From bb422a738f6566f7439cd347d54e321e4fe92a9f Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Mon, 18 Dec 2017 20:31:41 +0900 Subject: mm,vmscan: Make unregister_shrinker() no-op if register_shrinker() failed. Syzbot caught an oops at unregister_shrinker() because combination of commit 1d3d4437eae1bb29 ("vmscan: per-node deferred work") and fault injection made register_shrinker() fail and the caller of register_shrinker() did not check for failure. ---------- [ 554.881422] FAULT_INJECTION: forcing a failure. [ 554.881422] name failslab, interval 1, probability 0, space 0, times 0 [ 554.881438] CPU: 1 PID: 13231 Comm: syz-executor1 Not tainted 4.14.0-rc8+ #82 [ 554.881443] Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 [ 554.881445] Call Trace: [ 554.881459] dump_stack+0x194/0x257 [ 554.881474] ? arch_local_irq_restore+0x53/0x53 [ 554.881486] ? find_held_lock+0x35/0x1d0 [ 554.881507] should_fail+0x8c0/0xa40 [ 554.881522] ? fault_create_debugfs_attr+0x1f0/0x1f0 [ 554.881537] ? check_noncircular+0x20/0x20 [ 554.881546] ? find_next_zero_bit+0x2c/0x40 [ 554.881560] ? ida_get_new_above+0x421/0x9d0 [ 554.881577] ? find_held_lock+0x35/0x1d0 [ 554.881594] ? __lock_is_held+0xb6/0x140 [ 554.881628] ? check_same_owner+0x320/0x320 [ 554.881634] ? lock_downgrade+0x990/0x990 [ 554.881649] ? find_held_lock+0x35/0x1d0 [ 554.881672] should_failslab+0xec/0x120 [ 554.881684] __kmalloc+0x63/0x760 [ 554.881692] ? lock_downgrade+0x990/0x990 [ 554.881712] ? register_shrinker+0x10e/0x2d0 [ 554.881721] ? trace_event_raw_event_module_request+0x320/0x320 [ 554.881737] register_shrinker+0x10e/0x2d0 [ 554.881747] ? prepare_kswapd_sleep+0x1f0/0x1f0 [ 554.881755] ? _down_write_nest_lock+0x120/0x120 [ 554.881765] ? memcpy+0x45/0x50 [ 554.881785] sget_userns+0xbcd/0xe20 (...snipped...) [ 554.898693] kasan: CONFIG_KASAN_INLINE enabled [ 554.898724] kasan: GPF could be caused by NULL-ptr deref or user memory access [ 554.898732] general protection fault: 0000 [#1] SMP KASAN [ 554.898737] Dumping ftrace buffer: [ 554.898741] (ftrace buffer empty) [ 554.898743] Modules linked in: [ 554.898752] CPU: 1 PID: 13231 Comm: syz-executor1 Not tainted 4.14.0-rc8+ #82 [ 554.898755] Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 [ 554.898760] task: ffff8801d1dbe5c0 task.stack: ffff8801c9e38000 [ 554.898772] RIP: 0010:__list_del_entry_valid+0x7e/0x150 [ 554.898775] RSP: 0018:ffff8801c9e3f108 EFLAGS: 00010246 [ 554.898780] RAX: dffffc0000000000 RBX: 0000000000000000 RCX: 0000000000000000 [ 554.898784] RDX: 0000000000000000 RSI: ffff8801c53c6f98 RDI: ffff8801c53c6fa0 [ 554.898788] RBP: ffff8801c9e3f120 R08: 1ffff100393c7d55 R09: 0000000000000004 [ 554.898791] R10: ffff8801c9e3ef70 R11: 0000000000000000 R12: 0000000000000000 [ 554.898795] R13: dffffc0000000000 R14: 1ffff100393c7e45 R15: ffff8801c53c6f98 [ 554.898800] FS: 0000000000000000(0000) GS:ffff8801db300000(0000) knlGS:0000000000000000 [ 554.898804] CS: 0010 DS: 002b ES: 002b CR0: 0000000080050033 [ 554.898807] CR2: 00000000dbc23000 CR3: 00000001c7269000 CR4: 00000000001406e0 [ 554.898813] DR0: 0000000020000000 DR1: 0000000020000000 DR2: 0000000000000000 [ 554.898816] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000600 [ 554.898818] Call Trace: [ 554.898828] unregister_shrinker+0x79/0x300 [ 554.898837] ? perf_trace_mm_vmscan_writepage+0x750/0x750 [ 554.898844] ? down_write+0x87/0x120 [ 554.898851] ? deactivate_super+0x139/0x1b0 [ 554.898857] ? down_read+0x150/0x150 [ 554.898864] ? check_same_owner+0x320/0x320 [ 554.898875] deactivate_locked_super+0x64/0xd0 [ 554.898883] deactivate_super+0x141/0x1b0 ---------- Since allowing register_shrinker() callers to call unregister_shrinker() when register_shrinker() failed can simplify error recovery path, this patch makes unregister_shrinker() no-op when register_shrinker() failed. Also, reset shrinker->nr_deferred in case unregister_shrinker() was by error called twice. Signed-off-by: Tetsuo Handa Signed-off-by: Aliaksei Karaliou Reported-by: syzbot Cc: Glauber Costa Cc: Al Viro Signed-off-by: Al Viro --- mm/vmscan.