On Fri, Mar 07, 2008 at 01:58:22PM +0100, Aurelien Jarno wrote: > Martin Michlmayr a écrit : > > * Joey Hess <joeyh@debian.org> [2008-03-07 00:32]: > >> Looks like a somehow corrupt cramfs image. > >> > >> The mipsel images do not have this problem. I see tbm reported a > >> similar problem on debian-mips[1], and it seems to affect real hardware. > > > > Looks like the same issue indeed. It's good to know that this also > > happens in qemu since I wasn't sure whether it was specific to SGI. > > Anyway, I mostly care about mipsel these days (which works), so maybe > > someone else can take a look. > > > > And someone should put in the errata that mips d-i beta1 is broken. > > At a first glance, it looks like a problem during the initrd build > process. The image is shorter on mips than on mipsel, while I see now > reason for that. Also I have recently built d-i on mips a lot of time to > work on the mklibs bug, and I don't about remember seeing this problem. > > I will try to build the image myself to see if it changes something. The image I have generated works correctly here. You can find it on: http://temp.aurel32.net/d-i-lenny-qemu/ The md5sums are: 12082c31bae04e909ce2f8a082a96b35 initrd.gz 0b0c8a0e87338bf48bae59438a46bd54 vmlinux-2.6.22-3-qemu It has about the same size as the mipsel one, which confirms the problem is in the build process. Thiemo, could you please have a look at this? -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' aurel32@debian.org | aurelien@aurel32.net `- people.debian.org/~aurel32 | www.aurel32.net
Attachment:
signature.asc
Description: Digital signature