| Message ID | 20210508035533.23222-1-dinghui@sangfor.com.cn (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] mm/page_alloc: fix counting of free pages after take off from buddy | expand |
On Sat, May 08, 2021 at 11:55:33AM +0800, Ding Hui wrote: > Recently we found there is a lot MemFree left in /proc/meminfo after > do a lot of pages soft offline. > > I think it's incorrect since NR_FREE_PAGES should not contain HWPoison pages. > For offline free pages, after a successful call take_page_off_buddy(), the > page is no longer belong to buddy allocator, and will not be used any more, > but we missed accounting NR_FREE_PAGES in this situation. > > Do update like rmqueue() does. > > Signed-off-by: Ding Hui <dinghui@sangfor.com.cn> > --- > V2: > use __mod_zone_freepage_state instead of __mod_zone_page_state > > mm/page_alloc.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c > index cfc72873961d..e124a615303b 100644 > --- a/mm/page_alloc.c > +++ b/mm/page_alloc.c > @@ -8947,6 +8947,7 @@ bool take_page_off_buddy(struct page *page) > del_page_from_free_list(page_head, zone, page_order); > break_down_buddy_pages(zone, page_head, page, 0, > page_order, migratetype); > + __mod_zone_freepage_state(zone, -1, migratetype); Page offline code (see set_migratetype_isolate()) seems to handle NR_FREE_PAGES counter in its own way, so I think that it's more correct to call __mod_zone_freepage_state() only when is_migrate_isolate(migratetype)) is false. Otherwise, the patch looks good to me. Thanks, Naoya Horiguchi > ret = true; > break; > } > -- > 2.17.1 >
On 2021/5/25 16:32, HORIGUCHI NAOYA(堀口 直也) wrote: > On Sat, May 08, 2021 at 11:55:33AM +0800, Ding Hui wrote: >> Recently we found there is a lot MemFree left in /proc/meminfo after >> do a lot of pages soft offline. >> >> I think it's incorrect since NR_FREE_PAGES should not contain HWPoison pages. >> For offline free pages, after a successful call take_page_off_buddy(), the >> page is no longer belong to buddy allocator, and will not be used any more, >> but we missed accounting NR_FREE_PAGES in this situation. >> >> Do update like rmqueue() does. >> >> Signed-off-by: Ding Hui <dinghui@sangfor.com.cn> >> --- >> V2: >> use __mod_zone_freepage_state instead of __mod_zone_page_state >> >> mm/page_alloc.c | 1 + >> 1 file changed, 1 insertion(+) >> >> diff --git a/mm/page_alloc.c b/mm/page_alloc.c >> index cfc72873961d..e124a615303b 100644 >> --- a/mm/page_alloc.c >> +++ b/mm/page_alloc.c >> @@ -8947,6 +8947,7 @@ bool take_page_off_buddy(struct page *page) >> del_page_from_free_list(page_head, zone, page_order); >> break_down_buddy_pages(zone, page_head, page, 0, >> page_order, migratetype); >> + __mod_zone_freepage_state(zone, -1, migratetype); > > Page offline code (see set_migratetype_isolate()) seems to handle > NR_FREE_PAGES counter in its own way, so I think that it's more correct to > call __mod_zone_freepage_state() only when is_migrate_isolate(migratetype)) > is false. > > Otherwise, the patch looks good to me. > Thanks for reply and suggestion, I'll send v3 patch later. >> ret = true; >> break; >> } >> -- >> 2.17.1
diff --git a/mm/page_alloc.c b/mm/page_alloc.c index cfc72873961d..e124a615303b 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -8947,6 +8947,7 @@ bool take_page_off_buddy(struct page *page) del_page_from_free_list(page_head, zone, page_order); break_down_buddy_pages(zone, page_head, page, 0, page_order, migratetype); + __mod_zone_freepage_state(zone, -1, migratetype); ret = true; break; }
Recently we found there is a lot MemFree left in /proc/meminfo after do a lot of pages soft offline. I think it's incorrect since NR_FREE_PAGES should not contain HWPoison pages. For offline free pages, after a successful call take_page_off_buddy(), the page is no longer belong to buddy allocator, and will not be used any more, but we missed accounting NR_FREE_PAGES in this situation. Do update like rmqueue() does. Signed-off-by: Ding Hui <dinghui@sangfor.com.cn> --- V2: use __mod_zone_freepage_state instead of __mod_zone_page_state mm/page_alloc.c | 1 + 1 file changed, 1 insertion(+)