boot. In case you need a DTB to boot your linux kernel, you need the following step: bash$ make canyonlands.dtb In case you configured modules you will also need to compile the modules: make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- modules add install the modules (make sure to pass the correct root path for module installation): bash$ make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- INSTALL_MOD_PATH=/opt/eldk-4.2 /ppc_4xx modules_install 6.3. Installation For now it is sufficient to copy the Linux kernel image into the directory used by your [758]TFTP server: bash$ cp arch/powerpc/boot/uImage /tftpboot/uImage * [759]7. Booting Embedded Linux + [760]7.1. Introduction + [761]7.2. Flattened Device Tree Blob + [762]7.3. Passing Kernel Arguments + [763]7.4. Boot Arguments Unleashed + [764]7.5. Networked Operation with Root Filesystem over NFS o [765]7.5.1. Bootlog of tftp'd Linux kernel with Root Filesystem over NFS + [766]7.6. Boot from Flash Memory + [767]7.7. Standalone Operation with Ramdisk Image 7. Booting Embedded Linux 7.1. Introduction In principle, if you have a Linux kernel image and the flattened device tree blob somewhere in system memory (RAM, ROM, flash...), then all you need to boot the system is the bootm command. Assume a Linux kernel image has been stored at address 0xFC000000 and the flattened device tree blob has been stored at address 0xFC1E0000 - then you can boot this image with the following command: => bootm FC000000 - FC1E0000 7.2. Flattened Device Tree Blob Linux kernel expects certain information on the hardware that it runs on. For kernels compiled with fdt support, this information has the form of a device tree, which is based on the Open Firmware specification. Bootloaders like U-Boot that do not implement the Open Firmware API, are expected to pass to the kernel a binary form of the flattened device tree, commonly referred to as FDT blob or simply the blob. Device trees are defined in human-readable text files, which are part of the Linux 2.6 source tree. Device tree source for the canyonlands board is found in arch/powerpc/boot/dts/canyonlands.dts file. Before the device tree can be passed to the kernel, it has to be compiled to the binary form by the dtc compiler. The dtc compiler is included with the Linux kernel since 2.6.25. Since 2.6.26 there is also a simple makefile rule to generate the blob: make ARCH=powerpc CROSS_COMPILE=ppc_4xxFP- canyonlands.dtb After the blob has been compiled, it has to be transferred from where it was built ("arch/powerpc/boot/canyonlands.dtb") to target's memory, for example over the [768]TFTP protocol using U-Boot's tftp command. Then, the blob is passed to the kernel by the bootm command, and its address in memory is one of the arguments to bootm - refer to the description of this command in [769]UBootCmdGroupExec for more details. Note that U-Boot makes some automatic modifications to the blob before passing it to the kernel - mainly adding and modifying information that is learnt at run-time. U-Boot also has provisions to alter a flattened device tree in arbitrary ways from the command line, refer to the description of the fdt commands found in [770]UBootCmdGroupMisc. Notes: * Flattened Device Tree custodian's page at [771]http://www.denx.de/wiki/U-Boot/UBootFdtInfo contains useful information, and a number of references. * At the time of this writing (September 2007) blob handling is still a very fresh feature and undergonig frequent changes. Reader is encouraged to watch the u-boot-users and linuxppc-dev mailing lists for important news (required version of the dtc compiler, blob compilation options, flattened device tree source file structure, etc.). 7.3. Passing Kernel Arguments In nearly all cases, you will want to pass additional information to the Linux kernel; for instance, information about the root device or network configuration. In U-Boot, this is supported using the bootargs environment variable. Its contents are automatically passed to the Linux kernel as boot arguments (or "command line" arguments). This allows the use of the same Linux kernel image in a wide range of configurations. For instance, by just changing the contents of the bootargs variable you can use the very same Linux kernel image to boot with an initrd ramdisk image, with a root filesystem over NFS, with a [772]CompactFlash disk or from a flash filesystem. As one example, to boot the Linux kernel image at address0x400000 using the initrd ramdisk image at address0x600000 as root filesystem, and with the flattened device tree blob at address 0x800000, and with the flattened device tree blob at address 0xFC1E0000, you can use the following commands: => setenv bootargs root=/dev/ram rw => bootm0x400000 0x600000 0x800000 0xFC1E0000 To boot the same kernel image with a root filesystem over NFS, the following command sequence can be used. This example assumes that your NFS server has the IP address "192.168.1.1" and exports the directory "/opt/eldk-4.2/ppc_4xx" as root filesystem for the target. The target has been assigned the IP address "192.168.100.6" and the hostname "canyonlands". A netmask of "255.255.0.0" is used: => setenv bootargs root=/dev/nfs rw nfsroot=192.168.1.1:/opt/eldk-4.2/ppc_4xx i p=192.168.100.6:192.168.1.1:192.168.1.1:255.255.0.0:canyonlands::off => bootm0x400000 - 0x800000 - 0xFC1E0000 Please see also the files Documentation/initrd.txt and Documentation/nfsroot.txt in your Linux kernel source directory for more information about which options can be passed to the Linux kernel. ALERT! Note: Once your system is up and running, if you have a simple shell login, you can normally examine the boot arguments that were used by the kernel for the most recent boot with the command: $ cat /proc/cmdline 7.4. Boot Arguments Unleashed Passing command line arguments to the Linux kernel allows for very flexible and efficient configuration which is especially important in Embedded Systems. It is somewhat strange that these features are nearly undocumented everywhere else. One reason for that is certainly the very limited capabilities of other boot loaders. It is especially U-Boot's capability to easily define, store, and use environment variables that makes it such a powerful tool in this area. In the examples above we have already seen how we can use for instance the root and ip boot arguments to pass information about the root filesystem or network configuration. The ip argument is not only useful in configurations with root filesystem over NFS; if the Linux kernel has the CONFIG_IP_PNP configuration enabled (IP kernel level autoconfiguration), this can be used to enable automatic configuration of IP addresses of devices and of the routing table during kernel boot, based on either information supplied on the kernel command line or by [773]BOOTP or RARP protocols. The advantage of this mechanism is that you don't have to spend precious system memory (RAM and flash) for network configuration tools like ifconfig or route - especially in Embedded Systems where you seldom have to change the network configuration while the system is running. We can use U-Boot environment variables to store all necessary configuration parameters: => setenv ipaddr 192.168.100.6 => setenv serverip 192.168.1.1 => setenv netmask 255.255.0.0 => setenv hostname canyonlands => setenv rootpath /opt/eldk-4.2/ppc_4xx => saveenv Then you can use these variables to build the boot arguments to be passed to the Linux kernel: => setenv nfsargs 'root=/dev/nfs rw nfsroot=${serverip}:${rootpath}' Note how apostrophes are used to delay the substitution of the referenced environment variables. This way, the current values of these variables get inserted when assigning values to the "bootargs" variable itself later, i. e. when it gets assembled from the given parts before passing it to the kernel. This allows us to simply redefine any of the variables (say, the value of "ipaddr" if it has to be changed), and the changes will automatically propagate to the Linux kernel. ALERT! Note: You cannot use this method directly to define for example the "bootargs" environment variable, as the implicit usage of this variable by the "bootm" command will not trigger variable expansion - this happens only when using the "setenv" command. In the next step, this can be used for a flexible method to define the "bootargs" environment variable by using a function-like approach to build the boot arguments step by step: => setenv ramargs setenv bootargs root=/dev/ram rw => setenv nfsargs 'setenv bootargs root=/dev/nfs rw nfsroot=${serverip}:${rootp ath}' => setenv addip 'setenv bootargs ${bootargs} ip=${ipaddr}:${serverip}:${gateway ip}:${netmask}:${hostname}::off' => setenv ram_root 'run ramargs addip;bootm ${kernel_addr} ${ramdisk_addr} ${fd t_addr} ${fdt_addr} ' => setenv nfs_root 'run nfsargs addip;bootm ${kernel_addr} - ${fdt_addr} - ${f dt_addr} ' In this setup we define two variables, ram_root and nfs_root, to boot with root filesystem from a ramdisk image or over NFS, respecively. The variables can be executed using U-Boot's run command. These variables make use of the run command itself: * First, either run ramargs or run nfsargs is used to initialize the bootargs environment variable as needed to boot with ramdisk image or with root over NFS. * Then, in both cases, run addip is used to append the ip parameter to use the Linux kernel IP autoconfiguration mechanism for configuration of the network settings. * Finally, the bootm command is used with three resp. two address arguments %ENDIF0% to boot the Linux kernel image with resp. without a ramdisk image. (We assume here that the variables kernel_addr , ramdisk_addr and fdt_addr %ENDIF0% have already been set.) This method can be easily extended to add more customization options when needed. If you have used U-Boot's network commands before (and/or read the documentation), you will probably have recognized that the names of the U-Boot environment variables we used in the examples above are exactly the same as those used with the U-Boot commands to boot over a network using [774]DHCP or [775]BOOTP. That means that, instead of manually setting network configuration parameters like IP address, etc., these variables will be set automatically to the values retrieved with the network boot protocols. This is explained in detail in the sections about the respective U-Boot commands. 7.5. Networked Operation with Root Filesystem over NFS This section will show how to boot the target into Linux with no more than U-Boot residing on it. For this we will use the tftp command of U-Boot to transfer a Linux kernel and boot it with the NFS rootfilesystem provided by the [776]ELDK. For this to work, we rely on some U-Boot environment variables to be set up correctly, i.e. the network parameters, the names of files to transfer via tftp and last but not least some scripts easing the assembly of the Linux command line. The whole process is packaged up into one script shown before we actually execute it. Note that the Linux kernel will also output the command line used, so you can easily check if everything worked like expected. The command line in this example passes at least the following information to the: * root=/dev/nfs rw: the root filesystem will be mounted using NFS, and it will be writable. * nfsroot=192.168.1.1:/opt/eldk-4.2/ppc_4xx: the NFS server has the IP address 192.168.1.1, and exports the directory /opt/eldk-4.2/ppc_4xx for our system to use as root filesystem. * ip=192.168.100.6:192.168.1.1:192.168.1.1:255.255.0.0:canyonlands:: off: the target has the IP address 192.168.100.6; the NFS server is 192.168.1.1; there is a gateway at IP address 192.168.1.1; the netmask is 255.255.0.0 and our hostname is canyonlands. The first ethernet interface (eth0) willbe used, and the Linux kernel will immediately use this network configuration and not try to re-negotiate it (IP autoconfiguration is off). See Documentation/nfsroot.txt in you Linux kernel source directory for more information about these parameters and other options. 7.5.1. Bootlog of tftp'd Linux kernel with Root Filesystem over NFS => setenv fdt_file /tftpboot/duts/canyonlands/canyonlands.dtb => setenv bootfile /tftpboot/duts/canyonlands/uImage => setenv rootpath /opt/eldk-4.2/ppc_4xxFP/ => printenv net_nfs net_nfs=tftp ${kernel_addr_r} ${bootfile}; tftp ${fdt_addr_r} ${fdt_file}; run nfsargs addip addtty addmisc;bootm ${kernel_addr_r} - ${fdt_addr_r} => run net_nfs Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/uImage'. Load address: 0x400000 Loading: ################################################################# ################################################################# #### done Bytes transferred = 1958609 (1de2d1 hex) Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/canyonlands.dtb'. Load address: 0xb00000 Loading: # done Bytes transferred = 10190 (27ce hex) ## Booting kernel from Legacy Image at 00400000 ... Image Name: Linux-2.6.32.7-00007-g08eba26 Created: 2010-02-04 17:54:22 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1958545 Bytes = 1.9 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Flattened Device Tree blob at 00b00000 Booting using the fdt blob at 0xb00000 Uncompressing Kernel Image ... OK Using PowerPC 44x Platform machine description Linux version 2.6.32.7-00007-g08eba26 (stefan@stefan-desktop) (gcc version 4.2. 2) #1 Thu Feb 4 18:54:16 CET 2010 Zone PFN ranges: DMA 0x00000000 -> 0x00020000 Normal 0x00020000 -> 0x00020000 HighMem 0x00020000 -> 0x00020000 Movable zone start PFN for each node early_node_map[1] active PFN ranges 0: 0x00000000 -> 0x00020000 MMU: Allocated 1088 bytes of context maps for 255 contexts Built 1 zonelists in Zone order, mobility grouping on. Total pages: 130048 Kernel command line: root=/dev/nfs rw nfsroot=192.168.1.1:/opt/eldk-4.2/ppc_4xx FP/ ip=192.168.100.6:192.168.1.1:192.168.1.254:255.255.0.0:canyonlands:eth0:off panic=1 console=ttyS0,115200 PID hash table entries: 2048 (order: 1, 8192 bytes) Dentry cache hash table entries: 65536 (order: 6, 262144 bytes) Inode-cache hash table entries: 32768 (order: 5, 131072 bytes) Memory: 515328k/524288k available (3880k kernel code, 8740k reserved, 172k data , 134k bss, 168k init) Kernel virtual memory layout: * 0xfffef000..0xfffff000 : fixmap * 0xffc00000..0xffe00000 : highmem PTEs * 0xffa00000..0xffc00000 : consistent mem * 0xffa00000..0xffa00000 : early ioremap * 0xe1000000..0xffa00000 : vmalloc & ioremap SLUB: Genslabs=11, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1 Hierarchical RCU implementation. NR_IRQS:512 UIC0 (32 IRQ sources) at DCR 0xc0 UIC1 (32 IRQ sources) at DCR 0xd0 UIC2 (32 IRQ sources) at DCR 0xe0 UIC3 (32 IRQ sources) at DCR 0xf0 clocksource: timebase mult[3c0000] shift[22] registered Mount-cache hash table entries: 512 NET: Registered protocol family 16 256k L2-cache enabled PCIE0: Port disabled via device-tree PCIE1: Checking link... PCIE1: No device detected. PCI host bridge /plb/pciex@d20000000 (primary) ranges: MEM 0x0000000e80000000..0x0000000effffffff -> 0x0000000080000000 MEM 0x0000000f00100000..0x0000000f001fffff -> 0x0000000000000000 IO 0x0000000f80010000..0x0000000f8001ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000f00100000 4xx PCI DMA offset set to 0x00000000 /plb/pciex@d20000000: Legacy ISA memory support enabled PCIE1: successfully set as root-complex PCI host bridge /plb/pci@c0ec00000 (primary) ranges: MEM 0x0000000d80000000..0x0000000dffffffff -> 0x0000000080000000 MEM 0x0000000c0ee00000..0x0000000c0eefffff -> 0x0000000000000000 IO 0x0000000c08000000..0x0000000c0800ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000c0ee00000 4xx PCI DMA offset set to 0x00000000 /plb/pci@c0ec00000: Legacy ISA memory support enabled PCI: Probing PCI hardware PCI: Hiding 4xx host bridge resources 0000:80:00.0 pci 0000:80:00.0: PCI bridge, secondary bus 0000:81 pci 0000:80:00.0: IO window: disabled pci 0000:80:00.0: MEM window: disabled pci 0000:80:00.0: PREFETCH window: disabled bio: create slab at 0 vgaarb: loaded SCSI subsystem initialized usbcore: registered new interface driver usbfs usbcore: registered new interface driver hub usbcore: registered new device driver usb Switching to clocksource timebase NET: Registered protocol family 2 IP route cache hash table entries: 4096 (order: 2, 16384 bytes) TCP established hash table entries: 16384 (order: 5, 131072 bytes) TCP bind hash table entries: 16384 (order: 4, 65536 bytes) TCP: Hash tables configured (established 16384 bind 16384) TCP reno registered NET: Registered protocol family 1 RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. JFFS2 version 2.2. (NAND) © 2001-2006 Red Hat, Inc. msgmni has been set to 1006 io scheduler noop registered io scheduler anticipatory registered (default) io scheduler deadline registered io scheduler cfq registered Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled serial8250.0: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550A console [ttyS0] enabled serial8250.0: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550A serial8250.0: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550A serial8250.0: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550A 4ef600300.serial: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550 4ef600400.serial: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550 4ef600500.serial: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550 4ef600600.serial: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550 brd: module loaded sata-dwc sata-dwc.0: id 0, controller version 1.82 sata-dwc sata-dwc.0: DMA initialized sata-dwc sata-dwc.0: **** No neg speed (nothing attached?) scsi0 : sata-dwc ata1: SATA max UDMA/133 irq 23 4cc000000.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank Amd/Fujitsu Extended Query Table at 0x0040 4cc000000.nor_flash: CFI does not contain boot bank location. Assuming top. number of CFI chips: 1 cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness. RedBoot partition parsing not available Creating 7 MTD partitions on "4cc000000.nor_flash": 0x000000000000-0x0000001e0000 : "kernel" 0x0000001e0000-0x000000200000 : "dtb" 0x000000200000-0x000001600000 : "ramdisk" 0x000001600000-0x000001a00000 : "jffs2" 0x000001a00000-0x000003f60000 : "user" 0x000003f60000-0x000003fa0000 : "env" 0x000003fa0000-0x000004000000 : "u-boot" NAND device: Manufacturer ID: 0x20, Chip ID: 0xf1 (ST Micro NAND 128MiB 3,3V 8- bit) Scanning device for bad blocks Creating 2 MTD partitions on "4e0000000.ndfc.nand": 0x000000000000-0x000000100000 : "u-boot" 0x000000000000-0x000003f00000 : "user" e1000e: Intel(R) PRO/1000 Network Driver - 1.0.2-k2 e1000e: Copyright (c) 1999-2008 Intel Corporation. PPC 4xx OCP EMAC driver, version 3.54 MAL v2 /plb/mcmal, 2 TX channels, 16 RX channels ZMII /plb/opb/emac-zmii@ef600d00 initialized RGMII /plb/opb/emac-rgmii@ef601500 initialized with MDIO support TAH /plb/opb/emac-tah@ef601350 initialized TAH /plb/opb/emac-tah@ef601450 initialized /plb/opb/emac-rgmii@ef601500: input 0 in RGMII mode eth0: EMAC-0 /plb/opb/ethernet@ef600e00, MAC 00:10:ec:01:08:84 eth0: found Generic MII PHY (0x00) /plb/opb/emac-rgmii@ef601500: input 1 in RGMII mode eth1: EMAC-1 /plb/opb/ethernet@ef600f00, MAC 00:10:ec:81:08:84 eth1: found Generic MII PHY (0x01) ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver ppc-of-ehci 4bffd0400.ehci: OF EHCI ppc-of-ehci 4bffd0400.ehci: new USB bus registered, assigned bus number 1 ppc-of-ehci 4bffd0400.ehci: irq 38, io mem 0x4bffd0400 ppc-of-ehci 4bffd0400.ehci: USB 2.0 started, EHCI 1.00 usb usb1: New USB device found, idVendor=1d6b, idProduct=0002 usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb1: Product: OF EHCI usb usb1: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ehci_hcd usb usb1: SerialNumber: PPC-OF USB usb usb1: configuration #1 chosen from 1 choice hub 1-0:1.0: USB hub found hub 1-0:1.0: 1 port detected ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver ppc-of-ohci 4bffd0000.usb: OF OHCI ppc-of-ohci 4bffd0000.usb: new USB bus registered, assigned bus number 2 ppc-of-ohci 4bffd0000.usb: irq 39, io mem 0x4bffd0000 usb usb2: New USB device found, idVendor=1d6b, idProduct=0001 usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb2: Product: OF OHCI usb usb2: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ohci_hcd usb usb2: SerialNumber: PPC-OF USB usb usb2: configuration #1 chosen from 1 choice hub 2-0:1.0: USB hub found hub 2-0:1.0: 1 port detected Initializing USB Mass Storage driver... usbcore: registered new interface driver usb-storage USB Mass Storage support registered. usbcore: registered new interface driver libusual dwc_otg: version 2.60a 22-NOV-2006 dwc_otg: Shared Tx FIFO mode dwc_otg: Using DMA mode dwc_otg dwc_otg.0: DWC OTG Controller dwc_otg dwc_otg.0: new USB bus registered, assigned bus number 3 dwc_otg dwc_otg.0: irq 28, io mem 0x00000000 dwc_otg: Init: Port Power? op_state=4 usb usb3: New USB device found, idVendor=1d6b, idProduct=0002 usb usb3: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb3: Product: DWC OTG Controller usb usb3: Manufacturer: Linux 2.6.32.7-00007-g08eba26 dwc_otg_hcd usb usb3: SerialNumber: dwc_otg.0 usb usb3: configuration #1 chosen from 1 choice hub 3-0:1.0: USB hub found hub 3-0:1.0: 1 port detected i2c /dev entries driver ibm-iic 4ef600700.i2c: using standard (100 kHz) mode rtc-m41t80 0-0068: chip found, driver version 0.05 rtc-m41t80 0-0068: rtc core: registered m41t80 as rtc0 ibm-iic 4ef600800.i2c: using standard (100 kHz) mode ad7414 0-0048: chip found TCP cubic registered NET: Registered protocol family 17 rtc-m41t80 0-0068: setting system clock to 2010-02-08 16:33:33 UTC (1265646813) usb 1-1: new high speed USB device using ppc-of-ehci and address 2 usb 1-1: New USB device found, idVendor=1370, idProduct=2168 usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-1: Product: Memorybird P usb 1-1: Manufacturer: FSC usb 1-1: SerialNumber: 10005383BB000032 usb 1-1: configuration #1 chosen from 1 choice scsi1 : SCSI emulation for USB Mass Storage devices eth0: link is down IP-Config: Complete: device=eth0, addr=192.168.100.6, mask=255.255.0.0, gw=192.168.1.254, host=canyonlands, domain=, nis-domain=(none), bootserver=192.168.1.1, rootserver=192.168.1.1, rootpath= Looking up port of RPC 100003/2 on 192.168.1.1 eth0: link is up, 1000 FDX, pause enabled Looking up port of RPC 100005/1 on 192.168.1.1 VFS: Mounted root (nfs filesystem) on device 0:14. Freeing unused kernel memory: 168k init modprobe: FATAL: Could not load /lib/modules/2.6.32.7-00007-g08eba26/modules.de p: No such file or directory modprobe: FATAL: Could not load /lib/modules/2.6.32.7-00007-g08eba26/modules.de p: No such file or directory INIT: version 2.86 booting Welcome to DENX Embedded Linux Environment Press 'I' to enter interactive startup. Setting clock : Mon Feb 8 16:33:38 CET 2010 [ OK ] Building the cache scsi 1:0:0:0: Direct-Access FSC Memorybird P 2.00 PQ: 0 ANSI: 2 [ OK ] Setting hostname canyonlands: [ OK ] Mounting local filesystems: [ OK ] Enabling /etc/fstab swaps: [ OK ] INIT: Entering runlevel: 3 Entering non-interactive startup FATAL: Could not load /lib/modules/2.6.32.7-00007-g08eba26/modules.dep: No such file or directory Bringing up loopback interface: [ OK ] FATAL: Could not load /lib/modules/2.6.32.7-00007-g08eba26/modules.dep: No such file or directory Starting system logger: [ OK ] Starting kernel logger: [ OK ] Starting rpcbind: [ OK ] Mounting NFS filesystems: [ OK ] Mounting other filesystems: [ OK ] Starting xinetd: [ OK ] DENX ELDK version 4.2 build 2008-04-01 Linux 2.6.32.7-00007-g08eba26 on a ppc canyonlands login: root Last login: Sat Jan 3 19:28:07 from pollux -bash-3.2# 7.6. Boot from Flash Memory The previous section described how to load the Linux kernel image over ethernet using [777]TFTP. This is especially well suited for your development and test environment, when the kernel image is still undergoing frequent changes, for instance because you are modifying kernel code or configuration. Later in your development cycle you will work on application code or device drivers, which can be loaded dynamically as modules. If the Linux kernel remains the same then you can save the time needed for the [778]TFTP download and put the kernel image into the flash memory of your canyonlands board. The U-Boot command flinfo can be used to display information about the available on-board flash on your system: => fli Bank # 1: CFI conformant FLASH (16 x 16) Size: 64 MB in 512 Sectors AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E Erase timeout: 16384 ms, write timeout: 2 ms Buffer write timeout: 5 ms, buffer size: 32 bytes Sector Start Addresses: FC000000 FC020000 FC040000 FC060000 FC080000 FC0A0000 FC0C0000 FC0E0000 FC100000 FC120000 FC140000 FC160000 FC180000 FC1A0000 FC1C0000 E FC1E0000 FC200000 FC220000 FC240000 FC260000 FC280000 FC2A0000 FC2C0000 FC2E0000 FC300000 FC320000 FC340000 FC360000 FC380000 FC3A0000 FC3C0000 E FC3E0000 E FC400000 E FC420000 E FC440000 E FC460000 E FC480000 E FC4A0000 E FC4C0000 E FC4E0000 E FC500000 E FC520000 E FC540000 E FC560000 E FC580000 E FC5A0000 E FC5C0000 E FC5E0000 E FC600000 E FC620000 E FC640000 E FC660000 E FC680000 E FC6A0000 E FC6C0000 E FC6E0000 E FC700000 E FC720000 E FC740000 E FC760000 E FC780000 E FC7A0000 E FC7C0000 E FC7E0000 E FC800000 E FC820000 E FC840000 E FC860000 E FC880000 E FC8A0000 E FC8C0000 E FC8E0000 E FC900000 E FC920000 E FC940000 E FC960000 E FC980000 E FC9A0000 E FC9C0000 E FC9E0000 E FCA00000 E FCA20000 E FCA40000 E FCA60000 E FCA80000 E FCAA0000 E FCAC0000 E FCAE0000 E FCB00000 E FCB20000 E FCB40000 E FCB60000 E FCB80000 E FCBA0000 E FCBC0000 E FCBE0000 E FCC00000 E FCC20000 E FCC40000 E FCC60000 E FCC80000 E FCCA0000 E FCCC0000 E FCCE0000 E FCD00000 E FCD20000 E FCD40000 E FCD60000 E FCD80000 E FCDA0000 E FCDC0000 E FCDE0000 E FCE00000 E FCE20000 E FCE40000 E FCE60000 E FCE80000 E FCEA0000 E FCEC0000 E FCEE0000 E FCF00000 E FCF20000 E FCF40000 E FCF60000 E FCF80000 E FCFA0000 E FCFC0000 E FCFE0000 E FD000000 E FD020000 E FD040000 E FD060000 E FD080000 E FD0A0000 E FD0C0000 E FD0E0000 E FD100000 E FD120000 E FD140000 E FD160000 E FD180000 E FD1A0000 E FD1C0000 E FD1E0000 E FD200000 E FD220000 E FD240000 E FD260000 E FD280000 E FD2A0000 E FD2C0000 E FD2E0000 E FD300000 E FD320000 E FD340000 E FD360000 E FD380000 E FD3A0000 E FD3C0000 E FD3E0000 E FD400000 E FD420000 E FD440000 E FD460000 E FD480000 E FD4A0000 E FD4C0000 E FD4E0000 E FD500000 E FD520000 E FD540000 E FD560000 E FD580000 E FD5A0000 E FD5C0000 E FD5E0000 E FD600000 FD620000 E FD640000 E FD660000 E FD680000 E FD6A0000 E FD6C0000 E FD6E0000 E FD700000 E FD720000 E FD740000 E FD760000 E FD780000 E FD7A0000 E FD7C0000 E FD7E0000 E FD800000 E FD820000 E FD840000 E FD860000 E FD880000 E FD8A0000 E FD8C0000 E FD8E0000 E FD900000 E FD920000 E FD940000 E FD960000 E FD980000 E FD9A0000 E FD9C0000 E FD9E0000 E FDA00000 E FDA20000 E FDA40000 E FDA60000 E FDA80000 E FDAA0000 E FDAC0000 E FDAE0000 E FDB00000 E FDB20000 E FDB40000 E FDB60000 E FDB80000 E FDBA0000 E FDBC0000 E FDBE0000 E FDC00000 E FDC20000 E FDC40000 E FDC60000 E FDC80000 E FDCA0000 E FDCC0000 E FDCE0000 E FDD00000 E FDD20000 E FDD40000 E FDD60000 E FDD80000 E FDDA0000 E FDDC0000 E FDDE0000 E FDE00000 E FDE20000 E FDE40000 E FDE60000 E FDE80000 E FDEA0000 E FDEC0000 E FDEE0000 E FDF00000 E FDF20000 E FDF40000 E FDF60000 E FDF80000 E FDFA0000 E FDFC0000 E FDFE0000 E FE000000 E FE020000 E FE040000 E FE060000 E FE080000 E FE0A0000 E FE0C0000 E FE0E0000 E FE100000 E FE120000 E FE140000 E FE160000 E FE180000 E FE1A0000 E FE1C0000 E FE1E0000 E FE200000 E FE220000 E FE240000 E FE260000 E FE280000 E FE2A0000 E FE2C0000 E FE2E0000 E FE300000 E FE320000 E FE340000 E FE360000 E FE380000 E FE3A0000 E FE3C0000 E FE3E0000 E FE400000 E FE420000 E FE440000 E FE460000 E FE480000 E FE4A0000 E FE4C0000 E FE4E0000 E FE500000 E FE520000 E FE540000 E FE560000 E FE580000 E FE5A0000 E FE5C0000 E FE5E0000 E FE600000 E FE620000 E FE640000 E FE660000 E FE680000 E FE6A0000 E FE6C0000 E FE6E0000 E FE700000 E FE720000 E FE740000 E FE760000 E FE780000 E FE7A0000 E FE7C0000 E FE7E0000 E FE800000 E FE820000 E FE840000 E FE860000 E FE880000 E FE8A0000 E FE8C0000 E FE8E0000 E FE900000 E FE920000 E FE940000 E FE960000 E FE980000 E FE9A0000 E FE9C0000 E FE9E0000 E FEA00000 E FEA20000 E FEA40000 E FEA60000 E FEA80000 E FEAA0000 E FEAC0000 E FEAE0000 E FEB00000 E FEB20000 E FEB40000 E FEB60000 E FEB80000 E FEBA0000 E FEBC0000 E FEBE0000 E FEC00000 E FEC20000 E FEC40000 E FEC60000 E FEC80000 E FECA0000 E FECC0000 E FECE0000 E FED00000 E FED20000 E FED40000 E FED60000 E FED80000 E FEDA0000 E FEDC0000 E FEDE0000 E FEE00000 E FEE20000 E FEE40000 E FEE60000 E FEE80000 E FEEA0000 E FEEC0000 E FEEE0000 E FEF00000 E FEF20000 E FEF40000 E FEF60000 E FEF80000 E FEFA0000 E FEFC0000 E FEFE0000 E FF000000 E FF020000 E FF040000 E FF060000 E FF080000 E FF0A0000 E FF0C0000 E FF0E0000 E FF100000 E FF120000 E FF140000 E FF160000 E FF180000 E FF1A0000 E FF1C0000 E FF1E0000 E FF200000 E FF220000 E FF240000 E FF260000 E FF280000 E FF2A0000 E FF2C0000 E FF2E0000 E FF300000 E FF320000 E FF340000 E FF360000 E FF380000 E FF3A0000 E FF3C0000 E FF3E0000 E FF400000 E FF420000 E FF440000 E FF460000 E FF480000 E FF4A0000 E FF4C0000 E FF4E0000 E FF500000 E FF520000 E FF540000 E FF560000 E FF580000 E FF5A0000 E FF5C0000 E FF5E0000 E FF600000 E FF620000 E FF640000 E FF660000 E FF680000 E FF6A0000 E FF6C0000 E FF6E0000 E FF700000 E FF720000 E FF740000 E FF760000 E FF780000 E FF7A0000 E FF7C0000 E FF7E0000 E FF800000 E FF820000 E FF840000 E FF860000 E FF880000 E FF8A0000 E FF8C0000 E FF8E0000 E FF900000 E FF920000 E FF940000 E FF960000 E FF980000 E FF9A0000 E FF9C0000 E FF9E0000 E FFA00000 E FFA20000 E FFA40000 E FFA60000 E FFA80000 E FFAA0000 E FFAC0000 E FFAE0000 E FFB00000 E FFB20000 E FFB40000 E FFB60000 E FFB80000 E FFBA0000 E FFBC0000 E FFBE0000 E FFC00000 E FFC20000 E FFC40000 E FFC60000 E FFC80000 E FFCA0000 E FFCC0000 E FFCE0000 E FFD00000 E FFD20000 E FFD40000 E FFD60000 E FFD80000 E FFDA0000 E FFDC0000 E FFDE0000 E FFE00000 E FFE20000 E FFE40000 E FFE60000 E FFE80000 E FFEA0000 E FFEC0000 E FFEE0000 E FFF00000 E FFF20000 E FFF40000 E FFF60000 RO FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO => From this output you can see the total amount of flash memory, and how it is divided in blocks (Erase Units or Sectors). The RO markers show blocks of flash memory that are write protected (by software) - this is the area where U-Boot is stored. The remaining flash memory is available for other use. For instance, we can store the Linux kernel image in flash starting at the start address of the next free flash sector. Before we can do this we must make sure that the flash memory in that region is empty - a Linux kernel image is typically around 600...700 kB, so to be on the safe side we dedicate the whole area from 0xFC000000 to 0xFC17FFFF for the kernel image. Keep in mind that with flash memory only whole erase units can be cleared. After having deleted the target flash area, you can download the Linux image and write it to flash. Below is a transcript of the complete operation with a final iminfo command to check the newly placed Linux kernel image in the flash memory. => => setenv kernel_addr FC000000 => => prot off FC000000 FC17FFFF Un-Protected 12 sectors => => era FC000000 FC17FFFF ............ done Erased 12 sectors => => tftp 100000 /tftpboot/canyonlands/uImage-duts ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0) Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/canyonlands/uImage-duts'. Load address: 0x100000 Loading: *################################################################# ###################################################### done Bytes transferred = 1744326 (1a9dc6 hex) => => imi 100000 ## Checking Image at 00100000 ... Legacy image found Image Name: Linux-2.6.25-rc8-01016-g94bf13b- Created: 2008-04-10 9:50:08 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1744262 Bytes = 1.7 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => => setenv ram_ws 100000 => => cp.b ${ram_ws} ${kernel_addr} ${filesize} Copy to Flash... done => => iminfo ${kernel_addr} ## Checking Image at fc000000 ... Legacy image found Image Name: Linux-2.6.25-rc8-01016-g94bf13b- Created: 2008-04-10 9:50:08 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1744262 Bytes = 1.7 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => => saveenv Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => Note how the filesize variable (which gets set by the [779]TFTP transfer) is used to automatically adjust for the actual image size. Since kernel requires the flattened device tree blob to be passed at boot time, you have to also write the blob to the flash memory. Below is a transcript of this operation. => => setenv fdt_addr FC1E0000 => => prot off FC1E0000 FC1FFFFF Un-Protected 1 sectors => => era FC1E0000 FC1FFFFF . done Erased 1 sectors => => tftp 100000 /tftpboot/canyonlands/canyonlands.dtb Waiting for PHY auto negotiation to complete... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0) Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/canyonlands/canyonlands.dtb'. Load address: 0x100000 Loading: *T # done Bytes transferred = 10000 (2710 hex) => => md 100000 00100000: d00dfeed 00002710 000000b8 00001b08 ......'......... 00100010: 00000028 00000011 00000010 00000000 ...(............ 00100020: 000002f5 00001a50 00000000 00000000 .......P........ 00100030: 00000000 00000000 00000000 00000000 ................ 00100040: 00000000 00000000 00000000 00000000 ................ 00100050: 00000000 00000000 00000000 00000000 ................ 00100060: 00000000 00000000 00000000 00000000 ................ 00100070: 00000000 00000000 00000000 00000000 ................ 00100080: 00000000 00000000 00000000 00000000 ................ 00100090: 00000000 00000000 00000000 00000000 ................ 001000a0: 00000000 00000000 00000000 00000000 ................ 001000b0: 00000000 00000000 00000001 00000000 ................ 001000c0: 00000003 00000004 00000000 00000002 ................ 001000d0: 00000003 00000004 0000000f 00000001 ................ 001000e0: 00000003 00000011 0000001b 616d6363 ............amcc 001000f0: 2c63616e 796f6e6c 616e6473 00000000 ,canyonlands.... => => setenv ram_ws 100000 => => cp.b ${ram_ws} ${fdt_addr} ${filesize} Copy to Flash... done => => md ${fdt_addr} fc1e0000: d00dfeed 00002710 000000b8 00001b08 ......'......... fc1e0010: 00000028 00000011 00000010 00000000 ...(............ fc1e0020: 000002f5 00001a50 00000000 00000000 .......P........ fc1e0030: 00000000 00000000 00000000 00000000 ................ fc1e0040: 00000000 00000000 00000000 00000000 ................ fc1e0050: 00000000 00000000 00000000 00000000 ................ fc1e0060: 00000000 00000000 00000000 00000000 ................ fc1e0070: 00000000 00000000 00000000 00000000 ................ fc1e0080: 00000000 00000000 00000000 00000000 ................ fc1e0090: 00000000 00000000 00000000 00000000 ................ fc1e00a0: 00000000 00000000 00000000 00000000 ................ fc1e00b0: 00000000 00000000 00000001 00000000 ................ fc1e00c0: 00000003 00000004 00000000 00000002 ................ fc1e00d0: 00000003 00000004 0000000f 00000001 ................ fc1e00e0: 00000003 00000011 0000001b 616d6363 ............amcc fc1e00f0: 2c63616e 796f6e6c 616e6473 00000000 ,canyonlands.... => => setenv filesize => => saveenv Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => Now we can boot directly from flash. All we need to do is passing the in-flash address of the image (FC000000) and the in-flash address of the flattened device tree (FC1E0000) with the bootm command; we also make the definition of the bootargs variable permanent now: => setenv bootcmd bootm FC000000 - FC1E0000 => setenv bootargs root=/dev/nfs rw nfsroot=${serverip}:${rootpath} ip=${ipaddr }:${serverip}:${gatewayip}:${netmask}:${hostname}::off Use printenv to verify that everything is OK before you save the environment settings: => printenv bootdelay=5 baudrate=115200 stdin=serial stdout=serial stderr=serial bootcmd=bootm FC000000 - FC1E0000 bootargs=root=/dev/nfs rw nfsroot=192.168.1.1:/opt/eldk-4.2/ppc_4xx ip=192.168.100.6:192.168.1.1:192.168.1.1:255.255.0.0:canyonlands::off .... => saveenv To test booting from flash you can now reset the board (either by power-cycling it, or using the U-Boot command reset), or you can manually call the boot command which will run the commands in the bootcmd variable: => run flash_self ## Booting kernel from Legacy Image at fc000000 ... Image Name: Linux-2.6.32.7-00007-g08eba26 Created: 2010-02-04 17:54:22 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1958545 Bytes = 1.9 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Loading init Ramdisk from Legacy Image at fc200000 ... Image Name: Simple Embedded Linux Framework Created: 2008-04-01 19:52:43 UTC Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 1727937 Bytes = 1.6 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Flattened Device Tree blob at fc1e0000 Booting using the fdt blob at 0xfc1e0000 Uncompressing Kernel Image ... OK Loading Ramdisk to 1fca6000, end 1fe4bdc1 ... OK Loading Device Tree to 00ffa000, end 00fff7cd ... OK Using PowerPC 44x Platform machine description Linux version 2.6.32.7-00007-g08eba26 (stefan@stefan-desktop) (gcc version 4.2. 2) #1 Thu Feb 4 18:54:16 CET 2010 Found initrd at 0xdfca6000:0xdfe4bdc1 Zone PFN ranges: DMA 0x00000000 -> 0x00020000 Normal 0x00020000 -> 0x00020000 HighMem 0x00020000 -> 0x00020000 Movable zone start PFN for each node early_node_map[1] active PFN ranges 0: 0x00000000 -> 0x00020000 MMU: Allocated 1088 bytes of context maps for 255 contexts Built 1 zonelists in Zone order, mobility grouping on. Total pages: 130048 Kernel command line: root=/dev/ram rw ip=192.168.100.6:192.168.1.1:192.168.1.25 4:255.255.0.0:canyonlands:eth0:off panic=1 console=ttyS0,115200 PID hash table entries: 2048 (order: 1, 8192 bytes) Dentry cache hash table entries: 65536 (order: 6, 262144 bytes) Inode-cache hash table entries: 32768 (order: 5, 131072 bytes) Memory: 513536k/524288k available (3880k kernel code, 10428k reserved, 172k dat a, 134k bss, 168k init) Kernel virtual memory layout: * 0xfffef000..0xfffff000 : fixmap * 0xffc00000..0xffe00000 : highmem PTEs * 0xffa00000..0xffc00000 : consistent mem * 0xffa00000..0xffa00000 : early ioremap * 0xe1000000..0xffa00000 : vmalloc & ioremap SLUB: Genslabs=11, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1 Hierarchical RCU implementation. NR_IRQS:512 UIC0 (32 IRQ sources) at DCR 0xc0 UIC1 (32 IRQ sources) at DCR 0xd0 UIC2 (32 IRQ sources) at DCR 0xe0 UIC3 (32 IRQ sources) at DCR 0xf0 clocksource: timebase mult[3c0000] shift[22] registered Mount-cache hash table entries: 512 NET: Registered protocol family 16 256k L2-cache enabled PCIE0: Port disabled via device-tree PCIE1: Checking link... PCIE1: No device detected. PCI host bridge /plb/pciex@d20000000 (primary) ranges: MEM 0x0000000e80000000..0x0000000effffffff -> 0x0000000080000000 MEM 0x0000000f00100000..0x0000000f001fffff -> 0x0000000000000000 IO 0x0000000f80010000..0x0000000f8001ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000f00100000 4xx PCI DMA offset set to 0x00000000 /plb/pciex@d20000000: Legacy ISA memory support enabled PCIE1: successfully set as root-complex PCI host bridge /plb/pci@c0ec00000 (primary) ranges: MEM 0x0000000d80000000..0x0000000dffffffff -> 0x0000000080000000 MEM 0x0000000c0ee00000..0x0000000c0eefffff -> 0x0000000000000000 IO 0x0000000c08000000..0x0000000c0800ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000c0ee00000 4xx PCI DMA offset set to 0x00000000 /plb/pci@c0ec00000: Legacy ISA memory support enabled PCI: Probing PCI hardware PCI: Hiding 4xx host bridge resources 0000:80:00.0 pci 0000:80:00.0: PCI bridge, secondary bus 0000:81 pci 0000:80:00.0: IO window: disabled pci 0000:80:00.0: MEM window: disabled pci 0000:80:00.0: PREFETCH window: disabled bio: create slab at 0 vgaarb: loaded SCSI subsystem initialized usbcore: registered new interface driver usbfs usbcore: registered new interface driver hub usbcore: registered new device driver usb Switching to clocksource timebase NET: Registered protocol family 2 IP route cache hash table entries: 4096 (order: 2, 16384 bytes) TCP established hash table entries: 16384 (order: 5, 131072 bytes) TCP bind hash table entries: 16384 (order: 4, 65536 bytes) TCP: Hash tables configured (established 16384 bind 16384) TCP reno registered NET: Registered protocol family 1 RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. Trying to unpack rootfs image as initramfs... rootfs image is not initramfs (no cpio magic); looks like an initrd Freeing initrd memory: 1687k freed JFFS2 version 2.2. (NAND) © 2001-2006 Red Hat, Inc. msgmni has been set to 1006 io scheduler noop registered io scheduler anticipatory registered (default) io scheduler deadline registered io scheduler cfq registered Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled serial8250.0: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550A console [ttyS0] enabled serial8250.0: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550A serial8250.0: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550A serial8250.0: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550A 4ef600300.serial: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550 4ef600400.serial: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550 4ef600500.serial: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550 4ef600600.serial: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550 brd: module loaded sata-dwc sata-dwc.0: id 0, controller version 1.82 sata-dwc sata-dwc.0: DMA initialized sata-dwc sata-dwc.0: **** No neg speed (nothing attached?) scsi0 : sata-dwc ata1: SATA max UDMA/133 irq 23 4cc000000.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank Amd/Fujitsu Extended Query Table at 0x0040 4cc000000.nor_flash: CFI does not contain boot bank location. Assuming top. number of CFI chips: 1 cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness. RedBoot partition parsing not available Creating 7 MTD partitions on "4cc000000.nor_flash": 0x000000000000-0x0000001e0000 : "kernel" 0x0000001e0000-0x000000200000 : "dtb" 0x000000200000-0x000001600000 : "ramdisk" 0x000001600000-0x000001a00000 : "jffs2" 0x000001a00000-0x000003f60000 : "user" 0x000003f60000-0x000003fa0000 : "env" 0x000003fa0000-0x000004000000 : "u-boot" NAND device: Manufacturer ID: 0x20, Chip ID: 0xf1 (ST Micro NAND 128MiB 3,3V 8- bit) Scanning device for bad blocks Creating 2 MTD partitions on "4e0000000.ndfc.nand": 0x000000000000-0x000000100000 : "u-boot" 0x000000000000-0x000003f00000 : "user" e1000e: Intel(R) PRO/1000 Network Driver - 1.0.2-k2 e1000e: Copyright (c) 1999-2008 Intel Corporation. PPC 4xx OCP EMAC driver, version 3.54 MAL v2 /plb/mcmal, 2 TX channels, 16 RX channels ZMII /plb/opb/emac-zmii@ef600d00 initialized RGMII /plb/opb/emac-rgmii@ef601500 initialized with MDIO support TAH /plb/opb/emac-tah@ef601350 initialized TAH /plb/opb/emac-tah@ef601450 initialized /plb/opb/emac-rgmii@ef601500: input 0 in RGMII mode eth0: EMAC-0 /plb/opb/ethernet@ef600e00, MAC 00:10:ec:01:08:84 eth0: found Generic MII PHY (0x00) /plb/opb/emac-rgmii@ef601500: input 1 in RGMII mode eth1: EMAC-1 /plb/opb/ethernet@ef600f00, MAC 00:10:ec:81:08:84 eth1: found Generic MII PHY (0x01) ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver ppc-of-ehci 4bffd0400.ehci: OF EHCI ppc-of-ehci 4bffd0400.ehci: new USB bus registered, assigned bus number 1 ppc-of-ehci 4bffd0400.ehci: irq 38, io mem 0x4bffd0400 ppc-of-ehci 4bffd0400.ehci: USB 2.0 started, EHCI 1.00 usb usb1: New USB device found, idVendor=1d6b, idProduct=0002 usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb1: Product: OF EHCI usb usb1: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ehci_hcd usb usb1: SerialNumber: PPC-OF USB usb usb1: configuration #1 chosen from 1 choice hub 1-0:1.0: USB hub found hub 1-0:1.0: 1 port detected ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver ppc-of-ohci 4bffd0000.usb: OF OHCI ppc-of-ohci 4bffd0000.usb: new USB bus registered, assigned bus number 2 ppc-of-ohci 4bffd0000.usb: irq 39, io mem 0x4bffd0000 usb usb2: New USB device found, idVendor=1d6b, idProduct=0001 usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb2: Product: OF OHCI usb usb2: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ohci_hcd usb usb2: SerialNumber: PPC-OF USB usb usb2: configuration #1 chosen from 1 choice hub 2-0:1.0: USB hub found hub 2-0:1.0: 1 port detected Initializing USB Mass Storage driver... usbcore: registered new interface driver usb-storage USB Mass Storage support registered. usbcore: registered new interface driver libusual dwc_otg: version 2.60a 22-NOV-2006 dwc_otg: Shared Tx FIFO mode dwc_otg: Using DMA mode dwc_otg dwc_otg.0: DWC OTG Controller dwc_otg dwc_otg.0: new USB bus registered, assigned bus number 3 dwc_otg dwc_otg.0: irq 28, io mem 0x00000000 dwc_otg: Init: Port Power? op_state=4 usb usb3: New USB device found, idVendor=1d6b, idProduct=0002 usb usb3: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb3: Product: DWC OTG Controller usb usb3: Manufacturer: Linux 2.6.32.7-00007-g08eba26 dwc_otg_hcd usb usb3: SerialNumber: dwc_otg.0 usb usb3: configuration #1 chosen from 1 choice hub 3-0:1.0: USB hub found hub 3-0:1.0: 1 port detected i2c /dev entries driver ibm-iic 4ef600700.i2c: using standard (100 kHz) mode rtc-m41t80 0-0068: chip found, driver version 0.05 rtc-m41t80 0-0068: rtc core: registered m41t80 as rtc0 ibm-iic 4ef600800.i2c: using standard (100 kHz) mode ad7414 0-0048: chip found TCP cubic registered NET: Registered protocol family 17 rtc-m41t80 0-0068: setting system clock to 2010-02-08 16:34:03 UTC (1265646843) usb 1-1: new high speed USB device using ppc-of-ehci and address 2 usb 1-1: New USB device found, idVendor=1370, idProduct=2168 usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-1: Product: Memorybird P usb 1-1: Manufacturer: FSC usb 1-1: SerialNumber: 10005383BB000032 usb 1-1: configuration #1 chosen from 1 choice scsi1 : SCSI emulation for USB Mass Storage devices eth0: link is down IP-Config: Complete: device=eth0, addr=192.168.100.6, mask=255.255.0.0, gw=192.168.1.254, host=canyonlands, domain=, nis-domain=(none), bootserver=192.168.1.1, rootserver=192.168.1.1, rootpath= RAMDISK: gzip image found at block 0 VFS: Mounted root (ext2 filesystem) on device 1:0. Freeing unused kernel memory: 168k init init started: BusyBox v1.7.1 (2008-04-01 21:48:01 MEST) starting pid 944, tty '': '/etc/rc.sh' starting pid 950, tty '': '/bin/sh' ~ # starting pid 949, tty '': '/bin/application' ### Application running ... ~ # 7.7. Standalone Operation with Ramdisk Image When your application development is completed, you usually will want to run your Embedded System standalone, i. e. independent from external resources like NFS filesystems. Instead of mounting the root filesystem from a remote server you can use a compressed ramdisk image, which is stored in flash memory and loaded into RAM when the system boots. Ramdisk images for tests can be found in the [780]ftp://ftp.denx.de/pub/LinuxPPC/usr/src/SELF/images/ directories. Load the ramdisk image into RAM and write it to flash as follows: => setenv ramdisk_addr 0xFC200000 => setenv ram_ws 0x100000 => protect off 0xFC200000 0xFC3FFFFF Un-Protected 16 sectors => erase 0xFC200000 0xFC3FFFFF ................ done Erased 16 sectors => tftp 0x100000 /tftpboot/duts/canyonlands/uRamdisk Using ppc_4xx_eth0 device TFTP from server 192.168.1.1; our IP address is 192.168.100.6 Filename '/tftpboot/duts/canyonlands/uRamdisk'. Load address: 0x100000 Loading: ################################################################# ##################################################### done Bytes transferred = 1728001 (1a5e01 hex) => iminfo ${ram_ws} ## Checking Image at 00100000 ... Legacy image found Image Name: Simple Embedded Linux Framework Created: 2008-04-01 19:52:43 UTC Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 1727937 Bytes = 1.6 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => cp.b ${ram_ws} ${ramdisk_addr} ${filesize} Copy to Flash... done => iminfo ${ramdisk_addr} ## Checking Image at fc200000 ... Legacy image found Image Name: Simple Embedded Linux Framework Created: 2008-04-01 19:52:43 UTC Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 1727937 Bytes = 1.6 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK => setenv ram_ws => saveenv Saving Environment to Flash... Un-Protected 1 sectors Un-Protected 1 sectors Erasing Flash... . done Erased 1 sectors Writing to Flash... done Protected 1 sectors Protected 1 sectors => To tell the Linux kernel to use the ramdisk image as root filesystem you have to modify the command line arguments passed to the kernel, and pass ramdisk image address as the second argument to the bootm command (first argument is the memory address of the Linux kernel image, the third one is the memory address of the flattened device tree blob): => run flash_self ## Booting kernel from Legacy Image at fc000000 ... Image Name: Linux-2.6.32.7-00007-g08eba26 Created: 2010-02-04 17:54:22 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1958545 Bytes = 1.9 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Loading init Ramdisk from Legacy Image at fc200000 ... Image Name: Simple Embedded Linux Framework Created: 2008-04-01 19:52:43 UTC Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 1727937 Bytes = 1.6 MB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Flattened Device Tree blob at fc1e0000 Booting using the fdt blob at 0xfc1e0000 Uncompressing Kernel Image ... OK Loading Ramdisk to 1fca6000, end 1fe4bdc1 ... OK Loading Device Tree to 00ffa000, end 00fff7cd ... OK Using PowerPC 44x Platform machine description Linux version 2.6.32.7-00007-g08eba26 (stefan@stefan-desktop) (gcc version 4.2. 2) #1 Thu Feb 4 18:54:16 CET 2010 Found initrd at 0xdfca6000:0xdfe4bdc1 Zone PFN ranges: DMA 0x00000000 -> 0x00020000 Normal 0x00020000 -> 0x00020000 HighMem 0x00020000 -> 0x00020000 Movable zone start PFN for each node early_node_map[1] active PFN ranges 0: 0x00000000 -> 0x00020000 MMU: Allocated 1088 bytes of context maps for 255 contexts Built 1 zonelists in Zone order, mobility grouping on. Total pages: 130048 Kernel command line: root=/dev/ram rw ip=192.168.100.6:192.168.1.1:192.168.1.25 4:255.255.0.0:canyonlands:eth0:off panic=1 console=ttyS0,115200 PID hash table entries: 2048 (order: 1, 8192 bytes) Dentry cache hash table entries: 65536 (order: 6, 262144 bytes) Inode-cache hash table entries: 32768 (order: 5, 131072 bytes) Memory: 513536k/524288k available (3880k kernel code, 10428k reserved, 172k dat a, 134k bss, 168k init) Kernel virtual memory layout: * 0xfffef000..0xfffff000 : fixmap * 0xffc00000..0xffe00000 : highmem PTEs * 0xffa00000..0xffc00000 : consistent mem * 0xffa00000..0xffa00000 : early ioremap * 0xe1000000..0xffa00000 : vmalloc & ioremap SLUB: Genslabs=11, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1 Hierarchical RCU implementation. NR_IRQS:512 UIC0 (32 IRQ sources) at DCR 0xc0 UIC1 (32 IRQ sources) at DCR 0xd0 UIC2 (32 IRQ sources) at DCR 0xe0 UIC3 (32 IRQ sources) at DCR 0xf0 clocksource: timebase mult[3c0000] shift[22] registered Mount-cache hash table entries: 512 NET: Registered protocol family 16 256k L2-cache enabled PCIE0: Port disabled via device-tree PCIE1: Checking link... PCIE1: No device detected. PCI host bridge /plb/pciex@d20000000 (primary) ranges: MEM 0x0000000e80000000..0x0000000effffffff -> 0x0000000080000000 MEM 0x0000000f00100000..0x0000000f001fffff -> 0x0000000000000000 IO 0x0000000f80010000..0x0000000f8001ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000f00100000 4xx PCI DMA offset set to 0x00000000 /plb/pciex@d20000000: Legacy ISA memory support enabled PCIE1: successfully set as root-complex PCI host bridge /plb/pci@c0ec00000 (primary) ranges: MEM 0x0000000d80000000..0x0000000dffffffff -> 0x0000000080000000 MEM 0x0000000c0ee00000..0x0000000c0eefffff -> 0x0000000000000000 IO 0x0000000c08000000..0x0000000c0800ffff -> 0x0000000000000000 Removing ISA hole at 0x0000000c0ee00000 4xx PCI DMA offset set to 0x00000000 /plb/pci@c0ec00000: Legacy ISA memory support enabled PCI: Probing PCI hardware PCI: Hiding 4xx host bridge resources 0000:80:00.0 pci 0000:80:00.0: PCI bridge, secondary bus 0000:81 pci 0000:80:00.0: IO window: disabled pci 0000:80:00.0: MEM window: disabled pci 0000:80:00.0: PREFETCH window: disabled bio: create slab at 0 vgaarb: loaded SCSI subsystem initialized usbcore: registered new interface driver usbfs usbcore: registered new interface driver hub usbcore: registered new device driver usb Switching to clocksource timebase NET: Registered protocol family 2 IP route cache hash table entries: 4096 (order: 2, 16384 bytes) TCP established hash table entries: 16384 (order: 5, 131072 bytes) TCP bind hash table entries: 16384 (order: 4, 65536 bytes) TCP: Hash tables configured (established 16384 bind 16384) TCP reno registered NET: Registered protocol family 1 RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. Trying to unpack rootfs image as initramfs... rootfs image is not initramfs (no cpio magic); looks like an initrd Freeing initrd memory: 1687k freed JFFS2 version 2.2. (NAND) © 2001-2006 Red Hat, Inc. msgmni has been set to 1006 io scheduler noop registered io scheduler anticipatory registered (default) io scheduler deadline registered io scheduler cfq registered Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled serial8250.0: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550A console [ttyS0] enabled serial8250.0: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550A serial8250.0: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550A serial8250.0: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550A 4ef600300.serial: ttyS0 at MMIO 0x4ef600300 (irq = 19) is a 16550 4ef600400.serial: ttyS1 at MMIO 0x4ef600400 (irq = 20) is a 16550 4ef600500.serial: ttyS2 at MMIO 0x4ef600500 (irq = 29) is a 16550 4ef600600.serial: ttyS3 at MMIO 0x4ef600600 (irq = 21) is a 16550 brd: module loaded sata-dwc sata-dwc.0: id 0, controller version 1.82 sata-dwc sata-dwc.0: DMA initialized sata-dwc sata-dwc.0: **** No neg speed (nothing attached?) scsi0 : sata-dwc ata1: SATA max UDMA/133 irq 23 4cc000000.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank Amd/Fujitsu Extended Query Table at 0x0040 4cc000000.nor_flash: CFI does not contain boot bank location. Assuming top. number of CFI chips: 1 cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness. RedBoot partition parsing not available Creating 7 MTD partitions on "4cc000000.nor_flash": 0x000000000000-0x0000001e0000 : "kernel" 0x0000001e0000-0x000000200000 : "dtb" 0x000000200000-0x000001600000 : "ramdisk" 0x000001600000-0x000001a00000 : "jffs2" 0x000001a00000-0x000003f60000 : "user" 0x000003f60000-0x000003fa0000 : "env" 0x000003fa0000-0x000004000000 : "u-boot" NAND device: Manufacturer ID: 0x20, Chip ID: 0xf1 (ST Micro NAND 128MiB 3,3V 8- bit) Scanning device for bad blocks Creating 2 MTD partitions on "4e0000000.ndfc.nand": 0x000000000000-0x000000100000 : "u-boot" 0x000000000000-0x000003f00000 : "user" e1000e: Intel(R) PRO/1000 Network Driver - 1.0.2-k2 e1000e: Copyright (c) 1999-2008 Intel Corporation. PPC 4xx OCP EMAC driver, version 3.54 MAL v2 /plb/mcmal, 2 TX channels, 16 RX channels ZMII /plb/opb/emac-zmii@ef600d00 initialized RGMII /plb/opb/emac-rgmii@ef601500 initialized with MDIO support TAH /plb/opb/emac-tah@ef601350 initialized TAH /plb/opb/emac-tah@ef601450 initialized /plb/opb/emac-rgmii@ef601500: input 0 in RGMII mode eth0: EMAC-0 /plb/opb/ethernet@ef600e00, MAC 00:10:ec:01:08:84 eth0: found Generic MII PHY (0x00) /plb/opb/emac-rgmii@ef601500: input 1 in RGMII mode eth1: EMAC-1 /plb/opb/ethernet@ef600f00, MAC 00:10:ec:81:08:84 eth1: found Generic MII PHY (0x01) ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver ppc-of-ehci 4bffd0400.ehci: OF EHCI ppc-of-ehci 4bffd0400.ehci: new USB bus registered, assigned bus number 1 ppc-of-ehci 4bffd0400.ehci: irq 38, io mem 0x4bffd0400 ppc-of-ehci 4bffd0400.ehci: USB 2.0 started, EHCI 1.00 usb usb1: New USB device found, idVendor=1d6b, idProduct=0002 usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb1: Product: OF EHCI usb usb1: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ehci_hcd usb usb1: SerialNumber: PPC-OF USB usb usb1: configuration #1 chosen from 1 choice hub 1-0:1.0: USB hub found hub 1-0:1.0: 1 port detected ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver ppc-of-ohci 4bffd0000.usb: OF OHCI ppc-of-ohci 4bffd0000.usb: new USB bus registered, assigned bus number 2 ppc-of-ohci 4bffd0000.usb: irq 39, io mem 0x4bffd0000 usb usb2: New USB device found, idVendor=1d6b, idProduct=0001 usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb2: Product: OF OHCI usb usb2: Manufacturer: Linux 2.6.32.7-00007-g08eba26 ohci_hcd usb usb2: SerialNumber: PPC-OF USB usb usb2: configuration #1 chosen from 1 choice hub 2-0:1.0: USB hub found hub 2-0:1.0: 1 port detected Initializing USB Mass Storage driver... usbcore: registered new interface driver usb-storage USB Mass Storage support registered. usbcore: registered new interface driver libusual dwc_otg: version 2.60a 22-NOV-2006 dwc_otg: Shared Tx FIFO mode dwc_otg: Using DMA mode dwc_otg dwc_otg.0: DWC OTG Controller dwc_otg dwc_otg.0: new USB bus registered, assigned bus number 3 dwc_otg dwc_otg.0: irq 28, io mem 0x00000000 dwc_otg: Init: Port Power? op_state=4 usb usb3: New USB device found, idVendor=1d6b, idProduct=0002 usb usb3: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb3: Product: DWC OTG Controller usb usb3: Manufacturer: Linux 2.6.32.7-00007-g08eba26 dwc_otg_hcd usb usb3: SerialNumber: dwc_otg.0 usb usb3: configuration #1 chosen from 1 choice hub 3-0:1.0: USB hub found hub 3-0:1.0: 1 port detected i2c /dev entries driver ibm-iic 4ef600700.i2c: using standard (100 kHz) mode rtc-m41t80 0-0068: chip found, driver version 0.05 rtc-m41t80 0-0068: rtc core: registered m41t80 as rtc0 ibm-iic 4ef600800.i2c: using standard (100 kHz) mode ad7414 0-0048: chip found TCP cubic registered NET: Registered protocol family 17 rtc-m41t80 0-0068: setting system clock to 2010-02-08 16:34:03 UTC (1265646843) usb 1-1: new high speed USB device using ppc-of-ehci and address 2 usb 1-1: New USB device found, idVendor=1370, idProduct=2168 usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 usb 1-1: Product: Memorybird P usb 1-1: Manufacturer: FSC usb 1-1: SerialNumber: 10005383BB000032 usb 1-1: configuration #1 chosen from 1 choice scsi1 : SCSI emulation for USB Mass Storage devices eth0: link is down IP-Config: Complete: device=eth0, addr=192.168.100.6, mask=255.255.0.0, gw=192.168.1.254, host=canyonlands, domain=, nis-domain=(none), bootserver=192.168.1.1, rootserver=192.168.1.1, rootpath= RAMDISK: gzip image found at block 0 VFS: Mounted root (ext2 filesystem) on device 1:0. Freeing unused kernel memory: 168k init init started: BusyBox v1.7.1 (2008-04-01 21:48:01 MEST) starting pid 944, tty '': '/etc/rc.sh' starting pid 950, tty '': '/bin/sh' ~ # starting pid 949, tty '': '/bin/application' ### Application running ... ~ # 8. Building and Using Modules This section still needs to be written (this is a wiki, so please feel free to contribute!). In the meantime, please refer to file [781]Documentation/kbuild/modules.txt in the Linux source tree. * [782]9. Advanced Topics + [783]9.1. Flash Filesystems o [784]9.1.1. Memory Technology Devices o [785]9.1.2. Journalling Flash File System o [786]9.1.3. Second Version of JFFS o [787]9.1.4. Compressed ROM Filesystem + [788]9.2. The TMPFS Virtual Memory Filesystem o [789]9.2.1. Mount Parameters o [790]9.2.2. Kernel Support for tmpfs o [791]9.2.3. Usage of tmpfs in Embedded Systems + [792]9.3. Adding Swap Space + [793]9.4. Splash Screen Support in Linux + [794]9.5. Root File System: Design and Building o [795]9.5.1. Root File System on a Ramdisk o [796]9.5.2. Root File System on a JFFS2 File System o [797]9.5.3. Root File System on a cramfs File System o [798]9.5.4. Root File System on a Read-Only ext2 File System o [799]9.5.5. Root File System on a Flash Card o [800]9.5.6. Root File System in a Read-Only File in a FAT File System + [801]9.6. Root File System Selection + [802]9.7. Overlay File Systems + [803]9.8. The Persistent RAM File system (PRAMFS) o [804]9.8.1. Mount Parameters o [805]9.8.2. Example 9. Advanced Topics This section lists some advanced topics of interest to users of U-Boot and Linux. 9.1. Flash Filesystems 9.1.1. Memory Technology Devices All currently available flash filesystems are based on the Memory Technology Devices MTD layer, so you must enable (at least) the following configuration options to get flash filesystem support in your system: CONFIG_MTD=y CONFIG_MTD_PARTITIONS=y CONFIG_MTD_CHAR=y CONFIG_MTD_BLOCK=y CONFIG_MTD_CFI=y CONFIG_MTD_GEN_PROBE=y CONFIG_MTD_CFI_AMDSTD=y CONFIG_MTD_ROM=y CONFIG_MTD_canyonlands=y ALERT! Note: this configuration uses CFI conformant AMD flash chips; you may need to adjust these settings on other boards. The partition layout of the flash devices is contained in the flat device tree for the system (see [806]13.1. Flat Device Tree). Informational messages of the MTD subsystem can be found in the Linux bootlog, i.e. see section [807]7.5.1. Bootlog of tftp'd Linux kernel with Root Filesystem over NFS. One can discover this information in a running system using the proc filesystem: -bash-3.2# cat /proc/mtd dev: size erasesize name mtd0: 001e0000 00020000 "kernel" mtd1: 00020000 00020000 "dtb" mtd2: 01400000 00020000 "ramdisk" mtd3: 00400000 00020000 "jffs2" mtd4: 02560000 00020000 "user" mtd5: 00040000 00020000 "env" mtd6: 00060000 00020000 "u-boot" mtd7: 00100000 00020000 "u-boot" mtd8: 03f00000 00020000 "user" -bash-3.2# Now we can run some basic tests to verify that the flash driver routines and the partitioning works as expected: -bash-3.2# xxd /dev/mtd3 | head -4 0000000: ffff ffff ffff ffff ffff ffff ffff ffff ................ 0000010: ffff ffff ffff ffff ffff ffff ffff ffff ................ 0000020: ffff ffff ffff ffff ffff ffff ffff ffff ................ 0000030: ffff ffff ffff ffff ffff ffff ffff ffff ................ -bash-3.2# In the hex-dumps of the [808]MTD devices you can identify some strings that verify that we indeed see an U-Boot environment, a Linux kernel, a ramdisk image and an empty partition to play wih. The last output shows the partition to be empty. We can try write some data into it: -bash-3.2# date > /tmp/tempfile -bash-3.2# dd if=/dev/zero of=/tmp/tempfile bs=1 count=4096 seek=50 4096+0 records in 4096+0 records out 4096 bytes (4.1 kB) copied, 0.0107987 s, 379 kB/s -bash-3.2# dd if=/tmp/tempfile of=/dev/mtd3 bs=4096 count=1 1+0 records in 1+0 records out 4096 bytes (4.1 kB) copied, 0.0179843 s, 228 kB/s -bash-3.2# head -1 /dev/mtd3 Mon Feb 8 16:36:04 CET 2010 -bash-3.2# dd if=/tmp/tempfile of=/dev/mtd3 bs=4096 count=1 1+0 records in 1+0 records out 4096 bytes (4.1 kB) copied, 0.0179315 s, 228 kB/s -bash-3.2# As you can see it worked the first time. When we tried to write the (new date) again, we got an error. The reason is that the date has changed (probably at least the seconds) and flash memory cannot be simply overwritten - it has to be erased first. You can use the eraseall Linux commands to erase a whole [809]MTD partition: -bash-3.2# flash_eraseall /dev/mtd3 Erasing 128 Kibyte @ 0 -- 0 % complete.Erasing 128 Kibyte @ 20000 -- 3 % comp lete.Erasing 128 Kibyte @ 40000 -- 6 % complete.Erasing 128 Kibyte @ 60000 -- 9 % complete.Erasing 128 Kibyte @ 80000 -- 12 % complete.Erasing 128 Kibyte @ a0000 -- 15 % complete.Erasing 128 Kibyte @ c0000 -- 18 % complete.Erasing 128 Kibyte @ e0000 -- 21 % complete.Erasing 128 Kibyte @ 100000 -- 25 % complete.Er asing 128 Kibyte @ 120000 -- 28 % complete.Erasing 128 Kibyte @ 140000 -- 31 % complete.Erasing 128 Kibyte @ 160000 -- 34 % complete.Erasing 128 Kibyte @ 1800 00 -- 37 % complete.Erasing 128 Kibyte @ 1a0000 -- 40 % complete.Erasing 128 Ki byte @ 1c0000 -- 43 % complete.Erasing 128 Kibyte @ 1e0000 -- 46 % complete.Era sing 128 Kibyte @ 200000 -- 50 % complete.Erasing 128 Kibyte @ 220000 -- 53 % c omplete.Erasing 128 Kibyte @ 240000 -- 56 % complete.Erasing 128 Kibyte @ 26000 0 -- 59 % complete.Erasing 128 Kibyte @ 280000 -- 62 % complete.Erasing 128 Kib yte @ 2a0000 -- 65 % complete.Erasing 128 Kibyte @ 2c0000 -- 68 % complete.Eras ing 128 Kibyte @ 2e0000 -- 71 % complete.Erasing 128 Kibyte @ 300000 -- 75 % co mplete.Erasing 128 Kibyte @ 320000 -- 78 % complete.Erasing 128 Kibyte @ 340000 -- 81 % complete.Erasing 128 Kibyte @ 360000 -- 84 % complete.Erasing 128 Kiby te @ 380000 -- 87 % complete.Erasing 128 Kibyte @ 3a0000 -- 90 % complete.Erasi ng 128 Kibyte @ 3c0000 -- 93 % complete.Erasing 128 Kibyte @ 3e0000 -- 96 % com plete. -bash-3.2# date > /tmp/tempfile -bash-3.2# dd if=/dev/zero of=/tmp/tempfile bs=1 count=4096 seek=50 4096+0 records in 4096+0 records out 4096 bytes (4.1 kB) copied, 0.01086 s, 377 kB/s -bash-3.2# dd if=/tmp/tempfile of=/dev/mtd3 bs=4096 count=1 1+0 records in 1+0 records out 4096 bytes (4.1 kB) copied, 0.0180026 s, 228 kB/s -bash-3.2# head -1 /dev/mtd3 Mon Feb 8 16:35:35 CET 2010 -bash-3.2# We have now sufficient proof that the [810]MTD layer is working as expected, so we can try creating a flash filesystem. 9.1.2. Journalling Flash File System At the moment it seems that the Journalling Flash File System JFFS is the best choice for filesystems in flash memory of embedded devices. You must enable the following configuration options to get [811]JFFS support in your system: CONFIG_JFFS_FS=y CONFIG_JFFS_FS_VERBOSE=0 If the flash device is erased, we can simply mount it, and the creation of the [812]JFFS filesystem is performed automagically. TIP Note: For simple accesses like direct read or write operations or erasing you use the character device interface (/dev/mtd*) of the [813]MTD layer, while for filesystem operations like mounting we must use the block device interface (/dev/mtdblock*). # eraseall /dev/mtd2 Erased 4096 Kibyte @ 0 -- 100% complete. # mount -t jffs /dev/mtdblock2 /mnt # mount /dev/root on / type nfs (rw,v2,rsize=4096,wsize=4096,hard,udp,nolock,addr=10.0. 0.2) proc on /proc type proc (rw) devpts on /dev/pts type devpts (rw) /dev/mtdblock2 on /mnt type jffs (rw) # df Filesystem 1k-blocks Used Available Use% Mounted on /dev/root 2087212 1232060 855152 60% / /dev/mtdblock2 3584 0 3584 0% /mnt Now you can access the files in the [814]JFFS filesystem in the /mnt directory. 9.1.3. Second Version of [815]JFFS Probably even more interesting for embedded systems is the second version of [816]JFFS, JFFS2, since it not only fixes a few design issues with [817]JFFS, but also adds transparent compression, so that you can save a lot of precious flash memory. The mkfs.jffs2 tool is used to create a JFFS2 filesystem image; it populates the image with files from a given directory. For instance, to create a JFFS2 image for a flash partition of 3 MB total size and to populate it with the files from the /tmp/flashtools directory you would use: # mkfs.jffs2 --pad=3145728 --eraseblock=262144 \ --root=/tmp/flashtools/ --output image.jffs2 # eraseall /dev/mtd4 Erased 3072 Kibyte @ 0 -- 100% complete. \# dd if=image.jffs2 of=/dev/mtd4 bs=256k 12+0 records in 12+0 records out # mount -t jffs2 /dev/mtdblock4 /mnt # df /mnt Filesystem 1k-blocks Used Available Use% Mounted on /dev/mtdblock4 3072 2488 584 81% /mnt ALERT! Note: Especially when you are running time-critical applications on your system you should carefully study if the behaviour of the flash filesystem might have any negative impact on your application. After all, a flash device is not a normal harddisk. This is especially important when your flash filesystem gets full; JFFS2 acts a bit weird then: * You will note that an increasing amount of [818]CPU time is spent by the filesystem's garbage collection kernel thread. * Access times to the files on the flash filesystem may increase drastically. * Attempts to truncate a file (to free space) or to rename it may fail: ... # cp /bin/bash file cp: writing `file': No space left on device # >file bash: file: No space left on device # mv file foo mv: cannot create regular file `foo': No space left on device You will have to use rm to actually delete a file in this situation. This is especially critical when you are using the flash filesystem to store log files: when your application detects some abnormal condition and produces lots of log messages (which usually are especially important in this situation) the filesystem may fill up and cause extreme long delays - if your system crashes, the most important messages may never be logged at all. 9.1.4. Compressed ROM Filesystem In some cases it is sufficent to have read-only access to some files, and if the files are big enough it becomes desirable to use some method of compression. The Compressed ROM Filesystem CramFs might be a solution here. ALERT! Please note that [819]CramFs has - beside the fact that it is a read-only filesystem - some severe limitations (like missing support for timestamps, hard links, and 16/32 bit uid/gids), but there are many situations in Embedded Systems where it's still useful. To create a [820]CramFs filesystem a special tool mkcramfs is used to create a file which contains the [821]CramFs image. Note that the [822]CramFs filesystem can be written and read only by kernels with PAGE_CACHE_SIZE == 4096, and some versions of the mkcramfs program may have other restrictions like that the filesystem must be written and read with architectures of the same endianness. Especially the endianness requirement makes it impossible to build the [823]CramFs image on x86 PC host when you want to use it on a [824]Power Architecture® target. The endianness problem has been fixed in the version of mkcramfs that comes with the [825]ELDK. In some cases you can use a target system running with root filesystem mounted over NFS to create the [826]CramFs image on the native system and store it to flash for further use. ALERT! Note: The normal version of the mkcramfs program tries to initialize some entries in the filesystem's superblock with random numbers by reading /dev/random; this may hang permanently on your target because there is not enough input (like mouse movement) to the entropy pool. You may want to use a modified version of mkcramfs which does not depend on /dev/random. To create a [827]CramFs image, you put all files you want in the filesystem into one directory, and then use the mkcramfs= program as follows: $ mkdir /tmp/test $ cp ... /tmp/test $ du -sk /tmp/test 64 /tmp/test $ mkcramfs /tmp/test test.cramfs.img Super block: 76 bytes erase eraseall mkfs.jffs lock unlock Directory data: 176 bytes -54.96% (-4784 bytes) erase -55.46% (-5010 bytes) eraseall -51.94% (-8863 bytes) mkfs.jffs -58.76% (-4383 bytes) lock -59.68% (-4215 bytes) unlock Everything: 24 kilobytes $ ls -l test.cramfs.img -rw-r--r-- 1 wd users 24576 Nov 10 23:44 test.cramfs.img As you can see, the [828]CramFs image test.cramfs.img takes just 24 kB, while the input directory contained 64 kB of data. Savings of some 60% like in this case are typical [829]CramFs. Now we write the [830]CramFs image to a partition in flash and test it: # cp test.cramfs.img /dev/mtd3 # mount -t cramfs /dev/mtdblock3 /mnt # mount /dev/root on / type nfs (rw,v2,rsize=4096,wsize=4096,hard,udp,nolock,addr=10.0. 0.2) proc on /proc type proc (rw) devpts on /dev/pts type devpts (rw) /dev/mtdblock3 on /mnt type cramfs (rw) # ls -l /mnt total 54 -rwxr-xr-x 1 wd users 8704 Jan 9 16:32 erase -rwxr-xr-x 1 wd users 9034 Jan 1 01:00 eraseall -rwxr-xr-x 1 wd users 7459 Jan 1 01:00 lock -rwxr-xr-x 1 wd users 17063 Jan 1 01:00 mkfs.jffs -rwxr-xr-x 1 wd users 7063 Jan 1 01:00 unlock Note that all the timestamps in the [831]CramFs filesyste are bogus, and so is for instance the output of the df command for such filesystems: # df /mnt Filesystem 1k-blocks Used Available Use% Mounted on /dev/mtdblock3 0 0 0 - /mnt 9.2. The TMPFS Virtual Memory Filesystem The tmpfs filesystem, formerly known as shmfs, is a filesystem keeping all files in virtual memory. Everything in tmpfs is temporary in the sense that no files will be created on any device. If you unmount a tmpfs instance, everything stored therein is lost. tmpfs puts everything into the kernel internal caches and grows and shrinks to accommodate the files it contains and is able to swap unneeded pages out to swap space. It has maximum size limits which can be adjusted on the fly via 'mount -o remount ...' If you compare it to ramfs (which was the template to create tmpfs) you gain swapping and limit checking. Another similar thing is the RAM disk (/dev/ram*), which simulates a fixed size hard disk in physical RAM, where you have to create an ordinary filesystem on top. Ramdisks cannot swap and you do not have the possibility to resize them. 9.2.1. Mount Parameters tmpfs has a couple of mount options: * size: The limit of allocated bytes for this tmpfs instance. The default is half of your physical RAM without swap. If you oversize your tmpfs instances the machine will deadlock since the OOM handler will not be able to free that memory. * nr_blocks: The same as size, but in blocks of PAGECACHE_SIZE. * nr_inodes: The maximum number of inodes for this instance. The default is half of the number of your physical RAM pages. These parameters accept a suffix k, m or g for kilo, mega and giga and can be changed on remount. To specify the initial root directory you can use the following mount options: * mode: The permissions as an octal number * uid: The user id * gid: The group id These options do not have any effect on remount. You can change these parameters with chmod(1), chown(1) and chgrp(1) on a mounted filesystem. So the following mount command will give you a tmpfs instance on /mytmpfs which can allocate 12MB of RAM/SWAP and it is only accessible by root. mount -t tmpfs -o size=12M,mode=700 tmpfs /mytmpfs 9.2.2. Kernel Support for tmpfs In order to use a tmpfs filesystem, the CONFIG_TMPFS option has to be enabled for your kernel configuration. It can be found in the Filesystems configuration group. You can simply check if a running kernel supports tmpfs by searching the contents of /proc/fileysystems: bash# grep tmpfs /proc/filesystems nodev tmpfs bash# 9.2.3. Usage of tmpfs in Embedded Systems In embedded systems tmpfs is very well suited to provide read and write space (e.g. /tmp and /var) for a read-only root file system such as [832]CramFs described in section [833]9.1.4. Compressed ROM Filesystem. One way to achieve this is to use symbolic links. The following code could be part of the startup file /etc/rc.sh of the read-only ramdisk: #!/bin/sh ... # Won't work on read-only root: mkdir /tmpfs mount -t tmpfs tmpfs /tmpfs mkdir /tmpfs/tmp /tmpfs/var # Won't work on read-only root: ln -sf /tmpfs/tmp /tmpfs/var / ... The commented out sections will of course fail on a read-only root filesystem, so you have to create the /tmpfs mount-point and the symbolic links in your root filesystem beforehand in order to successfully use this setup. 9.3. Adding Swap Space If you are running out of system RAM, you can add virtual memory by using swap space. If you reserved a swap partition on your disk drive, you have to initialize it once using the mkswap command: # fdisk -l /dev/hda Disk /dev/hda: 16 heads, 63 sectors, 1575 cylinders Units = cylinders of 1008 * 512 bytes Device Boot Start End Blocks Id System /dev/hda1 1 5 2488+ 83 Linux /dev/hda2 6 10 2520 83 Linux /dev/hda3 11 141 66024 82 Linux swap /dev/hda4 142 1575 722736 83 Linux # mkswap /dev/hda3 Setting up swapspace version 1, size = 67604480 bytes Then, to activate it, you use the swapon command like this: # free total used free shared buffers cached Mem: 14628 14060 568 8056 100 11664 -/+ buffers/cache: 2296 12332 Swap: 0 0 0 # free total used free shared buffers cached Mem: 14628 14060 568 8056 100 11664 -/+ buffers/cache: 2296 12332 Swap: 0 0 0 # swapon /dev/hda3 Adding Swap: 66016k swap-space (priority -2) # free total used free shared buffers cached Mem: 14628 14084 544 8056 100 11648 -/+ buffers/cache: 2336 12292 Swap: 66016 0 66016 If you forgot to reserve (sufficient) space in a separate partition on your disk, you can still use an ordinary file for swap space. You only have to create a file of appropriate size, and initialize it as follows: # mount /dev/hda4 /mnt # df Filesystem 1k-blocks Used Available Use% Mounted on /dev/root 2087212 1378824 708388 67% / /dev/hda4 711352 20 675196 1% /mnt # dd if=/dev/zero of=/mnt/swapfile bs=1024k count=64 64+0 records in 64+0 records out # mkswap /mnt/swapfile Setting up swapspace version 1, size = 67104768 bytes Then activate it: # free total used free shared buffers cached Mem: 14628 14084 544 6200 96 11788 -/+ buffers/cache: 2200 12428 Swap: 0 0 0 # swapon /mnt/swapfile Adding Swap: 65528k swap-space (priority -3) # free total used free shared buffers cached Mem: 14628 14084 544 6200 96 11752 -/+ buffers/cache: 2236 12392 Swap: 65528 0 65528 9.4. Splash Screen Support in Linux To complement the [834]U-Boot Splash Screen feature the new configuration option "CONFIG_FB_PRE_INIT_FB" was added to the Linux kernel. This allows the Linux kernel to skip certain parts of the framebuffer initialization and to reuse the framebuffer contents that was set up by the U-Boot firmware. This allows to have an image displayed nearly immediately after power-on, so the delay needed to boot the Linux kernel is masked to the user. The current implementation has some limitations: * We did not succeed in reusing the previously allocated framebuffer contents directly. Instead, Linux will allocate a new framebuffer, copy the contents, and then switch the display. This adds a minimal delay to the boot time, but is otherwise invisible to the user. * Linux manages its own colormap, and we considered it too much effort to keep the same settings as used by U-Boot. Instead we use the "trick" that U-Boot will fill the color map table backwards (top down). This works pretty well for images which use no more than 200...225 colors. If the images uses more colors, a bad color mapping may result. TIP We strongly recommend to convert all images that will be loaded as Linux splash screens to use no more than 225 colors. The "ppmquant" tool can be used for this purpose (see [835]Bitmap Support in U-Boot for details). * Usually there will be a Linux device driver that is used to adjust the brightness and contrast of the display. When this driver starts, a visible change of brightness will happen if the default settings as used by U-Boot differ. TIP We recommend to store settings of brightness and contrast in U-Boot environment variables that can be shared between U-Boot and Linux. This way it is possible (assuming adequate driver support) to adjust the display settings correctly already in U-Boot and thus to avoid any flicker of the display when Linux takes over control. 9.5. Root File System: Design and Building It is not an easy task to design the root file system for an embedded system. There are three major problems to be solved: 1. what to put in it 2. which file system type to use 3. where to store and how to boot it For now we will assume that the contents of the root file system is aready known; for example, it is given to us as a directory tree or a tarball which contains all the required files. We will also assume that our system is a typical resource-limited embedded system so we will especially look for solutions where the root file system can be stored on on-board flash memory or other flash memory based devices like CompactFlash or SD cards, MMC or USB memory sticks. A widespread approach to build a root file system is to use some Linux distribution (like the [836]ELDK) and to remove things not needed. This approach may be pretty common, but it is almost always terribly wrong. You also don't build a family home by taking a skyscraper and removing parts. Like a house, a root file system should be built bottom up, starting from scratch and adding things you know you need. Never add anything where you don't exactly know what it's needed for. But our focus here is on the second item: the options we have for chosing a file system type and the consequences this has. In all cases we will base our experiments on the same content of the root filesystem; we use the images of the [837]SELF (Simple Embedded Linux Framework) that come with the [838]ELDK. In a first step we will transform the [839]SELF images into a tarball to meet the requirements mentioned above: In a [840]ELDK installation, the [841]SELF images can be found in the /opt/eldk//images/ directory. There is already a compressed ramdisk image in this directory, which we will use (ramdisk_image.gz): 1. Uncompress ramdisk image: bash$ gzip -d -c -v /opt/eldk/ppc_8xx/images/ramdisk_image.gz >/tmp/ramdisk_ima ge /opt/eldk/ppc_8xx/images/ramdisk_image.gz: 61.4% ALERT! Note: The following steps require root permissions! 2. Mount ramdisk image: bash# mount -o loop /tmp/ramdisk_image /mnt/tmp 3. Create tarball; to avoid the need for root permissions in the following steps we don't include the device files in our tarball: bash# cd /mnt/tmp bash# tar -zc --exclude='dev/*' -f /tmp/rootfs.tar.gz * 4. Instead, we create a separate tarball which contains only the device entries so we can use them when necessary (with cramfs): bash# tar -zcf /tmp/devices.tar.gz dev/ bash# cd /tmp 5. Unmount ramdisk image: bash# umount /mnt/tmp We will use the /tmp/rootfs.tar.gz tarball as master file in all following experiments. 9.5.1. Root File System on a Ramdisk Ram disks are used very often to hold the root file system of embedded systems. They have several advantages: * well-known * well-supported by the Linux kernel * simple to build * simple to use - you can even combine the ramdisk with the Linux kernel into a single image file * RAM based, thus pretty fast * writable file system * original state of file system after each reboot = easy recovery from accidental or malicious data corruption etc. On the other hand, there are several disadvantages, too: * big memory footprint: you always have to load the complete filesystem into RAM, even if only small parts of are actually used * slow boot time: you have to load (and uncompress) the whole image before the first application process can start * only the whole image can be replaced (not individual files) * additional storage needed for writable persistent data Actually there are only very few situations where a ramdisk image is the optimal solution. But because they are so easy to build and use we will discuss them here anyway. In almost all cases you will use an ext2 file system in your ramdisk image. The following steps are needed to create it: 1. Create a directory tree with the content of the target root filesystem. We do this by unpacking our master tarball: $ mkdir rootfs $ cd rootfs $ tar zxf /tmp/rootfs.tar.gz 2. We use the genext2fs tool to create the ramdisk image as this allows to use a simple text file to describe which devices shall be created in the generated file system image. That means that no root permissions are required at all. We use the following device table rootfs_devices.tab: # /dev d 755 0 0 - - - - - /dev/console c 640 0 0 5 1 - - - /dev/fb0 c 640 0 0 29 0 - - - /dev/full c 640 0 0 1 7 - - - /dev/hda b 640 0 0 3 0 - - - /dev/hda b 640 0 0 3 1 1 1 16 /dev/kmem c 640 0 0 1 2 - - - /dev/mem c 640 0 0 1 1 - - - /dev/mtd c 640 0 0 90 0 0 2 16 /dev/mtdblock b 640 0 0 31 0 0 1 16 /dev/mtdr c 640 0 0 90 1 0 2 16 /dev/nftla b 640 0 0 93 0 - - - /dev/nftla b 640 0 0 93 1 1 1 8 /dev/nftlb b 640 0 0 93 16 - - - /dev/nftlb b 640 0 0 93 17 1 1 8 /dev/null c 640 0 0 1 3 - - - /dev/ptyp c 640 0 0 2 0 0 1 10 /dev/ptypa c 640 0 0 2 10 - - - /dev/ptypb c 640 0 0 2 11 - - - /dev/ptypc c 640 0 0 2 12 - - - /dev/ptypd c 640 0 0 2 13 - - - /dev/ptype c 640 0 0 2 14 - - - /dev/ptypf c 640 0 0 2 15 - - - /dev/ram b 640 0 0 1 0 0 1 2 /dev/ram b 640 0 0 1 1 - - - /dev/rtc c 640 0 0 10 135 - - - /dev/tty c 640 0 0 4 0 0 1 4 /dev/tty c 640 0 0 5 0 - - - /dev/ttyS c 640 0 0 4 64 0 1 8 /dev/ttyp c 640 0 0 3 0 0 1 10 /dev/ttypa c 640 0 0 3 10 - - - /dev/ttypb c 640 0 0 3 11 - - - /dev/ttypc c 640 0 0 3 12 - - - /dev/ttypd c 640 0 0 3 13 - - - /dev/ttype c 640 0 0 3 14 - - - /dev/ttypf c 640 0 0 3 15 - - - /dev/zero c 640 0 0 1 5 - - - A description of the format of this table is part of the manual page for the genext2fs tool, genext2fs(8). 3. We can now create an ext2 file system image using the genext2fs tool: $ ROOTFS_DIR=rootfs # directory with root file system content $ ROOTFS_SIZE=3700 # size of file system image $ ROOTFS_FREE=100 # free space wanted $ ROOTFS_INODES=380 # number of inodes $ ROOTFS_DEVICES=rootfs_devices.tab # device description file $ ROOTFS_IMAGE=ramdisk.img # generated file system image $ genext2fs -U \ -d ${ROOTFS_DIR} \ -D ${ROOTFS_DEVICES} \ -b ${ROOTFS_SIZE} \ -r ${ROOTFS_FREE} \ -i ${ROOTFS_INODES} \ ${ROOTFS_IMAGE} 4. Compress the file system image: $ gzip -v9 ramdisk.img rootfs.img: 55.6% -- replaced with ramdisk.img.gz 5. Create an U-Boot image file from it: $ mkimage -T ramdisk -C gzip -n 'Test Ramdisk Image' \ > -d ramdisk.img.gz uRamdisk Image Name: Test Ramdisk Image Created: Sun Jun 12 16:58:06 2005 Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 1618547 Bytes = 1580.61 kB = 1.54 MB Load Address: 0x00000000 Entry Point: 0x00000000 We now have a root file system image uRamdisk that can be used with U-Boot. 9.5.2. Root File System on a JFFS2 File System JFFS2 (Journalling Flash File System version 2) was specifically designed for use on flash memory devices in embedded systems. It is a log-structured file system which means that it is robust against loss of power, crashes or other unorderly shutdowns of the system ("robust" means that data that is just being written when the system goes down may be lost, but the file system itself does not get corrupted and the system can be rebootet without need for any kind of file system check). Some of the advantages of using JFFS2 as root file system in embedded systems are: * file system uses compression, thus making efficient use of flash memory * log-structured file system, thus robust against unorderly shutdown * flash sector wear-leveling * writable flash file system Disadvantages are: * long mount times (especially older versions) * slow when reading: files to be read get uncompressed on the fly which eats [842]CPU cycles and takes time * slow when writing: files to be written get compressed, which eats [843]CPU cycles and takes time, but it may even take much longer until data gets actually stored in flash if the file system becomes full and blocks must be erased first or - even worse - if garbage collection becomes necessary * The garbage collector thread may run at any time, consuming [844]CPU cycles and blocking accesses to the file system. Despite the aforementioned disadvantages, systems using a JFFS2 based root file system are easy to build, make efficient use of the available resources and can run pretty reliably. To create a JFFS2 based root file system please proceed as follows: 1. Create a directory tree with the content of the target root filesystem. We do this by unpacking our master tarball: $ mkdir rootfs $ cd rootfs $ tar zxf /tmp/rootfs.tar.gz 2. We can now create a JFFS2 file system image using the mkfs.jffs2 tool: $ ROOTFS_DIR=rootfs # directory with root file system content $ ROOTFS_EBSIZE=0x20000 # erase block size of flash memory $ ROOTFS_ENDIAN=b # target system is big endian $ ROOTFS_DEVICES=rootfs_devices.tab # device description file $ ROOTFS_IMAGE=jffs2.img # generated file system image $ mkfs.jffs2 -U \ -d ${ROOTFS_DIR} \ -D ${ROOTFS_DEVICES} \ -${ROOTFS_ENDIAN} \ -e ${ROOTFS_EBSIZE} \ -o ${ROOTFS_IMAGE} mkfs.jffs2: skipping device_table entry '/dev': no parent directory! ALERT! Note: When you intend to write the JFFS2 file system image to a NAND flash device, you should also add the "-n" (or "--no-cleanmarkers") option, as cleanmarkers are not needed then. When booting the Linux kernel prints the following messages showing the default partition map which is used for the flash memory on the TQM8xxL boards: TQM flash bank 0: Using static image partition definition Creating 7 MTD partitions on "TQM8xxL0": 0x00000000-0x00040000 : "u-boot" 0x00040000-0x00100000 : "kernel" 0x00100000-0x00200000 : "user" 0x00200000-0x00400000 : "initrd" 0x00400000-0x00600000 : "cramfs" 0x00600000-0x00800000 : "jffs" 0x00400000-0x00800000 : "big_fs" We use U-Boot to load and store the JFFS2 image into the last partition and set up the Linux boot arguments to use this as root device: 1. Erase flash: => era 40400000 407FFFFF ................. done Erased 35 sectors 2. Download JFFS2 image: => tftp 100000 /tftpboot/TQM860L/jffs2.img Using FEC ETHERNET device TFTP from server 192.168.3.1; our IP address is 192.168.3.80 Filename '/tftpboot/TQM860L/jffs2.img'. Load address: 0x100000 Loading: ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ######## done Bytes transferred = 2033888 (1f08e0 hex) 3. Copy image to flash: => cp.b 100000 40400000 ${filesize} Copy to Flash... done 4. set up boot arguments to use flash partition 6 as root device: => setenv mtd_args setenv bootargs root=/dev/mtdblock6 rw rootfstype=jffs2 => printenv addip addip=setenv bootargs ${bootargs} ip=${ipaddr}:${serverip}:${gatewayip}:${netma sk}:${hostname}:${netdev}:off panic=1 => setenv flash_mtd 'run mtd_args addip;bootm ${kernel_addr}' => run flash_mtd Using FEC ETHERNET device TFTP from server 192.168.3.1; our IP address is 192.168.3.80 Filename '/tftpboot/TQM860L/uImage'. Load address: 0x200000 Loading: ################################################################# ################################################################# ########### done Bytes transferred = 719233 (af981 hex) ## Booting image at 40040000 ... Image Name: Linux-2.4.25 Created: 2005-06-12 16:32:24 UTC Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 782219 Bytes = 763.9 kB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK Uncompressing Kernel Image ... OK Linux version 2.4.25 (wd@xpert) (gcc version 3.3.3 (DENX ELDK 3.1.1 3.3.3-9)) # 1 Sun Jun 12 18:32:18 MEST 2005 On node 0 totalpages: 4096 zone(0): 4096 pages. zone(1): 0 pages. zone(2): 0 pages. Kernel command line: root=/dev/mtdblock6 rw rootfstype=jffs2 ip=192.168.3.80:19 2.168.3.1::255.255.255.0:tqm860l:eth1:off panic=1 Decrementer Frequency = 187500000/60 Calibrating delay loop... 49.86 BogoMIPS ... NET4: Unix domain sockets 1.0/SMP for Linux NET4.0. VFS: Mounted root (jffs2 filesystem). Freeing unused kernel memory: 56k init BusyBox v0.60.5 (2005.03.07-06:54+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. # ### Application running ... # mount rootfs on / type rootfs (rw) /dev/mtdblock6 on / type jffs2 (rw) /proc on /proc type proc (rw) # df / Filesystem 1k-blocks Used Available Use% Mounted on rootfs 4096 2372 1724 58% / 9.5.3. Root File System on a cramfs File System cramfs is a compressed, read-only file system. Advantages are: * file system uses compression, thus making efficient use of flash memory * Allows for quick boot times as only used files get loaded and uncompressed Disadvantages are: * only the whole image can be replaced (not individual files) * additional storage needed for writable persistent data * mkcramfs tool does not support device table, so we need root permissions to create the required device files To create a cramfs based root file system please proceed as follows: 1. Create a directory tree with the content of the target root filesystem. We do this by unpacking our master tarball: $ mkdir rootfs $ cd rootfs $ tar -zxf /tmp/rootfs.tar.gz 2. Create the required device files. We do this here by unpacking a special tarball which holds only the device file entries. ALERT! Note: this requires root permissions! # cd rootfs # tar -zxf /tmp/devices.tar.gz 3. Many tools require some storage place in a filesystem, so we must provide at least one (small) writable filesystem. For all data which may be lost when the system goes down, a "tmpfs" filesystem is the optimal choice. To create such a writable tmpfs filesystem we add the following lines to the /etc/rc.sh script: # mount TMPFS because root-fs is readonly /bin/mount -t tmpfs -o size=2M tmpfs /tmpfs Some tools require write permissions on some device nodes (for example, to change ownership and permissions), or dynamically (re-) create such files (for example, /dev/log which is usually a Unix Domain socket). The files are placed in a writable filesystem; in the root filesystem symbolic links are used to point to their new locations: dev/ptyp0 -> /tmpfs/dev/ptyp0 dev/ttyp0 -> /tmpfs/dev/ttyp0 dev/ptyp1 -> /tmpfs/dev/ptyp1 dev/ttyp1 -> /tmpfs/dev/ttyp1 dev/ptyp2 -> /tmpfs/dev/ptyp2 dev/ttyp2 -> /tmpfs/dev/ttyp2 dev/ptyp3 -> /tmpfs/dev/ptyp3 dev/ttyp3 -> /tmpfs/dev/ttyp3 dev/ptyp4 -> /tmpfs/dev/ptyp4 dev/ttyp4 -> /tmpfs/dev/ttyp4 dev/ptyp5 -> /tmpfs/dev/ptyp5 dev/ttyp5 -> /tmpfs/dev/ttyp5 dev/ptyp6 -> /tmpfs/dev/ptyp6 dev/ttyp6 -> /tmpfs/dev/ttyp6 dev/ptyp7 -> /tmpfs/dev/ptyp7 dev/ttyp7 -> /tmpfs/dev/ttyp7 dev/ptyp8 -> /tmpfs/dev/ptyp8 dev/ttyp8 -> /tmpfs/dev/ttyp8 dev/ptyp9 -> /tmpfs/dev/ptyp9 dev/ttyp9 -> /tmpfs/dev/ttyp9 dev/ptypa -> /tmpfs/dev/ptypa dev/ttypa -> /tmpfs/dev/ttypa dev/ptypb -> /tmpfs/dev/ptypb dev/ttypb -> /tmpfs/dev/ttypb dev/ptypc -> /tmpfs/dev/ptypc dev/ttypc -> /tmpfs/dev/ttypc dev/ptypd -> /tmpfs/dev/ptypd dev/ttypd -> /tmpfs/dev/ttypd dev/ptype -> /tmpfs/dev/ptype dev/ttype -> /tmpfs/dev/ttype dev/ptypf -> /tmpfs/dev/ptypf dev/ttypf -> /tmpfs/dev/ttypf tmp -> /tmpfs/tmp var -> /tmpfs/var dev/log -> /var/log/log In case you use dhclient also: etc/dhclient.conf -> /tmpfs/var/lib/dhclient.conf etc/resolv.conf -> /tmpfs/var/lib/resolv.conf To place the corresponding directories and device files in the tmpfs file system, the following code is added to the /etc/rc.sh script: mkdir -p /tmpfs/tmp /tmpfs/dev \ /tmpfs/var/lib/dhcp /tmpfs/var/lock /tmpfs/var/run while read name minor do mknod /tmpfs/dev/ptyp$name c 2 $minor mknod /tmpfs/dev/ttyp$name c 3 $minor done <<__EOD__ 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 a 10 b 11 c 12 d 13 e 14 f 15 __EOD__ chmod 0666 /tmpfs/dev/* 4. We can now create a cramfs file system image using the mkcramfs tool: $ ROOTFS_DIR=rootfs # directory with root file system content $ ROOTFS_ENDIAN="-r" # target system has reversed (big) endianes s $ ROOTFS_IMAGE=cramfs.img # generated file system image PATH=/opt/eldk/usr/bin:$PATH mkcramfs ${ROOTFS_ENDIAN} ${DEVICES} ${ROOTFS_DIR} ${ROOTFS_IMAGE} Swapping filesystem endian-ness bin dev etc ... -48.78% (-86348 bytes) in.ftpd -46.02% (-16280 bytes) in.telnetd -45.31% (-74444 bytes) xinetd Everything: 1864 kilobytes Super block: 76 bytes CRC: c166be6d warning: gids truncated to 8 bits. (This may be a security concern.) 