From 0b28179a6138a5edd9d82ad2687c05b3773c387b Mon Sep 17 00:00:00 2001 From: Vasily Averin Date: Fri, 5 Nov 2021 13:38:02 -0700 Subject: mm, oom: pagefault_out_of_memory: don't force global OOM for dying tasks Patch series "memcg: prohibit unconditional exceeding the limit of dying tasks", v3. Memory cgroup charging allows killed or exiting tasks to exceed the hard limit. It can be misused and allowed to trigger global OOM from inside a memcg-limited container. On the other hand if memcg fails allocation, called from inside #PF handler it triggers global OOM from inside pagefault_out_of_memory(). To prevent these problems this patchset: (a) removes execution of out_of_memory() from pagefault_out_of_memory(), becasue nobody can explain why it is necessary. (b) allow memcg to fail allocation of dying/killed tasks. This patch (of 3): Any allocation failure during the #PF path will return with VM_FAULT_OOM which in turn results in pagefault_out_of_memory which in turn executes out_out_memory() and can kill a random task. An allocation might fail when the current task is the oom victim and there are no memory reserves left. The OOM killer is already handled at the page allocator level for the global OOM and at the charging level for the memcg one. Both have much more information about the scope of allocation/charge request. This means that either the OOM killer has been invoked properly and didn't lead to the allocation success or it has been skipped because it couldn't have been invoked. In both cases triggering it from here is pointless and even harmful. It makes much more sense to let the killed task die rather than to wake up an eternally hungry oom-killer and send him to choose a fatter victim for breakfast. Link: https://lkml.kernel.org/r/0828a149-786e-7c06-b70a-52d086818ea3@virtuozzo.com Signed-off-by: Vasily Averin Suggested-by: Michal Hocko Acked-by: Michal Hocko Cc: Johannes Weiner Cc: Mel Gorman Cc: Roman Gushchin Cc: Shakeel Butt Cc: Tetsuo Handa Cc: Uladzislau Rezki Cc: Vladimir Davydov Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/oom_kill.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'mm/oom_kill.c') diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 989f35a2bbb1..d89545d1a5b8 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1137,6 +1137,9 @@ void pagefault_out_of_memory(void) if (mem_cgroup_oom_synchronize(true)) return; + if (fatal_signal_pending(current)) + return; + if (!mutex_trylock(&oom_lock)) return; out_of_memory(&oc); -- cgit From 60e2793d440a3ec95abb5d6d4fc034a4b480472d Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Fri, 5 Nov 2021 13:38:06 -0700 Subject: mm, oom: do not trigger out_of_memory from the #PF Any allocation failure during the #PF path will return with VM_FAULT_OOM which in turn results in pagefault_out_of_memory. This can happen for 2 different reasons. a) Memcg is out of memory and we rely on mem_cgroup_oom_synchronize to perform the memcg OOM handling or b) normal allocation fails. The latter is quite problematic because allocation paths already trigger out_of_memory and the page allocator tries really hard to not fail allocations. Anyway, if the OOM killer has been already invoked there is no reason to invoke it again from the #PF path. Especially when the OOM condition might be gone by that time and we have no way to find out other than allocate. Moreover if the allocation failed and the OOM killer hasn't been invoked then we are unlikely to do the right thing from the #PF context because we have already lost the allocation context and restictions and therefore might oom kill a task from a different NUMA domain. This all suggests that there is no legitimate reason to trigger out_of_memory from pagefault_out_of_memory so drop it. Just to be sure that no #PF path returns with VM_FAULT_OOM without allocation print a warning that this is happening before we restart the #PF. [VvS: #PF allocation can hit into limit of cgroup v1 kmem controller. This is a local problem related to memcg, however, it causes unnecessary