linux/mm/page_alloc.c, branch v5.0

mm, page_alloc: fix a division by zero error when boosting watermarks v2

2019-02-21T17:01:00Z

Yury Norov reported that an arm64 KVM instance could not boot since after v5.0-rc1 and could addressed by reverting the patches 1c30844d2dfe272d58c ("mm: reclaim small amounts of memory when an external 73444bc4d8f92e46a20 ("mm, page_alloc: do not wake kswapd with zone lock held") The problem is that a division by zero error is possible if boosting occurs very early in boot if the system has very little memory. This patch avoids the division by zero error. Link: http://lkml.kernel.org/r/20190213143012.GT9565@techsingularity.net Fixes: 1c30844d2dfe ("mm: reclaim small amounts of memory when an external fragmentation event occurs") Signed-off-by: Mel Gorman Reported-by: Yury Norov Tested-by: Yury Norov Tested-by: Will Deacon Acked-by: Vlastimil Babka Cc: Andrea Arcangeli Cc: David Rientjes Cc: Michal Hocko Cc: Catalin Marinas Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm: Use fixed constant in page_frag_alloc instead of size + 1

2019-02-17T23:48:43Z

This patch replaces the size + 1 value introduced with the recent fix for 1 byte allocs with a constant value. The idea here is to reduce code overhead as the previous logic would have to read size into a register, then increment it, and write it back to whatever field was being used. By using a constant we can avoid those memory reads and arithmetic operations in favor of just encoding the maximum value into the operation itself. Fixes: 2c2ade81741c ("mm: page_alloc: fix ref bias in page_frag_alloc() for 1-byte allocs") Signed-off-by: Alexander Duyck Signed-off-by: David S. Miller

mm: page_alloc: fix ref bias in page_frag_alloc() for 1-byte allocs

2019-02-14T17:12:17Z

The basic idea behind ->pagecnt_bias is: If we pre-allocate the maximum number of references that we might need to create in the fastpath later, the bump-allocation fastpath only has to modify the non-atomic bias value that tracks the number of extra references we hold instead of the atomic refcount. The maximum number of allocations we can serve (under the assumption that no allocation is made with size 0) is nc->size, so that's the bias used. However, even when all memory in the allocation has been given away, a reference to the page is still held; and in the `offset < 0` slowpath, the page may be reused if everyone else has dropped their references. This means that the necessary number of references is actually `nc->size+1`. Luckily, from a quick grep, it looks like the only path that can call page_frag_alloc(fragsz=1) is TAP with the IFF_NAPI_FRAGS flag, which requires CAP_NET_ADMIN in the init namespace and is only intended to be used for kernel testing and fuzzing. To test for this issue, put a `WARN_ON(page_ref_count(page) == 0)` in the `offset < 0` path, below the virt_to_page() call, and then repeatedly call writev() on a TAP device with IFF_TAP|IFF_NO_PI|IFF_NAPI_FRAGS|IFF_NAPI, with a vector consisting of 15 elements containing 1 byte each. Signed-off-by: Jann Horn Signed-off-by: David S. Miller

Revert "mm, memory_hotplug: initialize struct pages for the full memory section"

2019-01-28T18:35:22Z

This reverts commit 2830bf6f05fb3e05bc4743274b806c821807a684. The underlying assumption that one sparse section belongs into a single numa node doesn't hold really. Robert Shteynfeld has reported a boot failure. The boot log was not captured but his memory layout is as follows: Early memory node ranges node 1: [mem 0x0000000000001000-0x0000000000090fff] node 1: [mem 0x0000000000100000-0x00000000dbdf8fff] node 1: [mem 0x0000000100000000-0x0000001423ffffff] node 0: [mem 0x0000001424000000-0x0000002023ffffff] This means that node0 starts in the middle of a memory section which is also in node1. memmap_init_zone tries to initialize padding of a section even when it is outside of the given pfn range because there are code paths (e.g. memory hotplug) which assume that the full worth of memory section is always initialized. In this particular case, though, such a range is already intialized and most likely already managed by the page allocator. Scribbling over those pages corrupts the internal state and likely blows up when any of those pages gets used. Reported-by: Robert Shteynfeld Fixes: 2830bf6f05fb ("mm, memory_hotplug: initialize struct pages for the full memory section") Cc: stable@kernel.org Signed-off-by: Michal Hocko Signed-off-by: Linus Torvalds

