From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.0 required=3.0 tests=INCLUDES_PATCH, MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 0BD67C46475 for ; Thu, 25 Oct 2018 08:24:29 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id A71E3204FD for ; Thu, 25 Oct 2018 08:24:28 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org A71E3204FD Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726992AbeJYQz5 (ORCPT ); Thu, 25 Oct 2018 12:55:57 -0400 Received: from mail-ed1-f67.google.com ([209.85.208.67]:38039 "EHLO mail-ed1-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726473AbeJYQz4 (ORCPT ); Thu, 25 Oct 2018 12:55:56 -0400 Received: by mail-ed1-f67.google.com with SMTP id c1-v6so7614379ede.5 for ; Thu, 25 Oct 2018 01:24:13 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=e/av0pB8zJCwcJxHsePcmzWL1U7Z3G4tDndbbOHv8Og=; b=ZYnPUYx85isNjBvMBE4wP6K+7P1pgj2FysIh4wVyjjRWd2YaYVZnLOMnWWNsu9UYuc NiTmFOu3QlySQl5a/k8xDeG9LNE+0SDAl/+Bu6bPgGOUoY17AdaKHpYb7G0Gvf8vIBVH dF+MB6upL0IO5wk46GNNN2n/9I22qugHh6dcMjUjoPutTp5VHMcbDonh3CeWiNEuaFby uHC4OOtDLRoBexJEKCAlh42qwJexzFxuKs3uMhVOmvXWODe1YcYk3pG+EQIc9RID9/zz Bt9vM8o+Cc7nWe+1tMwqBo2yyXboMHqAi+IDiVd5JIdVTJCs4g7SvEHBpw/khVmwUTwz xprw== X-Gm-Message-State: AGRZ1gKcEas/fKR4QX/0cMX1NSKZ6CUbe8wiGPw3OLXnT1Z1d9/YpCbT zGZszvcdlxk99mv4KpvlhCc= X-Google-Smtp-Source: AJdET5e0QqkEcd4GBDzyoq1FkDzf2Fs/LbjX69f2r2oP9Z/TgnX0nNGmzCcvRPQgpQdJdzfGWCNp1A== X-Received: by 2002:aa7:d4c9:: with SMTP id t9-v6mr741268edr.256.1540455853240; Thu, 25 Oct 2018 01:24:13 -0700 (PDT) Received: from tiehlicka.suse.cz (prg-ext-pat.suse.com. [213.151.95.130]) by smtp.gmail.com with ESMTPSA id m24-v6sm2628277edd.31.2018.10.25.01.24.12 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Thu, 25 Oct 2018 01:24:12 -0700 (PDT) From: Michal Hocko To: Cc: Tetsuo Handa , Roman Gushchin , David Rientjes , Andrew Morton , LKML , Michal Hocko Subject: [RFC PATCH v2 2/3] mm, oom: keep retrying the oom_reap operation as long as there is substantial memory left Date: Thu, 25 Oct 2018 10:24:02 +0200 Message-Id: <20181025082403.3806-3-mhocko@kernel.org> X-Mailer: git-send-email 2.19.1 In-Reply-To: <20181025082403.3806-1-mhocko@kernel.org> References: <20181025082403.3806-1-mhocko@kernel.org> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Michal Hocko oom_reaper is not able to reap all types of memory. E.g. mlocked mappings or page tables. In some cases this might be a lot of memory and we do rely on exit_mmap to release that memory. Yet we cannot rely on exit_mmap to set MMF_OOM_SKIP right now because there are several places when sleeping locks are taken. This patch adds a simple heuristic to check for the amount of memory the mm is sitting on after oom_reaper is done with it. If this is still a considerable amount of the original memory (more than 1/4 of the original badness) then simply keep retrying oom_reap_task_mm. The portion is quite arbitrary and a subject of future tuning based on real life usecases but it should catch some outliars when the vast portion of the memory is consumed by non-reapable memory. Changes since RFC - do not use hardcoded threshold for retry and rather use a portion of the original badness instead Signed-off-by: Michal