Hi Henry. I recently migrated to ARCH=powerpc on 2.6.24 from ARCH=ppc on 2.6.16 (having obtained a distribution from the ltib project).
The ppc architecture is nearing end-of-life as far as development goes, so I recommend jumping to powerpc now. In mey recent experience I encountered better driver support, and I generally noticed that future patching/development would support ARCH=powerpc. I think the old ppc implied embedded (or non-mac) architectures, while the powerpc was intended for mac desktops. I might be wrong, but it looks like the two have merged. This architecture issue doesn't speak directly to your question, but if you could get 2.6.25 working with ARCH=powerpc on your lite5200b then you and I would be working with basically the same set-up and I could help you better. For now, here are a few questions: 1) Have you successfully NFS mounted your rootfs and not woried about the RAMdisk at this point? 2) Maybe you are using ARCH=powerpc, but your comment in /arch/ppc/configs/ has me confused. Did you mean in /arch/powerpc/configs/ ? Please clarify if you are using ARCH=powerpc and I think I can be of more assistance. For example, on my 2.6.24 kernel I needed to... 1) Halve the memory sive specified in lite5200b.dts. It specified twice as much memory as is actually on my lite5200b, causing a strange, long delay at boot time as page initialization attempted to operate on a bunch of non-existent memory, near as I can tell. The symptom was around 30-45 seconds of delay between U-boot boot and any console output. Weird, but I figured it had something to do with RAM initialization only after getting my BDI2000 JTAG debugger going and breaking to see that the kernel appears stuck on something related to initializing the RAM. This drove me to discover the size spec is wrong in the dts file. 2) If you need edge-sensitive interrupts you'll need a set interrupt type callback. Mpc52xx_pic.c is missing this callback. Check another powerpc architecture for set_interrupt_type. I would upload a patch but it turns out I am fine with level-sensitive only so I didn't go through the trouble of formalizing a fix for this. 3) SPI driver for SPI peripheral (not PSC): this is missing but there is a patch for it. Tell me if you need it and I'll direct you). 4) IN GENERAL, THE DEVICE TREE CONCEPT IS COOL, BUT IT HAS A LEARNING CURVE. I can help you. For example, just requesting an external IRQ requires creating a "device" node so that you can associate a specific IRQ number with said device for easy access from your module driver code at run time. Basically, ALL register access is done by run-time parsing of the device tree information, NOT by simply direct register access, given a bunch of defined address constants in some header file. Sorry if none of this helps, but I just recently clawed my way up the learning curve so some of the wounds are fresh. Hopefully I can provide a bit of guidance on the easy concepts. -Mike ________________________________ From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Henry Chen Sent: Tuesday, June 03, 2008 9:55 PM To: linuxppc-embedded@ozlabs.org Subject: Is the linux-2.6.25.4 compatible with freescale lite5200b? Hi, all expert! Thank you for your patience. I want to use Linux as the OS of freescale lite5200b develop board. Then I have tried two version of kernel, 2.6.25.4 and 2.4.25. We succeed in 2.4.25, failed in 2.6.25.4. We use the default config file for the board, lite5200_defconfig, which would be found both in /arch/ppc/configs/. And we use the same ramdisk. The following info is about the error of 2.6.15.4. => bootm fe000000 fe200000 ## Booting image at fe000000 ... Image Name: Linux-2.6.25 Image Type: PowerPC Linux Kernel Image (gzip compressed) Data Size: 802989 Bytes = 784.2 kB Load Address: 00000000 Entry Point: 00000000 Uncompressing Kernel Image ... OK ## Loading RAMDisk Image at fe200000 ... Image Name: rootfs in flash Image Type: PowerPC Linux RAMDisk Image (gzip compressed) Data Size: 5055519 Bytes = 4.8 MB Load Address: fe5a0000 Entry Point: fe5a0000 Loading Ramdisk to 0fa7c000, end 0ff4e41f ... OK id mach(): done MMU:enter MMU:hw init MMU:mapin MMU:setio MMU:exit setup_arch: enter setup_arch: bootmem arch: exit Would you please help us to solve it and explain the cause? Best regards! Sincerely yours, Henry
_______________________________________________ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