global OOM kills that are repeated over and over again and escalate into a real disaster. This has been broken since kmem accounting has been introduced for cgroup v1 (3.8). There was no kmem specific reclaim for the separate limit so the only way to handle kmem hard limit was to return with ENOMEM. In upstream the problem will be fixed by removing the outdated kmem limit, however stable and LTS kernels cannot do it and are still affected. This patch fixes the problem and should be backported into stable/LTS.] Link: https://lkml.kernel.org/r/f5fd8dd8-0ad4-c524-5f65-920b01972a42@virtuozzo.com Signed-off-by: Michal Hocko Signed-off-by: Vasily Averin Acked-by: Michal Hocko Cc: Johannes Weiner Cc: Mel Gorman Cc: Roman Gushchin Cc: Shakeel Butt Cc: Tetsuo Handa Cc: Uladzislau Rezki Cc: Vladimir Davydov Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/oom_kill.c | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) (limited to 'mm/oom_kill.c') diff --git a/mm/oom_kill.c b/mm/oom_kill.c index d89545d1a5b8..bfa9e348c3a3 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1120,19 +1120,15 @@ bool out_of_memory(struct oom_control *oc) } /* - * The pagefault handler calls here because it is out of memory, so kill a - * memory-hogging task. If oom_lock is held by somebody else, a parallel oom - * killing is already in progress so do nothing. + * The pagefault handler calls here because some allocation has failed. We have + * to take care of the memcg OOM here because this is the only safe context without + * any locks held but let the oom killer triggered from the allocation context care + * about the global OOM. */ void pagefault_out_of_memory(void) { - struct oom_control oc = { - .zonelist = NULL, - .nodemask = NULL, - .memcg = NULL, - .gfp_mask = 0, - .order = 0, - }; + static DEFINE_RATELIMIT_STATE(pfoom_rs, DEFAULT_RATELIMIT_INTERVAL, + DEFAULT_RATELIMIT_BURST); if (mem_cgroup_oom_synchronize(true)) return; @@ -1140,10 +1136,8 @@ void pagefault_out_of_memory(void) if (fatal_signal_pending(current)) return; - if (!mutex_trylock(&oom_lock)) - return; - out_of_memory(&oc); - mutex_unlock(&oom_lock); + if (__ratelimit(&pfoom_rs)) + pr_warn("Huh VM_FAULT_OOM leaked out to the #PF handler. Retrying PF\n"); } SYSCALL_DEFINE2(process_mrelease, int, pidfd, unsigned int, flags) -- cgit From 3723929eb0f50e2101de739cdb66458a4f1f4b27 Mon Sep 17 00:00:00 2001 From: Sultan Alsawaf Date: Fri, 5 Nov 2021 13:43:25 -0700 Subject: mm: mark the OOM reaper thread as freezable The OOM reaper alters user address space which might theoretically alter the snapshot if reaping is allowed to happen after the freezer quiescent state. To this end, the reaper kthread uses wait_event_freezable() while waiting for any work so that it cannot run while the system freezes. However, the current implementation doesn't respect the freezer because all kernel threads are created with the PF_NOFREEZE flag, so they are automatically excluded from freezing operations. This means that the OOM reaper can race with system snapshotting if it has work to do while the system is being frozen. Fix this by adding a set_freezable() call which will clear the PF_NOFREEZE flag and thus make the OOM reaper visible to the freezer. Please note that the OOM reaper altering the snapshot this way is mostly a theoretical concern and has not been observed in practice. Link: https://lkml.kernel.org/r/20210921165758.6154-1-sultan@kerneltoast.com Link: https://lkml.kernel.org/r/20210918233920.9174-1-sultan@kerneltoast.com Fixes: aac453635549 ("mm, oom: introduce oom reaper") Signed-off-by: Sultan Alsawaf Acked-by: Michal Hocko Cc: David Rientjes Cc: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/oom_kill.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'mm/oom_kill.c') diff --git a/mm/oom_kill.c b/mm/oom_kill.c index bfa9e348c3a3..d365cc84a486 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -641,6 +641,8 @@ done: static int oom_reaper(void *unused) { + set_freezable(); + while (true) { struct task_struct *tsk = NULL; -- cgit