Hocko --- mm/oom_kill.c | 32 ++++++++++++++++++++++++++------ 1 file changed, 26 insertions(+), 6 deletions(-) diff --git a/mm/oom_kill.c b/mm/oom_kill.c index b3b2c2bbd8ab..ab42717661dc 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -189,6 +189,16 @@ static bool is_dump_unreclaim_slabs(void) return (global_node_page_state(NR_SLAB_UNRECLAIMABLE) > nr_lru); } +/* + * Rough memory consumption of the given mm which should be theoretically freed + * when the mm is removed. + */ +static unsigned long oom_badness_pages(struct mm_struct *mm) +{ + return get_mm_rss(mm) + get_mm_counter(mm, MM_SWAPENTS) + + mm_pgtables_bytes(mm) / PAGE_SIZE; +} + /** * oom_badness - heuristic function to determine which candidate task to kill * @p: task struct of which task we should calculate @@ -230,8 +240,7 @@ unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *memcg, * The baseline for the badness score is the proportion of RAM that each * task's rss, pagetable and swap space use. */ - points = get_mm_rss(p->mm) + get_mm_counter(p->mm, MM_SWAPENTS) + - mm_pgtables_bytes(p->mm) / PAGE_SIZE; + points = oom_badness_pages(p->mm); task_unlock(p); /* Normalize to oom_score_adj units */ @@ -488,7 +497,7 @@ static DECLARE_WAIT_QUEUE_HEAD(oom_reaper_wait); static struct task_struct *oom_reaper_list; static DEFINE_SPINLOCK(oom_reaper_lock); -static bool __oom_reap_task_mm(struct mm_struct *mm) +static bool __oom_reap_task_mm(struct mm_struct *mm, unsigned long original_badness) { struct vm_area_struct *vma; bool ret = true; @@ -532,6 +541,16 @@ static bool __oom_reap_task_mm(struct mm_struct *mm) } } + /* + * If we still sit on a noticeable amount of memory even after successfully + * reaping the address space then keep retrying until exit_mmap makes some + * further progress. + * TODO: add a flag for a stage when the exit path doesn't block anymore + * and hand over MMF_OOM_SKIP handling there in that case + */ + if (oom_badness_pages(mm) > (original_badness >> 2)) + ret = false; + return ret; } @@ -541,7 +560,7 @@ static bool __oom_reap_task_mm(struct mm_struct *mm) * Returns true on success and false if none or part of the address space * has been reclaimed and the caller should retry later. */ -static bool oom_reap_task_mm(struct task_struct *tsk, struct mm_struct *mm) +static bool oom_reap_task_mm(struct task_struct *tsk, struct mm_struct *mm, unsigned long original_badness) { bool ret = true; @@ -564,7 +583,7 @@ static bool oom_reap_task_mm(struct task_struct *tsk, struct mm_struct *mm) trace_start_task_reaping(tsk->pid); /* failed to reap part of the address space. Try again later */ - ret = __oom_reap_task_mm(mm); + ret = __oom_reap_task_mm(mm, original_badness); if (!ret) goto out_finish; @@ -586,9 +605,10 @@ static void oom_reap_task(struct task_struct *tsk) { int attempts = 0; struct mm_struct *mm = tsk->signal->oom_mm; + unsigned long original_badness = oom_badness_pages(mm); /* Retry the down_read_trylock(mmap_sem) a few times */ - while (attempts++ < MAX_OOM_REAP_RETRIES && !oom_reap_task_mm(tsk, mm)) + while (attempts++ < MAX_OOM_REAP_RETRIES && !oom_reap_task_mm(tsk, mm, original_badness)) schedule_timeout_idle(HZ/10); if (attempts <= MAX_OOM_REAP_RETRIES || -- 2.19.1