5. We can use the same setup as before for the JFFS2 filesystem, just changing the bootargument to "rootfstype=cramfs" 9.5.4. Root File System on a Read-Only ext2 File System When storing the root file system in on-board flash memory it seems only natural to look for special flash filesystems like JFFS2, or for other file system types that are designed for such environments like cramfs. It seems to be a bad idea to use a standard ext2 file system because it contains neither any type of wear leveling which is needed for writable file systems in flash memory, nor is it robust against unorderly shutdowns. The situation changes if we use an ext2 file system which we mount read-only. Such a configuration can be very useful in some situations. Advantages: * very fast * low RAM memory footprint Disadvantages: * high flash memory footprint because no compression To create an ext2 image that can be used as a read-only root file system the following steps are necessary: 1. Create a directory tree with the content of the target root filesystem. We do this by unpacking our master tarball: $ mkdir rootfs $ cd rootfs $ tar -zxf /tmp/rootfs.tar.gz 2. Like with the cramfs root file system, we use "tmpfs" for cases where a writable file system is needed and add the following lines to the /etc/rc.sh script: # mount TMPFS because root-fs is readonly /bin/mount -t tmpfs -o size=2M tmpfs /tmpfs We also create the same symbolic links for device files that must be placed in a writable filesystem: dev/ptyp0 -> /tmpfs/dev/ptyp0 dev/ttyp0 -> /tmpfs/dev/ttyp0 dev/ptyp1 -> /tmpfs/dev/ptyp1 dev/ttyp1 -> /tmpfs/dev/ttyp1 dev/ptyp2 -> /tmpfs/dev/ptyp2 dev/ttyp2 -> /tmpfs/dev/ttyp2 dev/ptyp3 -> /tmpfs/dev/ptyp3 dev/ttyp3 -> /tmpfs/dev/ttyp3 dev/ptyp4 -> /tmpfs/dev/ptyp4 dev/ttyp4 -> /tmpfs/dev/ttyp4 dev/ptyp5 -> /tmpfs/dev/ptyp5 dev/ttyp5 -> /tmpfs/dev/ttyp5 dev/ptyp6 -> /tmpfs/dev/ptyp6 dev/ttyp6 -> /tmpfs/dev/ttyp6 dev/ptyp7 -> /tmpfs/dev/ptyp7 dev/ttyp7 -> /tmpfs/dev/ttyp7 dev/ptyp8 -> /tmpfs/dev/ptyp8 dev/ttyp8 -> /tmpfs/dev/ttyp8 dev/ptyp9 -> /tmpfs/dev/ptyp9 dev/ttyp9 -> /tmpfs/dev/ttyp9 dev/ptypa -> /tmpfs/dev/ptypa dev/ttypa -> /tmpfs/dev/ttypa dev/ptypb -> /tmpfs/dev/ptypb dev/ttypb -> /tmpfs/dev/ttypb dev/ptypc -> /tmpfs/dev/ptypc dev/ttypc -> /tmpfs/dev/ttypc dev/ptypd -> /tmpfs/dev/ptypd dev/ttypd -> /tmpfs/dev/ttypd dev/ptype -> /tmpfs/dev/ptype dev/ttype -> /tmpfs/dev/ttype dev/ptypf -> /tmpfs/dev/ptypf dev/ttypf -> /tmpfs/dev/ttypf tmp -> /tmpfs/tmp var -> /tmpfs/var dev/log -> /var/log/log In case you use dhclient also: etc/dhclient.conf -> /tmpfs/var/lib/dhclient.conf etc/resolv.conf -> /tmpfs/var/lib/resolv.conf To place the corresponding directories and device files in the tmpfs file system, the following code is added to the /etc/rc.sh script: mkdir -p /tmpfs/tmp /tmpfs/dev \ /tmpfs/var/lib/dhcp /tmpfs/var/lock /tmpfs/var/run while read name minor do mknod /tmpfs/dev/ptyp$name c 2 $minor mknod /tmpfs/dev/ttyp$name c 3 $minor done <<__EOD__ 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 a 10 b 11 c 12 d 13 e 14 f 15 __EOD__ chmod 0666 /tmpfs/dev/* 3. Like we did for the ramdisk, we now create an ext2 file system image using the genext2fs tool: $ ROOTFS_DIR=rootfs # directory with root file system content $ ROOTFS_SIZE=3700 # size of file system image $ ROOTFS_FREE=100 # free space wanted $ ROOTFS_INODES=380 # number of inodes $ ROOTFS_DEVICES=rootfs_devices.tab # device description file $ ROOTFS_IMAGE=ext2.img # generated file system image $ genext2fs -U \ -d ${ROOTFS_DIR} \ -D ${ROOTFS_DEVICES} \ -b ${ROOTFS_SIZE} \ -r ${ROOTFS_FREE} \ -i ${ROOTFS_INODES} \ ${ROOTFS_IMAGE} 4. We can again use the same setup as before for the JFFS2 filesystem, just changing the boot argument to "rootfstype=ext2" (or simply omit it completely as this is the default anyway), and we must change the "rw" argument into "ro" to mount our root file system really read-only: ... Linux version 2.4.25 (wd@xpert) (gcc version 3.3.3 (DENX ELDK 3.1.1 3.3.3-9)) # 1 Sun Jun 12 18:32:18 MEST 2005 On node 0 totalpages: 4096 zone(0): 4096 pages. zone(1): 0 pages. zone(2): 0 pages. Kernel command line: root=/dev/mtdblock6 ro rootfstype=ext2 ip=192.168.3.80:192 .168.3.1::255.255.255.0:tqm860l:eth1:off panic=1 Decrementer Frequency = 187500000/60 Calibrating delay loop... 49.86 BogoMIPS ... 9.5.5. Root File System on a Flash Card Using an ext2 file system on a flash memory card (like CompactFlash, SD, MMC or a USB memory stick) is standard technology. To avoid unnecessary flash wear it is a good idea to mount the root file system read-only, or at least using the "noatime" mount option. For our test we can use the "ext2.img" file from the previous step without changes: 1. In this test we use a standard CompactFlash card which comes with a single partition on it. We use U-Boot to copy the ext2 file system image into this partition: => tftp 100000 /tftpboot/TQM860L/ext2.img Using FEC ETHERNET device TFTP from server 192.168.3.1; our IP address is 192.168.3.80 Filename '/tftpboot/TQM860L/ext2.img'. Load address: 0x100000 Loading: ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ################################################################# ########################## done Bytes transferred = 3788800 (39d000 hex) => ide part Partition Map for IDE device 0 -- Partition Type: DOS Partition Start Sector Num Sectors Type 1 32 500704 6 => ide write 100000 20 1ce8 IDE write: device 0 block # 32, count 7400 ... 7400 blocks written: OK Note that the "ide write" command takes parameters as hex numbers, and the write count is in terms of disk blocks of 512 bytes each. So we have to use 0x20 for the starts sector of the first partition, and 3788800 / 512 = 7400 = 0x1CE8 for the block count. 2. We now prepare the Linux boot arguments to take this partition as read-only root device: => setenv cf_args setenv bootargs root=/dev/hda1 ro => setenv flash_cf 'run cf_args addip;bootm ${kernel_addr} - ${fdt_addr}' => setenv bootcmd run flash_cf 3. ...and boot the system: ... Linux version 2.4.25 (wd@xpert) (gcc version 3.3.3 (DENX ELDK 3.1.1 3.3.3-9)) # 1 Sun Jun 12 18:32:18 MEST 2005 On node 0 totalpages: 4096 zone(0): 4096 pages. zone(1): 0 pages. zone(2): 0 pages. Kernel command line: root=/dev/hda1 ro ip=192.168.3.80:192.168.3.1::255.255.255 .0:tqm860l:eth1:off panic=1 Decrementer Frequency = 187500000/60 Calibrating delay loop... 49.86 BogoMIPS ... 9.5.6. Root File System in a Read-Only File in a FAT File System This is a more complicated example that shows that - depending on project requirements - many other alternatives for chosing a root file system for your embedded system exist. The scenario is as follows: on your embedded device you use a cheap and popular storage medium like CompactFlash, MMC or SD cards or USB memory sticks to store both the Linux kernel and your root file system. You want to distribute software updates over the internet: your customers can download the file from your web site, or you sent the images by email. Your customers may use any flash card or memory stick they happen to find, so you have no information about brand or size of the storage device. Unfortunately most of your customers use Windows systems. And they don't want to be bothered with long instructions how to create special partitions on the storage device or how to write binary images or things like that. A simple "copy file" operation is nearly exhausting their capabilities. What to do? Well, if copying a file is all your customers can do we should not ask for more. Storage devices like CompactFlash cards etc. typically come with a single partition on it, which holds a FAT or VFAT file system. This cannot be used as a Linux root file system directly, so we have to use some trickery. Here is one possible solution: Your software distribution consistes of two files: The first file is the Linux kernel with a minimal ramdisk image attached (using the multi-file image format for U-Boot); U-Boot can load and boot such files from a FAT or VFAT file system. The second file is your root file system. For convenience and speed we use again an image of an ext2 file system. When Linux boots, it will initially use the attached ramdisk as root file system. The programs in this ramdisk will mount the FAT or VFAT file system - read-only. Then we can use a loop device (see losetup(8)) to associate the root file system image with a block device which can be used as a mount point. And finally we use pivot_root(8) to change the root file system to our image on the CF card. This sounds not so complicated, and actually it is quite simple once you understand what needs to be done. Here is a more detailed description: 1. The root file system image is easy: as mantioned before, we will use an ext2 file system image, and to avoid wearing the flash storage device we will use it in read-only mode - we did a read-only ext2 root file system image before, and here we can just re-use the existing image file. 2. The initial ramdisk image that performs the pivot_root step must be created from scratch, but we already know how to create ramdisk images, so we just have to figure out what to put in it. The most important tool here is nash, a script interpreter that was specifically designed for such purposes (see nash(8)). We don't need any additional tools, and if we use static linking, then the nash binary plus a small script to control it is all we need for our initial ramdisk. To be precise, we need a couple of (empty) directories (bin, dev, etc, lib, loopfs, mnt, proc, and sysroot), the bin/nash binary, the linuxrc script and a symbolic link sbin pointing to bin: drwxr-xr-x 2 wd users 4096 Apr 13 01:11 bin -rwxr-xr-x 1 wd users 469512 Apr 11 22:47 bin/nash drwxr-xr-x 2 wd users 4096 Apr 12 00:04 dev drwxr-xr-x 2 wd users 4096 Apr 12 00:04 etc drwxr-xr-x 2 wd users 4096 Apr 12 00:04 lib -rwxr-xr-x 1 wd users 511 Apr 13 01:28 linuxrc drwxr-xr-x 2 wd users 4096 Apr 12 00:04 loopfs drwxr-xr-x 2 wd users 4096 Apr 12 00:09 mnt drwxr-xr-x 2 wd users 4096 Apr 12 00:04 proc lrwxrwxrwx 1 wd users 3 Jun 12 18:54 sbin -> bin drwxr-xr-x 2 wd users 4096 Apr 12 00:04 sysroot 3. We also need only a minimal device table for creating the initial ramdisk: # /dev d 755 0 0 - - - - - /dev/console c 640 0 0 5 1 - - - /dev/hda b 640 0 0 3 0 - - - /dev/hda b 640 0 0 3 1 1 1 8 /dev/loop b 640 0 0 7 0 0 1 4 /dev/null c 640 0 0 1 3 - - - /dev/ram b 640 0 0 1 0 0 1 2 /dev/ram b 640 0 0 1 1 - - - /dev/tty c 640 0 0 4 0 0 1 4 /dev/tty c 640 0 0 5 0 - - - /dev/ttyS c 640 0 0 4 64 0 1 4 /dev/zero c 640 0 0 1 5 - - - 4. To create the initial ramdisk we perform the usual steps: $ INITRD_DIR=initrd $ INITRD_SIZE=490 $ INITRD_FREE=0 $ INITRD_INODES=54 $ INITRD_DEVICES=initrd_devices.tab $ INITRD_IMAGE=initrd.img $ genext2fs -U \ -d ${INITRD_DIR} \ -D ${INITRD_DEVICES} \ -b ${INITRD_SIZE} \ -r ${INITRD_FREE} \ -i ${INITRD_INODES} \ ${INITRD_IMAGE} $ gzip -v9 ${INITRD_IMAGE} The result is a really small (233 kB) compressed ramdisk image. 5. Assuming you already have your Linux kernel image, you can now use mkimage to build an U-Boot multi-file image that combines the Linux kernel and the initial ramdisk: $ LINUX_KERNEL=linuxppc_2_4_devel/arch/ppc/boot/images/vmlinux.gz $ mkimage -A ppc -O Linux -T multi -C gzip \ > -n 'Linux with Pivot Root Helper' \ > -d ${LINUX_KERNEL}:${INITRD_IMAGE}.gz linux.img Image Name: Linux with Pivot Root Helper Created: Mon Jun 13 01:48:11 2005 Image Type: PowerPC Linux Multi-File Image (gzip compressed) Data Size: 1020665 Bytes = 996.74 kB = 0.97 MB Load Address: 0x00000000 Entry Point: 0x00000000 Contents: Image 0: 782219 Bytes = 763 kB = 0 MB Image 1: 238433 Bytes = 232 kB = 0 MB The newly created file linux.img is the second image we have to copy to the CF card. We are done. But wait - one essential part was not mentioned yet: the linuxrc script in our initial ramdisk image which contains all the magic. This script is quite simple: #!/bin/nash echo Mounting /proc filesystem mount -t proc /proc /proc echo Creating block devices mkdevices /dev echo Creating root device mkrootdev /dev/root echo 0x0100 > /proc/sys/kernel/real-root-dev echo Mounting flash card mount -o noatime -t vfat /dev/hda1 /mnt echo losetup for filesystem image losetup /dev/loop0 /mnt/rootfs.img echo Mounting root filesystem image mount -o defaults --ro -t ext2 /dev/loop0 /sysroot echo Running pivot_root pivot_root /sysroot /sysroot/initrd umount /initrd/proc Let's go though it step by step: * The first line says that it's a script file for the /bin/nash interpreter. ALERT! Note: even if this file looks like a shell script it is NOT interpreted by a shell, but by the nash interpreter. For a complete list of available nash commands and their syntax please refer to the manual page, nash(8). * The first action is to mount the /proc pseudo file system which is needed to find out some required information. * Then we create block device entries for all partitions listed in /proc/partitions (mkdevices command). * In the next step a block device for our new root file system is created (mkrootdev command). * Then we mount the CF card. We assume that there is only a single partition on it (/dev/hda1) which is of type VFAT (which also will work with FAT file systems). These assumptions work fine with basicly all memory devices used under Windows. * We further assume that the file name of the root file system image on the CF card is "rootfs.img" - this file now gets mounted using a loop device (losetup and mount commands). * Our file system image, is now mounted on the /sysroot directory. In the last step we use pivot_root to make this the new root file system. * As a final cleanup we unmount the /proc file system which is not needed any more. There is one tiny flaw in this method: since we mount the CF card on a directory in the ramdisk to be able to access to root file system image. This means that we cannot unmount the CF card, which in turn prevents us from freeing the space for the inital ramdisk. The consequence is that you permanently lose approx. 450 kB of RAM for the ramdisk. [We could of course re-use this ramdisk space for temporary data, but such optimization is beyond the scope of this document.] And how does this work on our target? 1. First we copy the two images to the CF card; we do this on the target under Linux: bash-2.05b# fdisk -l /dev/hda Disk /dev/hda: 256 MB, 256376832 bytes 16 heads, 32 sectors/track, 978 cylinders Units = cylinders of 512 * 512 = 262144 bytes Device Boot Start End Blocks Id System /dev/hda1 * 1 978 250352 6 FAT16 bash-2.05b# mkfs.vfat /dev/hda1 mkfs.vfat 2.8 (28 Feb 2001) bash-2.05b# mount -t vfat /dev/hda1 /mnt bash-2.05b# cp -v linux.img rootfs.img /mnt/ `linux.img' -> `/mnt/linux.img' `rootfs.img' -> `/mnt/rootfs.img' bash-2.05b# ls -l /mnt total 4700 -rwxr--r-- 1 root root 1020729 Jun 14 05:36 linux.img -rwxr--r-- 1 root root 3788800 Jun 14 05:36 rootfs.img bash-2.05b# umount /mnt 2. We now prepare U-Boot to load the "uMulti" file (combined Linux kernel and initial ramdisk) from the CF card and boot it: => setenv fat_args setenv bootargs rw => setenv fat_boot 'run fat_args addip;fatload ide 0:1 200000 linux.img;bootm' => setenv bootcmd run fat_boot 3. And finally we try it out: U-Boot 1.1.3 (Jun 13 2005 - 02:24:00) CPU: XPC86xxxZPnnD4 at 50 MHz: 4 kB I-Cache 4 kB D-Cache FEC present Board: TQM860LDB0A3-T50.202 DRAM: 16 MB FLASH: 8 MB In: serial Out: serial Err: serial Net: SCC ETHERNET, FEC ETHERNET [PRIME] PCMCIA: 3.3V card found: Transcend 256M Fixed Disk Card IDE interface [silicon] [unique] [single] [sleep] [standby] [idle] [low power] Bus 0: OK Device 0: Model: Transcend 256M Firm: 1.1 Ser#: SSSC256M04Z27A25906T Type: Removable Hard Disk Capacity: 244.5 MB = 0.2 GB (500736 x 512) Type "run flash_nfs" to mount root filesystem over NFS Hit any key to stop autoboot: 0 reading linux.img 1025657 bytes read ## Booting image at 00200000 ... Image Name: Linux with Pivot Root Helper Created: 2005-06-13 0:32:41 UTC Image Type: PowerPC Linux Multi-File Image (gzip compressed) Data Size: 1025593 Bytes = 1001.6 kB Load Address: 00000000 Entry Point: 00000000 Contents: Image 0: 787146 Bytes = 768.7 kB Image 1: 238433 Bytes = 232.8 kB Verifying Checksum ... OK Uncompressing Multi-File Image ... OK Loading Ramdisk to 00f3d000, end 00f77361 ... OK Linux version 2.4.25 (wd@xpert) (gcc version 3.3.3 (DENX ELDK 3.1.1 3.3.3-9)) # 1 Mon Jun 13 02:32:10 MEST 2005 On node 0 totalpages: 4096 zone(0): 4096 pages. zone(1): 0 pages. zone(2): 0 pages. Kernel command line: rw ip=192.168.3.80:192.168.3.1::255.255.255.0:tqm860l:eth1 :off panic=1 Decrementer Frequency = 187500000/60 Calibrating delay loop... 49.86 BogoMIPS ... NET4: Unix domain sockets 1.0/SMP for Linux NET4.0. RAMDISK: Compressed image found at block 0 Freeing initrd memory: 232k freed VFS: Mounted root (ext2 filesystem). Red Hat nash version 4.1.18 starting Mounting /proc filesystem Creating block devices Creating root device Mounting flash card hda: hda1 hda: hda1 losetup for filesystem image Mounting root filesystem image Running pivot_root Freeing unused kernel memory: 60k init BusyBox v0.60.5 (2005.03.07-06:54+0000) Built-in shell (msh) Enter 'help' for a list of built-in commands. # ### Application running ... Kernel with a Flattened Device Tree Blob When booting an arch/powerpc kernel that requires a flattened device tree blob, the above procedure must be slightly modified. Namely, the multi-image file has to include the blob as the thrid image. Here's an example of the mkimage command to create it: mkimage -A ppc -O Linux -T multi -C gzip -n 'Kernel + Pivot Root Helper initrd + FDT blob' -d vmlinux.bin.gz:ramdisk_image.gz:canyonlands.dtb kernel+initrd+bl ob.img Image Name: Kernel + Pivot Root Helper initrd + FDT blob Created: Fri Sep 14 18:24:29 2007 Image Type: PowerPC Linux Multi-File Image (gzip compressed) Data Size: 2894576 Bytes = 2826.73 kB = 2.76 MB Load Address: 0x00000000 Entry Point: 0x00000000 Contents: Image 0: 1351205 Bytes = 1319 kB = 1 MB Image 1: 1531063 Bytes = 1495 kB = 1 MB Image 2: 12288 Bytes = 12 kB = 0 MB The newly created file kernel+initrd+blob.img needs to be copied to the CF card. 9.6. Root File System Selection Now we know several options for file systems we can use, and know how to create the corresponding images. But how can we decide which one to chose? For practical purposes in embedded systems the following criteria are often essential: * boot time (i. e. time needed from power on until application code is running) * flash memory footprint * RAM memory footprint * effects on software updates The following data was measured for the different configurations. All measurements were performed on the same TQM860L board (MPC860 [845]CPU at 50 MHz, 16 MB RAM, 8 MB flash, 256 MB CompactFlash card): [846]File System Type [847]Boot Time [848]Free Mem [849]Updates [850]while running ramdisk 16.3 sec 6.58 MB whole image yes JFFS2 21.4 sec 10.3 MB per file only non-active files cramfs 10.8 sec 10.3 MB whole image no ext2 (ro) 9.1 sec 10.8 MB whole image no ext2 on CF (ro) 9.3 sec 10.9 MB whole image no File on FAT fs 11.4 sec 7.8 MB whole image yes As you can see, the ramdisk solution is the worst of all in terms of RAM memory footprint; also it takes a pretty long time to boot. However, it is one of the few solutions that allow an in-situ update while the system is running. JFFS2 is easy to use as it's a writable file system but it takes a long time to boot. A read-only ext2 file system shines when boot time and RAM memory footprint are important; you pay for this with an increased flash memory footprint. External flash memory devices like CompactFlash cards or USB memory sticks can be cheap and efficient solutions especially when lots of data need to be stored or when easy update procedures are required. - 9.7. Overlay File Systems Introduction Overlay File Systems provide an interesting approach to several frequent problems in Embedded Systems. For example, mini_fo is a virtual kernel file system that can make read-only file systems writable. This is done by redirecting modifying operations to a writeable location called "storage directory", and leaving the original data in the "base directory" untouched. When reading, the file system merges the modifed and original data so that only the newest versions will appear. This occurs transparently to the user, who can access the data like on any other read-write file system. What it is good for? In embedded systems the main use of mini_fo is to overlay the root file system. This means it is mounted on top of the regular root file system, thereby allowing applications or users to transparently make modifications to it but redirecting these to a different location. Some examples of why this is usefull are explained in the following sections. Making a read-only root filesystem writeable Root file systems stored in flash are often read only, such as [851]cramfs or [852]read only ext2. While this offers major advantages in terms of speed and flash memory footprint, it nevertheless is often desireable to be able to modify the root file system, for example to * apply (small) software updates without having to burn a whole new root file system image to flash * make modifications during developement when frequent changes to the root file system occur. This can be achieved by mounting mini_fo on top of the root file system and using a (probably small) writeable partition as the storage file system. This could be either a JFFS2 flash file system, or during development even an external hard disk. This has the following advantages: * read-only file systems (fast, small memory footprint) can be used like persistent writable file systems (in contrast to a ramdisk) * slow flash journalling file systems with large flash memory footprint can be avoided. Non persistant changes Ramdisks are often used when the root file system needs to be modified non-persistantly. This works well, but downsides are the large RAM memory footprint and the time costly operation of copying the ramdisk into RAM during startup. These can be avoided by overlaying the root file system as in the previous example but with the difference that the [853]tmpfs file system is used as storage. Thus only modified files are stored in RAM, and can even be swapped out if neccessary. This saves boot time and RAM! Resetable changes Mini_fo can be easily used to implement a "reset to factory defaults" function by overlaying the default root file system. When configuration changes are made, these are automatically directed to the storage file system and take precedence over the original files. Now, to restore the system to factory defaults, all that needs to be done is delete the contents of the storage directory. This will remove all changes made to the root file system and return it to the original state. ALERT! Note: Deleting the contents of the storage directory should only be done when the overlay file system is unmounted. Examples Generally, there are two different ways of overlaying the root file system, which both make sense in different scenarios. Starting a single application in a chrooted overlayed environment This is easy. Let's assume "/" is the read-only root file system and /dev/mtdblock5 contains a small JFFS2 flash partition that shall be used to store modifications made by application "/usr/bin/autoPilot": # mount -t jffs2 /dev/mtdblock5 /tmp/sto # insmod mini_fo.o # mount -t mini_fo -o base=/,sto=/tmp/sto/ / /mnt/mini_fo/ # cd /mnt/mini_fo/ # chroot . /usr/bin/autoPilot The mini_fo file system is mounted with "/" as base directory, "/tmp/sto/" as storage directory to the mount point "/mnt/mini_fo". After that, chroot(1) is used to start the application with the new file system root "/mnt/mini_fo". All modifications made by the application will be stored to the JFFS2 file system in /tmp/sto. Starting the whole system system in chrooted overlayed environment This is more interesting, and a bit trickier, as mounting needs to be done during system startup after the root file system has been mounted, but before init is started. The best way to do this is to have a script that mounts the mini_fo file system on top of root and then starts init in the chrooted overlayed environment. For example assume the following script "overlay_init", stored in /sbin/: #!/bin/bash # # mount mini_fo overlay file system and execute init # # make sure these exist in the read-only file system STORAGE=/tmp/sto MOUNT_POINT=/mnt/mini_fo/ # mount tmpfs as storage file system with a maximum size of 32MB mount -t tmpfs -o rw,size=32M none $STORAGE /sbin/modprobe mini_fo mount -t mini_fo -o base=/,sto=$STORAGE / $MOUNT_POINT exec /usr/sbin/chroot $MOUNT_POINT /sbin/init echo "exec chroot failed, bad!" exec /bin/sh exit 1 Now its easy to choose between a mini_fo overlayed and the regular non overlayed system just by setting the "init" kernel parameter in the boot loader to "init=/sbin/overlay_init". Tips * pivot_root(1) can be used with chroot if there is need to access the original non overlayed root file system from the chrooted overlayed environment. Performance overhead The mini_fo file system is inserted as an additional layer between the VFS and the native file system, and thus creates some overhead that varies strongly depending of the operation performed. 1. modifying a regular file for the first time This results in a copy of the original file beeing created in the storage directory, that is then modified. Overhead depends on the size of the modified file. 2. Reading from files, creating new files, modifying already modified files These operations are passed directly through to the lower native layer, and only impose an overhead of 1-2%. Further information This section discusses how the mini_fo overlay file system can be used in embedded systems. More general information is available at the mini_fo project page: [854]http://www.denx.de/wiki/Know/MiniFOHome. 9.8. The Persistent RAM File system (PRAMFS) The pramfs file system supports persistent memory devices such as SRAM. Instead of having a block emulation layer over such a memory area and using a normal file system on top of that, pramfs seeks to induce minimal overhead in this situation. Most important in this respect is that the normal block layer caching of the Linux kernel is circumvented in pramfs. 9.8.1. Mount Parameters The most important parameters for normal usage are * physaddr: The physical address of the static memory. * init: When given, it will initialize the file system to that size. 9.8.2. Example We will show a sample usage of pramfs in this section using normal DRAM on a board with at least 256MB of memory. For pramfs we reserve the upper 32MB by appending mem=224M to the kernel command line. First off we generate some testdata on a persistent file system (/tmp) to demonstrate that pramfs survives a reboot (of course with power always applied to keep the DRAM refreshed): bash-3.00# dd if=/dev/urandom bs=1M count=8 of=/tmp/testdata 8+0 records in 8+0 records out bash-3.00# Next we mount the 32MB that we reserved and initialize it to be 32MB in size and copy the testfile. A final compare shows that the copy was indeed successful so we can reboot: bash-3.00# mount -t pramfs -o physaddr=0xe000000,init=0x2000000 none /mnt bash-3.00# cp /tmp/testdata /mnt bash-3.00# cmp /tmp/testdata /mnt/testdata bash-3.00# reboot Having rebooted (using mem=224M on the kernel command line again of course) we mount the file system but this time without the init parameter because it is preinitialized. We then check the contents again: bash-3.00# mount -t pramfs -o physaddr=0xe000000 none /mnt bash-3.00# ls /mnt testdata bash-3.00# cmp /tmp/testdata /mnt/testdata bash-3.00# * [855]10. Debugging + [856]10.1. Debugging of U-Boot o [857]10.1.1. Debugging of U-Boot Before Relocation o [858]10.1.2. Debugging of U-Boot After Relocation + [859]10.2. Linux Kernel Debugging o [860]10.2.1. Linux Kernel and Statically Linked Device Drivers o [861]10.2.2. Dynamically Loaded Device Drivers (Modules) o [862]10.2.3. GDB Macros to Simplify Module Loading + [863]10.3. GDB Startup File and Utility Scripts + [864]10.4. Tips and Tricks + [865]10.5. Application Debugging o [866]10.5.1. Local Debugging o [867]10.5.2. Remote Debugging + [868]10.6. Debugging with Graphical User Interfaces 10. Debugging The purpose of this document is not to provide an introduction into programming and debugging in general. We assume that you know how to use the GNU debugger [869]gdb and probably it's graphical frontends like [870]ddd. We also assume that you have access to adequate tools for your work, i. e. a BDI2000 [871]BDM/JTAG debugger. The following discussion assumes that the host name of your BDI2000 is bdi. Please note that there are several limitations in earlier versions of GDB. The version of GDB as distributed with the [872]ELDK contains several bug fixes and extensions. If you find that your GDB behaves differently, have a look at the GDB sources and patches that come with the [873]ELDK source. 10.1. Debugging of U-Boot When U-Boot starts it is running from ROM space. Running from flash would make it nearly impossible to read from flash while executing code from flash not to speak of updating the U-Boot image in flash itself. To be able to do just that, U-Boot relocates itself to RAM. We therefore have two phases with different program addresses. The following sections show how to debug U-Boot in both phases. 