Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754167AbbKXMjo (ORCPT ); Tue, 24 Nov 2015 07:39:44 -0500 Received: from mx2.suse.de ([195.135.220.15]:51664 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753478AbbKXMgq (ORCPT ); Tue, 24 Nov 2015 07:36:46 -0500 From: Vlastimil Babka To: linux-mm@kvack.org Cc: linux-kernel@vger.kernel.org, Andrew Morton , Joonsoo Kim , Minchan Kim , Sasha Levin , "Kirill A. Shutemov" , Mel Gorman , Michal Hocko , Vlastimil Babka Subject: [PATCH v2 7/9] mm, page_owner: dump page owner info from dump_page() Date: Tue, 24 Nov 2015 13:36:19 +0100 Message-Id: <1448368581-6923-8-git-send-email-vbabka@suse.cz> X-Mailer: git-send-email 2.6.3 In-Reply-To: <1448368581-6923-1-git-send-email-vbabka@suse.cz> References: <1448368581-6923-1-git-send-email-vbabka@suse.cz> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5197 Lines: 142 The page_owner mechanism is useful for dealing with memory leaks. By reading /sys/kernel/debug/page_owner one can determine the stack traces leading to allocations of all pages, and find e.g. a buggy driver. This information might be also potentially useful for debugging, such as the VM_BUG_ON_PAGE() calls to dump_page(). So let's print the stored info from dump_page(). Example output: page:ffffea0002868a00 count:1 mapcount:0 mapping:ffff8800bba8e958 index:0x63a22c flags: 0x1fffff80000060(lru|active) page dumped because: VM_BUG_ON_PAGE(1) page->mem_cgroup:ffff880138efdc00 page allocated via order 0, migratetype Movable, gfp_mask 0x2420848(GFP_NOFS|GFP_NOFAIL|GFP_HARDWALL|GFP_MOVABLE) [] __alloc_pages_nodemask+0x15a/0xa30 [] alloc_pages_current+0x88/0x120 [] __page_cache_alloc+0xe6/0x120 [] pagecache_get_page+0x56/0x200 [] __getblk_slow+0xd2/0x2b0 [] __getblk_gfp+0x40/0x50 [] jbd2_journal_get_descriptor_buffer+0x3e/0x90 [jbd2] [] jbd2_journal_commit_transaction+0x8e3/0x1870 [jbd2] page has been migrated, last migrate reason: compaction Signed-off-by: Vlastimil Babka --- include/linux/page_owner.h | 9 +++++++++ mm/debug.c | 2 ++ mm/page_alloc.c | 1 + mm/page_owner.c | 25 +++++++++++++++++++++++++ 4 files changed, 37 insertions(+) diff --git a/include/linux/page_owner.h b/include/linux/page_owner.h index 555893b..46f1b93 100644 --- a/include/linux/page_owner.h +++ b/include/linux/page_owner.h @@ -13,6 +13,7 @@ extern void __set_page_owner(struct page *page, extern gfp_t __get_page_owner_gfp(struct page *page); extern void __copy_page_owner(struct page *oldpage, struct page *newpage); extern void __set_page_owner_migrate_reason(struct page *page, int reason); +extern void __dump_page_owner(struct page *page); static inline void reset_page_owner(struct page *page, unsigned int order) { @@ -44,6 +45,11 @@ static inline void set_page_owner_migrate_reason(struct page *page, int reason) if (static_branch_unlikely(&page_owner_inited)) __set_page_owner_migrate_reason(page, reason); } +static inline void dump_page_owner(struct page *page) +{ + if (static_branch_unlikely(&page_owner_inited)) + __dump_page_owner(page); +} #else static inline void reset_page_owner(struct page *page, unsigned int order) { @@ -62,5 +68,8 @@ static inline void copy_page_owner(struct page *oldpage, struct page *newpage) static inline void set_page_owner_migrate_reason(struct page *page, int reason) { } +static inline void dump_page_owner(struct page *page) +{ +} #endif /* CONFIG_PAGE_OWNER */ #endif /* __LINUX_PAGE_OWNER_H */ diff --git a/mm/debug.c b/mm/debug.c index 1a71a3b..c3300ee 100644 --- a/mm/debug.c +++ b/mm/debug.c @@ -10,6 +10,7 @@ #include #include #include +#include static const struct trace_print_flags pageflag_names[] = { {1UL << PG_locked, "locked" }, @@ -118,6 +119,7 @@ void dump_page_badflags(struct page *page, const char *reason, void dump_page(struct page *page, const char *reason) { dump_page_badflags(page, reason, 0); + dump_page_owner(page); } EXPORT_SYMBOL(dump_page); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 61a023a..f806a1a 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -448,6 +448,7 @@ static void bad_page(struct page *page, const char *reason, printk(KERN_ALERT "BUG: Bad page state in process %s pfn:%05lx\n", current->comm, page_to_pfn(page)); dump_page_badflags(page, reason, bad_flags); + dump_page_owner(page); print_modules(); dump_stack(); diff --git a/mm/page_owner.c b/mm/page_owner.c index 59fd6f1..a81cfa0 100644 --- a/mm/page_owner.c +++ b/mm/page_owner.c @@ -193,6 +193,31 @@ print_page_owner(char __user *buf, size_t count, unsigned long pfn, return -ENOMEM; } +void __dump_page_owner(struct page *page) +{ + struct page_ext *page_ext = lookup_page_ext(page); + struct stack_trace trace = { + .nr_entries = page_ext->nr_entries, + .entries = &page_ext->trace_entries[0], + }; + gfp_t gfp_mask = page_ext->gfp_mask; + int mt = gfpflags_to_migratetype(gfp_mask); + + if (!test_bit(PAGE_EXT_OWNER, &page_ext->flags)) { + pr_alert("page_owner info is not active (free page?)\n"); + return; + } + ; + pr_alert("page allocated via order %u, migratetype %s, gfp_mask 0x%x", + page_ext->order, migratetype_names[mt], gfp_mask); + dump_gfpflag_names(gfp_mask); + print_stack_trace(&trace, 0); + + if (page_ext->last_migrate_reason != -1) + pr_alert("page has been migrated, last migrate reason: %s\n", + migrate_reason_names[page_ext->last_migrate_reason]); +} + static ssize_t read_page_owner(struct file *file, char __user *buf, size_t count, loff_t *ppos) { -- 2.6.3 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/