From 7ace99170789bc53cbb7e9e352d7a3851208fbcf Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Thu, 6 Aug 2015 15:46:13 -0700 Subject: mm, meminit: allow early_pfn_to_nid to be used during runtime early_pfn_to_nid() historically was inherently not SMP safe but only used during boot which is inherently single threaded or during hotplug which is protected by a giant mutex. With deferred memory initialisation there was a thread-safe version introduced and the early_pfn_to_nid would trigger a BUG_ON if used unsafely. Memory hotplug hit that check. This patch makes early_pfn_to_nid introduces a lock to make it safe to use during hotplug. Signed-off-by: Mel Gorman Reported-by: Alex Ng Tested-by: Alex Ng Acked-by: Peter Zijlstra (Intel) Cc: Nicolai Stange Cc: Dave Hansen Cc: Fengguang Wu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index ef19f22b2b7d..ea0e6a6e7a11 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -981,21 +981,21 @@ static void __init __free_pages_boot_core(struct page *page, #if defined(CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID) || \ defined(CONFIG_HAVE_MEMBLOCK_NODE_MAP) -/* Only safe to use early in boot when initialisation is single-threaded */ + static struct mminit_pfnnid_cache early_pfnnid_cache __meminitdata; int __meminit early_pfn_to_nid(unsigned long pfn) { + static DEFINE_SPINLOCK(early_pfn_lock); int nid; - /* The system will behave unpredictably otherwise */ - BUG_ON(system_state != SYSTEM_BOOTING); - + spin_lock(&early_pfn_lock); nid = __early_pfn_to_nid(pfn, &early_pfnnid_cache); - if (nid >= 0) - return nid; - /* just returns 0 */ - return 0; + if (nid < 0) + nid = 0; + spin_unlock(&early_pfn_lock); + + return nid; } #endif -- cgit v1.2.3 From d3cd131d935ab3bab700491edbbd7cad4040ce50 Mon Sep 17 00:00:00 2001 From: Nicolai Stange Date: Thu, 6 Aug 2015 15:46:16 -0700 Subject: mm, meminit: replace rwsem with completion Commit 0e1cc95b4cc7 ("mm: meminit: finish initialisation of struct pages before basic setup") introduced a rwsem to signal completion of the initialization workers. Lockdep complains about possible recursive locking: ============================================= [ INFO: possible recursive locking detected ] 4.1.0-12802-g1dc51b8 #3 Not tainted --------------------------------------------- swapper/0/1 is trying to acquire lock: (pgdat_init_rwsem){++++.+}, at: [] page_alloc_init_late+0xc7/0xe6 but task is already holding lock: (pgdat_init_rwsem){++++.+}, at: [] page_alloc_init_late+0x3e/0xe6 Replace the rwsem by a completion together with an atomic "outstanding work counter". [peterz@infradead.org: Barrier removal on the grounds of being pointless] [mgorman@suse.de: Applied review feedback] Signed-off-by: Nicolai Stange Signed-off-by: Mel Gorman Acked-by: Peter Zijlstra (Intel) Cc: Dave Hansen Cc: Alex Ng Cc: Fengguang Wu Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index ea0e6a6e7a11..322628278ae4 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -18,7 +18,6 @@ #include #include #include -#include #include #include #include @@ -1060,7 +1059,15 @@ static void __init deferred_free_range(struct page *page, __free_pages_boot_core(page, pfn, 0); } -static __initdata DECLARE_RWSEM(pgdat_init_rwsem); +/* Completion tracking for deferred_init_memmap() threads */ +static atomic_t pgdat_init_n_undone __initdata; +static __initdata DECLARE_COMPLETION(pgdat_init_all_done_comp); + +static inline void __init pgdat_init_report_one_done(void) +{ + if (atomic_dec_and_test(&pgdat_init_n_undone)) + complete(&pgdat_init_all_done_comp); +} /* Initialise remaining memory on a node */ static int __init deferred_init_memmap(void *data) @@ -1077,7 +1084,7 @@ static int __init deferred_init_memmap(void *data) const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); if (first_init_pfn == ULONG_MAX) { - up_read(&pgdat_init_rwsem); + pgdat_init_report_one_done(); return 0; } @@ -1177,7 +1184,8 @@ free_range: pr_info("node %d initialised, %lu pages in %ums\n", nid, nr_pages, jiffies_to_msecs(jiffies - start)); - up_read(&pgdat_init_rwsem); + + pgdat_init_report_one_done(); return 0; } @@ -1185,14 +1193,14 @@ void __init page_alloc_init_late(void) { int nid; + /* There will be num_node_state(N_MEMORY) threads */ + atomic_set(&pgdat_init_n_undone, num_node_state(N_MEMORY)); for_each_node_state(nid, N_MEMORY) { - down_read(&pgdat_init_rwsem); kthread_run(deferred_init_memmap, NODE_DATA(nid), "pgdatinit%d", nid); } /* Block until all are initialised */ - down_write(&pgdat_init_rwsem); - up_write(&pgdat_init_rwsem); + wait_for_completion(&pgdat_init_all_done_comp); } #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */ -- cgit v1.2.3 From 4248b0da460839e30eaaad78992b9a1dd3e63e21 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Thu, 