mm, page_alloc: do not wake kswapd with zone lock held

2019-01-09T01:15:11Z

syzbot reported the following regression in the latest merge window and it was confirmed by Qian Cai that a similar bug was visible from a different context. ====================================================== WARNING: possible circular locking dependency detected 4.20.0+ #297 Not tainted ------------------------------------------------------ syz-executor0/8529 is trying to acquire lock: 000000005e7fb829 (&pgdat->kswapd_wait){....}, at: __wake_up_common_lock+0x19e/0x330 kernel/sched/wait.c:120 but task is already holding lock: 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: spin_lock include/linux/spinlock.h:329 [inline] 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: rmqueue_bulk mm/page_alloc.c:2548 [inline] 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: __rmqueue_pcplist mm/page_alloc.c:3021 [inline] 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: rmqueue_pcplist mm/page_alloc.c:3050 [inline] 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: rmqueue mm/page_alloc.c:3072 [inline] 000000009bb7bae0 (&(&zone->lock)->rlock){-.-.}, at: get_page_from_freelist+0x1bae/0x52a0 mm/page_alloc.c:3491 It appears to be a false positive in that the only way the lock ordering should be inverted is if kswapd is waking itself and the wakeup allocates debugging objects which should already be allocated if it's kswapd doing the waking. Nevertheless, the possibility exists and so it's best to avoid the problem. This patch flags a zone as needing a kswapd using the, surprisingly, unused zone flag field. The flag is read without the lock held to do the wakeup. It's possible that the flag setting context is not the same as the flag clearing context or for small races to occur. However, each race possibility is harmless and there is no visible degredation in fragmentation treatment. While zone->flag could have continued to be unused, there is potential for moving some existing fields into the flags field instead. Particularly read-mostly ones like zone->initialized and zone->contiguous. Link: http://lkml.kernel.org/r/20190103225712.GJ31517@techsingularity.net Fixes: 1c30844d2dfe ("mm: reclaim small amounts of memory when an external fragmentation event occurs") Reported-by: syzbot+93d94a001cfbce9e60e1@syzkaller.appspotmail.com Signed-off-by: Mel Gorman Acked-by: Vlastimil Babka Tested-by: Qian Cai Cc: Dmitry Vyukov Cc: Vlastimil Babka Cc: Michal Hocko Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm/page_alloc.c: allow error injection

2018-12-28T20:11:51Z

Model call chain after should_failslab(). Likewise, we can now use a kprobe to override the return value of should_fail_alloc_page() and inject allocation failures into alloc_page*(). This will allow injecting allocation failures using the BCC tools even without building kernel with CONFIG_FAIL_PAGE_ALLOC and booting it with a fail_page_alloc= parameter, which incurs some overhead even when failures are not being injected. On the other hand, this patch adds an unconditional call to should_fail_alloc_page() from page allocation hotpath. That overhead should be rather negligible with CONFIG_FAIL_PAGE_ALLOC=n when there's no kprobe attached, though. [vbabka@suse.cz: changelog addition] Link: http://lkml.kernel.org/r/20181214074330.18917-1-bpoirier@suse.com Signed-off-by: Benjamin Poirier Acked-by: Vlastimil Babka Cc: Arnd Bergmann Cc: Michal Hocko Cc: Pavel Tatashin Cc: Oscar Salvador Cc: Mike Rapoport Cc: Joonsoo Kim Cc: Alexander Duyck Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm, page_alloc: enable pcpu_drain with zone capability

2018-12-28T20:11:51Z

drain_all_pages is documented to drain per-cpu pages for a given zone (if non-NULL). The current implementation doesn't match the description though. It will drain all pcp pages for all zones that happen to have cached pages on the same cpu as the given zone. This will lead to premature pcp cache draining for zones that are not of any interest to the caller - e.g. compaction, hwpoison or memory offline. This forces the page allocator to take locks and potential lock contention as a result. There is no real reason for this sub-optimal implementation. Replace per-cpu work item with a dedicated structure which contains a pointer to the zone and pass it over to the worker. This will get the zone information all the way down to the worker function and do the right job. [akpm@linux-foundation.org: avoid 80-col tricks] [mhocko@suse.com: refactor the whole changelog] Link: http://lkml.kernel.org/r/20181212142550.61686-1-richard.weiyang@gmail.com Signed-off-by: Wei Yang Acked-by: Michal Hocko Reviewed-by: Oscar Salvador Reviewed-by: David Hildenbrand Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm/page_alloc.c: don't call kasan_free_pages() at deferred mem init

