From e3bba3c3c90cd434c1ccb9e5dc704a96baf9541c Mon Sep 17 00:00:00 2001 From: Naoya Horiguchi Date: Thu, 23 Jan 2014 15:52:53 -0800 Subject: fs/proc/page.c: add PageAnon check to surely detect thp stable_page_flags() checks !PageHuge && PageTransCompound && PageLRU to know that a specified page is thp or not. But sometimes it's not enough and we fail to detect thp when the thp is on pagevec. This happens only for a few seconds after LRU list operations, but it makes it difficult to control our applications depending on this flag. So this patch adds another check PageAnon to detect thps on pagevec. It might not give the future extensibility for thp pagecache, but it's OK at least for now. Signed-off-by: Naoya Horiguchi Cc: David Rientjes Cc: KOSAKI Motohiro Cc: Wu Fengguang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/page.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs/proc/page.c') diff --git a/fs/proc/page.c b/fs/proc/page.c index b8730d9ebaee..cab84b6272ed 100644 --- a/fs/proc/page.c +++ b/fs/proc/page.c @@ -118,10 +118,12 @@ u64 stable_page_flags(struct page *page) /* * PageTransCompound can be true for non-huge compound pages (slab * pages or pages allocated by drivers with __GFP_COMP) because it - * just checks PG_head/PG_tail, so we need to check PageLRU to make - * sure a given page is a thp, not a non-huge compound page. + * just checks PG_head/PG_tail, so we need to check PageLRU/PageAnon + * to make sure a given page is a thp, not a non-huge compound page. */ - else if (PageTransCompound(page) && PageLRU(compound_trans_head(page))) + else if (PageTransCompound(page) && + (PageLRU(compound_trans_head(page)) || + PageAnon(compound_trans_head(page)))) u |= 1 << KPF_THP; /* -- cgit v1.2.3 From abaf3787ac26ba33e2f75e76b1174c32254c25b0 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Thu, 23 Jan 2014 15:55:45 -0800 Subject: fs/proc: don't use module_init for non-modular core code PROC_FS is a bool, so this code is either present or absent. It will never be modular, so using module_init as an alias for __initcall is rather misleading. Fix this up now, so that we can relocate module_init from init.h into module.h in the future. If we don't do this, we'd have to add module.h to obviously non-modular code, and that would be ugly at best. Note that direct use of __initcall is discouraged, vs. one of the priority categorized subgroups. As __initcall gets mapped onto device_initcall, our use of fs_initcall (which makes sense for fs code) will thus change these registrations from level 6-device to level 5-fs (i.e. slightly earlier). However no observable impact of that small difference has been observed during testing, or is expected. Also note that this change uncovers a missing semicolon bug in the registration of vmcore_init as an initcall. Signed-off-by: Paul Gortmaker Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/page.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/proc/page.c') diff --git a/fs/proc/page.c b/fs/proc/page.c index cab84b6272ed..02174a610315 100644 --- a/fs/proc/page.c +++ b/fs/proc/page.c @@ -219,4 +219,4 @@ static int __init proc_page_init(void) proc_create("kpageflags", S_IRUSR, NULL, &proc_kpageflags_operations); return 0; } -module_init(proc_page_init); +fs_initcall(proc_page_init); -- cgit v1.2.3