>> 3) kernel size is uninfluent: I tested 930KiB-lzma kernels and 1240KiB 
>> gzip-kernels.

>It sounds to me like this is because of the 1.4M (cannot
>remember the exact size) constraint of the kernel size
>on NAND that the bootloader is able to load?

Unfortunately not :(
Kernel did not exceed the max size: I'm aware of that Zaurus
limitation: 1294336B = 1264KiB  (and only 1048576 for collie)
The only patch really necessary for Zaurus to boot from nand is [2].
Any hope to have it upstream?

Back to compressors, I say it must be the commit e7db7b427 because
kernel 2.6.33-rc4 was booting just fine.

Beside this, a month ago I booted succesfully 2.6.32 with default gzip
compression but when I however tried the previous version of the
LZO/LZMA patches [1] the same kernel stopped to boot, even the gzip
compressed zImage. I ended up disabling the patches...

So there is something wrong in that patches...what surprises me is the
very same kernels boot when launched by kexec.

Regards

Andrea


[1] c/o Albin Tonnerre
v3-1-4-Add-support-for-LZO-compressed-kernels.patch
v3-2-4-Add-support-for-LZO-compressed-kernels-for-ARM.patch
v3-3-4-Add-support-for-LZO-compressed-kernels-on-x86.patch
v3-4-4-Add-LZO-compression-support-for-initramfs-and-old-style-initrd.patch
ARM-Add-support-for-LZMA-compressed-kernel-images.patch


[2] c/o Richard Purdie
http://www.rpsys.net/openzaurus/patches/archive/pxa-linking-bug-r1.patch

_______________________________________________
Zaurus-devel mailing list
Zaurus-devel@lists.linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/zaurus-devel

Reply via email to