2018-12-28T20:11:51Z

When CONFIG_KASAN is enabled on large memory SMP systems, the deferrred pages initialization can take a long time. Below were the reported init times on a 8-socket 96-core 4TB IvyBridge system. 1) Non-debug kernel without CONFIG_KASAN [ 8.764222] node 1 initialised, 132086516 pages in 7027ms 2) Debug kernel with CONFIG_KASAN [ 146.288115] node 1 initialised, 132075466 pages in 143052ms So the page init time in a debug kernel was 20X of the non-debug kernel. The long init time can be problematic as the page initialization is done with interrupt disabled. In this particular case, it caused the appearance of following warning messages as well as NMI backtraces of all the cores that were doing the initialization. [ 68.240049] rcu: INFO: rcu_sched detected stalls on CPUs/tasks: [ 68.241000] rcu: 25-...0: (100 ticks this GP) idle=b72/1/0x4000000000000000 softirq=915/915 fqs=16252 [ 68.241000] rcu: 44-...0: (95 ticks this GP) idle=49a/1/0x4000000000000000 softirq=788/788 fqs=16253 [ 68.241000] rcu: 54-...0: (104 ticks this GP) idle=03a/1/0x4000000000000000 softirq=721/825 fqs=16253 [ 68.241000] rcu: 60-...0: (103 ticks this GP) idle=cbe/1/0x4000000000000000 softirq=637/740 fqs=16253 [ 68.241000] rcu: 72-...0: (105 ticks this GP) idle=786/1/0x4000000000000000 softirq=536/641 fqs=16253 [ 68.241000] rcu: 84-...0: (99 ticks this GP) idle=292/1/0x4000000000000000 softirq=537/537 fqs=16253 [ 68.241000] rcu: 111-...0: (104 ticks this GP) idle=bde/1/0x4000000000000000 softirq=474/476 fqs=16253 [ 68.241000] rcu: (detected by 13, t=65018 jiffies, g=249, q=2) The long init time was mainly caused by the call to kasan_free_pages() to poison the newly initialized pages. On a 4TB system, we are talking about almost 500GB of memory probably on the same node. In reality, we may not need to poison the newly initialized pages before they are ever allocated. So KASAN poisoning of freed pages before the completion of deferred memory initialization is now disabled. Those pages will be properly poisoned when they are allocated or freed after deferred pages initialization is done. With this change, the new page initialization time became: [ 21.948010] node 1 initialised, 132075466 pages in 18702ms This was still about double the non-debug kernel time, but was much better than before. Link: http://lkml.kernel.org/r/1544459388-8736-1-git-send-email-longman@redhat.com Signed-off-by: Waiman Long Reviewed-by: Andrew Morton Cc: Andrey Ryabinin Cc: Alexander Potapenko Cc: Dmitry Vyukov Cc: Michal Hocko Cc: Pasha Tatashin Cc: Oscar Salvador Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm/pageblock: throw compile error if pageblock_bits cannot hold MIGRATE_TYPES

2018-12-28T20:11:51Z

Currently, NR_PAGEBLOCK_BITS and MIGRATE_TYPES are not associated by code. If someone adds extra migrate type, then he may forget to enlarge the NR_PAGEBLOCK_BITS. Hence it requires some way to fix. NR_PAGEBLOCK_BITS depends on MIGRATE_TYPES, while these macro spread on two different .h file with reverse dependency, it is a little hard to refer to MIGRATE_TYPES in pageblock-flag.h. This patch tries to remind such relation in compiling-time. Link: http://lkml.kernel.org/r/1544508709-11358-1-git-send-email-kernelfans@gmail.com Signed-off-by: Pingfan Liu Reviewed-by: Andrew Morton Acked-by: Vlastimil Babka Cc: Michal Hocko Cc: Pavel Tatashin Cc: Oscar Salvador Cc: Mike Rapoport Cc: Joonsoo Kim Cc: Alexander Duyck Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

mm/page_alloc.c: drop uneeded meminit and meminitdata

2018-12-28T20:11:49Z

Since commit 03e85f9d5f1 ("mm/page_alloc: Introduce free_area_init_core_hotplug"), some functions changed to only be called during system initialization. Concretly, free_area_init_node() and the functions that hang from it. Also, some variables are no longer used after the system has gone through initialization. So this could be considered as a late clean-up for that patch. This patch changes the functions from __meminit to __init, and the variables from __meminitdata to __initdata. In return, we get some KBs back: Before: Freeing unused kernel image memory: 2472K After: Freeing unused kernel image memory: 2480K Link: http://lkml.kernel.org/r/20181204111507.4808-1-osalvador@suse.de Signed-off-by: Oscar Salvador Reviewed-by: Wei Yang Cc: Michal Hocko Cc: Pavel Tatashin Cc: Vlastimil Babka Cc: Alexander Duyck Cc: David Hildenbrand Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

linux/mm/page_alloc.c, branch v5.0

mm, page_alloc: fix a division by zero error when boosting watermarks v2

mm: Use fixed constant in page_frag_alloc instead of size + 1

mm: page_alloc: fix ref bias in page_frag_alloc() for 1-byte allocs

Revert "mm, memory_hotplug: initialize struct pages for the full memory section"

mm, page_alloc: do not wake kswapd with zone lock held

mm/page_alloc.c: allow error injection

mm, page_alloc: enable pcpu_drain with zone capability

mm/page_alloc.c: don't call kasan_free_pages() at deferred mem init

mm/pageblock: throw compile error if pageblock_bits cannot hold MIGRATE_TYPES

mm/page_alloc.c: drop uneeded __meminit and __meminitdata

mm/page_alloc.c: drop uneeded meminit and meminitdata