6 Aug 2015 15:46:20 -0700 Subject: fs, file table: reinit files_stat.max_files after deferred memory initialisation Dave Hansen reported the following; My laptop has been behaving strangely with 4.2-rc2. Once I log in to my X session, I start getting all kinds of strange errors from applications and see this in my dmesg: VFS: file-max limit 8192 reached The problem is that the file-max is calculated before memory is fully initialised and miscalculates how much memory the kernel is using. This patch recalculates file-max after deferred memory initialisation. Note that using memory hotplug infrastructure would not have avoided this problem as the value is not recalculated after memory hot-add. 4.1: files_stat.max_files = 6582781 4.2-rc2: files_stat.max_files = 8192 4.2-rc2 patched: files_stat.max_files = 6562467 Small differences with the patch applied and 4.1 but not enough to matter. Signed-off-by: Mel Gorman Reported-by: Dave Hansen Cc: Nicolai Stange Cc: Dave Hansen Cc: Alex Ng Cc: Fengguang Wu Cc: Peter Zijlstra (Intel) Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dcache.c | 13 +++---------- fs/file_table.c | 24 +++++++++++++++--------- include/linux/fs.h | 5 +++-- init/main.c | 2 +- mm/page_alloc.c | 3 +++ 5 files changed, 25 insertions(+), 22 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/fs/dcache.c b/fs/dcache.c index 5c8ea15e73a5..9b5fe503f6cb 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -3442,22 +3442,15 @@ void __init vfs_caches_init_early(void) inode_init_early(); } -void __init vfs_caches_init(unsigned long mempages) +void __init vfs_caches_init(void) { - unsigned long reserve; - - /* Base hash sizes on available memory, with a reserve equal to - 150% of current kernel size */ - - reserve = min((mempages - nr_free_pages()) * 3/2, mempages - 1); - mempages -= reserve; - names_cachep = kmem_cache_create("names_cache", PATH_MAX, 0, SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); dcache_init(); inode_init(); - files_init(mempages); + files_init(); + files_maxfiles_init(); mnt_init(); bdev_cache_init(); chrdev_init(); diff --git a/fs/file_table.c b/fs/file_table.c index 7f9d407c7595..ad17e05ebf95 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -25,6 +25,7 @@ #include #include #include +#include #include @@ -308,19 +309,24 @@ void put_filp(struct file *file) } } -void __init files_init(unsigned long mempages) +void __init files_init(void) { - unsigned long n; - filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0, SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL); + percpu_counter_init(&nr_files, 0, GFP_KERNEL); +} - /* - * One file with associated inode and dcache is very roughly 1K. - * Per default don't use more than 10% of our memory for files. - */ +/* + * One file with associated inode and dcache is very roughly 1K. Per default + * do not use more than 10% of our memory for files. + */ +void __init files_maxfiles_init(void) +{ + unsigned long n; + unsigned long memreserve = (totalram_pages - nr_free_pages()) * 3/2; + + memreserve = min(memreserve, totalram_pages - 1); + n = ((totalram_pages - memreserve) * (PAGE_SIZE / 1024)) / 10; - n = (mempages * (PAGE_SIZE / 1024)) / 10; files_stat.max_files = max_t(unsigned long, n, NR_FILE); - percpu_counter_init(&nr_files, 0, GFP_KERNEL); } diff --git a/include/linux/fs.h b/include/linux/fs.h index cc008c338f5a..84b783f277f7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -55,7 +55,8 @@ struct vm_fault; extern void __init inode_init(void); extern void __init inode_init_early(void); -extern void __init files_init(unsigned long); +extern void __init files_init(void); +extern void __init files_maxfiles_init(void); extern struct files_stat_struct files_stat; extern unsigned long get_max_files(void); @@ -2245,7 +2246,7 @@ extern int ioctl_preallocate(struct file *filp, void __user *argp); /* fs/dcache.c */ extern void __init vfs_caches_init_early(void); -extern void __init vfs_caches_init(unsigned long); +extern void __init vfs_caches_init(void); extern struct kmem_cache *names_cachep; diff --git a/init/main.c b/init/main.c index c5d5626289ce..56506553d4d8 100644 --- a/init/main.c +++ b/init/main.c @@ -656,7 +656,7 @@ asmlinkage __visible void __init start_kernel(void) key_init(); security_init(); dbg_late_init(); - vfs_caches_init(totalram_pages); + vfs_caches_init(); signals_init(); /* rootfs populating might need page-writeback */ page_writeback_init(); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 322628278ae4..cb61f44eb3fc 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1201,6 +1201,9 @@ void __init page_alloc_init_late(void) /* Block until all are initialised */ wait_for_completion(&pgdat_init_all_done_comp); + + /* Reinit limits that are based