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'mm') diff --git a/mm/vmscan.c b/mm/vmscan.c index c02c850ea349..47d5ced51f2d 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -297,10 +297,13 @@ EXPORT_SYMBOL(register_shrinker); */ void unregister_shrinker(struct shrinker *shrinker) { + if (!shrinker->nr_deferred) + return; down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); kfree(shrinker->nr_deferred); + shrinker->nr_deferred = NULL; } EXPORT_SYMBOL(unregister_shrinker); -- cgit v1.2.3 From e8c24773d6b2cd9bc8b36bd6e60beff599be14be Mon Sep 17 00:00:00 2001 From: Dave Young Date: Thu, 4 Jan 2018 16:17:45 -0800 Subject: mm: check pfn_valid first in zero_resv_unavail With latest kernel I get below bug while testing kdump: BUG: unable to handle kernel paging request at ffffea00034b1040 IP: zero_resv_unavail+0xbd/0x126 PGD 37b98067 P4D 37b98067 PUD 37b97067 PMD 0 Oops: 0002 [#1] SMP Modules linked in: CPU: 0 PID: 0 Comm: swapper Not tainted 4.15.0-rc1+ #316 Hardware name: LENOVO 20ARS1BJ02/20ARS1BJ02, BIOS GJET92WW (2.42 ) 03/03/2017 task: ffffffff81a0e4c0 task.stack: ffffffff81a00000 RIP: 0010:zero_resv_unavail+0xbd/0x126 RSP: 0000:ffffffff81a03d88 EFLAGS: 00010006 RAX: 0000000000000000 RBX: ffffea00034b1040 RCX: 0000000000000010 RDX: 0000000000000000 RSI: 0000000000000092 RDI: ffffea00034b1040 RBP: 00000000000d2c41 R08: 00000000000000c0 R09: 0000000000000a0d R10: 0000000000000002 R11: 0000000000007f01 R12: ffffffff81a03d90 R13: ffffea0000000000 R14: 0000000000000063 R15: 0000000000000062 FS: 0000000000000000(0000) GS:ffffffff81c73000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffea00034b1040 CR3: 0000000037609000 CR4: 00000000000606b0 Call Trace: ? free_area_init_nodes+0x640/0x664 ? zone_sizes_init+0x58/0x72 ? setup_arch+0xb50/0xc6c ? start_kernel+0x64/0x43d ? secondary_startup_64+0xa5/0xb0 Code: c1 e8 0c 48 39 d8 76 27 48 89 de 48 c1 e3 06 48 c7 c7 7a 87 79 81 e8 b0 c0 3e ff 4c 01 eb b9 10 00 00 00 31 c0 48 89 df 49 ff c6 ab eb bc 6a 00 49 c7 c0 f0 93 d1 81 31 d2 83 ce ff 41 54 49 RIP: zero_resv_unavail+0xbd/0x126 RSP: ffffffff81a03d88 CR2: ffffea00034b1040 ---[ end trace f5ba9e8f73c7ee26 ]--- This is introduced by commit a4a3ede2132a ("mm: zero reserved and unavailable struct pages"). The reason is some efi reserved boot ranges is not reported in E820 ram. In my case it is a bgrt buffer: efi: mem00: [Boot Data |RUN| | | | | | | |WB|WT|WC|UC] range=[0x00000000d2c41000-0x00000000d2c85fff] (0MB) Use "add_efi_memmap" can workaround the problem with another fix: http://lkml.kernel.org/r/20171130052327.GA3500@dhcp-128-65.nay.redhat.com In zero_resv_unavail it would be better to check pfn_valid first before zero the page struct. This fixes the problem and potential other similar problems. Also as Pavel Tatashin suggested checks pfn_valid at the beginning of the section. The range is backed by real memory. The memory range is efi "Boot Service Data", that means after ExitBootServices() these ranges can be used as system ram. But some of them need to be reserved, for example the