On Fri, Oct 14, 2022 at 09:42:20AM +1100, Dave Chinner wrote:
> On Sun, Oct 02, 2022 at 11:19:52AM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <djwong@xxxxxxxxxx>
> >
> > Convert all the online scrub code to use the Linux slab allocator
> > functions directly instead of going through the kmem wrappers.
> >
> > Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx>
>
> .....
>
> > ---
> > diff --git a/fs/xfs/scrub/btree.c b/fs/xfs/scrub/btree.c
> > index 2f4519590dc1..566a093b2cf3 100644
> > --- a/fs/xfs/scrub/btree.c
> > +++ b/fs/xfs/scrub/btree.c
> > @@ -431,10 +431,10 @@ xchk_btree_check_owner(
> > * later scanning.
> > */
> > if (cur->bc_btnum == XFS_BTNUM_BNO || cur->bc_btnum == XFS_BTNUM_RMAP) {
> > - co = kmem_alloc(sizeof(struct check_owner),
> > - KM_MAYFAIL);
> > + co = kmalloc(sizeof(struct check_owner), XCHK_GFP_FLAGS);
> > if (!co)
> > return -ENOMEM;
> > + INIT_LIST_HEAD(&co->list);
>
> Fixes some other bug?
>
> It's obvious that it should be initialised, so I'm not concerned
> about it being here, just checking that you intended to fix this
> here and it doesn't belong to some other patchset?
Yeah, it'll fix list debugging tripping over the unininitialized
list_head. I probably should've made that a separate oneliner but...
eh. There are already too many patches.
--D
> Otherwise:
>
> Reviewed-by: Dave Chinner <dchinner@xxxxxxxxxx>
>
> --
> Dave Chinner
> david@xxxxxxxxxxxxx
