On Thu, 4 Mar 2004, Sonia Martinez wrote:
The kernel and file system image (u-boot multi image) is loaded into RAM with kermit at 0xa3100000. But, when we try to boot linux with the bootm command, it stops at "Starting Kernel" (We're working with Linux kernel 2.4.19 patched with -rmk7-pxa1 in eLinOS distribution).
In order to see if something went wrong, we've compared vmlinux.bin before compression and the kernel decompressed in RAM, in position Load_address. We've found that the information is the same but one is in big endian (in RAM) and the other one in little endian (vmlinux.bin). The platform works in little endian, so this is probably the problem. How can we solve it?? Is it a gunzip problem??
It seems very unlikely to me that your u-boot has decompressed your kernel into a wrong endianess. Have double checked that you're really looking at what you're think you're looking at?
It is much more likely that your kernel might not be properly supporting your board in the early boot phase and just hangs there. Depending how far it got, you may try to check the contents of the kernel's log_buf array after resetting the hardware. All the early printk()s go there before the serial driver is started.
Regards, Marius