В Thu, 16 Apr 2015 07:51:50 +0300
Toomas Soome <tso...@me.com> пишет:
>
> > On 16.04.2015, at 7:20, Andrei Borzenkov <arvidj...@gmail.com> wrote:
> >
> > В Wed, 15 Apr 2015 23:51:16 +0300
> > Toomas Soome <tso...@me.com> пишет:
> >
> >>
> >> hi!
> >>
> >> yep, this old bug is not fixed in grub. cpy can (theoretically?) overflow.
> >
> > You mean "length"? Or do you really mean pointer overflow?
> >
> > Anyway in both cases it seems more reasonable to check when length is
> > computed, not after overflow, when it is already too late.
>
> integer overflow. from 0xFFFFFFFF to 0x0.
signed integer overflow is undefined behavior in C. It is too late to
check for it after it happened.
>
> >
> >>
> >> ---
> >> grub-core/fs/zfs/zfs_lz4.c | 2 ++
> >> 1 file changed, 2 insertions(+)
> >>
> >> diff --git a/grub-core/fs/zfs/zfs_lz4.c b/grub-core/fs/zfs/zfs_lz4.c
> >> index 1212a89..ca6445d 100644
> >> --- a/grub-core/fs/zfs/zfs_lz4.c
> >> +++ b/grub-core/fs/zfs/zfs_lz4.c
> >> @@ -185,6 +185,8 @@ LZ4_uncompress_unknownOutputSize(const char *source,
> >> }
> >> /* copy literals */
> >> cpy = op + length;
> >> + if (cpy < op)
> >> + goto _output_error;
> >> if ((cpy > oend - COPYLENGTH) ||
> >> (ip + length > iend - COPYLENGTH)) {
> >> if (cpy > oend)
> >
> >
> > _______________________________________________
> > Grub-devel mailing list
> > Grub-devel@gnu.org
> > https://lists.gnu.org/mailman/listinfo/grub-devel
>
>
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> https://lists.gnu.org/mailman/listinfo/grub-devel
_______________________________________________
Grub-devel mailing list
Grub-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/grub-devel