bgrt image address in an acpi table, if the image memory is freed then kexec reboot will fail because kexec inherit same acpi table to initialize the driver. Link: http://lkml.kernel.org/r/20171201095048.GA3084@dhcp-128-65.nay.redhat.com Fixes: a4a3ede2132a ("mm: zero reserved and unavailable struct pages") Signed-off-by: Dave Young Cc: Michal Hocko Cc: Pavel Tatashin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'mm') diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 7e5e775e97f4..76c9688b6a0a 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -6260,6 +6260,8 @@ void __paginginit zero_resv_unavail(void) pgcnt = 0; for_each_resv_unavail_range(i, &start, &end) { for (pfn = PFN_DOWN(start); pfn < PFN_UP(end); pfn++) { + if (!pfn_valid(ALIGN_DOWN(pfn, pageblock_nr_pages))) + continue; mm_zero_struct_page(pfn_to_page(pfn)); pgcnt++; } -- cgit v1.2.3 From 4991c09c7c812dba13ea9be79a68b4565bb1fa4e Mon Sep 17 00:00:00 2001 From: Anshuman Khandual Date: Thu, 4 Jan 2018 16:17:52 -0800 Subject: mm/mprotect: add a cond_resched() inside change_pmd_range() While testing on a large CPU system, detected the following RCU stall many times over the span of the workload. This problem is solved by adding a cond_resched() in the change_pmd_range() function. INFO: rcu_sched detected stalls on CPUs/tasks: 154-....: (670 ticks this GP) idle=022/140000000000000/0 softirq=2825/2825 fqs=612 (detected by 955, t=6002 jiffies, g=4486, c=4485, q=90864) Sending NMI from CPU 955 to CPUs 154: NMI backtrace for cpu 154 CPU: 154 PID: 147071 Comm: workload Not tainted 4.15.0-rc3+ #3 NIP: c0000000000b3f64 LR: c0000000000b33d4 CTR: 000000000000aa18 REGS: 00000000a4b0fb44 TRAP: 0501 Not tainted (4.15.0-rc3+) MSR: 8000000000009033 CR: 22422082 XER: 00000000 CFAR: 00000000006cf8f0 SOFTE: 1 GPR00: 0010000000000000 c00003ef9b1cb8c0 c0000000010cc600 0000000000000000 GPR04: 8e0000018c32b200 40017b3858fd6e00 8e0000018c32b208 40017b3858fd6e00 GPR08: 8e0000018c32b210 40017b3858fd6e00 8e0000018c32b218 40017b3858fd6e00 GPR12: ffffffffffffffff c00000000fb25100 NIP [c0000000000b3f64] plpar_hcall9+0x44/0x7c LR [c0000000000b33d4] pSeries_lpar_flush_hash_range+0x384/0x420 Call Trace: flush_hash_range+0x48/0x100 __flush_tlb_pending+0x44/0xd0 hpte_need_flush+0x408/0x470 change_protection_range+0xaac/0xf10 change_prot_numa+0x30/0xb0 task_numa_work+0x2d0/0x3e0 task_work_run+0x130/0x190 do_notify_resume+0x118/0x120 ret_from_except_lite+0x70/0x74 Instruction dump: 60000000 f8810028 7ca42b78 7cc53378 7ce63b78 7d074378 7d284b78 7d495378 e9410060 e9610068 e9810070 44000022 <7d806378> e9810028 f88c0000 f8ac0008 Link: http://lkml.kernel.org/r/20171214140551.5794-1-khandual@linux.vnet.ibm.com Signed-off-by: Anshuman Khandual Suggested-by: Nicholas Piggin Acked-by: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/mprotect.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'mm') diff --git a/mm/mprotect.c b/mm/mprotect.c index ec39f730a0bf..58b629bb70de 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -166,7 +166,7 @@ static inline unsigned long change_pmd_range(struct vm_area_struct *vma, next = pmd_addr_end(addr, end); if (!is_swap_pmd(*pmd) && !pmd_trans_huge(*pmd) && !pmd_devmap(*pmd) && pmd_none_or_clear_bad(pmd)) - continue; + goto next; /* invoke the mmu notifier if the pmd is populated */ if (!mni_start) { @@ -188,7 +188,7 @@ static inline unsigned long change_pmd_range(struct vm_area_struct *vma, } /* huge pmd was handled */ - continue; + goto next; } } /* fall through, the trans huge pmd just split */ @@ -196,6 +196,8 @@ static inline unsigned long change_pmd_range(struct vm_area_struct *vma, this_pages = change_pte_range(vma, pmd, addr, next, newprot, dirty_accountable, prot_numa); pages += this_pages; +next: + cond_resched(); } while (pmd++, addr = next, addr != end); if (mni_start) -- cgit v1.2.3 From 152a2d199e1385c6ccef17c24555103b30447c91 Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Thu, 4 Jan 2018 16:17:59 -0800 Subject: mm/debug.c: provide useful debugging information for VM_BUG With the recent addition of hashed kernel pointers, places which need to produce useful debug output have to specify %px, not %p. This patch fixes all the VM debug to use %px. This is appropriate because it's debug output that the user should never be able to trigger, and kernel developers need to see the actual pointers. Link: http://lkml.kernel.org/r/20171219133236.GE13680@bombadil.infradead.org Signed-off-by: Matthew Wilcox Acked-by: Michal Hocko Cc: "Tobin C. Harding" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/debug.c | 28 ++++++++++++++-------------- 1 file changed, 14 insertions(+), 14 deletions(-) (limited to 'mm') diff --git a/mm/debug.c b/mm/debug.c index d947f3e03b0d..56e2d9125ea5 100644 --- a/mm/debug.c +++ b/mm/debug.c @@ -50,7 +50,7 @@ void __dump_page(struct page *page, const char *reason) */ int mapcount = PageSlab(page) ? 0 : page_mapcount(page); - pr_emerg("page:%p count:%d mapcount:%d mapping:%p index:%#lx", + pr_emerg("page:%px count:%d mapcount:%d mapping:%px index:%#lx", page, page_ref_count(page), mapcount, page->mapping, page_to_pgoff(page)); if (PageCompound(page)) @@ -69,7 +69,7 @@ void __dump_page(struct page *page, const char *reason) #ifdef CONFIG_MEMCG if (page->mem_cgroup) - pr_alert("page->mem_cgroup:%p\n", page->mem_cgroup); + pr_alert("page->mem_cgroup:%px\n", page->mem_cgroup); #endif } @@ -84,10 +84,10 @@ EXPORT_SYMBOL(dump_page); void dump_vma(const struct vm_area_struct *vma) { - pr_emerg("vma %p start %p end %p\n" - "next %p prev %p mm %p\n" - "prot %lx anon_vma %p vm_ops %p\n" - "pgoff %lx file %p private_data %p\n" + pr_emerg("vma %px start %px end %px\n" + "next %px prev %px mm %px\n" + "prot %lx anon_vma %px vm_ops %px\n" + "pgoff %lx file %px private_data %px\n" "flags: %#lx(%pGv)\n", vma, (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_next, vma->vm_prev, vma->vm_mm, @@ -100,27 +100,27 @@ EXPORT_SYMBOL(dump_vma); void dump_mm(const struct mm_struct *mm) { - pr_emerg("mm %p mmap %p seqnum %d task_size %lu\n" + pr_emerg("mm %px mmap %px seqnum %d task_size %lu\n" #ifdef CONFIG_MMU - "get_unmapped_area %p\n" + "get_unmapped_area %px\n" #endif "mmap_base %lu mmap_legacy_base %lu highest_vm_end %lu\n" - "pgd %p mm_users %d mm_count %d pgtables_bytes %lu map_count %d\n" + "pgd %px mm_users %d mm_count %d pgtables_bytes %lu map_count %d\n" "hiwater_rss %lx hiwater_vm %lx total_vm %lx locked_vm %lx\n" "pinned_vm %lx data_vm %lx exec_vm %lx stack_vm %lx\n" "start_code %lx end_code %lx start_data %lx end_data %lx\n" "start_brk %lx brk %lx start_stack %lx\n" "arg_start %lx arg_end %lx env_start %lx env_end %lx\n" - "binfmt %p flags %lx core_state %p\n" + "binfmt %px flags %lx core_state %px\n" #ifdef CONFIG_AIO - "ioctx_table %p\n" + "ioctx_table %px\n" #endif #ifdef CONFIG_MEMCG - "owner %p " + "owner %px " #endif - "exe_file %p\n" + "exe_file %px\n" #ifdef CONFIG_MMU_NOTIFIER - "mmu_notifier_mm %p\n" + "mmu_notifier_mm %px\n" #endif #ifdef CONFIG_NUMA_BALANCING "numa_next_scan %lu numa_scan_offset %lu numa_scan_seq %d\n" -- cgit v1.2.3 From cdc346b36e1dfec201b24eddb7bdbcff6727db04 