10.1.1. Debugging of U-Boot Before Relocation Before relocation, the addresses in the ELF file can be used without any problems, so debugging U-Boot in this phase with the BDI2000 is quite easy: bash[0]$ ${CROSS_COMPILE}gdb u-boot GNU gdb 5.1.1 Copyright 2002 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "--host=i386-redhat-linux --target=ppc-linux"... (gdb) target remote bdi:2001 Remote debugging using bdi:2001 0xfffffffc in ?? () (gdb) b cpu_init_f Breakpoint 1 at 0xfffd3310: file cpu_init.c, line 136. (gdb) c Continuing. Breakpoint 1, cpu_init_f () at cpu_init.c:136 136 asm volatile(" bl 0f" ::: "lr"); (gdb) s 137 asm volatile("0: mflr 3" ::: "r3"); (gdb) 138 asm volatile(" addi 4, 0, 14" ::: "r4"); (gdb) cpu_init_f is the first C function called from the code in start.C. 10.1.2. Debugging of U-Boot After Relocation For debugging U-Boot after relocation we need to know the address to which U-Boot relocates itself to. When no exotic features like PRAM are used, this address usually is - CONFIG_SYS_MONITOR_LEN. In our example with 16MB RAM and CONFIG_SYS_MONITOR_LEN = 192KB this yields the address 0x1000000 - 0x30000 = 0xFD0000. In other cases, check the source code, and apply some common sense. For example, on Power Architecture® we use "r2" to hold a pointer to the "global data" structure ("struct global_data"); this structure contains a field unsigned long reloc_off; /* Relocation Offset */ which is the offset between the image addresses in flash and in RAM. You can easily print this value in gdb like that: (gdb) print/x ((gd_t *)$r2)->reloc_off Then add this value to the value of TEXT_BASE as defined in your board's config.mk file, and you get the start address of the U-Boot image in RAM. With this knowledge, we can instruct gdb to forget the old symbol table and reload the symbols with our calculated offset: (gdb) symbol-file Discard symbol table from `/home/dzu/denx/cvs-trees/u-boot/u-boot'? (y or n) y No symbol file now. (gdb) add-symbol-file u-boot 0xfd0000 add symbol table from file "u-boot" at .text_addr = 0xfd0000 (y or n) y Reading symbols from u-boot...done. (gdb) b board_init_r Breakpoint 2 at 0xfd99ac: file board.c, line 533. (gdb) c Continuing. Breakpoint 2, board_init_r (id=0xfbb1f0, dest_addr=16495088) at board.c:533 533 { (gdb) board_init_r is the first C routine running in the newly relocated C friendly RAM environment. The simple example above relocates the symbols of only one section, .text. Other sections of the executable image (like .data, .bss, etc.) are not relocated and this prevents gdb from accessing static and global variables by name. See more sophisticated examples in section [874]10.3. GDB Startup File and Utility Scripts. 10.2. Linux Kernel Debugging 10.2.1. Linux Kernel and Statically Linked Device Drivers 10.2.2. Dynamically Loaded Device Drivers (Modules) First start GDB in the root directory of your Linux kernel, using the vmlinux kernel image as file to debug: bash$ cd bash$ ${CROSS_COMPILE}gdb vmlinux GNU gdb 5.1.1 Copyright 2002 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "--host=i386-redhat-linux --target=ppc-linux". (gdb) Now attach to the target and start execution with the commands: (gdb) target remote bdi:2001 Remote debugging using bdi:2001 0x00000100 in ?? () (gdb) c Continuing. Now the target should boot Linux as usual. Next you need to load your kernel module on the target: bash# insmod -m ex_sw.o Sections: Size Address Align .this 00000060 cf030000 2**2 .text 000002f4 cf030060 2**2 .rodata 00000134 cf030354 2**2 .data 00000000 cf030488 2**0 .sdata 0000000c cf030488 2**2 .kstrtab 00000085 cf030494 2**0 .bss 00000000 cf030519 2**0 .sbss 00000008 cf03051c 2**2 ... The option -m prints out the addresses of the various code and data segments ( .text, .data, .sdata, .bss, .sbss ) after relocation. GDB needs these addresses to know where all the symbols are located. We now interrupt GDB to load the symbol table of the module as follows: (gdb) ^C Program received signal SIGSTOP, Stopped (signal). ... (gdb) add-symbol-file /ex_sw.o 0xcf030060\ -s .rodata 0xcf030354\ -s .data 0xcf030488\ -s .sdata 0xcf030488\ -s .bss 0xcf030519\ -s .sbss 0xcf03051c add symbol table from file "/ex_sw.o" at .text_addr = 0xcf030060 .rodata_addr = 0xcf030354 .data_addr = 0xcf030488 .sdata_addr = 0xcf030488 .bss_addr = 0xcf030519 .sbss_addr = 0xcf03051c (y or n) y Reading symbols from /ex_sw.o...done. Now you can list the source code of the module, set break points or inspect variables as usual: (gdb) l fun 61 static RT_TASK *thread; 62 63 static int cpu_used[NR_RT_CPUS]; 64 65 static void fun(int t) 66 { 67 unsigned int loops = LOOPS; 68 while(loops--) { 69 cpu_used[hard_cpu_id()]++; 70 rt_leds_set_mask(1,t); (gdb) (gdb) b ex_sw.c:69 Breakpoint 1 at 0xcf03007c: file ex_sw.c, line 69. (gdb) c Continuing. Breakpoint 1, fun (t=1) at ex_sw.c:69 69 cpu_used[hard_cpu_id()]++; (gdb) p ntasks $1 = 16 (gdb) p stack_size $2 = 3000 The next section demonstrates a way to automate the symbol table loading procedure. 10.2.3. GDB Macros to Simplify Module Loading The following GDB macros and scripts help you to load kernel modules into GDB in a half-automatic way. It assumes, that the module on the target has been installed with the command: bash# insmod -m my_module.o > my_module.o.map In your $HOME directory you need the scripts add-symbol-file.sh and the GDB startup file .gdbinit, which are listed in [875]10.3. GDB Startup File and Utility Scripts below. Now you can include the symbol definition into GDB with: bash$ ${CROSS_COMPILE}gdb vmlinux GNU gdb 5.1.1 Copyright 2002 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "--host=i386-redhat-linux --target=ppc-linux". 0x00000100 in ?? () c Continuing. ^C Program received signal SIGSTOP, Stopped (signal). 0xcf02a91c in ?? () (gdb) add-module rtai4/examples/sw/ex_sw.o add symbol table from file "/HHL/8xx/target/home/wolf/rtai4/examples/sw/ex_sw.o " at .text_addr = 0xcf030060 .rodata_addr = 0xcf030340 .data_addr = 0xcf030464 .sdata_addr = 0xcf030464 .bss_addr = 0xcf0304f5 .sbss_addr = 0xcf0304f8 (gdb) b ex_sw.c:69 Breakpoint 1 at 0xcf03007c: file ex_sw.c, line 69. (gdb) c Continuing. Breakpoint 1, fun (t=0x1) at ex_sw.c:69 69 cpu_used[hard_cpu_id()]++; (gdb) p/d loops $2 = 999986939 (gdb) p t $3 = 0x1 (gdb) d b Delete all breakpoints? (y or n) y (gdb) c Continuing. 10.3. GDB Startup File and Utility Scripts In addition to the add-module macro, the followin example GDB startup file contains a few other useful settings and macros, which you may want to adjust to your local environment: set output-radix 16 target remote bdi:2001 define reset detach target remote bdi:2001 end define add-module shell ~/add-symbol-file.sh $arg0 source ~/add-symbol-file.gdb end document add-module Usage: add-module Do add-symbol-file for module automatically. Note: A map file with the extension ".map" must have been created with "insmod -m > .map" in advance. end The following shell script ~/add-symbol-file.sh is used to run the GDB add-symbol-file command automatically: #!/bin/sh # # Constructs the GDB "add-symbol-file" command string # from the map file of the specified kernel module. add_sect() { ADDR=`awk '/^'$1' / {print $3}' $MAPFILE` if [ "$ADDR" != "" ]; then echo "-s $1 0x`awk '/^'$1' / {print $3}' $MAPFILE`" fi } [ $# == 1 ] && [ -r "$1" ] || { echo "Usage: $0 " >&2 ; exit 1 ; } MAPFILE=$1.map ARGS="0x`awk '/^.text / {print $3}' $MAPFILE`\ `add_sect .rodata`\ `add_sect .data`\ `add_sect .sdata`\ `add_sect .bss`\ `add_sect .sbss`\ " echo "add-symbol-file $1 $ARGS" > ~/add-symbol-file.gdb 10.4. Tips and Tricks * To prevent GDB from jumping around in the code when trying to single step, i. e. when it seems as if the code is not executing line by line, you can recompile your code with the following additional compiler options: -fno-schedule-insns -fno-schedule-insns2 * On some systems (like the MPC8xx or MPC8260) you can only define one hardware breakpoint. Therefore you must delete an existing breakpoint before you can define a new one: (gdb) d b Delete all breakpoints? (y or n) y (gdb) b ex_preempt.c:63 Breakpoint 2 at 0xcf030080: file ex_preempt.c, line 63. 10.5. Application Debugging 10.5.1. Local Debugging In case there is a native GDB available for your target you can use it for application debugging as usual: bash$ gcc -Wall -g -o hello hello.c bash$ gdb hello ... (gdb) l 1 #include 2 3 int main(int argc, char* argv[]) 4 { 5 printf ("Hello world\n"); 6 return 0; 7 } (gdb) break 5 Breakpoint 1 at 0x8048466: file hello.c, line 5. (gdb) run Starting program: /opt/eldk/ppc_8xx/tmp/hello Breakpoint 1, main (argc=0x1, argv=0xbffff9f4) at hello.c:5 5 printf ("Hello world\n"); (gdb) c Continuing. Hello world Program exited normally. 10.5.2. Remote Debugging gdbserver allows you to connect your program with a remote GDB using the "target remote" command. On the target machine, you need to have a copy of the program you want to debug. gdbserver does not need your program's symbol table, so you can strip the program if necessary to save space. GDB on the host system does all the symbol handling. Here is an example: bash$ ${CROSS_COMPILE}gcc -Wall -g -o hello hello.c bash$ cp -p hello /hello-stripped bash$ ${CROSS_COMPILE}strip /hello-stripped To use the server, you must tell it how to communicate with GDB, the name of your program, and the arguments for your program. To start a debugging session via network type on the target: bash$ cd bash$ gdbserver 192.168.1.1:12345 hello-stripped Process hello-stripped created; pid = 353 And then on the host: bash$ ${CROSS_COMPILE}gdb hello ... (gdb) set solib-absolute-prefix /opt/eldk/$CROSS_COMPILE (gdb) dir /opt/eldk/$CROSS_COMPILE Source directories searched: /opt/eldk/$CROSS_COMPILE:$cdir:$cwd (gdb) target remote 192.168.1.99:12345 Remote debugging using 192.168.1.99:12345 0x30012748 in ?? () ... (gdb) l 1 #include 2 3 int main(int argc, char* argv[]) 4 { 5 printf ("Hello world\n"); 6 return 0; 7 } (gdb) break 5 Breakpoint 1 at 0x10000498: file hello.c, line 5. (gdb) continue Continuing. Breakpoint 1, main (argc=1, argv=0x7ffffbe4) at hello.c:5 5 printf ("Hello world\n"); (gdb) p argc $1 = 1 (gdb) continue Continuing. Program exited normally. ALERT! If the target program you want to debug is linked against shared libraries, you must tell GDB where the proper target libraries are located. This is done using the set solib-absolute-prefix GDB command. If this command is omitted, then, apparently, GDB loads the host versions of the libraries and gets crazy because of that. 10.6. Debugging with Graphical User Interfaces It is convenient to use DDD, a Graphical User Interface to GDB, for debugging as it allows to define and execute frequently used commands via buttons. You can start DDD with the command: bash$ ddd --debugger ${CROSS_COMPILE}gdb & If DDD is not already installed on your Linux system, have a look at your distribution media. 11. Simple Embedded Linux Framework 12. Books, Mailing Lists, Links, etc. This section provides references on where to find more information Contents: * [876]12. Books, Mailing Lists, Links, etc. + [877]12.1. Application Notes + [878]12.2. Further Reading o [879]12.2.1. License Issues o [880]12.2.2. Linux kernel o [881]12.2.3. General Linux / Unix programming o [882]12.2.4. Network Programming o [883]12.2.5. C++ programming o [884]12.2.6. Java programming o [885]12.2.7. Power Architecture® Programming o [886]12.2.8. Embedded Topics + [887]12.3. Mailing Lists + [888]12.4. Links + [889]12.5. Tools 12.1. Application Notes A collection of [890]Application Notes relevant for embedded computing can be found on the DENX web server. 12.2. Further Reading 12.2.1. License Issues Articles * [891]http://www.gnu.org/licenses/gpl-2.0.html: GNU General Public License, version 2 * [892]http://www.gnu.org/copyleft/gpl.html: GNU General Public License * [893]http://www.softwarefreedom.org/resources/2008/compliance-guid e.html: A Practical Guide to [894]GPL Compliance * [895]http://article.gmane.org/gmane.comp.video.dri.devel/52751 Alan Cox about combining [896]GPL device drivers with closed source user space libraries 12.2.2. Linux kernel Books * Karim Yaghmour, Jon Masters, Gilad Ben-Yossef, Philippe Gerum: "Building Embedded Linux Systems 2nd edition", Paperback: 462 pages, O'Reilly & Associates; (August 2008); ISBN 10: 0-596-52968-6; ISBN 13: 9780596529680 ISBN 059600222X - IMHO the best book about Embedded Linux so far. An absolute must have. * Greg Kroah-Hartman: "Linux Kernel in a Nutshell", 198 pages, O'Reilly ("In Nutshell" series), (December 2006), ISBN 10: 0-596-10079-5; ISBN 13: 9780596100797 - Tarball of PDF files (3 MB): [897]http://www.kernel.org/pub/linux/kernel/people/gregkh/lkn/lkn_ pdf.tar.bz2 - Tarball of DocBook files (1 MB): [898]http://www.kernel.org/pub/linux/kernel/people/gregkh/lkn/lkn_ xml.tar.bz2 * Craig Hollabaugh: "Embedded Linux: Hardware, Software, and Interfacing", Paperback: 432 pages; Addison Wesley Professional; (March 7, 2002); ISBN 0672322269 * Christopher Hallinan: "Embedded Linux Primer: A Practical Real-World Approach", 576 pages, Prentice Hall, September 2006, ISBN-10: 0-13-167984-8; ISBN-13: 978-0-13-167984-9 * Jonathan Corbet, Alessandro Rubini, Greg Kroah-Hartman: [899]"Linux Device Drivers", 3rd Edition ; Paperback: 636 pages; O'Reilly & Associates; 3rd edition (February 2005); ISBN: 0-596-00590-31 - The reference book for writing Linux device drivers. An absolute must have. [900]=> Read online * Jürgen Quade, Eva-Katharina Kunst: [901]"Linux-Treiber entwickeln"; Broschur: 436 pages; dpunkt.verlag, Juni 2004; ISBN 3898642380 - focused on kernel 2.6, unfortunately German only - [902]=> Read online * Sreekrishnan Venkateswaran: "Essential Linux Device Drivers", 744 pages, Prentice Hall, March 2008, ISBN-10: 0-13-239655-6; ISBN-13: 978-0-13-239655-4 - [903]=> Read online Articles * [904]The Linux Kernel - describing most aspects of the Linux Kernel. Probably, the first reference for beginners. Lots of illustrations explaining data structures use and relationships. In short: a must have. * [905]Linux Kernel Module Programming Guide - Very nice 92 pages [906]GPL book on the topic of modules programming. Lots of examples. * LWN: Porting device drivers to the 2.6 kernel - Series of articles (37) in Linux Weekly News: [907]http://lwn.net/Articles/driver-porting/ * MIPS Linux Porting Guide: [908]http://linux.junsun.net/porting-howto/porting-howto.html * Andries Brouwers remarks to the linux kernel: [909]http://www.win.tue.nl/~aeb/linux/lk/lk.html 12.2.3. General Linux / Unix programming Books * W. Richard Stevens: "Advanced Programming in the UNIX Environment", Addision Wesley, ISBN 0-201-56317-7 * Eric S. Raymond: "The Art of Unix Programming", Addision Wesley, ISBN 0131429019 [910]=> Read online. If you don't want to read the whole book then at least look at the [911]Basics of the Unix philosophy condensing lots of experience into a few rules. This is essential reading. * David R. Butenhof: "Programming with POSIX Threads", Addision Wesley, ISBN 0-201-63392-2. * Bradford Nichols, Dick Buttlar and Jacqueline Proulx Farrell: "Pthreads Programming", O'Reilly & Associates * "Git Community Book" See [912]http://book.git-scm.com/ or download the [913]PDF version. Articles * The GNU C Library: [914]http://www.linuxselfhelp.com/gnu/glibc/html_chapter/libc_toc. html General Linux Programming: [915]http://www.linuxselfhelp.com/cats/programming.html * Multi-Threaded Programming With POSIX Threads: [916]http://users.actcom.co.il/~choo/lupg/tutorials/multi-thread/m ulti-thread.html * Brad Hards: [917]The Linux USB Input Subsystem, Part I [918]http://www.linuxjournal.com/article/6396 * Brad Hards: [919]Using the Input Subsystem, Part II [920]http://www.linuxjournal.com/article/6429 * Ulrich Drepper: Position Independent Binaries: [921]"Text Relocations" [922]http://people.redhat.com/drepper/textrelocs.html * Ulrich Drepper: [923]"How to Write Shared Libraries" [924]http://people.redhat.com/drepper/dsohowto.pdf * Ulrich Drepper: [925]"What Every Programmer Should Know About Memory" [926]http://people.redhat.com/drepper/cpumemory.pdf * David Goldberg: [927]"What Every Computer Scientist Should Know About Floating-Point Arithmetic" [928]http://www.physics.ohio-state.edu/~dws/grouplinks/floating_po int_math.pdf * More Ulrich Drepper stuff: [929]http://people.redhat.com/drepper/ * How to [930]optimize DSOs by identifying unused non-exported functions and data. [931]http://blog.flameeyes.eu/articles/2008/01/17/today-how-to-ide ntify-unused-exported-functions-and-variables * A quite complete history of the UNIX family can be found here: [932]http://www.levenez.com/unix/ * Unix Manual, first edition, 3 November 1971: [933]http://cm.bell-labs.com/cm/cs/who/dmr/1stEdman.html * John Graham-Cumming: Debugging Makefiles [934]http://newsletter.embedded.com/cgi-bin4/DM/y/e4Kd0G1ErD0FrY0E 3FS0EZ * Binutils / ld documentation: [935]Linker Scripts * git ready - learn one git command at a time: [936]http://gitready.com/ Standards: * POSIX.1-2008, IEEE Std 1003.1-2008, The Open Group Technical Standard Base Specifications, Issue 7: [937]http://pubs.opengroup.org/onlinepubs/9699919799/mindex.html * Linux Standard Base: [938]http://refspecs.freestandards.org/lsb.shtml * [939]Single UNIX Specification, Version 3 (needs registration even for online viewing) * [940]Single UNIX Specification, Version 2 * PCI Bus Bindings - Standard for Boot Firmware: [941]http://playground.sun.com/1275/bindings/pci/pci2_1.pdf * International standardization working group for the programming language C: [942]http://www.open-std.org/jtc1/sc22/WG14/ 12.2.4. Network Programming Books * W. Richard Stevens: "TCP/IP Illustrated, Volume 1 - The Protocols", Addision Wesley, ISBN 0-201-63346-9 * Gary R. Wright, W. Richard Stevens: "TCP/IP Illustrated, Volume 2 - The Implementation", Addision Wesley, ISBN 0-201-63354-X * W. Richard Stevens: "TCP/IP Illustrated, Volume 3 - TCP for Transactions", Addision Wesley, ISBN 0-201-63495-3 * W. Richard Stevens: "UNIX Network Programming, Volume 1 - Networking APIs: Sockets and XTI", 2nd ed., Prentice Hall, ISBN-0-13-490012-X * W. Richard Stevens: "UNIX Network Programming, Volume 2 - Interprocess Communication", 2nd ed., Prentice Hall, ISBN-0-13-081081-9 Articles * Linux Networking topics (like [943]NAPI, [944]GSO, [945]VLAN, [946]IPsec etc.): [947]http://linux-net.osdl.org/index.php/Main_Page 12.2.5. C++ programming Books * Scott Meyers: "Effective C++: 55 Specific Ways to Improve Your Programs and Designs (3rd Edition)", Addison-Wesley, May 20, 2005, ISBN: 0321334876 12.2.6. Java programming Books * Joshua Bloch: "Effective Java -- Programming Language Guide", 2001, Addison Wesley, ISBN 0-201-31005-8, 250 pages 12.2.7. Power Architecture® Programming Books * Programming Environments Manual for 32-Bit Implementations of the [948]PowerPC architecture: [949]http://www.freescale.com/files/product/doc/MPCFPE32B.pdf * IBM PDF file (600+ page book) on PowerPC assembly language: [950]http://www-3.ibm.com/chips/techlib/techlib.nsf/techdocs/85256 9B20050FF778525699600719DF2 * Power.org(TM) Standard for Embedded Power Architecture(TM) Platform Requirements (ePAPR): [951]http://www.power.org/resources/downloads/Power_ePAPR_APPROVED _v1.0.pdf Articles * Introduction to Assembly on the PowerPC: [952]http://www-106.ibm.com/developerworks/library/l-ppc/?t=gr,lnx w09=PowPC * IBM PDF compiler writers guide on PPC asm tuning etc.: [953]http://www-3.ibm.com/chips/techlib/techlib.nsf/techdocs/85256 9B20050FF7785256996007558C6 * A developer's guide to the POWER architecture: [954]http://www-128.ibm.com/developerworks/linux/library/l-powarch /index.html * PowerPC [955]EABI Calling Sequence: [956]ftp://sourceware.redhat.com/pub/binutils/ppc-docs/ppc-eabi-ca lling-sequence * PowerPC Embedded Application Binary Interface (32-Bit Implementation): [957]ftp://sourceware.redhat.com/pub/binutils/ppc-docs/ppc-eabi-19 95-01.pdf * Developing PowerPC Embedded Application Binary Interface ([958]EABI) Compliant Programs [959]http://www-306.ibm.com/chips/techlib/techlib.nsf/techdocs/852 569B20050FF77852569970071B0D6 * System V Application Binary Interface - PowerPC Processor Supplement: [960]http://refspecs.freestandards.org/elf/elfspec_ppc.pdf * Device Tree Wiki: [961]http://devicetree.org/Main_Page * Device Tree Usage: [962]http://devicetree.org/Device_Tree_Usage * Linux for PowerPC Embedded Systems HOWTO (very old): [963]http://penguinppc.org/embedded/howto/PowerPC-Embedded-HOWTO.h tml * Linux for PowerPC Embedded Systems HOWTO (old): [964]http://www.denx.de/twiki/bin/view/PPCEmbedded * Understanding MPC5200 Bestcomm Firmware: [965]Posting on [966]linuxppc-embedded@ozlabs.org mailing list (see also the mailing list [967]archive entry), source code [968]disasm.c for a disassember, and [969]"SmartDMA Hand-Assembly Guides" document. 12.2.8. Embedded Topics Articles * Things you always wanted to know about NAND flash but never dared to ask: [970]Micron Application Note * The ultimate goal of [971]Embedded C++ is to provide embedded systems programmers with a subset of C++ that is easy for the average C programmer to understand and use. * Our contribution to the Darwin year 2009: Hardware designs that will not replicate: [972]Topic in DENX Wiki 12.3. Mailing Lists These are some mailing lists of interest. If you are new to mailing lists then please take the time to read at least [973]RFC 1855. * [974]linux-arm-kernel - Communications among developers and users of Linux on arm boards * [975]linuxppc-embedded - Communications among developers and users of Linux on embedded [976]Power Architecture® boards ALERT! This mailing list has been merged into the linuxppc-dev mailing list below and thus does not exist anymore. * [977]linuxppc-dev - Communications among active developers of Linux on 32 bit [978]Power Architecture® plattforms. Not intended for user support. * [979]u-boot - Support for "U-Boot" Universal Bootloader * [980]ELDK - Support for DENX Embedded Linux Development Kit 12.4. Links Linux Kernel Resources: * The Linux Documentation Project : [981]http://www.tldp.org/ * Generic ("official") Linux Kernel sources: git: [982]http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6. git;a=tree [983]FTP: [984]ftp://ftp.kernel.org/pub/linux/kernel/v2.6/ * Full git history of Linux: [985]http://thread.gmane.org/gmane.linux.kernel/690811 * Generic kernel sources for [986]Power Architecture(TM) systems: [987]http://penguinppc.org/dev/kernel.shtml * DENX kernel sources: [988]http://git.denx.de/?p=linux-denx.git;a=summary * Cross-Referencing the Linx Kernel: [989]http://lxr.linux.no/source/?a=ppc * Starting point for Linux based asm (mostly x86): [990]http://linuxassembly.org/ Realtime, Xenomai, [991]RTAI: * Xenomai Home Page: [992]http://www.xenomai.org/ * * [993]Hackbench, a commonly used system stress tool * [994]Calibrator, determines cache * [995]RTAI Home Page: [996]http://www.rtai.org/ * DENX [997]RTAI Patches: [998]ftp://ftp.denx.de/pub/RTAI/ sizes at runtime, rendering it another useful system stress tool U-Boot: * U-Boot Project Page: [999]http://www.denx.de/wiki/U-Boot/WebHome. ALERT! Note that the old [1000]SourceForge page is not maintained anymore. * DENX U-Boot and Linux Guide: [1001]http://www.denx.de/twiki/bin/view/DULG Cross Development Tools: * DENX Embedded Linux Development Kit: [1002]http://www.denx.de/twiki/bin/view/DULG/ELDK Miscalleneous or unsorted material: * BDI2000 [1003]List of supported Flash Memories: This document not only lists the currently supported flash chips, but also the required settings in the BDI config file. * BDI2000 configuration files: [1004]ftp://78.31.64.234/bdigdb/config/ 12.5. Tools * [1005]http://lxr.linux.no/source/ - Cross-Referencing the Linux Kernel - using a versatile hypertext cross-referencing tool for the Linux Kernel source tree (the Linux Cross-Reference project) * [1006]ftp://ftp.denx.de/pub/tools/backtrace - Decode Stack Backtrace - Perl script to decode the Stack Backtrace printed by the Linux Kernel when it panics * [1007]ftp://ftp.denx.de/pub/tools/clone_tree - "Clone" a Source Tree - Perl script to create a working copy of a source tree (for example the Linux Kernel) which contains mainly symbolic links (and automagically omits "unwanted" files like [1008]CVS repository data, etc.) * [1009]13. Appendix + [1010]13.1. Flat Device Tree + [1011]13.2. Flat Device Tree + [1012]13.3. BDI2000 Configuration file 13. Appendix 13.1. Flat Device Tree /* * Device Tree Source for AMCC Canyonlands (460EX) * * Copyright 2008-2009 DENX Software Engineering, Stefan Roese * * This file is licensed under the terms of the GNU General Public * License version 2. This program is licensed "as is" without * any warranty of any kind, whether express or implied. */ /dts-v1/; / { #address-cells = <2>; #size-cells = <1>; model = "amcc,canyonlands"; compatible = "amcc,canyonlands"; dcr-parent = <&{/cpus/cpu@0}>; aliases { ethernet0 = &EMAC0; ethernet1 = &EMAC1; serial0 = &UART0; serial1 = &UART1; }; cpus { #address-cells = <1>; #size-cells = <0>; cpu@0 { device_type = "cpu"; model = "PowerPC,460EX"; reg = <0x00000000>; clock-frequency = <0>; /* Filled in by U-Boot */ timebase-frequency = <0>; /* Filled in by U-Boot */ i-cache-line-size = <32>; d-cache-line-size = <32>; i-cache-size = <32768>; d-cache-size = <32768>; dcr-controller; dcr-access-method = "native"; next-level-cache = <&L2C0>; }; }; memory { device_type = "memory"; reg = <0x00000000 0x00000000 0x00000000>; /* Filled in by U-Boot */ }; UIC0: interrupt-controller0 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <0>; dcr-reg = <0x0c0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; }; UIC1: interrupt-controller1 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <1>; dcr-reg = <0x0d0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0x1e 0x4 0x1f 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; UIC2: interrupt-controller2 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <2>; dcr-reg = <0x0e0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0xa 0x4 0xb 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; UIC3: interrupt-controller3 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <3>; dcr-reg = <0x0f0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0x10 0x4 0x11 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; SDR0: sdr { compatible = "ibm,sdr-460ex"; dcr-reg = <0x00e 0x002>; }; CPR0: cpr { compatible = "ibm,cpr-460ex"; dcr-reg = <0x00c 0x002>; }; L2C0: l2c { compatible = "ibm,l2-cache-460ex", "ibm,l2-cache"; dcr-reg = <0x020 0x008 /* Internal SRAM DCR's */ 0x030 0x008>; /* L2 cache DCR's */ cache-line-size = <32>; /* 32 bytes */ cache-size = <262144>; /* L2, 256K */ interrupt-parent = <&UIC1>; interrupts = <11 1>; }; plb { compatible = "ibm,plb-460ex", "ibm,plb4"; #address-cells = <2>; #size-cells = <1>; ranges; clock-frequency = <0>; /* Filled in by U-Boot */ SDRAM0: sdram { compatible = "ibm,sdram-460ex", "ibm,sdram-405gp"; dcr-reg = <0x010 0x002>; }; CRYPTO: crypto@180000 { compatible = "amcc,ppc460ex-crypto", "amcc,ppc4xx-crypto"; reg = <4 0x00180000 0x80400>; interrupt-parent = <&UIC0>; interrupts = <0x1d 0x4>; }; MAL0: mcmal { compatible = "ibm,mcmal-460ex", "ibm,mcmal2"; dcr-reg = <0x180 0x062>; num-tx-chans = <2>; num-rx-chans = <16>; #address-cells = <0>; #size-cells = <0>; interrupt-parent = <&UIC2>; interrupts = < /*TXEOB*/ 0x6 0x4 /*RXEOB*/ 0x7 0x4 /*SERR*/ 0x3 0x4 /*TXDE*/ 0x4 0x4 /*RXDE*/ 0x5 0x4>; }; USB0: ehci@bffd0400 { compatible = "ibm,usb-ehci-460ex", "usb-ehci"; interrupt-parent = <&UIC2>; interrupts = <0x1d 4>; reg = <4 0xbffd0400 0x90 4 0xbffd0490 0x70>; }; USB1: usb@bffd0000 { compatible = "ohci-le"; reg = <4 0xbffd0000 0x60>; interrupt-parent = <&UIC2>; interrupts = <0x1e 4>; }; USBOTG0: usbotg@bff80000 { compatible = "amcc,usb-otg-460ex", "amcc,usb-otg"; reg = <4 0xbff80000 0x10000>; interrupt-parent = <&USBOTG0>; interrupts = <0 1 2>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; interrupt-map-mask = <0xffffffff>; }; SATA0: sata@bffd1000 { compatible = "amcc,sata-460ex"; reg = <4 0xbffd1000 0x800 /* SATA */ 4 0xbffd0800 0x400>; /* AHBDMA */ interrupt-parent = <&UIC3>; interrupts = <0 4 /* SATA */ 5 4>; /* AHBDMA */ }; POB0: opb { compatible = "ibm,opb-460ex", "ibm,opb"; #address-cells = <1>; #size-cells = <1>; ranges = <0xb0000000 0x00000004 0xb0000000 0x50000000>; clock-frequency = <0>; /* Filled in by U-Boot */ EBC0: ebc { compatible = "ibm,ebc-460ex", "ibm,ebc"; dcr-reg = <0x012 0x002>; #address-cells = <2>; #size-cells = <1>; clock-frequency = <0>; /* Filled in by U-Boot */ /* ranges property is supplied by U-Boot */ interrupts = <0x6 0x4>; interrupt-parent = <&UIC1>; nor_flash@0,0 { compatible = "amd,s29gl512n", "cfi-flash"; bank-width = <2>; reg = <0x00000000 0x00000000 0x04000000>; #address-cells = <1>; #size-cells = <1>; partition@0 { label = "kernel"; reg = <0x00000000 0x001e0000>; }; partition@1e0000 { label = "dtb"; reg = <0x001e0000 0x00020000>; }; partition@200000 { label = "ramdisk"; reg = <0x00200000 0x01400000>; }; partition@1600000 { label = "jffs2"; reg = <0x01600000 0x00400000>; }; partition@1a00000 { label = "user"; reg = <0x01a00000 0x02560000>; }; partition@3f60000 { label = "env"; reg = <0x03f60000 0x00040000>; }; partition@3fa0000 { label = "u-boot"; reg = <0x03fa0000 0x00060000>; }; }; ndfc@3,0 { compatible = "ibm,ndfc"; reg = <0x00000003 0x00000000 0x00002000>; ccr = <0x00001000>; bank-settings = <0x80002222>; #address-cells = <1>; #size-cells = <1>; nand { #address-cells = <1>; #size-cells = <1>; partition@0 { label = "u-boot"; reg = <0x00000000 0x00100000>; }; partition@100000 { label = "user"; reg = <0x00000000 0x03f00000>; }; }; }; }; UART0: serial@ef600300 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600300 0x00000008>; virtual-reg = <0xef600300>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1 0x4>; }; UART1: serial@ef600400 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600400 0x00000008>; virtual-reg = <0xef600400>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC0>; interrupts = <0x1 0x4>; }; UART2: serial@ef600500 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600500 0x00000008>; virtual-reg = <0xef600500>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1d 0x4>; }; UART3: serial@ef600600 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600600 0x00000008>; virtual-reg = <0xef600600>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1e 0x4>; }; IIC0: i2c@ef600700 { compatible = "ibm,iic-460ex", "ibm,iic"; reg = <0xef600700 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x2 0x4>; #address-cells = <1>; #size-cells = <0>; rtc@68 { compatible = "stm,m41t80"; reg = <0x68>; interrupt-parent = <&UIC2>; interrupts = <0x19 0x8>; }; sttm@48 { compatible = "ad,ad7414"; reg = <0x48>; interrupt-parent = <&UIC1>; interrupts = <0x14 0x8>; }; }; IIC1: i2c@ef600800 { compatible = "ibm,iic-460ex", "ibm,iic"; reg = <0xef600800 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x3 0x4>; }; ZMII0: emac-zmii@ef600d00 { compatible = "ibm,zmii-460ex", "ibm,zmii"; reg = <0xef600d00 0x0000000c>; }; RGMII0: emac-rgmii@ef601500 { compatible = "ibm,rgmii-460ex", "ibm,rgmii"; reg = <0xef601500 0x00000008>; has-mdio; }; TAH0: emac-tah@ef601350 { compatible = "ibm,tah-460ex", "ibm,tah"; reg = <0xef601350 0x00000030>; }; TAH1: emac-tah@ef601450 { compatible = "ibm,tah-460ex", "ibm,tah"; reg = <0xef601450 0x00000030>; }; EMAC0: ethernet@ef600e00 { device_type = "network"; compatible = "ibm,emac-460ex", "ibm,emac4sync"; interrupt-parent = <&EMAC0>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; reg = <0xef600e00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <0>; mal-rx-channel = <0>; cell-index = <0>; max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; rx-fifo-size-gige = <16384>; phy-mode = "rgmii"; phy-map = <0x00000000>; rgmii-device = <&RGMII0>; rgmii-channel = <0>; tah-device = <&TAH0>; tah-channel = <0>; has-inverted-stacr-oc; has-new-stacr-staopc; }; EMAC1: ethernet@ef600f00 { device_type = "network"; compatible = "ibm,emac-460ex", "ibm,emac4sync"; interrupt-parent = <&EMAC1>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; reg = <0xef600f00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <1>; mal-rx-channel = <8>; cell-index = <1>; max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; rx-fifo-size-gige = <16384>; phy-mode = "rgmii"; phy-map = <0x00000000>; rgmii-device = <&RGMII0>; rgmii-channel = <1>; tah-device = <&TAH1>; tah-channel = <1>; has-inverted-stacr-oc; has-new-stacr-staopc; mdio-device = <&EMAC0>; }; }; PCIX0: pci@c0ec00000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pcix-460ex", "ibm,plb-pcix"; primary; large-inbound-windows; enable-msi-hole; reg = <0x0000000c 0x0ec00000 0x00000008 /* Config space access */ 0x00000000 0x00000000 0x00000000 /* no IACK cycles */ 0x0000000c 0x0ed00000 0x00000004 /* Special cycles */ 0x0000000c 0x0ec80000 0x00000100 /* Internal registers */ 0x0000000c 0x0ec80100 0x000000fc>; /* Internal messaging reg isters */ /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000d 0x80000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000c 0x0ee00000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000c 0x08000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 0 to 0x3f */ bus-range = <0x0 0x3f>; /* All PCI interrupts are routed to ext IRQ 2 -> UIC1-0 */ interrupt-map-mask = <0x0 0x0 0x0 0x0>; interrupt-map = < 0x0 0x0 0x0 0x0 &UIC1 0x0 0x8 >; }; PCIE0: pciex@d00000000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pciex-460ex", "ibm,plb-pciex"; primary; port = <0x0>; /* port number */ reg = <0x0000000d 0x00000000 0x20000000 /* Config space access */ 0x0000000c 0x08010000 0x00001000>; /* Registers */ dcr-reg = <0x100 0x020>; sdr-base = <0x300>; /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x00000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000f 0x00000000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 40 to 0x7f */ bus-range = <0x40 0x7f>; /* Legacy interrupts (note the weird polarity, the bridge seems * to invert PCIe legacy interrupts). * We are de-swizzling here because the numbers are actually for * port of the root complex virtual P2P bridge. But I want * to avoid putting a node for it in the tree, so the numbers * below are basically de-swizzled numbers. * The real slot is on idsel 0, so the swizzling is 1:1 */ interrupt-map-mask = <0x0 0x0 0x0 0x7>; interrupt-map = < 0x0 0x0 0x0 0x1 &UIC3 0xc 0x4 /* swizzled int A */ 0x0 0x0 0x0 0x2 &UIC3 0xd 0x4 /* swizzled int B */ 0x0 0x0 0x0 0x3 &UIC3 0xe 0x4 /* swizzled int C */ 0x0 0x0 0x0 0x4 &UIC3 0xf 0x4 /* swizzled int D */>; }; PCIE1: pciex@d20000000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pciex-460ex", "ibm,plb-pciex"; primary; port = <0x1>; /* port number */ reg = <0x0000000d 0x20000000 0x20000000 /* Config space access */ 0x0000000c 0x08011000 0x00001000>; /* Registers */ dcr-reg = <0x120 0x020>; sdr-base = <0x340>; /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x80000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000f 0x00100000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80010000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 80 to 0xbf */ bus-range = <0x80 0xbf>; /* Legacy interrupts (note the weird polarity, the bridge seems * to invert PCIe legacy interrupts). * We are de-swizzling here because the numbers are actually for * port of the root complex virtual P2P bridge. But I want * to avoid putting a node for it in the tree, so the numbers * below are basically de-swizzled numbers. * The real slot is on idsel 0, so the swizzling is 1:1 */ interrupt-map-mask = <0x0 0x0 0x0 0x7>; interrupt-map = < 0x0 0x0 0x0 0x1 &UIC3 0x10 0x4 /* swizzled int A */ 0x0 0x0 0x0 0x2 &UIC3 0x11 0x4 /* swizzled int B */ 0x0 0x0 0x0 0x3 &UIC3 0x12 0x4 /* swizzled int C */ 0x0 0x0 0x0 0x4 &UIC3 0x13 0x4 /* swizzled int D */>; }; }; }; 13.2. Flat Device Tree /* * Device Tree Source for AMCC Canyonlands (460EX) * * Copyright 2008-2009 DENX Software Engineering, Stefan Roese * * This file is licensed under the terms of the GNU General Public * License version 2. This program is licensed "as is" without * any warranty of any kind, whether express or implied. */ /dts-v1/; / { #address-cells = <2>; #size-cells = <1>; model = "amcc,canyonlands"; compatible = "amcc,canyonlands"; dcr-parent = <&{/cpus/cpu@0}>; aliases { ethernet0 = &EMAC0; ethernet1 = &EMAC1; serial0 = &UART0; serial1 = &UART1; }; cpus { #address-cells = <1>; #size-cells = <0>; cpu@0 { device_type = "cpu"; model = "PowerPC,460EX"; reg = <0x00000000>; clock-frequency = <0>; /* Filled in by U-Boot */ timebase-frequency = <0>; /* Filled in by U-Boot */ i-cache-line-size = <32>; d-cache-line-size = <32>; i-cache-size = <32768>; d-cache-size = <32768>; dcr-controller; dcr-access-method = "native"; next-level-cache = <&L2C0>; }; }; memory { device_type = "memory"; reg = <0x00000000 0x00000000 0x00000000>; /* Filled in by U-Boot */ }; UIC0: interrupt-controller0 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <0>; dcr-reg = <0x0c0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; }; UIC1: interrupt-controller1 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <1>; dcr-reg = <0x0d0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0x1e 0x4 0x1f 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; UIC2: interrupt-controller2 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <2>; dcr-reg = <0x0e0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0xa 0x4 0xb 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; UIC3: interrupt-controller3 { compatible = "ibm,uic-460ex","ibm,uic"; interrupt-controller; cell-index = <3>; dcr-reg = <0x0f0 0x009>; #address-cells = <0>; #size-cells = <0>; #interrupt-cells = <2>; interrupts = <0x10 0x4 0x11 0x4>; /* cascade */ interrupt-parent = <&UIC0>; }; SDR0: sdr { compatible = "ibm,sdr-460ex"; dcr-reg = <0x00e 0x002>; }; CPR0: cpr { compatible = "ibm,cpr-460ex"; dcr-reg = <0x00c 0x002>; }; L2C0: l2c { compatible = "ibm,l2-cache-460ex", "ibm,l2-cache"; dcr-reg = <0x020 0x008 /* Internal SRAM DCR's */ 0x030 0x008>; /* L2 cache DCR's */ cache-line-size = <32>; /* 32 bytes */ cache-size = <262144>; /* L2, 256K */ interrupt-parent = <&UIC1>; interrupts = <11 1>; }; plb { compatible = "ibm,plb-460ex", "ibm,plb4"; #address-cells = <2>; #size-cells = <1>; ranges; clock-frequency = <0>; /* Filled in by U-Boot */ SDRAM0: sdram { compatible = "ibm,sdram-460ex", "ibm,sdram-405gp"; dcr-reg = <0x010 0x002>; }; CRYPTO: crypto@180000 { compatible = "amcc,ppc460ex-crypto", "amcc,ppc4xx-crypto"; reg = <4 0x00180000 0x80400>; interrupt-parent = <&UIC0>; interrupts = <0x1d 0x4>; }; MAL0: mcmal { compatible = "ibm,mcmal-460ex", "ibm,mcmal2"; dcr-reg = <0x180 0x062>; num-tx-chans = <2>; num-rx-chans = <16>; #address-cells = <0>; #size-cells = <0>; interrupt-parent = <&UIC2>; interrupts = < /*TXEOB*/ 0x6 0x4 /*RXEOB*/ 0x7 0x4 /*SERR*/ 0x3 0x4 /*TXDE*/ 0x4 0x4 /*RXDE*/ 0x5 0x4>; }; USB0: ehci@bffd0400 { compatible = "ibm,usb-ehci-460ex", "usb-ehci"; interrupt-parent = <&UIC2>; interrupts = <0x1d 4>; reg = <4 0xbffd0400 0x90 4 0xbffd0490 0x70>; }; USB1: usb@bffd0000 { compatible = "ohci-le"; reg = <4 0xbffd0000 0x60>; interrupt-parent = <&UIC2>; interrupts = <0x1e 4>; }; USBOTG0: usbotg@bff80000 { compatible = "amcc,usb-otg-460ex", "amcc,usb-otg"; reg = <4 0xbff80000 0x10000>; interrupt-parent = <&USBOTG0>; interrupts = <0 1 2>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; interrupt-map-mask = <0xffffffff>; }; SATA0: sata@bffd1000 { compatible = "amcc,sata-460ex"; reg = <4 0xbffd1000 0x800 /* SATA */ 4 0xbffd0800 0x400>; /* AHBDMA */ interrupt-parent = <&UIC3>; interrupts = <0 4 /* SATA */ 5 4>; /* AHBDMA */ }; POB0: opb { compatible = "ibm,opb-460ex", "ibm,opb"; #address-cells = <1>; #size-cells = <1>; ranges = <0xb0000000 0x00000004 0xb0000000 0x50000000>; clock-frequency = <0>; /* Filled in by U-Boot */ EBC0: ebc { compatible = "ibm,ebc-460ex", "ibm,ebc"; dcr-reg = <0x012 0x002>; #address-cells = <2>; #size-cells = <1>; clock-frequency = <0>; /* Filled in by U-Boot */ /* ranges property is supplied by U-Boot */ interrupts = <0x6 0x4>; interrupt-parent = <&UIC1>; nor_flash@0,0 { compatible = "amd,s29gl512n", "cfi-flash"; bank-width = <2>; reg = <0x00000000 0x00000000 0x04000000>; #address-cells = <1>; #size-cells = <1>; partition@0 { label = "kernel"; reg = <0x00000000 0x001e0000>; }; partition@1e0000 { label = "dtb"; reg = <0x001e0000 0x00020000>; }; partition@200000 { label = "ramdisk"; reg = <0x00200000 0x01400000>; }; partition@1600000 { label = "jffs2"; reg = <0x01600000 0x00400000>; }; partition@1a00000 { label = "user"; reg = <0x01a00000 0x02560000>; }; partition@3f60000 { label = "env"; reg = <0x03f60000 0x00040000>; }; partition@3fa0000 { label = "u-boot"; reg = <0x03fa0000 0x00060000>; }; }; ndfc@3,0 { compatible = "ibm,ndfc"; reg = <0x00000003 0x00000000 0x00002000>; ccr = <0x00001000>; bank-settings = <0x80002222>; #address-cells = <1>; #size-cells = <1>; nand { #address-cells = <1>; #size-cells = <1>; partition@0 { label = "u-boot"; reg = <0x00000000 0x00100000>; }; partition@100000 { label = "user"; reg = <0x00000000 0x03f00000>; }; }; }; }; UART0: serial@ef600300 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600300 0x00000008>; virtual-reg = <0xef600300>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1 0x4>; }; UART1: serial@ef600400 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600400 0x00000008>; virtual-reg = <0xef600400>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC0>; interrupts = <0x1 0x4>; }; UART2: serial@ef600500 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600500 0x00000008>; virtual-reg = <0xef600500>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1d 0x4>; }; UART3: serial@ef600600 { device_type = "serial"; compatible = "ns16550"; reg = <0xef600600 0x00000008>; virtual-reg = <0xef600600>; clock-frequency = <0>; /* Filled in by U-Boot */ current-speed = <0>; /* Filled in by U-Boot */ interrupt-parent = <&UIC1>; interrupts = <0x1e 0x4>; }; IIC0: i2c@ef600700 { compatible = "ibm,iic-460ex", "ibm,iic"; reg = <0xef600700 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x2 0x4>; #address-cells = <1>; #size-cells = <0>; rtc@68 { compatible = "stm,m41t80"; reg = <0x68>; interrupt-parent = <&UIC2>; interrupts = <0x19 0x8>; }; sttm@48 { compatible = "ad,ad7414"; reg = <0x48>; interrupt-parent = <&UIC1>; interrupts = <0x14 0x8>; }; }; IIC1: i2c@ef600800 { compatible = "ibm,iic-460ex", "ibm,iic"; reg = <0xef600800 0x00000014>; interrupt-parent = <&UIC0>; interrupts = <0x3 0x4>; }; ZMII0: emac-zmii@ef600d00 { compatible = "ibm,zmii-460ex", "ibm,zmii"; reg = <0xef600d00 0x0000000c>; }; RGMII0: emac-rgmii@ef601500 { compatible = "ibm,rgmii-460ex", "ibm,rgmii"; reg = <0xef601500 0x00000008>; has-mdio; }; TAH0: emac-tah@ef601350 { compatible = "ibm,tah-460ex", "ibm,tah"; reg = <0xef601350 0x00000030>; }; TAH1: emac-tah@ef601450 { compatible = "ibm,tah-460ex", "ibm,tah"; reg = <0xef601450 0x00000030>; }; EMAC0: ethernet@ef600e00 { device_type = "network"; compatible = "ibm,emac-460ex", "ibm,emac4sync"; interrupt-parent = <&EMAC0>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; reg = <0xef600e00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <0>; mal-rx-channel = <0>; cell-index = <0>; max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; rx-fifo-size-gige = <16384>; phy-mode = "rgmii"; phy-map = <0x00000000>; rgmii-device = <&RGMII0>; rgmii-channel = <0>; tah-device = <&TAH0>; tah-channel = <0>; has-inverted-stacr-oc; has-new-stacr-staopc; }; EMAC1: ethernet@ef600f00 { device_type = "network"; compatible = "ibm,emac-460ex", "ibm,emac4sync"; interrupt-parent = <&EMAC1>; interrupts = <0x0 0x1>; #interrupt-cells = <1>; #address-cells = <0>; #size-cells = <0>; interrupt-map = ; reg = <0xef600f00 0x000000c4>; local-mac-address = [000000000000]; /* Filled in by U-Boot */ mal-device = <&MAL0>; mal-tx-channel = <1>; mal-rx-channel = <8>; cell-index = <1>; max-frame-size = <9000>; rx-fifo-size = <4096>; tx-fifo-size = <2048>; rx-fifo-size-gige = <16384>; phy-mode = "rgmii"; phy-map = <0x00000000>; rgmii-device = <&RGMII0>; rgmii-channel = <1>; tah-device = <&TAH1>; tah-channel = <1>; has-inverted-stacr-oc; has-new-stacr-staopc; mdio-device = <&EMAC0>; }; }; PCIX0: pci@c0ec00000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pcix-460ex", "ibm,plb-pcix"; primary; large-inbound-windows; enable-msi-hole; reg = <0x0000000c 0x0ec00000 0x00000008 /* Config space access */ 0x00000000 0x00000000 0x00000000 /* no IACK cycles */ 0x0000000c 0x0ed00000 0x00000004 /* Special cycles */ 0x0000000c 0x0ec80000 0x00000100 /* Internal registers */ 0x0000000c 0x0ec80100 0x000000fc>; /* Internal messaging reg isters */ /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000d 0x80000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000c 0x0ee00000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000c 0x08000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 0 to 0x3f */ bus-range = <0x0 0x3f>; /* All PCI interrupts are routed to ext IRQ 2 -> UIC1-0 */ interrupt-map-mask = <0x0 0x0 0x0 0x0>; interrupt-map = < 0x0 0x0 0x0 0x0 &UIC1 0x0 0x8 >; }; PCIE0: pciex@d00000000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pciex-460ex", "ibm,plb-pciex"; primary; port = <0x0>; /* port number */ reg = <0x0000000d 0x00000000 0x20000000 /* Config space access */ 0x0000000c 0x08010000 0x00001000>; /* Registers */ dcr-reg = <0x100 0x020>; sdr-base = <0x300>; /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x00000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000f 0x00000000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80000000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 40 to 0x7f */ bus-range = <0x40 0x7f>; /* Legacy interrupts (note the weird polarity, the bridge seems * to invert PCIe legacy interrupts). * We are de-swizzling here because the numbers are actually for * port of the root complex virtual P2P bridge. But I want * to avoid putting a node for it in the tree, so the numbers * below are basically de-swizzled numbers. * The real slot is on idsel 0, so the swizzling is 1:1 */ interrupt-map-mask = <0x0 0x0 0x0 0x7>; interrupt-map = < 0x0 0x0 0x0 0x1 &UIC3 0xc 0x4 /* swizzled int A */ 0x0 0x0 0x0 0x2 &UIC3 0xd 0x4 /* swizzled int B */ 0x0 0x0 0x0 0x3 &UIC3 0xe 0x4 /* swizzled int C */ 0x0 0x0 0x0 0x4 &UIC3 0xf 0x4 /* swizzled int D */>; }; PCIE1: pciex@d20000000 { device_type = "pci"; #interrupt-cells = <1>; #size-cells = <2>; #address-cells = <3>; compatible = "ibm,plb-pciex-460ex", "ibm,plb-pciex"; primary; port = <0x1>; /* port number */ reg = <0x0000000d 0x20000000 0x20000000 /* Config space access */ 0x0000000c 0x08011000 0x00001000>; /* Registers */ dcr-reg = <0x120 0x020>; sdr-base = <0x340>; /* Outbound ranges, one memory and one IO, * later cannot be changed */ ranges = <0x02000000 0x00000000 0x80000000 0x0000000e 0x80000000 0x000 00000 0x80000000 0x02000000 0x00000000 0x00000000 0x0000000f 0x00100000 0x00000000 0x00100000 0x01000000 0x00000000 0x00000000 0x0000000f 0x80010000 0x00000000 0x00010000>; /* Inbound 2GB range starting at 0 */ dma-ranges = <0x42000000 0x0 0x0 0x0 0x0 0x0 0x80000000>; /* This drives busses 80 to 0xbf */ bus-range = <0x80 0xbf>; /* Legacy interrupts (note the weird polarity, the bridge seems * to invert PCIe legacy interrupts). * We are de-swizzling here because the numbers are actually for * port of the root complex virtual P2P bridge. But I want * to avoid putting a node for it in the tree, so the numbers * below are basically de-swizzled numbers. * The real slot is on idsel 0, so the swizzling is 1:1 */ interrupt-map-mask = <0x0 0x0 0x0 0x7>; interrupt-map = < 0x0 0x0 0x0 0x1 &UIC3 0x10 0x4 /* swizzled int A */ 0x0 0x0 0x0 0x2 &UIC3 0x11 0x4 /* swizzled int B */ 0x0 0x0 0x0 0x3 &UIC3 0x12 0x4 /* swizzled int C */ 0x0 0x0 0x0 0x4 &UIC3 0x13 0x4 /* swizzled int D */>; }; }; }; 13.3. BDI2000 Configuration file ; bdiGDB configuration file for AMCC 460EX Evaluation Kit "Canyonlands" ; --------------------------------------------------------------------- ; ; [INIT] ; Setup TLB WTLB 0xF0000095 0x4F00003F ;Boot Space 256MB WTLB 0x00000094 0x0000003F ;SDRAM 256MB @ 0x00000000 WTLB 0x90000095 0x4000003F ;ISRAM/OCM ; Setup Peripheral Bus ;;;;;;;;;;;;;;;;;;;;;;;;;;;;; WDCR 0x12 0x00000010 ;Select EBC0_B0AP WDCR 0x13 0x10055e00 WDCR 0x12 0x00000000 ;Select EBC0_B0CR WDCR 0x13 0xfc0da000 ; 64MByte ; OCM ;WDCR 0x20 0x40004580 ;16k [TARGET] JTAGCLOCK 1 ;use 8 MHz JTAG clock CPUTYPE 440 ;the used target CPU type WAKEUP 500 ;wakeup time after reset BREAKMODE HARD ;SOFT or HARD, HARD uses PPC hardware breakpoint STEPMODE HWBP ;JTAG or HWBP, HWBP uses one or two hardware breakpoints [HOST] IP 192.168.1.1 FILE /tftpboot/canyonlands/u-boot.bin FORMAT BIN DUMP /tftpboot/canyonlands/dump.bin PROMPT 460EX> [FLASH] ;WORKSPACE 0xe3040000 ;workspace in OCM ;WORKSPACE 0x70000000 ;workspace in OCM ;WORKSPACE 0x90040000 ;workspace in OCM CHIPTYPE MIRRORX16 ;Flash type CHIPSIZE 0x1000000 ;The size of one flash chip in bytes BUSWIDTH 16 ;The width of the flash memory bus in bits (8 | 16 | 32) FILE /tftpboot/canyonlands/u-boot.bin ;FORMAT BIN 0xFFF80000 FORMAT BIN 0xFFFA0000 ;ERASE 0xFFF80000 ;erase sector 4 ERASE 0xFFFA0000 ;erase sector 4 ERASE 0xFFFC0000 ;erase sector 4 ERASE 0xFFFE0000 ;erase sector 6 [REGS] FILE /tftpboot/BDI2000/reg460ex.def * [1013]14. FAQ - Frequently Asked Questions + [1014]14.1. ELDK o [1015]14.1.1. ELDK Installation under FreeBSD o [1016]14.1.2. ELDK Installation Hangs o [1017]14.1.3. .gvfs: Permission Denied o [1018]14.1.4. Installation on Local Harddisk o [1019]14.1.5. System Include Files Missing o [1020]14.1.6. patch: command not found o [1021]14.1.7. ELDK Include Files Missing o [1022]14.1.8. Using the ELDK on a 64 bit platform o [1023]14.1.9. GDB Problems with BDI2000/BDI3000 on e500 Cores o [1024]14.1.10. How can I check if Floating Point support is working? o [1025]14.1.11. ELDK 2.x Installation Aborts o [1026]14.1.12. Enable SSH Access + [1027]14.2. U-Boot o [1028]14.2.1. Can U-Boot be configured such that it can be started in RAM? o [1029]14.2.2. Relocation cannot be done when using -mrelocatable o [1030]14.2.3. Source object has EABI version 4, but target has EABI version 0 o [1031]14.2.4. U-Boot crashes after relocation to RAM o [1032]14.2.5. Warning - bad CRC, using default environment o [1033]14.2.6. Net: No ethernet found o [1034]14.2.7. Wrong debug symbols after relocation o [1035]14.2.8. Decoding U-Boot Crash Dumps o [1036]14.2.9. Porting Problem: cannot move location counter backwards o [1037]14.2.10. U-Boot Doesn't Run after Upgrading my Compiler o [1038]14.2.11. How Can I Reduce The Image Size? o [1039]14.2.12. Erasing Flash Fails o [1040]14.2.13. Ethernet Does Not Work o [1041]14.2.14. Where Can I Get a Valid MAC Address from? o [1042]14.2.15. Why do I get TFTP timeouts? o [1043]14.2.16. Why is my Ethernet operation not reliable? o [1044]14.2.17. How the Command Line Parsing Works # [1045]14.2.17.1. Old, simple command line parser # [1046]14.2.17.2. Hush shell # [1047]14.2.17.3. Hush shell scripts # [1048]14.2.17.4. General rules o [1049]14.2.18. How can I load and uncompress a compressed image o [1050]14.2.19. How can I create an uImage from a ELF file o [1051]14.2.20. My standalone program does not work o [1052]14.2.21. Linux hangs after uncompressing the kernel o [1053]14.2.22. How can I implement automatic software updates? + [1054]14.3. Linux o [1055]14.3.1. Linux crashes randomly o [1056]14.3.2. Linux crashes when uncompressing the kernel o [1057]14.3.3. Linux Post Mortem Analysis o [1058]14.3.4. Linux kernel register usage o [1059]14.3.5. Linux Kernel Ignores my bootargs o [1060]14.3.6. Cannot configure Root Filesystem over NFS o [1061]14.3.7. Linux Kernel Panics because "init" process dies o [1062]14.3.8. Unable to open an initial console o [1063]14.3.9. System hangs when entering User Space (ARM) o [1064]14.3.10. Mounting a Filesystem over NFS hangs forever o [1065]14.3.11. Ethernet does not work in Linux o [1066]14.3.12. Loopback interface does not work o [1067]14.3.13. Linux kernel messages are not printed on the console o [1068]14.3.14. Linux ignores input when using the framebuffer driver o [1069]14.3.15. How to switch off the screen saver and the blinking cursor? o [1070]14.3.16. BogoMIPS Value too low o [1071]14.3.17. Linux Kernel crashes when using a ramdisk image o [1072]14.3.18. Ramdisk Greater than 4 MB Causes Problems o [1073]14.3.19. Combining a Kernel and a Ramdisk into a Multi-File Image o [1074]14.3.20. Adding Files to Ramdisk is Non Persistent o [1075]14.3.21. Kernel Configuration for PCMCIA o [1076]14.3.22. Configure Linux for PCMCIA Cards using the Card Services package o [1077]14.3.23. Configure Linux for PCMCIA Cards without the Card Services package # [1078]14.3.23.1. Using a MacOS Partition Table # [1079]14.3.23.2. Using a MS-DOS Partition Table o [1080]14.3.24. Boot-Time Configuration of MTD Partitions o [1081]14.3.25. Use NTP to synchronize system time against RTC o [1082]14.3.26. Configure Linux for XIP (Execution In Place) # [1083]14.3.26.1. XIP Kernel # [1084]14.3.26.2. Cramfs Filesystem # [1085]14.3.26.3. Hints and Notes # [1086]14.3.26.4. Space requirements and RAM saving, an example o [1087]14.3.27. Use SCC UART with Hardware Handshake o [1088]14.3.28. How can I access U-Boot environment variables in Linux? o [1089]14.3.29. The =appWeb= server hangs *OR* /dev/random hangs o [1090]14.3.30. Swapping over NFS o [1091]14.3.31. Using NFSv3 for NFS Root Filesystem o [1092]14.3.32. Using and Configuring the SocketCAN Driver o [1093]14.3.33. Telnet / SSH (dropbear) server not working + [1094]14.4. Self o [1095]14.4.1. How to Add Files to a SELF Ramdisk o [1096]14.4.2. How to Increase the Size of the Ramdisk + [1097]14.5. RTAI o [1098]14.5.1. Conflicts with asm clobber list + [1099]14.6. BDI2000 o [1100]14.6.1. Where can I find BDI2000 Configuration Files? o [1101]14.6.2. How to Debug Linux Exceptions o [1102]14.6.3. How to single step through "RFI" instruction o [1103]14.6.4. Setting a breakpoint doesn't work o [1104]14.6.5. Remote 'g' packet reply is too long + [1105]14.7. Motorola LITE5200 Board o [1106]14.7.1. LITE5200 Installation Howto o [1107]14.7.2. USB does not work on Lite5200 board 14. FAQ - Frequently Asked Questions This is a collection of questions which came up repeatedly. Give me more feedback and I will add more stuff here. The items are categorized whether they concern U-Boot itself, the Linux kernel or the [1108]SELF framework. 14.1. [1109]ELDK 14.1.1. [1110]ELDK Installation under [1111]FreeBSD Question: How can I install [1112]ELDK on a [1113]FreeBSD system? Answer: [Thanks to Rafal Jaworowski for these detailed instructions.] This is a short tutorial how to host [1114]ELDK on [1115]FreeBSD 5.x and 6.x. The procedure described below was tested on 5.2.1, 5.3 and 6-current releases; we assume the reader is equipped with the [1116]ELDK 3.x CDROM or ISO image for installation, and is familiar with [1117]FreeBSD basic administration tasks like ports/packages installation. 1. Prerequisites: 1. Install linux_base The first step is to install the Linux compatibility layer from ports /usr/ports/emulators/linux_base/ or packages ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages/em ulators/ ALERT! Please make sure to install version 7.1_5 (linux_base-7.1_5.tbz) or later; in particular, version 6.1.5 which can also be found in the ports tree does not work properly! The compatibility layer is activated by # kldload linux 2. Install bash Since [1118]ELDK and Linux build scripts are organised around bash while [1119]FreeBSD does not have it in base, this shell needs to be installed either from ports /usr/ports/shells/bash2/ or packages collection ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages/sh ells/ The installation puts the bash binary in /usr/local/bin. It is a good idea to create a symlink in /bin so that hash bang from scripts (#!/bin/bash) works without modifications: # cd /bin # ln -s /usr/local/bin/bash 2. Prepare [1120]ELDK ALERT! This step is only needed for [1121]ELDK release 3.1 and older versions. Copy the install files from the CDROM or ISO image to a writable location. Brand the [1122]ELDK installer as Linux ELF file: # cd # brandelf -t Linux ./install TIP Note: The following workaround might be a good alternative for the tedious copying of the installation CDROM to a writable location and manual branding: you can set a fallback branding in [1123]FreeBSD - when the loader cannot recognise the ELF brand it will switch to the last resort defined. # sysctl -w kern.elf32.fallback_brand=3 kern.elf32.fallback_brand: -1 -> 3 With this setting, the normal [1124]ELDK CDROM images should work. 3. Install [1125]ELDK normally as described in [1126]3.5.3. Initial Installation 4. Set envrionment variables and PATH as needed for [1127]ELDK (in bash); for example: bash$ export CROSS_COMPILE=ppc_8xx- bash$ export PATH=${PATH}:/opt/eldk/bin:/opt/eldk/usr/bin 5. Hints for building U-Boot: [1128]FreeBSD normally uses BSD-style 'make' in base, but in order to compile U-Boot 'gmake' (GNU make) has to be used; this is installed as part of the 'linux_base' package (see above). U-Boot should build according to standard [1129]ELDK instructions, for example: bash$ cd /opt/eldk/ppc_8xx/usr/src/u-boot-1.1.2 bash$ gmake TQM823L_config bash$ gmake all 6. Hints for building Linux: There are three issues with the Makefile in the Linux kernel source tree: o GNU make has to be used. o The 'expr' utility in [1130]FreeBSD base behaves differently from the version than is used in Linux so we need to modify the Makefile to explicitly use the Linux version (which is part of the Linux compatibility package). This is best achieved with defining "EXPR = /compat/linux/usr/bin/expr" somewhere at =Makefile='s beginning and replacing all references to 'expr' with the variable ${EXPR). o Some build steps (like when running 'scripts/mkdep' can generate very long arguments lists (especially is the Linux kernel tree is in a directory with long absolute filenames). A solution is to use xargs to split such long commands into several with shorter argument lists. The Linux kernel can then be built following the standard instructions, for example: bash$ cd /opt/eldk/ppc_8xx/usr/src/linux-2.4.25/ bash$ gmake mrproper bash$ gmake TQM823L_config bash$ gmake oldconfig bash$ gmake dep bash$ gmake -j6 uImage [1131]ELDK Installation Hangs Question: I try to install the [1132]ELDK on a Linux PC, and the installation hangs. It starts fine, but then it freezes like this: ... Preparing... ########################################### [100%] 1:db4-devel-ppc_4xx ########################################### [100%] Preparing... ########################################### [100%] 1:db4-utils-ppc_4xx ########################################### [100%] Preparing... ########################################### [100%] 1:glib2-ppc_4xx ########################################### [100%] Preparing... ########################################### [100%] 1:glib2-devel-ppc_4xx ########################################### [100%] Preparing... ########################################### [100%] Answer: This is almost certainly a FUTEX problem. To verify this, please wait until the process grinds to a halt, then use ps to find the pid of the "rpm" process that was started by the "install" program (use "ps -axf" which gives you a nice hierarchy, look for the "install" process, then for "rpm") and then attach to it with "strace -p". Most probably you will see the something like this: # strace -p 21197 Process 21197 attached - interrupt to quit futex(0x96fe17c, FUTEX_WAIT_PRIVATE, 1, NULL i. e. the process is hanging in a futex call. We have seen this more than once with differing Linux systems, but unfortunately we don't know a clean and reliable way to fix it yet. We suspect that it is a kernel/libc combination problem because it usually went away usually after changing the exact used kernel version. The only workaround we can recommend so far is to update your host system and install more recent versions of the Linux kernel and/or the glibc C library (assuming such are available for your Linux distribution; if not, falling back to a previous kernel version may help, too). Note: This is only needed for the installer, the problem does not happen with the regular use of the [1133]ELDK. 14.1.3. .gvfs: Permission Denied Question: When trying to install the [1134]ELDK, I get error messages like this for each and every package that gets installed: Preparing... ################### 100% 1: rpm... ################### 100% Error: Failed to stat /home/wd/.gvfs: Permission Denied This happens even though I run the installer as root. Answer: Even though flagged as an error, these messages are harmless warnings that can be safely ignored. Before the RPM tool starts to install a package, it checks if there is sufficient space for it in the file system. Unfortunately it is dumb and checks all mounted file systems for space, but the permissions of the ".gvfs" directory (the mount point for the Gnome Virtual File System) do not permit this. Note: Actually the messages are not printed despite the fact that you are running as root, but because you run as root. You have permissions to check the "$HOME/.gvfs" directory, while root gets an error: $ df -h /home/wd/.gvfs Filesystem Size Used Avail Use% Mounted on gvfs-fuse-daemon 0 0 0 - /home/wd/.gvfs $ sudo df -h /home/wd/.gvfs df: `/home/wd/.gvfs': Permission denied df: no file systems processed 14.1.4. Installation on Local Harddisk Question: I have a local harddisk drive connected to my target board. Can I install the [1135]ELDK on it and run it like a standard Linux distribution? Answer: Yes, this is possible. It requires only minor adjustments. The following example assumes you are using a SCSI disk drive, but the same can be done with standard SATA or PATA drives, too: 1. Boot the target with root file system over NFS. 2. Create the necessary partitions on your disk drive: you need at last a swap partition and a file system partition. bash-3.00# fdisk -l Disk /dev/sda: 36.9 GB, 36951490048 bytes 64 heads, 32 sectors/track, 35239 cylinders Units = cylinders of 2048 * 512 = 1048576 bytes Device Boot Start End Blocks Id System /dev/sda1 1 978 1001456 82 Linux swap / Solaris /dev/sda2 979 12423 11719680 83 Linux /dev/sda3 12424 23868 11719680 83 Linux /dev/sda4 23869 35239 11643904 83 Linux 3. Format the partititons: bash-3.00# mkswap /dev/sda1 bash-3.00# mke2fs -j -m1 /dev/sda2 4. Mount the file system: bash-3.00# mount /dev/sda2 /mnt 5. Copy the content of the (NFS) root file system into the mounted file system: bash-3.00# tar --one-file-system -c -f - / | ( cd /mnt ; tar xpf - ) 6. Adjust /etc/fstab for the disk file system: bash-3.00# vi /mnt/etc/fstab bash-3.00# cat /mnt/etc/fstab /dev/sda2 / ext3 defaults 1 1 /dev/sda1 swap swap defaults 0 0 proc /proc proc defaults 0 0 sysfs /sys sysfs defaults 0 0 7. Adjust /etc/rc.sysinit for running from local disk; remove the following comments: bash-3.00# diff -u /mnt/etc/rc.sysinit.ORIG /mnt/etc/rc.sysinit --- /mnt/etc/rc.sysinit.ORIG 2007-01-21 04:37:00.000000000 +0100 +++ /mnt/etc/rc.sysinit 2007-03-02 10:58:22.000000000 +0100 @@ -460,9 +460,9 @@ # Remount the root filesystem read-write. update_boot_stage RCmountfs -#state=`LC_ALL=C awk '/ \/ / && ($3 !