The reason that folio_referenced() returns negative is because
rwc.contended is 1, vm_flags is 0 at this time and so there is
no need to detect VM_EXEC. And only when referenced > 0, pra->vm_flags
will be filled with vma->vm_flags in folio_referenced_one().
Signed-off-by: Zhiguo Jiang <[email protected]>
---
mm/vmscan.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index e587dafeef94..f0b4325cccd4 100755
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2095,7 +2095,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
/* Referenced or rmap lock contention: rotate */
if (folio_referenced(folio, 0, sc->target_mem_cgroup,
- &vm_flags) != 0) {
+ &vm_flags) > 0) {
/*
* Identify referenced, file-backed active folios and
* give them one more trip around the active list. So
--
2.39.0
On Wed, 25 Oct 2023 23:44:32 +0800 Zhiguo Jiang <[email protected]> wrote:
> The reason that folio_referenced() returns negative is because
> rwc.contended is 1, vm_flags is 0 at this time and so there is
> no need to detect VM_EXEC. And only when referenced > 0, pra->vm_flags
> will be filled with vma->vm_flags in folio_referenced_one().
>
> ...
>
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2095,7 +2095,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
>
> /* Referenced or rmap lock contention: rotate */
> if (folio_referenced(folio, 0, sc->target_mem_cgroup,
> - &vm_flags) != 0) {
> + &vm_flags) > 0) {
> /*
> * Identify referenced, file-backed active folios and
> * give them one more trip around the active list. So
I suspect this email did not make it through the linux-mm server. I
can't find it in my archive or at https://lore.kernel.org/linux-mm/.
So please resend? Prior to doing so, please update the changelog to
describe the before- and after- userspace visible effects of the
change.
Thanks.