On Mon, Jul 15, 2024 at 4:10 AM Xiubo Li <xiubli@xxxxxxxxxx> wrote:
>
>
> On 7/11/24 14:47, ethanwu wrote:
> > The kmalloc size of pagevec mempool is incorrectly calculated.
> > It misses the size of page pointer and only accounts the number for the array.
> >
> > Fixes: a0102bda5bc0 ("ceph: move sb->wb_pagevec_pool to be a global mempool")
> > Signed-off-by: ethanwu <ethanwu@xxxxxxxxxxxx>
> > ---
> > fs/ceph/super.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/fs/ceph/super.c b/fs/ceph/super.c
> > index 885cb5d4e771..46f640514561 100644
> > --- a/fs/ceph/super.c
> > +++ b/fs/ceph/super.c
> > @@ -961,7 +961,8 @@ static int __init init_caches(void)
> > if (!ceph_mds_request_cachep)
> > goto bad_mds_req;
> >
> > - ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10, CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT);
> > + ceph_wb_pagevec_pool = mempool_create_kmalloc_pool(10,
> > + (CEPH_MAX_WRITE_SIZE >> PAGE_SHIFT) * sizeof(struct page *));
> > if (!ceph_wb_pagevec_pool)
> > goto bad_pagevec_pool;
> >
>
> Good cache, LGTM.
I don't think this matters in practice since 64M / 4K = 16K * 10 = 160K
of memory reserved just for page pointers is already quite a lot...
However, since commit a0102bda5bc0 is talking about moving a pool to be
global without adjusting its parameters, I agree that this fix is valid.
Thanks,
Ilya
