| Message ID | 20220624025309.1033400-4-ying.huang@intel.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [1/7] migrate: fix syscall move_pages() return value for failure | expand |
On 6/24/2022 10:53 AM, Huang Ying wrote: > In unmap_and_move(), if the new THP cannot be allocated, -ENOMEM will > be returned, and migrate_pages() will try to split the THP unless > "reason" is MR_NUMA_MISPLACED (that is, nosplit == true). But when > nosplit == true, the THP migration failure will not be counted. This > is incorrect. So in this patch, the THP migration failure will be > counted for -ENOMEM regardless of nosplit is true or false. The > nr_failed counting is fixed too, although that is not used actually. No strong opinion about the 'nr_failed' updating, like you said, we did not use it in this case. Maybe just add some comments instead of adding some unused code? Otherwise looks good to me. Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com> > > Signed-off-by: "Huang, Ying" <ying.huang@intel.com> > Fixes: 5984fabb6e82 ("mm: move_pages: report the number of non-attempted pages") > Cc: Baolin Wang <baolin.wang@linux.alibaba.com> > Cc: Zi Yan <ziy@nvidia.com> > Cc: Yang Shi <shy828301@gmail.com> > --- > mm/migrate.c | 8 +++++--- > 1 file changed, 5 insertions(+), 3 deletions(-) > > diff --git a/mm/migrate.c b/mm/migrate.c > index 82444e7df9f1..542533e4e3cf 100644 > --- a/mm/migrate.c > +++ b/mm/migrate.c > @@ -1425,11 +1425,11 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page, > /* > * When memory is low, don't bother to try to migrate > * other pages, just exit. > - * THP NUMA faulting doesn't split THP to retry. > */ > - if (is_thp && !nosplit) { > + if (is_thp) { > nr_thp_failed++; > - if (!try_split_thp(page, &thp_split_pages)) { > + /* THP NUMA faulting doesn't split THP to retry. */ > + if (!nosplit && !try_split_thp(page, &thp_split_pages)) { > nr_thp_split++; > goto retry; > } > @@ -1446,6 +1446,8 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page, > */ > list_splice_init(&thp_split_pages, from); > nr_thp_failed += thp_retry; > + if (!no_subpage_counting) > + nr_failed += retry; > goto out; > case -EAGAIN: > if (is_thp)
diff --git a/mm/migrate.c b/mm/migrate.c index 82444e7df9f1..542533e4e3cf 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1425,11 +1425,11 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page, /* * When memory is low, don't bother to try to migrate * other pages, just exit. - * THP NUMA faulting doesn't split THP to retry. */ - if (is_thp && !nosplit) { + if (is_thp) { nr_thp_failed++; - if (!try_split_thp(page, &thp_split_pages)) { + /* THP NUMA faulting doesn't split THP to retry. */ + if (!nosplit && !try_split_thp(page, &thp_split_pages)) { nr_thp_split++; goto retry; } @@ -1446,6 +1446,8 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page, */ list_splice_init(&thp_split_pages, from); nr_thp_failed += thp_retry; + if (!no_subpage_counting) + nr_failed += retry; goto out; case -EAGAIN: if (is_thp)
In unmap_and_move(), if the new THP cannot be allocated, -ENOMEM will be returned, and migrate_pages() will try to split the THP unless "reason" is MR_NUMA_MISPLACED (that is, nosplit == true). But when nosplit == true, the THP migration failure will not be counted. This is incorrect. So in this patch, the THP migration failure will be counted for -ENOMEM regardless of nosplit is true or false. The nr_failed counting is fixed too, although that is not used actually. Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Fixes: 5984fabb6e82 ("mm: move_pages: report the number of non-attempted pages") Cc: Baolin Wang <baolin.wang@linux.alibaba.com> Cc: Zi Yan <ziy@nvidia.com> Cc: Yang Shi <shy828301@gmail.com> --- mm/migrate.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)