Mon Sep 17 00:00:00 2001 From: Sergey Senozhatsky Date: Thu, 4 Jan 2018 16:18:02 -0800 Subject: mm/zsmalloc.c: include fs.h `struct file_system_type' and alloc_anon_inode() function are defined in fs.h, include it directly. Link: http://lkml.kernel.org/r/20171219104219.3017-1-sergey.senozhatsky@gmail.com Signed-off-by: Sergey Senozhatsky Cc: Minchan Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/zsmalloc.c | 1 + 1 file changed, 1 insertion(+) (limited to 'mm') diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c index 685049a9048d..683c0651098c 100644 --- a/mm/zsmalloc.c +++ b/mm/zsmalloc.c @@ -53,6 +53,7 @@ #include #include #include +#include #define ZSPAGE_MAGIC 0x58 -- cgit v1.2.3 From d09cfbbfa0f761a97687828b5afb27b56cbf2e19 Mon Sep 17 00:00:00 2001 From: Baoquan He Date: Thu, 4 Jan 2018 16:18:06 -0800 Subject: mm/sparse.c: wrong allocation for mem_section In commit 83e3c48729d9 ("mm/sparsemem: Allocate mem_section at runtime for CONFIG_SPARSEMEM_EXTREME=y") mem_section is allocated at runtime to save memory. It allocates the first dimension of array with sizeof(struct mem_section). It costs extra memory, should be sizeof(struct mem_section *). Fix it. Link: http://lkml.kernel.org/r/1513932498-20350-1-git-send-email-bhe@redhat.com Fixes: 83e3c48729 ("mm/sparsemem: Allocate mem_section at runtime for CONFIG_SPARSEMEM_EXTREME=y") Signed-off-by: Baoquan He Tested-by: Dave Young Acked-by: Kirill A. Shutemov Cc: Kirill A. Shutemov Cc: Ingo Molnar Cc: Andy Lutomirski Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Atsushi Kumagai Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/sparse.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm') diff --git a/mm/sparse.c b/mm/sparse.c index 7a5dacaa06e3..2609aba121e8 100644 --- a/mm/sparse.c +++ b/mm/sparse.c @@ -211,7 +211,7 @@ void __init memory_present(int nid, unsigned long start, unsigned long end) if (unlikely(!mem_section)) { unsigned long size, align; - size = sizeof(struct mem_section) * NR_SECTION_ROOTS; + size = sizeof(struct mem_section*) * NR_SECTION_ROOTS; align = 1 << (INTERNODE_CACHE_SHIFT); mem_section = memblock_virt_alloc(size, align); } -- cgit v1.2.3 From d9570ee3bd1d4f20ce63485f5ef05663866fe6c0 Mon Sep 17 00:00:00 2001 From: Dmitry Vyukov Date: Fri, 12 Jan 2018 16:53:10 -0800 Subject: kmemleak: allow to coexist with fault injection kmemleak does one slab allocation per user allocation. So if slab fault injection is enabled to any degree, kmemleak instantly fails to allocate and turns itself off. However, it's useful to use kmemleak with fault injection to find leaks on error paths. On the other hand, checking kmemleak itself is not so useful because (1) it's a debugging tool and (2) it has a very regular allocation pattern (basically a single allocation site, so it either works or not). Turn off fault injection for kmemleak allocations. Link: http://lkml.kernel.org/r/20180109192243.19316-1-dvyukov@google.com Signed-off-by: Dmitry Vyukov Cc: Catalin Marinas Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/kmemleak.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm') diff --git a/mm/kmemleak.c b/mm/kmemleak.c index d73c14294f3a..f656ca27f6c2 100644 --- a/mm/kmemleak.c +++ b/mm/kmemleak.c @@ -127,7 +127,7 @@ /* GFP bitmask for kmemleak internal allocations */ #define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \ __GFP_NORETRY | __GFP_NOMEMALLOC | \ - __GFP_NOWARN) + __GFP_NOWARN | __GFP_NOFAIL) /* scanning area inside a memory block */ struct kmemleak_scan_area { -- cgit v1.2.3