Thread: [SSI-devel] Re: initramfs
Brought to you by:
brucewalker,
rogertsang
From: Brian J. W. <Bri...@hp...> - 2005-04-29 00:06:39
|
En Chiang Lee wrote: > Hi Brian, > > I've been adding the SSI changes to the FC3 mkinitrd, and it's at a > stage where the initrd created looks right. The initrd created is > actually a cpio archive (initramfs) for 2.6 kernels. > > The problem now is that when I try to boot with this initrd, if fails with > > Kernel panic - not syncing: No init found > > I'm not sure how the initramfs stuff works, so I've attached the > mkinitrd to see if there's anything that's seriously wrong. > > Does the kernel have to be built with some option for cpio archives to > work? > > En Chiang Hi En Chiang, I'm not exactly sure what's wrong. It's possible that some of the OpenSSI changes to init/main.c are interfering with how initramfs is supposed to work. Can we delay on this issue by forcing mkinitrd to always build an old-style initrd if $ssi_boot is set? If so, please add an SSI_XXX comment next to that change, to explain why initramfs isn't being used. BTW, another test that should be run is whether mkinitrd without the --cfs option can successfully build an initramfs that is virtually identical to what the base mkinitrd would produce. Thanks, Brian |