On Wed, Aug 21, 2024 at 10:33:55AM GMT, Lizhi Xu wrote:
> On Tue, 20 Aug 2024 19:34:18 -0400, Kent Overstreet wrote:
> > > When journal v2 entry nr overflow, it will cause the value of ja->nr to
> > > be incorrect, this will result in the allocated memory to ja->buckets
> > > being too small, leading to out of bounds access in bch2_dev_journal_init.
> > >
> > > Reported-by: syzbot+47ecc948aadfb2ab3...@syzkaller.appspotmail.com
> > > Signed-off-by: Lizhi Xu <lizhi...@windriver.com>
> > > ---
> > > fs/bcachefs/journal_sb.c | 5 +++++
> > > 1 file changed, 5 insertions(+)
> > >
> > > diff --git a/fs/bcachefs/journal_sb.c b/fs/bcachefs/journal_sb.c
> > > index db80e506e3ab..db2b2100e4e5 100644
> > > --- a/fs/bcachefs/journal_sb.c
> > > +++ b/fs/bcachefs/journal_sb.c
> > > @@ -119,6 +119,11 @@ static int bch2_sb_journal_v2_validate(struct bch_sb
> > > *sb, struct bch_sb_field *f
> > > for (i = 0; i < nr; i++) {
> > > b[i].start = le64_to_cpu(journal->d[i].start);
> > > b[i].end = b[i].start + le64_to_cpu(journal->d[i].nr);
> > > + if (le64_to_cpu(journal->d[i].nr) > UINT_MAX) {
> > > + prt_printf(err, "journal v2 entry d[%u].nr %llu
> > > overflow\n",
> > > + i, le64_to_cpu(journal->d[i].nr));
> > > + goto err;
> > > + }
> >
> > no, you need to sum up _all_ the entries and verify the total doesn't
> > overflow UINT_MAX
> The overflow value of le64_to_cpu(journal->d[i].nr) is
> 18446744073709551615(for u64),
> or in other words, it is -1 for s64.
>
> Therefore, the existing check for single entry is retained, and an overflow
> check for the total value of all entry is will added.
No, this is completely broken.