~ /rootfs/) { print $4 }' /proc/mounts` -#[ "$state" != "rw" -a "$READONLY" != "yes" ] && \ -# action $"Remounting root filesystem in read-write mode: " mount -n -o remou nt,rw / +state=`LC_ALL=C awk '/ \/ / && ($3 !~ /rootfs/) { print $4 }' /proc/mounts` +[ "$state" != "rw" -a "$READONLY" != "yes" ] && \ + action $"Remounting root filesystem in read-write mode: " mount -n -o remoun t,rw / # Clean up SELinux labels if [ -n "$SELINUX" ]; then 8. Unmount disk: bash-3.00# umount /mnt 9. Reboot, and adjust boot arguments to use disk partition as root file system => setenv diskargs setenv bootargs root=/dev/sda2 ro => setenv net_disk 'tftp ${loadaddr} ${bootfile};run diskargs addip addcons;boo tm' => saveenv 10. Boot with these settings => run net_disk 14.1.5. System Include Files Missing Question: when installing [1136]ELDK on Ubuntu 6.06 dapper drake I get the following error messages.... Preparing... ########################################### [100%] 1:kernel-source-ppc_6xx ########################################### [100%] Configuring kernel... scripts/basic/fixdep.c:107:23: error: sys/types.h: No such file or directory scripts/basic/fixdep.c:108:22: error: sys/stat.h: No such file or directory scripts/basic/fixdep.c:109:22: error: sys/mman.h: No such file or directory ... Answer: The Linux installation on your host is missing essential files that are needed to perform software development and use a C compiler. On Ubuntu, check for example if you miss a "libc6-dev" package. The specific package name differs from distribution to distribution; on Fedora, you need for example the "glibc-headers" package. If you want to work with a Linux kernel you will probably also need other packages. 14.1.6. patch: command not found Question: When installing [1137]ELDK on Ubuntu I get the following error message: ...prepare-kernel.sh: line 376: patch: command not found Answer: The error message contains clear hints for the solution: the "patch" command cannot be found on your system, so most probably it has not been installed yet. Please try: $ sudo apt-get install patch 14.1.7. [1138]ELDK Include Files Missing Question: After configuring and compiling a Linux kernel in the kernel source tree that comes with the [1139]ELDK, I cannot compile user space programs any more - I get error messages because many #include file like etc. are missing. This is with [1140]ELDK 4.0 or 4.1. Answer: This problem is caused by the way how the [1141]ELDK is packaged. At the moment, the [1142]ELDK kernel headers are not packed into a separate "kernel-headers" RPM to avoid duplication, because the kernel source tree is always installed. Instead, the [1143]ELDK "kernel-headers" package is just a set of symlinks. This worked fine in the past, but fails with the new support for ARCH=powerpc systems. The next version of the [1144]ELDK will contain a real kernel-headers RPM, which will fix this problem. As a workaround on current systems, you can install the real kernel include files into the "include/asm", "include/linux" and "include/mtd" directories. To do this, the following commands can be used: bash$ /bin/rpm -e kernel-headers-ppc_ bash$ cd /ppc_ bash$ rm usr/include/asm bash$ tar -xvzf kernel-headers-powerpc.tar.gz The tarball mentioned above can be downloaded [1145]here. It contains the include files that get installed by running the "make ARCH=powerpc headers_install" command in the Linux kernel tree. This problem is fixed in [1146]ELDK 4.2 and later releases. 14.1.8. Using the [1147]ELDK on a 64 bit platform As the [1148]ELDK is compiled for 32-bit host systems, a compatibility layer is required on 64-bit systems. This package is usually called ia32-libs. So on a Debian or Ubuntu system a sudo apt-get install ia32-libs should be enough to make the [1149]ELDK work. On the U-Boot mailing list, it was reported that for a 64 bit Fedora 11 the following should be enough: sudo yum -y install glibc.i686 zlib.i686 14.1.9. GDB Problems with BDI2000/BDI3000 on e500 Cores Question: I am using the cross-gdb from [1150]ELDK together with the BDI [1151]JTAG hardware debugger from Abatron to debug my e500 based system. When I try to attach to the BDI from within gdb, I get an error message like this: $ ppc_85xxDP-gdb u-boot ....... (gdb) target remote bdi14:2001 Remote debugging using bdi14:2001 Remote 'g' packet reply is too long: 000000600fb7bd380fb7bf78e000450000000001000000000ffdd8c00ffd912400001c010000000 00ffdd9fc0000000148000084ffffde790ffd8d000ffdd9fc0ffc25dc0ffee1400fb7bda00ffee0 410fb7bd980fb7bd9c0ffdd9fc0ffce6e40ffd4154000000010ffd4154000000000ffee0400fb81 3d80ffd95fc00000000b0d58ab4000000605f3fefe50fb7bd38f9a645f60fb7bf78ef5efef5e000 4500be7b3f9f0000000157dfffaf00000000e959f23e0ffdd8c09e726d7b0ffd912489bbb30b000 01c01a73b7ff600000000773ae2de0ffdd9fc2bddd17e000000019b9bbe3748000084fd7bebf6ff ffde796c2e392f0ffd8d0066fb55fb0ffdd9fc2b66bef40ffc25dcff7fffce0ffee1401aac69e00 fb7bda0dacddcff0ffee0415497fbbd0fb7bd989d661b970fb7bd9c647efdde0ffdd9fcd9603cb3 0ffce6e4f42113670ffd415475bdf5db00000001de6cd7f70ffd4154fef8b18f00000000ba2decd 30ffee0409b8beafe0fb813d8722c7bff0ffd95fcddf79dcd000000000ff9890c00029200280220 480ff9862c0ff9864000000000deadbeef (gdb) Answer: If you are using at least BDI firmware v1.09 then most probably you forgot to include the following directive in the BDI config file: [TARGET] REGLIST E500 ;send registers in E500 sequence to GDB Also be sure that the gdb really thinks that it debugs an e500 core: (gdb) show architecture The target architecture is set automatically (currently powerpc:e500) (gdb) If this is not the case, then fix this problem first. It might just be that you are not using the right cross debugger in the first place. 14.1.10. How can I check if Floating Point support is working? Question: The floating point performance of my P2020 QorIQ processor is really poor. I am not using the [1152]ELDK, but a tool chain from FOOBAR. Can this be a problem? What can I do to verify this? Answer: The P20xx QorIQ processors use an e500v2 core which does not include a normal Floating Point Unit (FPU), but instead a Signal Processing Engine (SPE Version 2). You can run FP calculations on the SPE, but there are no special FP registers available as on a normal FPU, so General Purpose Registers must be used for passing of FP operands. While this is still much faster than pure soft-float emulation, it is missing the advantages and the speed of a full-blown, separate standard FPU with a full FP register set. Also, your tool chain needs to be aware of this feature, and must contain support for it. Eventually special compiler options are needed - check your documentation. With the [1153]ELDK, the needed settings are automatically pre-set when you just chose the correct target architecture packages, cf. [1154]3.4. Supported Target Architectures To test what your tool chain is doing, you best compile a smal test program and check the generated code. The following examples were done with [1155]ELDK 4.2 for [1156]Power Architecture® targets: 1. Test Program: $ cat fp_test.c double foo (double x, double y) { double z; z = (x + y) / (x * y); return z; } 2. Build for normal FPU support (using the ppc_6xx target architecture): $ export CROSS_COMPILE=ppc_6xx- $ ppc_6xx-gcc -S -O fp_test.c Check results: $ cat fp_test.s .file "fp_test.c" .section ".text" .align 2 .globl foo .type foo, @function foo: fadd 0,1,2 fmul 1,1,2 fdiv 1,0,1 blr .size foo, .-foo .ident "GCC: (GNU) 4.2.2" .section .note.GNU-stack,"",@progbits The use of floating point machine instructions ("fadd", "fmul", "fdiv") and the fact that no additional register use is needed is a clear indication that full support for the hardware FPU is available in this configuration. 3. Build for soft-float emulation (using the ppc_8xx target architecure): $ export CROSS_COMPILE=ppc_8xx- $ ppc_8xx-gcc -S -O fp_test.c $ cat fp_test.s .file "fp_test.c" .globl __adddf3 .globl __muldf3 .globl __divdf3 .section ".text" .align 2 .globl foo .type foo, @function foo: stwu 1,-48(1) mflr 0 stw 24,16(1) stw 25,20(1) stw 26,24(1) stw 27,28(1) stw 28,32(1) stw 29,36(1) stw 0,52(1) mr 28,3 mr 29,4 mr 26,5 mr 27,6 bl __adddf3 mr 24,3 mr 25,4 mr 3,28 mr 4,29 mr 5,26 mr 6,27 bl __muldf3 mr 5,3 mr 6,4 mr 3,24 mr 4,25 bl __divdf3 lwz 0,52(1) mtlr 0 lwz 24,16(1) lwz 25,20(1) lwz 26,24(1) lwz 27,28(1) lwz 28,32(1) lwz 29,36(1) addi 1,1,48 blr .size foo, .-foo .ident "GCC: (GNU) 4.2.2" .section .note.GNU-stack,"",@progbits The fact that the compiler is calling helper functions (__adddf3, __muldf3, __divdf3) combined with heavy use of the General Purpose Registers is a clear indication for software-emulated FP support - and explains why this is so slow compared to a real FPU. 4. Build for SPE v2 support (as needed for example for a P2020 QorIQ processor, using the ppc_85xxDP target architecture): $ export CROSS_COMPILE=ppc_85xxDP- $ ppc_85xxDP-gcc -S -O fp_test.c $ cat fp_test.s .file "fp_test.c" .section ".text" .align 2 .globl foo .type foo, @function foo: stwu 1,-48(1) stw 3,8(1) stw 4,12(1) stw 5,16(1) stw 6,20(1) evmergelo 0,3,4 evmergelo 9,5,6 efdadd 11,0,9 efdmul 0,0,9 efddiv 11,11,0 evstdd 11,24(1) evmergehi 9,11,11 mr 10,11 stw 9,32(1) stw 10,36(1) mr 3,9 mr 4,10 addi 1,1,48 blr .size foo, .-foo .ident "GCC: (GNU) 4.2.2" .section .note.GNU-stack,"",@progbits Here we can see moderate use of General Purpos Registers combined with the use of SPE machine instructions (evmergelo, efdadd, efdmul, efddiv, evstdd, evmergehi) which proves that the compiler really generates code that supports the SPE. 14.1.11. [1157]ELDK 2.x Installation Aborts Question: I tried to install [1158]ELDK version 2.x on a [1159]SuSE 8.2 / [1160]SuSE 9 / [1161]RedHat-9 Linux host but failed - it terminated without installing any packages. Why? Answer: Newer Linux distributions use libraries that are incompatible to those used by the [1162]ELDK's installation tools. This problem was fixed in later releases of the [1163]ELDK (version 3.0 and later). It is therefore recommended to use a more recent version of the [1164]ELDK. If you really want to install an old version, the following back-port is available: Please download the file [1165]ftp://ftp.denx.de/pub/tmp/ELDK-update-2.2.0.tar.bz2 Then change into the source tree with the [1166]ELDK files and perform the following operations: bash$ rm RPMS/rpm-4.0.3-1.03b_2.i386.rpm \ RPMS/rpm-build-4.0.3-1.03b_2.i386.rpm \ RPMS/rpm-devel-4.0.3-1.03b_2.i386.rpm \ tools/usr/lib/rpm/rpmpopt-4.0.3 bash$ tar jxf /tmp/ELDK-update-2.2.0.tar.bz2 Then build the ISO image as documented, and try again. 14.1.12. Enable SSH Access Question: How can I enable SSH access to the target system when running with the [1167]ELDK file system mounted over NFS? Answer: The [1168]ELDK includes the dropbear SSH server and client packages (see [1169]http://matt.ucc.asn.au/dropbear/dropbear.html). To enable SSH access, you must first generate the SSH host keys for RSA and DSS: # dropbearkey -t rsa -f /etc/dropbear/dropbear_rsa_host_key Will output 1024 bit rsa secret key to '/etc/dropbear/dropbear_rsa_host_key' Generating key, this may take a while... Public key portion is: ssh-rsa ... Fingerprint: md5 ... # dropbearkey -t dss -f /etc/dropbear/dropbear_dss_host_key Will output 1024 bit dss secret key to '/etc/dropbear/dropbear_dss_host_key' Generating key, this may take a while... Public key portion is: ssh-dss ... Fingerprint: md5 ... Then you can start the dropbear daemon: # dropbear Now you should be able to access the target system through SSH. 14.2. U-Boot 14.2.1. Can U-Boot be configured such that it can be started in RAM? Question: I don't want to erase my flash memory because I'm not sure if my new U-Boot image will work. Is it possible to configure U-Boot such that I can load it into RAM instead of flash, and start it from my old boot loader? Answer: No. (Unless you're using a Blackfin processor, but you probably aren't.) Question: But I've been told it is possible?? Answer: Well, yes. Of course this is possible. This is software, so everything is possible. But it is difficult, unsupported, and fraught with peril. You are on your own if you choose to do it. And it will not help you to solve your problem. Question: Why? Answer: U-Boot expects to see a virgin [1170]CPU, i. e. the [1171]CPU state must match what you see if the processor starts executing the first instructions when it comes out of reset. If you want to start U-Boot from another boot loader, you must disable a lot of code, i. e. all initialization parts that already have been performed by this other boot loader, like setting up the memory controller, initializing the [1172]SDRAM, initializing the serial port, setting up a stack frame etc. Also you must disable the relocation to RAM and adjust the link addresses etc. This requires a lot of experience with U-Boot, and the fact that you had to ask if this can be done means that you are not in a position to do this. The code you have to disable contains the most critical parts in U-Boot, i. e. these are the areas where 99% or more of all errors are located when you port U-Boot to a new hardware. In the result, your RAM image may work, but in the end you will need a full image to program the flash memory with it, and then you will have to enable all this highly critical and completely untested code. You see? You cannot use a RAM version of U-Boot to avoid testing a flash version, so you can save all this effort and just burn your image to flash. Question: So how can I test an U-Boot image and recover my system if it doesn't work? Answer: Attach a BDI2000 (or any appropriate [1173]JTAG ICE) to your board, burn the image to flash, and debug it in its natural environment, i.e. U-Boot being the boot loader of the system and taking control over the [1174]CPU right as it comes out of reset. If something goes wrong, erase the flash and program a new image. This is a routine job using a BDI2000. 14.2.2. Relocation cannot be done when using -mrelocatable Question: I use [1175]ELDK version 3.0. When I build U-Boot I get error messages like this: {standard input}: Assembler messages: {standard input}:4998: Error: Relocation cannot be done when using -mrelocatabl e ... Answer: [1176]ELDK 3.0 uses GCC-3.2.2; your U-Boot sources are too old for this compiler. GCC-3.x requires a few adaptions which were added in later versions of U-Boot. Use for example the source tree (1.0.2) which is included with the [1177]ELDK, or download the latest version from [1178]CVS. 14.2.3. Source object has [1179]EABI version 4, but target has [1180]EABI version 0 Question: When trying to build U-Boot with an [1181]EABI compliant tool chain, I get such error messages: arm-ld: ERROR: Source object ... has EABI version 4, but target ... has EABI ve rsion 0 What does that mean, and how can I fix that? Answer: "EABI version 0" means the "apcs-gnu" [1182]ABI, while "EABI version 4" is the "aapcs-linux" [1183]ABI, aka "gnueabi". All U-Boot ARM sources are built with "-mapcs-gnu" option set in "cpu/arm/config.mk", while libgcc.a modules are built in "gnueabi" format, which is for example the ARM GCC default in [1184]ELDK Release 4.2. So the real problem is compatibility between toolchain [1185]ABI and U-Boot ARM [1186]ABI. In the Linux kernel there is a special kernel config option for [1187]EABI-enabled tool chains (CONFIG_AEABI), which enables special pieces of code in ARM assembler modules. We could follow this approach, reworking existing assembler sources and respective config.mk files in U-Boot. Alternatively, the tool chain could provide a separate version of libgcc.a built with old [1188]ABI. This could be done using the multilib approach. The advantage here is that no U-boot changes will be required. 14.2.4. U-Boot crashes after relocation to RAM Question: I have ported U-Boot to a custom board. It starts OK, but crashes or hangs after relocating itself to RAM. Why? Answer: Your [1189]SDRAM initialization is bad, and the system crashes when it tries to fetch instructions from RAM. Note that simple read and write accesses may still work, it's the burst mode that is failing. This only shows up when caches are enabled because cache is the primary (or only) user of burst operations in U-Boot. In Linux, burst accesses may also result from [1190]DMA. For example, it is typical that a system may crash under heavy network load if the Ethernet controller uses [1191]DMA to memory. ALERT! It is NOT sufficient to program the memory controller of your [1192]CPU; each [1193]SDRAM chip also requires a specific initialization sequence which you must adhere to to the letter - check with the chip manufacturer's manual. It has been observed that some operating systems like pSOS+ or VxWorks do not stress the memory subsystem as much as Linux or other UNIX systems like LynxOS do, so just because your board appears to work running another OS does not mean it is 100% OK. Standard memory tests are not effective in identifying this type of problem because they do not cause stressful cache burst read/write operations. With this caveat in mind, reportedly this program has found memory problems before: [1194]http://pyropus.ca/software/memtester/ Argument: But my board ran fine with bootloader XYZ and/or operating system ABC. Answer: Double-check your configuration that you claim runs properly... 1. Are you sure the [1195]SDRAM is initialized using the same init sequence and values? 2. Are you sure the memory controlling registers are set the same? 3. Are you sure your other configuration uses caches and/or [1196]DMA? If it doesn't, it isn't a valid comparison. 14.2.5. Warning - bad CRC, using default environment Question: I have ported U-Boot to a custom board. It seems to boot OK, but it prints: *** Warning - bad CRC, using default environment Why? Answer: Most probably everything is OK. The message is printed because the flash sector or ERPROM containing the environment variables has never been initialized yet. The message will go away as soon as you save the envrionment variables using the saveenv command. 14.2.6. Net: No ethernet found Question: I have ported U-Boot to a custom board. It seems to boot OK, but it prints: Net: No ethernet found. Why? Answer: Some network drivers (especially on ppc4xx) respond with such a message when no valid [1197]MAC address has been defined. Please make sure that a valid [1198]MAC address has been defined in the environment (using the "setenv" command). Then store the environment (using the "saveenv" command). After the next reboot, the Ethernet interface should be available. 14.2.7. Wrong debug symbols after relocation Question: I want to debug U-Boot after relocation to RAM, but it doesn't work since all the symbols are at wrong addresses now. Answer: To debug parts of U-Boot that are running from ROM/flash, i. e. before relocation, just use a command like "powerpc-linux-gdb uboot" as usual. For parts of U-Boot that run from RAM, i. e. after relocation, use "powerpc-linux-gdb" without arguments, and use the add-symbol-file command in GDB to load the symbol table at the relocation address in RAM. The only problem is that you need to know that address, which depends on RAM size, length reserved for U-Boot, size of "protected RAM" area, etc. If in doubt, enable DEBUG mode when building U-Boot so it prints the address to the console. TIP Hint: I use definitions like these in my .gdbinit file: define rom symbol-file file u-boot end define ram symbol-file add-symbol-file u-boot 0x01fe0000 end Note: when you want to switch modes during one debug session (i. e. without restarting GDB) you can "delete" the current symbol information by using the symbol-file command without arguments, and then either using "symbol-file u-boot" for code before relocation, or "add-symbol-file u-boot _offset_" for code after relocation. 14.2.8. Decoding U-Boot Crash Dumps When you are porting U-Boot to new hardware, or implementing extensions, you might run into situations where U-Boot crashes and prints a register dump and a stack trace, for example like this: Bus Fault @ 0x00f8d70c, fixup 0x00000000 Machine check in kernel mode. Caused by (from msr): regs 00f52cf8 Unknown values in msr NIP: 00F8D70C XER: 0000005F LR: 00F8D6F4 REGS: 00f52cf8 TRAP: 0200 DAR: F9F68C0 0 MSR: 00009002 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 00 GPR00: 00016ACC 00F52DE8 00000000 F9F68C00 00FA38EC 00000001 F9F68BF8 0000000B GPR08: 00000002 00F55470 00000000 00F52D94 44004024 00000000 00FA2F00 C0F75000 GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 GPR24: 00000000 00FA38EC 00F553C0 00F55480 00000000 00F52F80 00FA41C0 00000001 Call backtrace: 00000000 00F8F998 00F8FA88 00F8FAF8 00F90B5C 00F90CF8 00F8385C 00F79E6C 00F773B0 machine check To find out what happened, you can try to decode the stack backtrace (the list of addresses printed after the "Call backtrace:" line. The [1199]backtrace tool can be used for this purpose. However, there is a little problem: the addresses printed for the stack backtrace are after relocation of the U-Boot code to RAM; to use the backtrace tool you need to know U-Boot's address offset (the difference between the start address of U-Boot in flash and its relocation address in RAM). The easiest way to find out the relocation address is to enable debugging for the U-Boot source file lib_*/board.c - U-Boot will then print some debug messages ... Now running in RAM - U-Boot at: 00f75000 ... Now you have to calculate the address offset between your link address (The value of the TEXT_BASE definition in your board/?/config.mk file). In our case this value is 0x40000000, so the address offset is 0x40000000 - 0x00f75000 = 0x3f08b000 Now we use the backtrace script with the System.map file in the U-Boot source tree and this address offset: -> backtrace System.map 0x3f08b000 Reading symbols from System.map Using Address Offset 0x3f08b000 0x3f08b000 -- unknown address 0x4001a998 -- 0x4001a8d0 + 0x00c8 free_pipe 0x4001aa88 -- 0x4001aa2c + 0x005c free_pipe_list 0x4001aaf8 -- 0x4001aad0 + 0x0028 run_list 0x4001bb5c -- 0x4001ba68 + 0x00f4 parse_stream_outer 0x4001bcf8 -- 0x4001bcd8 + 0x0020 parse_file_outer 0x4000e85c -- 0x4000e6f8 + 0x0164 main_loop 0x40004e6c -- 0x40004b9c + 0x02d0 board_init_r 0x400023b0 -- 0x400023b0 + 0x0000 trap_init In this case the last "good" entry on the stack was in free_pipe... 14.2.9. Porting Problem: cannot move location counter backwards Question: I'm trying to port U-Boot to a new board and the linker throws an error message like this: board//u-boot.lds:75 cannot move location counter backwards (from 0 0000000b0008010 to 00000000b0008000) Answer: Check your linker script board/your_board/u-boot.lds which controls how the object files are linked together to build the U-Boot image. It looks as if your board uses an "embedded" environment, i. e. the flash sector containing the environment variables is surrounded by code. The u-boot.lds tries to collect as many as possible code in the first part, making the gap between this first part and the environment sector as small as possible. Everything that does not fit is then placed in the second part, after the environment sector. Some your modifications caused the code that was put in this first part to grow, so that the linker finds that it would have to overwrite space that is already used. Try commenting out one (or more) line(s) before the line containing the "common/environment.o" statement. [ "lib_generic/zlib.o" is usually a good candidate for testing as it's big ]. Once you get U-Boot linked, you can check in the u-boot.map file how big the gap is, and which object files could be used to fill it up again. 14.2.10. U-Boot Doesn't Run after Upgrading my Compiler Question: I encountered a big problem that U-Boot 1.1.4 compiled by [1200]ELDK 4.1 for MPC82xx crashed. But if I build it using gcc-3.4.6 based cross tools, U-Boot on my board boots correctly. The same U-Boot code built by [1201]ELDK 4.1 (gcc-4.0) failed, nothing occurs on the serial port. Answer: This is often a missing volatile attribute on shared variable references, particularly hardware registers. Newer compiler versions optimize more aggressively, making missing volatile attributes visible. If you use -O0 (no optimization) does it fix the problem? If it does, it most likely is an optimization/volatile issue. The hard part is figuring out where. Device handling and board-specific code is the place to start. 14.2.11. How Can I Reduce The Image Size? Question: I am trying to reduce the size of the u-boot.bin file so that it fits into 256 KB. I disabled all the drivers that I didn't need but the binary size is still 512KB, it seems to be a hard number coded in somewhere. Where can the image size be altered from? Answer: Some processors have a fixed reset vector address at 0xFFFFFFFC, so the U-Boot image has to include that address, i. e. it covers the full range from the start address to the end of the 32 bit address space. In such a case, the start address must be changed - check the setting of TEXT_BASE in your board//config.mk file. 14.2.12. Erasing Flash Fails Question: I tried to erase the flash memory like erase 40050000 40050100 It fails. What am I doing wrong? Answer: Remember that flash memory cannot be erased in arbitrary areas, but only in so called "erase regions" or "sectors". If you have U-Boot running you can use the flinfo (Flash information, short fli) command to print information about the flash memory on your board, for instance: => fli Bank # 1: AMD AM29LV160B (16 Mbit, bottom boot sect) Size: 4 MB in 35 Sectors Sector Start Addresses: 40000000 (RO) 40008000 (RO) 4000C000 (RO) 40010000 (RO) 40020000 (RO) 40040000 40060000 40080000 400A0000 400C0000 400E0000 40100000 40120000 40140000 40160000 40180000 401A0000 401C0000 401E0000 40200000 40220000 40240000 40260000 40280000 402A0000 402C0000 402E0000 40300000 40320000 40340000 40360000 40380000 403A0000 403C0000 403E0000 In the example above, the area 40050000 ... 40050100 lies right in the middle of a erase unit (40040000 ... 4005FFFF), so you cannot erase it without erasing the whole sector, i. e. you have to type => erase 40040000 4005FFFF Also note that there are some sectors marked as read-only ((RO)); you cannot erase or overwrite these sectors without un-protecting the sectors first (see the U-Boot protect command). 14.2.13. Ethernet Does Not Work Question: Ethernet does not work on my board. Answer: Maybe you forgot to set a [1202]MAC address? Check if the "ethaddr" environment variable is defined, and if it has a sane value. If there are more than one Ethernet interfaces on your board, you may also have to check the [1203]MAC addresses for these, i. e. check the "eth1addr", "eth2addr", etc. variables, too. Question: I have configured a [1204]MAC address of 01:02:03:04:05:06, and I can see that an ARP packet is sent by U-Boot, and that an ARP reply is sent by the server, but U-Boot never receives any packets. What's wrong? Answer: You have chosen a [1205]MAC address which, according to the ANSI/IEEE 802-1990 standard, has the multicast bit set. Under normal conditions a network interface discards such packets, and this is what U-Boot is doing. This is not a bug, but correct behaviour. Please use only valid [1206]MAC addresses that were assigned to you. For bring-up testing in the lab you can also use so-called locally administered ethernet addresses. These are addresses that have the 2nd LSB in the most significant byte of [1207]MAC address set. The gen_eth_addr tool that comes with U-Boot (see "tools/gen_eth_addr" ) can be used to generate random addresses from this pool. 14.2.14. Where Can I Get a Valid [1208]MAC Address from? Question: Where can I get a valid [1209]MAC address from? Answer: You have to buy a block of 4096 [1210]MAC addresses (IAB = Individual Address Block) or a block of 16M [1211]MAC addresses (OUI = Organizationally Unique Identifier, also referred to as 'company id') from IEEE Registration Authority. The current cost of an IAB is $550.00, the cost of an OUI is $1,650.00. See [1212]http://standards.ieee.org/regauth/oui/index.shtml You can buy Eproms containing [1213]MAC addresses from: [1214]Maxim or [1215]Microchip. You can set the "locally administered" bit to make your own [1216]MAC address (no guarantee of uniqueness, but pretty good odds if you don't do something dumb). Ref: [1217]Wikipedia Universally administered and locally administered addresses are distinguished by setting the second least significant bit of the most significant byte of the address. If the bit is 0, the address is universally administered. If it is 1, the address is locally administered. The bit is 0 in all OUIs. For example, 02-00-00-00-00-01. The most significant byte is 02h. The binary is 00000010 and the second least significant bit is 1. Therefore, it is a locally administered address. In U-Boot, you can use the "gen_eth_addr" tool to generate a random "locally administered" [1218]MAC address. Here are the needed commands: $ make tools/gen_eth_addr cc tools/gen_eth_addr.c -o tools/gen_eth_addr $ tools/gen_eth_addr ba:d0:4a:9c:4e:ce 14.2.15. Why do I get [1219]TFTP timeouts? Question 1:: When trying to download a file from the [1220]TFTP server I always get timeouts like these: ... Loading: #######T ##################################T###################T ####T ##T # ###T #T #########T ########T #############T ##T #############T ########T #############T #####T ###T ######T #######T #######T #############T ##T ##############T ########### ########### done If the target is connected directly to the host PC (i. e. without a switch inbetween) the problem goes away or is at least less incisive. What's wrong? Answer 1:: Most probably you have a full duplex/half duplex problem. Verify that U-Boot is setting the ethernet interface on your board to the proper duplex mode (full/half). I'm guessing your board is half duplex but your switch is full (typical of a switch ;-). The switch sends traffic to your board while your board is transmitting... that is a collision (late collision at that) to your board but is OK to the switch. This doesn't happen nearly as much with a direct link to your PC since then you have a dedicated link without much asynchronous traffic. The software (U-Boot/Linux) needs to poll the [1221]PHY chip for duplex mode and then (re)configure the [1222]MAC chip (separate or built into the [1223]CPU) to match. If the poll isn't happening or has a bug, you have problems like described above. Question 2:: When I use tftp, there are some problems. My terminal always displays "Loading: T T T T T T T T T T T T T T T T T T T T". The whole information as follows: U-Boot 1.1.4_XT (Jun 6 2006 - 17:36:18) U-Boot code: 0C300000 -> 0C31AD70 BSS: -> 0C31EF98 RAM Configuration: Bank #0: 0c000000 8 MB Bank #1: 0c800000 8 MB Flash: 2 MB *** Warning - bad CRC, using default environment In: serial Out: serial Err: serial Hit any key to stop autoboot: 0 XT=> help tftp tftpboot [loadAddress] [bootfilename] XT=> tftpboot 0x0c700000 image.bin TFTP from server 192.168.0.23; our IP address is 192.168.0.70 Filename 'image.bin'. Load address: 0xc700000 Loading: T T T T T T T T T T T T T T T T T T T T Retry count exceeded; starting again TFTP from server 192.168.0.23; our IP address is 192.168.0.70 Would someone give me some suggestions? Answer 2:: (1) Verify your [1224]TFTP server is working. On a machine (not the [1225]TFTP server nor your development board) use tftp to read the target file. $ tftp 192.168.0.23 get image.bin If this doesn't work, fix your [1226]TFTP server configuration and make sure it is running. (2) If your [1227]TFTP server is working, run ethereal (or equivalent ethernet sniffing) to see what ethernet packets are being sent by your development board. It usually works best to run ethereal on your [1228]TFTP server (if you run it on a different machine and you use an ethernet switch, the third machine likely won't see the tftp packets). 