We're supposed to be holding the "vmf->ptl" spin_lock when we goto
out_map. The lock is dropped after if finishes cleaning up.
Fixes: 9aff7b33c74a ("mm: thp: refactor NUMA fault handling")
Signed-off-by: Dan Carpenter <[email protected]>
---
mm/huge_memory.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index e353bbc6cee3..caa0148f15bb 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1465,10 +1465,8 @@ vm_fault_t do_huge_pmd_numa_page(struct vm_fault *vmf)
pmd = pmd_modify(oldpmd, vma->vm_page_prot);
page = vm_normal_page_pmd(vma, haddr, pmd);
- if (!page) {
- spin_unlock(vmf->ptl);
+ if (!page)
goto out_map;
- }
/* See similar comment in do_numa_page for explanation */
if (!was_writable)
--
2.30.2
On Tue, Jun 01, 2021 at 12:24:09PM +0300, Dan Carpenter wrote:
> We're supposed to be holding the "vmf->ptl" spin_lock when we goto
> out_map. The lock is dropped after if finishes cleaning up.
>
> Fixes: 9aff7b33c74a ("mm: thp: refactor NUMA fault handling")
> Signed-off-by: Dan Carpenter <[email protected]>
Ouch.
Acked-by: Mel Gorman <[email protected]>
However, that git commit is not stable. Instead of Fixes: I would
suggest renaming the patch to "mm: thp: refactor NUMA fault handling
-fix" and replacing Fixes with "This patch is a fix to the mmotm patch
mm-thp-refactor-numa-fault-handling.patch". Andrew usually slots that
into the correct place in his quilt series and collapses the fixes before
sending to Linus which works better with bisection.
--
Mel Gorman
SUSE Labs
On Tue, Jun 01, 2021 at 11:08:49AM +0100, Mel Gorman wrote:
> On Tue, Jun 01, 2021 at 12:24:09PM +0300, Dan Carpenter wrote:
> > We're supposed to be holding the "vmf->ptl" spin_lock when we goto
> > out_map. The lock is dropped after if finishes cleaning up.
> >
> > Fixes: 9aff7b33c74a ("mm: thp: refactor NUMA fault handling")
> > Signed-off-by: Dan Carpenter <[email protected]>
>
> Ouch.
>
> Acked-by: Mel Gorman <[email protected]>
>
> However, that git commit is not stable. Instead of Fixes: I would
> suggest renaming the patch to "mm: thp: refactor NUMA fault handling
> -fix" and replacing Fixes with "This patch is a fix to the mmotm patch
> mm-thp-refactor-numa-fault-handling.patch". Andrew usually slots that
> into the correct place in his quilt series and collapses the fixes before
> sending to Linus which works better with bisection.
I know that these normally get folded in, but I assumed that Andrew
would want the Fixes tag so that he could fold them in automatically
using a mutt alias. #OneClickShopping
regards,
dan carpenter
On Tue, Jun 1, 2021 at 2:24 AM Dan Carpenter <[email protected]> wrote:
>
> We're supposed to be holding the "vmf->ptl" spin_lock when we goto
> out_map. The lock is dropped after if finishes cleaning up.
>
> Fixes: 9aff7b33c74a ("mm: thp: refactor NUMA fault handling")
> Signed-off-by: Dan Carpenter <[email protected]>
Thanks for catching this. Acked-by: Yang Shi <[email protected]>
> ---
> mm/huge_memory.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> index e353bbc6cee3..caa0148f15bb 100644
> --- a/mm/huge_memory.c
> +++ b/mm/huge_memory.c
> @@ -1465,10 +1465,8 @@ vm_fault_t do_huge_pmd_numa_page(struct vm_fault *vmf)
>
> pmd = pmd_modify(oldpmd, vma->vm_page_prot);
> page = vm_normal_page_pmd(vma, haddr, pmd);
> - if (!page) {
> - spin_unlock(vmf->ptl);
> + if (!page)
> goto out_map;
> - }
>
> /* See similar comment in do_numa_page for explanation */
> if (!was_writable)
> --
> 2.30.2
>