on free pages after the kernel is up */ + files_maxfiles_init(); } #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */ -- cgit v1.2.3 From f4c18e6f7b5bbb5b528b3334115806b0d76f50f9 Mon Sep 17 00:00:00 2001 From: Naoya Horiguchi Date: Thu, 6 Aug 2015 15:47:08 -0700 Subject: mm: check __PG_HWPOISON separately from PAGE_FLAGS_CHECK_AT_* The race condition addressed in commit add05cecef80 ("mm: soft-offline: don't free target page in successful page migration") was not closed completely, because that can happen not only for soft-offline, but also for hard-offline. Consider that a slab page is about to be freed into buddy pool, and then an uncorrected memory error hits the page just after entering __free_one_page(), then VM_BUG_ON_PAGE(page->flags & PAGE_FLAGS_CHECK_AT_PREP) is triggered, despite the fact that it's not necessary because the data on the affected page is not consumed. To solve it, this patch drops __PG_HWPOISON from page flag checks at allocation/free time. I think it's justified because __PG_HWPOISON flags is defined to prevent the page from being reused, and setting it outside the page's alloc-free cycle is a designed behavior (not a bug.) For recent months, I was annoyed about BUG_ON when soft-offlined page remains on lru cache list for a while, which is avoided by calling put_page() instead of putback_lru_page() in page migration's success path. This means that this patch reverts a major change from commit add05cecef80 about the new refcounting rule of soft-offlined pages, so "reuse window" revives. This will be closed by a subsequent patch. Signed-off-by: Naoya Horiguchi Cc: Andi Kleen Cc: Dean Nelson Cc: Tony Luck Cc: "Kirill A. Shutemov" Cc: Hugh Dickins Cc: David Rientjes Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/page-flags.h | 10 +++++++--- mm/huge_memory.c | 7 +------ mm/migrate.c | 5 ++++- mm/page_alloc.c | 4 ++++ 4 files changed, 16 insertions(+), 10 deletions(-) (limited to 'mm/page_alloc.c') diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index f34e040b34e9..41c93844fb1d 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -631,15 +631,19 @@ static inline void ClearPageSlabPfmemalloc(struct page *page) 1 << PG_private | 1 << PG_private_2 | \ 1 << PG_writeback | 1 << PG_reserved | \ 1 << PG_slab | 1 << PG_swapcache | 1 << PG_active | \ - 1 << PG_unevictable | __PG_MLOCKED | __PG_HWPOISON | \ + 1 << PG_unevictable | __PG_MLOCKED | \ __PG_COMPOUND_LOCK) /* * Flags checked when a page is prepped for return by the page allocator. - * Pages being prepped should not have any flags set. It they are set, + * Pages being prepped should not have these flags set. It they are set, * there has been a kernel bug or struct page corruption. + * + * __PG_HWPOISON is exceptional because it needs to be kept beyond page's + * alloc-free cycle to prevent from reusing the page. */ -#define PAGE_FLAGS_CHECK_AT_PREP ((1 << NR_PAGEFLAGS) - 1) +#define PAGE_FLAGS_CHECK_AT_PREP \ + (((1 << NR_PAGEFLAGS) - 1) & ~__PG_HWPOISON) #define PAGE_FLAGS_PRIVATE \ (1 << PG_private | 1 << PG_private_2) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index c107094f79ba..097c7a4bfbd9 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1676,12 +1676,7 @@ static void __split_huge_page_refcount(struct page *page, /* after clearing PageTail the gup refcount can be released */ smp_mb__after_atomic(); - /* - * retain hwpoison flag of the poisoned tail page: - * fix for the unsuitable process killed on Guest Machine(KVM) - * by the memory-failure. - */ - page_tail->flags &= ~PAGE_FLAGS_CHECK_AT_PREP | __PG_HWPOISON; + page_tail->flags &= ~PAGE_FLAGS_CHECK_AT_PREP; page_tail->flags |= (page->flags & ((1L << PG_referenced) | (1L << PG_swapbacked) | diff --git a/mm/migrate.c b/mm/migrate.c index ee401e4e5ef1..f2415be7d93b 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -950,7 +950,10 @@ out: list_del(&page->lru); dec_zone_page_state(page, NR_ISOLATED_ANON + page_is_file_cache(page)); - if (reason != MR_MEMORY_FAILURE) + /* Soft-offlined page shouldn't go through lru cache list */ + if (reason == MR_MEMORY_FAILURE) + put_page(page); + else putback_lru_page(page); } diff --git a/mm/page_alloc.c b/mm/page_alloc.c index cb61f44eb3fc..beda41710802 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1296,6 +1296,10 @@ static inline int check_new_page(struct page *page) bad_reason = "non-NULL mapping"; if (unlikely(atomic_read(&page->_count) != 0)) bad_reason = "nonzero _count"; + if (unlikely(page->flags & __PG_HWPOISON)) { + bad_reason = "HWPoisoned (hardware-corrupted)"; + bad_flags = __PG_HWPOISON; + } if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_PREP)) { bad_reason = "PAGE_FLAGS_CHECK_AT_PREP flag set"; bad_flags = PAGE_FLAGS_CHECK_AT_PREP; -- cgit v1.2.3