On 21.07.2011 00:28, Samuel Thibault wrote:
> tags 634799 + upstream patch
> thanks
>
> Hello,
>
I've looked into it and arrived to the similar conclusion that the
grub.cfg.new with inode = 0 is the problem. However I think no such
entries should be there in the first place. In any case GRUB must be
robust to such corruptions but I'd recommend to fix Hurd ext2
implementation as well.
> Digging a bit with the issue, it seems grub2 does not ignore entries
> with inode == 0 while it should. See linux' ext2_readdir() in
> fs/ext2/dir.c doing it:
>
> for ( ;(char*)de <= limit; de = ext2_next_entry(de)) {
> if (de->inode) {
> ...
> }
> }
>
> The attached patch fixes it. Shall I commit it to upstream grub2?
>
> Samuel-- Regards Vladimir 'φ-coder/phcoder' Serbinenko
signature.asc
Description: OpenPGP digital signature