14.2.16. Why is my Ethernet operation not reliable? Question: My ethernet connection is not working reliable. On one switch it works fine, but on another one it doesn't. or: Question: I always see transmit errors or timeouts for the first packet of a download, but then it works well. or: Question: I cannot mount the Linux root file system over NFS; especially not with recent Linux kernel versions (older kernel versions work better). Specifying "proto=tcp" as mount option greatly improves the situation. etc. Answer: There are many possible explanations for such problems. After eliminating the obvious sources (like broken cables etc.) you should check the configuration of your Ethernet [1229]PHY. One common cause of problems is if your [1230]PHY is hard configured in duplex mode (for example 100baseTX Full Duplex or 10baseT Full Duplex). If such a setup is combined with a autonegotiating switch, then trouble is ahead. [1231]Jerry Van Baren explained this as follows: Ignoring the configuration where both ends are (presumably correctly) manually configured, you end up with five cases, two of them misconfigured and WRONG: 1) Autonegotiation <-> autonegotiation - reliable. 2) 10bT half duplex <-> autonegotiation - reliable. 3) 100bT half duplex <-> autonegotiation - reliable. 4) 10bT *FULL* duplex <-> autonegotiation - *UNreliable*. 5) 100bT *FULL* duplex <-> autonegotiation - *UNreliable*. The problem that I've observed is that the *humans* (the weak links) that do the manual configuration don't understand that "parallel detection" *must be* half duplex by definition in the spec (it is hard to define a reliable algorithm to detect full duplex capability so the spec writers punted). As a result, the human invariably picks "full duplex" because everybody knows full duplex is better... and end up as case (4) or (5). They inadvertently end up with a slower unreliable link (lots of "collisions" resulting in runt packets) rather than the faster better link they thought they were picking (d'oh!). The really bad thing is that the network works fine in testing on an isolated LAN with no traffic and absolutely craps its pants when it hits the real world. That is my reasoning behind my statement that we can generally ignore the autonegotiation <-> fixed configuration case because the odds of it working properly are poor anyway. Rule: Always try to set up your [1232]PHY for autonegotiation. If you must use some fixed setting, then set it to half duplex mode. If you really must use a fixed full-duplex setting, then you absolutley must make sure that the link partner is configured exactly the same. 14.2.17. How the Command Line Parsing Works There are two different command line parsers available with U-Boot: the old "simple" one, and the much more powerful "hush" shell: 14.2.17.1. Old, simple command line parser * supports environment variables (through setenv / saveenv commands) * several commands on one line, separated by ';' * variable substitution using "... ${_variablename_} ..." syntax ALERT! NOTE: Older versions of U-Boot used "$(...)" for variable substitution. Support for this syntax is still present in current versions, but will be removed soon. Please use "${...}" instead, which has the additional benefit that your environment definitions are compatible with the Hush shell, too. * special characters ('$', ';') can be escaped by prefixing with '\', for example: setenv bootcmd bootm \${address} * You can also escape text by enclosing in single apostrophes, for example: setenv addip 'setenv bootargs ${bootargs} ip=${ipaddr}:${serverip}:${ga tewayip}:${netmask}:${hostname}:${netdev}:off' 14.2.17.2. Hush shell * similar to Bourne shell, with control structures like if...then...else...fi, for...do...done, while...do...done, until...do...done, ... * supports environment ("global") variables (through setenv / saveenv commands) and local shell variables (through standard shell syntax name=value ); only environment variables can be used with the run command, especially as the variable to run (i. e. the first argument). * In the current implementation, the local variables space and global environment variables space are separated. Local variables are those you define by simply typing like name=value. To access a local variable later on, you have to write '$name' or '${name}'; to execute the contents of a variable directly you can type '$name' at the command prompt. Note that local variables can only be used for simple commands, not for compound commands etc. * Global environment variables are those you can set and print using setenv and printenv. To run a command stored in such a variable, you need to use the run command, and you must not use the '$' sign to access them. * To store commands and special characters in a variable, use single quotation marks surrounding the whole text of the variable, instead of the backslashes before semicolons and special symbols. * Be careful when using the hash ('#') character - like with a "real" Bourne shell it is the comment character, so you have to escape it when you use it in the value of a variable. Examples: setenv bootcmd bootm \$address setenv addip 'setenv bootargs $bootargs ip=$ipaddr:$serverip:$gatewayip :$netmask:$hostname:$netdev:off' 14.2.17.3. Hush shell scripts Here are a few examples for the use of the advanced capabilities of the hush shell in U-Boot environment variables or scripts: Example: => setenv check 'if imi $addr; then echo Image OK; else echo Image corrupted!!; fi' => print check check=if imi $addr; then echo Image OK; else echo Image corrupted!!; fi => addr=0 ; run check ## Checking Image at 00000000 ... Bad Magic Number Image corrupted!! => addr=40000 ;run check ## Checking Image at 00040000 ... Image Name: ARM Linux-2.4.18 Created: 2003-06-02 14:10:54 UTC Image Type: ARM Linux Kernel Image (gzip compressed) Data Size: 801609 Bytes = 782.8 kB Load Address: 0c008000 Entry Point: 0c008000 Verifying Checksum ... OK Image OK Instead of "echo Image OK" there could be a command (sequence) to boot or otherwise deal with the correct image; instead of the "echo Image corrupted!!" there could be a command (sequence) to (load and) boot an alternative image, etc. Example: => addr1=0 => addr2=10 => bootm $addr1 || bootm $addr2 || tftpboot $loadaddr $loadfile && bootm ## Booting image at 00000000 ... Bad Magic Number ## Booting image at 00000010 ... Bad Magic Number TFTP from server 192.168.3.1; our IP address is 192.168.3.68 Filename '/tftpboot/TRAB/uImage'. Load address: 0xc400000 Loading: ################################################################# ################################################################# ########################### done Bytes transferred = 801673 (c3b89 hex) ## Booting image at 0c400000 ... Image Name: ARM Linux-2.4.18 This will check if the image at (flash?) address "addr1" is ok and boot it; if the image is not ok, the alternative image at address "addr2" will be checked and booted if it is found to be OK. If both images are missing or corrupted, a new image will be loaded over [1233]TFTP and booted. 14.2.17.4. General rules 1. If a command line (or an environment variable executed by a run command) contains several commands separated by semicolons, and one of these commands fails, the remaining commands will still be executed. 2. If you execute several variables with one call to run (i. e. calling run with a list of variables as arguments), any failing command will cause run to terminate, i. e. the remaining variables are not executed. 14.2.18. How can I load and uncompress a compressed image Question: Can I use U-Boot to load and uncompress a compressed image from flash into RAM? And can I choose whether I want to automatically run it at that time, or wait until later? Answer: Yes to both questions. First, you should generate your image as type "standalone" (using "mkimage ... -T standalone ..."). When you use the bootm command for such an image, U-Boot will automatically uncompress the code while it is storing it at that image's load address in RAM (given by the -a option to the mkimage command). As to the second question, by default, unless you say differently, U-Boot will automatically start the image by jumping to its entry point (given by the -e option to mkimage) after loading it. If you want to prevent automatic execution, just set the environment variable "autostart" to "no" ("setenv autostart no") before running bootm. 14.2.19. How can I create an uImage from a ELF file Question: I would like to run a standard distribution kernel on my target, but I can find only ELF kernel images or even RPM files. How can I use these? Answer: If you have just the kernel ELF file, this may be difficult, as you will usually also need a bunch of kernel modules that the distribution of your choice probably bundles with this kernel file. Try to locate and install these first. If you have a kernel RPM, this usually includes both the kernel ELF file and the required modules. Install these in the [1234]ELDK root file system so you can use this for example mounted over NFS. The following example uses a Fedora kernel RPM on a 4xxFP target: $ cd /tmp/ $ wget http://download.fedora.redhat.com/pub/fedora/linux/updates/11/ppc/kernel -2.6.30.9-90.fc11.ppc.rpm After downloading the RPM we install it (manually using "rpm2cpio" and "cpio" in the root of the [1235]ELDK file system, "/opt/eldk/ppc_4xxFP/" : $ cd /opt/eldk/ppc_4xxFP/ $ rpm2cpio /tmp/kernel-2.6.30.9-90.fc11.ppc.rpm | sudo cpio -vidum This installs a lot of kernel modules in "./lib/modules/" and a kernel ELF file in "./boot" : $ ls -l boot total 8792 -rw-r--r-- 1 root root 1226119 Oct 17 17:31 System.map-2.6.30.9-90.fc11.ppc -rw-r--r-- 1 root root 96224 Oct 17 17:31 config-2.6.30.9-90.fc11.ppc -rwxr-xr-x 1 root root 7673768 Oct 17 18:20 vmlinuz-2.6.30.9-90.fc11.ppc Now convert the ELF kernel image into an uImage file: $ ppc_4xxFP-objcopy -O binary boot/vmlinuz-2.6.30.9-90.fc11.ppc /tmp/vmlinux.bi n $ gzip -v9 /tmp/vmlinux.bin /tmp/vmlinux.bin: 58.1% -- replaced with /tmp/vmlinux.bin.gz $ mkimage -A ppc -O linux -T kernel -C gzip \ > -a 0x00000000 -e 0x00000000 \ > -n Linux-2.6.30.9-90.fc11.ppc \ > -d /tmp/vmlinux.bin.gz /tftpboot/uImage-2.6.30.9-90.fc11.ppc Image Name: Linux-2.6.30.9-90.fc11.ppc Created: Sun Nov 1 17:00:37 2009 Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 3187431 Bytes = 3112.73 kB = 3.04 MB Load Address: 0x00000000 Entry Point: 0x00000000 There you go. ALERT! Note: you still need the Device Tree Blob for your specific target board. This usually does not come with any of the standard distributions. Also, you may find that you need a ramdisk image to get some modules loaded that might be needed to mount your root file system. 14.2.20. My standalone program does not work Question: I tried adding some new code to the hellow_world.c demo program. This works well as soon as I only add code to the existing hello_world() function, but as soon as I add some functions of my own, things go all haywire: the code of the hello_world() function does not get executed correctly, and my new function gets called with unexpected arguments. What's wrong? Answer: You probably failed to notice that any code you add to the example program may shift the entry point address. You should check this using the nm program: $ ${CROSS_COMPILE}nm -n examples/hello_world 0000000000040004 T testfunc 0000000000040058 T hello_world 000000000004016c t dummy ... As you can see, the entry point (function hello_world()) is no longer at 0x40004 as it was before and as it's documented. Instead, it is now at 0x40058. So you have to start your standalone program at this address, and everything should work well. 14.2.21. Linux hangs after uncompressing the kernel Question: I am using U-Boot with a Linux kernel version Y (Y < 2.4.5-pre5), but the last message I see is Uncompressing Kernel Image ... OK Then the system hangs. Answer: Most probably you pass bad parameters to the Linux kernel. There are several possible reasons: + Bad device tree; for example, check that the memory map set up by the boot loader (like mapping of IMMR, PCI addresses etc.) is consistent with what is encoded in your device tree specification. Here some possible reasons for older Linux kernel versions: Linux: arch/ppc: + arch/ppc: Bad definition of the bd_info structure You must make sure that your machine specific header file (for instance include/asm-ppc/tqm8xx.h) includes the same definition of the Board Information structure as we define in include/ppcboot.h, and make sure that your definition of IMAP_ADDR uses the same value as your U-Boot configuration in CFG_IMMR. + Bad clock information Before kernel version 2.4.5-pre5 (BitKeeper Patch 1.1.1.6, 22MAY2001) the kernel expected the clock information in MHz, but recent kernels expect it in Hz instead. U-Boot passes the clock information in Hz by default. To switch to the old behaviour, you can set the environment variable "clocks_in_mhz" in U-Boot: => setenv clocks_in_mhz 1 => saveenv For recent kernel the "clocks_in_mhz" variable must not be set. If it is present in your environment, you can delete it as follows: => setenv clocks_in_mhz => saveenv ALERT! A common error is to try "setenv clocks_in_mhz 0" or to some other value - this will not work, as the value of the variable is not important at all. It is the existence of the variable that will be checked. * + Inconsistent memory map Some boards may need corrct mappings for some special hardware devices like BCSR (Board Control and Status Registers) etc. Verify that the mappings expected by Linux match those created by U-Boot. 14.2.22. How can I implement automatic software updates? Question: How I can use U-Boot to make it easy for my end users to upgrade the firmware of the device? Answer: This depends mainly on how you intend to distribute your software updates, and which physical interfaces are present (or usable for this purpose) on your device. Typically you will either distribute the software over the network, or you can use soMe type of storage device like USB mass storage device (memory stick etc.), SD cards etc. U-Boot already supports such a feature on several boards. You will probably have to adapt the code and/or the actual behaviour to your specific hardware and/or requirements. Please see here for a starting point: Network: When auto-update support over [1236]TFTP is enabled, U-Boot will test in the initialization sequence if a specific image file is present on the [1237]TFTP server. If this is the case, the image will be downloaded and, if it is considered ok, installed into flash. For details, please read [1238]doc/README.update and/or see commits [1239]4bae9090 and [1240]e83cc063. USB: Several boards implement this feature, all in a slightly different way; see [1241]board/trab/auto_update.c, [1242]board/mcc200/auto_update.c and [1243]board/esd/common/auto_update.c. With this feature enabled, U-Boot will check during the init sequence if a USB mass storage device is plugged in, if this contains a readable file system, and check if this contains one or more known image files. Additionally it is possible to check if the image versions on the USB device are more recent than those already stored in flash. If all programmed criteria match, and if the images can be read without error, the content of the on-board storage (flash, NAND, etc.) gets automatically updated. Adaption for other storage devices (like SD card etc.) should be trivial to implement. 14.3. Linux 14.3.1. Linux crashes randomly Question: On my board, Linux crashes randomly or has random exceptions (especially floating point exceptions if it is a [1244]Power Architecture® processor). Why? Answer: Quite likely your [1245]SDRAM initialization is bad. See [1246]UBootCrashAfterRelocation for more information. On a [1247]Power Architecture®, the instructions beginning with 0xFF are floating point instructions. When your memory subsystem fails, the [1248]Power Architecture® is reading bad values (0xFF) and thus executing illegal floating point instructions. 14.3.2. Linux crashes when uncompressing the kernel Question: When I try to boot Linux, it crashes during uncompressing the kernel image: => bootm 100000 ## Booting image at 00100000 ... Image Name: Linux-2.4.25 Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 1003065 Bytes = 979.6 kB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK Uncompressing Kernel Image ... Error: inflate() returned -3 GUNZIP ERROR - must RESET board to recover Answer: Your kernel image is quite big - nearly 1 MB compressed; when it gets uncompressed it will need 2.5 ... 3 MB, starting at address 0x0000. But your compressed image was stored at 1 MB (0x100000), so the uncompressed code will overwrite the (remaining) compressed image. The solution is thus simple: just use a higher address to download the compressed image into RAM. For example, try: => bootm 400000 14.3.3. Linux Post Mortem Analysis You may find yourself in a situation where the Linux kernel crashes or hangs without any output on the console. The first attempt to get more information in such a situation is a Post Mortem dump of the log buffer - often the Linux kernel has already collected useful information in its console I/O buffer which just does not get printed because the kernel does not run until successful initialization of the console port. Proceed as follows: 1. Find out the virtual address of the log buffer; For 2.4 Linux kernels search for "log_buf": 2.4 Linux: bash$ grep log_buf System.map c0182f54 b log_buf Here the virtual address of the buffer is 0xC0182F54 For 2.6 kernels "__log_buf" must be used: bash$ grep __log_buf System.map c02124c4 b __log_buf Here the virtual address of the buffer is 0xC02124C4 2. Convert to physical address: on [1249]Power Architecture® systems, the kernel is usually configured for a virtual address of kernel base (CONFIG_KERNEL_START) of 0xC0000000. Just subtract this value from the address you found. In our case we get: physical address = 0xC0182F54 - 0xC0000000 = 0x00182F54 3. Reset your board - do not power-cycle it! 4. Use your boot loader (you're running U-Boot, right?) to print a memory dump of that memory area: => md 0x00182F54 This whole operation is based on the assumption that your boot loader does not overwrite the RAM contents - U-Boot will take care not to destroy such valuable information. 14.3.4. Linux kernel register usage For the [1250]Power Architecture® architecture, the Linux kernel uses the following registers: R1: stack pointer R2: pointer to task_struct for the current task R3-R4: parameter passing and return values R5-R10: parameter passing R13: small data area pointer R30: GOT pointer R31: frame pointer A function can use r0 and r3 - r12 without saving and restoring them. r13 - r31 have to be preserved so they must be saved and restored when you want to use them. Also, cr2 - cr4 must be preserved, while cr0, cr1, cr5 - cr7, lr, ctr and xer can be used without saving & restoring them. [ Posted Tue, 15 Jul 2003 by Paul Mackerras to [1251]linuxppc-embedded@lists.linuxppc.org ]. See also the (E)ABI specifications for the [1252]Power Architecture® architecture, [1253]Developing PowerPC Embedded Application Binary Interface (EABI) Compliant Programs 14.3.5. Linux Kernel Ignores my bootargs Question: Why doesn't the kernel use the command-line options I set in the "bootargs" environment variable in U-Boot when I boot my target system? Answer: This problem is typical for ARM systems only. The following discussion is ARM-centric: First, check to ensure that you have configured your U-Boot build so that CONFIG_CMDLINE_TAG is enabled. (Other tags like CONFIG_SETUP_MEMORY_TAGS or CONFIG_INITRD_TAG may be needed, too.) This ensures that u-boot will boot the kernel with a command-line tag that incorporates the kernel options you set in the "bootargs" environment variable. If you have the CONFIG_CMDLINE_TAG option configured, the problem is almost certainly with your kernel build. You have to instruct the kernel to pick up the boot tags at a certain address. This is done in the machine descriptor macros, which are found in the processor start-up C code for your architecture. For the Intel DBPXA250 "Lubbock" development board, the machine descriptor macros are located at the bottom of the file arch/arm/mach-pxa/lubbock.c, and they look like this: MACHINE_START(LUBBOCK, "Intel DBPXA250 Development Platform") MAINTAINER("MontaVista Software Inc.") BOOT_MEM(0xa0000000, 0x40000000, io_p2v(0x40000000)) FIXUP(fixup_lubbock) MAPIO(lubbock_map_io) INITIRQ(lubbock_init_irq) MACHINE_END The machine descriptor macros for your machine will be located in a similar file in your kernel source tree. Having located your machine descriptor macros, the next step is to find out where U-Boot puts the kernel boot tags in memory for your architecture. On the Lubbock, this address turns out to be the start of physical RAM plus 0x100, or 0xa0000100. Add the "BOOT_PARAMS" macro with this address to your machine descriptor macros; the result should look something like this: MACHINE_START(LUBBOCK, "Intel DBPXA250 Development Platform") MAINTAINER("MontaVista Software Inc.") BOOT_PARAMS(0xa0000100) BOOT_MEM(0xa0000000, 0x40000000, io_p2v(0x40000000)) FIXUP(fixup_lubbock) MAPIO(lubbock_map_io) INITIRQ(lubbock_init_irq) MACHINE_END If there is already a BOOT_PARAMS macro in your machine descriptor macros, modify it so that it has the correct address. Then, rebuild your kernel and re-install it on your target. Now the kernel should be able to pick up the kernel options you have set in the "bootargs" environment variable. 14.3.6. Cannot configure Root Filesystem over NFS Question: I want to configure my system with root filesystem over NFS, but I cannot find any such configuration option. Answer: What you are looking for is the CONFIG_ROOT_NFS configuration option, which depends on CONFIG_IP_PNP. To enable root filesystem over NFS you must enable the "IP: kernel level autoconfiguration" option in the "Networking options" menu first. 14.3.7. Linux Kernel Panics because "init" process dies Question: I once had a running system but suddenly, without any changes, the Linux kernel started crashing because the "init" process was dying each time I tried to boot the system, for example like that: ... VFS: Mounted root (nfs filesystem). Freeing unused kernel memory: 140k init init has generated signal 11 but has no handler for it Kernel panic - not syncing: Attempted to kill init! Answer: You probably run your system with the root file system mounted over NFS. Change into the root directory of your target file system, and remove the file "etc/ld.so.cache". That should fix this problem: # cd /opt/eldk/ppc_6xx/ # rm -f etc/ld.so.cache Explanation: Normally, the file "etc/ld.so.cache" contains a compiled list of system libraries. This file is used by the dynamic linker/loader ld.so to cache library information. If it does not exist, rebuilt automatically. For some reason, a corrupted or partial file was written to your root file system. This corrupt file then confused the dynamic linker so that it crashed when trying to start the init process. Question: I cannot boot into my freshly installed [1254]ELDK Root-NFS because init dies with an unhandled signal like this: Freeing unused kernel memory: 124k init PHY: e0103320:00 - Link is Up - 100/Full init has generated signal 4 but has no handler for it Kernel panic - not syncing: Attempted to kill init! Rebooting in 1 seconds.. Answer: Your [1255]CPU does not have a floating point unit, your kernel has no math emulation (CONFIG_MATH_EMULATION) enabled but you still try to boot into a rootfilesystem intended for FPU systems. This is to be expected for example if you try to use a ppc_6xx rootfilesystem on an 8xx system. 14.3.8. Unable to open an initial console Question: The Linux kernel boots, but then hangs after printing: "Warning: unable to open an initial console". Answer: Most probably you have one or missing entries in the /dev directory in your root filesystem. If you are using the [1256]ELDK's root filesystem over NFS, you probably forgot to run the ELDK_MAKEDEV and ELDK_FIXOWNER scripts as described in [1257]3.7. Mounting Target Components via NFS. 14.3.9. System hangs when entering User Space (ARM) Question: The Linux kernel boots, but then the system hangs after printing: "Freeing init memory: 120K". I'm using the root file system from [1258]ELDK 4.2 (or later) on an ARM system. Answer: [1259]ELDK 4.2 (and later) for ARM provide an [1260]EABI compliant User Space environment. You must enable [1261]EABI support in your Linux kernel configuration, or the system will hang as described. Your kernel's ".config" file should contain "CONFIG_AEABI=y" for this. 14.3.10. Mounting a Filesystem over NFS hangs forever Question: We use the SELF ramdisk image that comes with the [1262]ELDK. When we try to mount a filesystem over NFS from the server, for example: # mount -t nfs 192.168.1.1:/target/home /home the command waits nearly 5 minutes in uninterruptable sleep. Then the mount finally succeeds. What's wrong? Answer: The default configuration of the [1263]SELF was not designed to mount additional filesystems with file locking over NFS, so no portmap deamon is running, which is causing your problems. There are two solutions for the problem: 1. Add the portmap deamon (/sbin/portmap) to the target filesystem and start it as part of the init scripts. 2. Tell the "mount" program and the kernel that you don't need file locking by passing the "nolock" option to the mount call, i. e. use # mount -o nolock -t nfs 192.168.1.1:/target/home /home Explanation: If you call the mount command like above (i. e. without the "nolock" option) an RPC call to the "portmap" deamon will be attempted which is required to start a lockd kernel thread which is necessary if you want to use file locking on the NFS filesystem. This call will fail only after a very long timeout. 14.3.11. Ethernet does not work in Linux Question: Ethernet does not work on my board. But everything is fine when I use the ethernet interface in U-Boot (for example by performing a [1264]TFTP download). This is a bug in U-Boot, right? Answer: No. It's a bug in the Linux ethernet driver. In some cases the Linux driver fails to set the [1265]MAC address. That's a buggy driver then - Linux ethernet drivers are supposed to read the [1266]MAC address at startup. On ->open, they are supposed to reprogram the [1267]MAC address back into the chip (but not the EEPROM, if any) whether or not the address has been changed. In general, a Linux driver shall not make any assumptions about any initialization being done (or not done) by a boot loader; instead, that driver is responsible for performing all of the necessary initialization itself. And U-Boot shall not touch any hardware it does not access itself. If you don't use the ethernet interface in U-Boot, it won't be initialized by U-Boot. A pretty extensive discussion of this issue can be found in the thread ATAG for [1268]MAC address on the ARM Linux mailing list. [1269]archive 1 [1270]archive 2 Some current methods for handling the [1271]MAC address programming: * use custom ATAGs (ARM systems) * use a [1272]Flattened Device Tree (if your arch/port supports it) * parse the U-Boot environment directly * pass it via the command line If your device driver does not support one of these sources directly, then do it yourself: * add an init board hook * program it from user space (`ifconfig hw ...`) * for people who need to do NFS root or similar, then use initramfs -- this is what it was designed for ! 14.3.12. Loopback interface does not work Question: When I boot Linux I get a "socket: Address family not supported by protocol" error message when I try to configure the loopback interface. What's wrong? Answer: This is most probably a problem with your kernel configuration. Make sure that the CONFIG_PACKET option is selected. 14.3.13. Linux kernel messages are not printed on the console Question: I expect to see some Linux kernel messages on the console, but there aren't any. Answer: This is absolutely normal when using the [1273]ELDK with root filesystem over NFS. The [1274]ELDK startup routines will start the syslog daemon, which will collect all kernel messages and write them into a logfile ( /var/log/messages ). If you want to see the messages at the console, either run "tail -f /var/log/messages &" on the console window, or stop the syslog daemon by issuing a "/etc/rc.d/init.d/syslog stop" command. Another alternative is to increase the console_loglevel of the kernel (any message with log level less than console_loglevel will be printed to the console). With the following command the console_loglevel could be set at runtime: "echo 8 > /proc/sys/kernel/printk". Now all messages are displayed on the console. 14.3.14. Linux ignores input when using the framebuffer driver Question: When using the framebuffer driver the console output goes to the LCD display, but I cannot input anything. What's wrong? Answer: You can define "console devices" using the console= boot argument. Add something like this to your bootargs setting: ... console=tty0 console=ttyS0,${baudrate} ... This will