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=-8.4 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,SPF_PASS, T_DKIMWL_WL_MED,URIBL_BLOCKED,USER_IN_DEF_DKIM_WL 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 4C4B7C433EF for ; Tue, 19 Jun 2018 05:13:50 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id D6C942083A for ; Tue, 19 Jun 2018 05:13:49 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="UQKgqUza" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org D6C942083A Authentication-Results: mail.kernel.org; dmarc=fail (p=reject dis=none) header.from=google.com 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 S937345AbeFSFNs (ORCPT ); Tue, 19 Jun 2018 01:13:48 -0400 Received: from mail-pf0-f195.google.com ([209.85.192.195]:47099 "EHLO mail-pf0-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S937238AbeFSFNo (ORCPT ); Tue, 19 Jun 2018 01:13:44 -0400 Received: by mail-pf0-f195.google.com with SMTP id q1-v6so9310135pff.13 for ; Mon, 18 Jun 2018 22:13:44 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=MqmCD10RwYEL9UtGfpqsQpvbtnr+99X+2SVsogA+HbA=; b=UQKgqUzav1CUNvMMvmaU3NyeD13FZKaVcoxdmKfZgZV4KVaxPcnD3E0svKMRxThx1P rkATmYdPRDTipCukiN3Ma1jVohMWcsgJBi7Fra96GTPgYisP8dPWhjWxcB44VQiVXPkn /1UCOVXAVtTnU3JKOavvZ2FwrMFR0Jc3T9KsOtsX7D75zqgOSsC1vR/q/HfzlZTPdiWH QEcVG9660l/iuX7KmPEOLFhDfFRmmRyjknyNx5p7o6PqukAIebGZBw9DtMOGtUYUeYBm 9HdN3eqcfzvg2e3ihjxO/NzXYGy3msuK8Oqh44ZogfPDOwQXcYKaM7SFv3VF0x9rMWTB U4/w== 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; bh=MqmCD10RwYEL9UtGfpqsQpvbtnr+99X+2SVsogA+HbA=; b=LCRolZS/y2BpdfKvk6dRnmYpWjiQaF9c4ypbVEdWaOc8qTsLknGAlvloWb1itp4qyC kLatwVJB6wzn0L1kD0ipjVHynh/pC9687Rh6BqdNnG7L1XFt8Wg4Vh5sMMrFBXaP1GiK xZypt8KfLB3y5/+7JRaoaXbWoCgYcMKSdpSb26mdx3tI+DDZXhbz8UR0Ux4h49GaSpYo gZU42QAU05VhK2BiHjoncITkafE8zg4O2DWbbSNr6UybqtBS1YPzd7a4MiUBaBpbJxig q0VMqHLBuuaxsjPc8ASnhucvs2C72LtDftN5tTp7IhCkZTmBsn+JQz3PYaGYSSgQnKD4 bG2w== X-Gm-Message-State: APt69E2oA7KVlIT62mDihOAfJwq4UOSa4qJCJ1dw1Akziw7OST6E5V4/ RyY4Xxoo8gXXXgwfpMwd9XjyCQ== X-Google-Smtp-Source: ADUXVKJNv+l8nqxG3So+pCP7h0SGLU6MaeeFdgFGKnXn1Cf3Anqmv31Ma6gtTlCKplnfpHLDVayMmg== X-Received: by 2002:a65:55c6:: with SMTP id k6-v6mr13406087pgs.6.1529385223576; Mon, 18 Jun 2018 22:13:43 -0700 (PDT) Received: from shakeelb.mtv.corp.google.com ([2620:15c:2cb:201:3a5f:3a4f:fa44:6b63]) by smtp.gmail.com with ESMTPSA id a23-v6sm28390504pfj.117.2018.06.18.22.13.41 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Mon, 18 Jun 2018 22:13:42 -0700 (PDT) From: Shakeel Butt To: Andrew Morton Cc: Michal Hocko , Johannes Weiner , Vladimir Davydov , Jan Kara , Greg Thelen , linux-kernel@vger.kernel.org, cgroups@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-mm@kvack.org, Shakeel Butt , Jan Kara , Amir Goldstein , Christoph Lameter , Pekka Enberg , David Rientjes , Joonsoo Kim , Mel Gorman , Vlastimil Babka , Alexander Viro Subject: [PATCH 1/3] mm: memcg: remote memcg charging for kmem allocations Date: Mon, 18 Jun 2018 22:13:25 -0700 Message-Id: <20180619051327.149716-2-shakeelb@google.com> X-Mailer: git-send-email 2.18.0.rc1.244.gcf134e6275-goog In-Reply-To: <20180619051327.149716-1-shakeelb@google.com> References: <20180619051327.149716-1-shakeelb@google.com> Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Introduce the memcg variant for kmalloc[_node] and kmem_cache_alloc[_node]. For kmem_cache_alloc, the kernel switches the root kmem cache with the memcg specific kmem cache for __GFP_ACCOUNT allocations to charge those allocations to the memcg. However, the memcg to charge is extracted from the current task_struct. This patch introduces the variant of kmem cache allocation functions where the memcg can be provided explicitly by the caller instead of deducing the memcg from the current task. The kmalloc allocations are underlying served using the kmem caches unless the size of the allocation request is larger than KMALLOC_MAX_CACHE_SIZE, in which case, the kmem caches are bypassed and the request is routed directly to page allocator. So, for __GFP_ACCOUNT kmalloc allocations, the memcg of current task is charged. This patch introduces memcg variant of kmalloc functions to allow callers to provide memcg for charging. These functions are useful for use-cases where the allocations should be charged to the memcg different from the memcg of the caller. One such concrete use-case is the allocations for fsnotify event objects where the objects should be charged to the listener instead of the producer. One requirement to call these functions is that the caller must have the reference to the memcg provided to these functions. If reference acquition on the given memcg is failed (it can fail if memcg is offline) then the current's memcg is tried. These functions implicitly assumes that the caller wants a __GFP_ACCOUNT allocation. This patch also introduces scope API for targeted memcg charging. All the __GFP_ACCOUNT allocations between memalloc_memcg_save(target_memcg) and memalloc_memcg_restore(old_memcg) will be charged to target_memcg. Traditionally kmem charging is skipped for allocations by kthreads and allocations during interrupts. The reason is that the current's memcg might not be the right owner for such allocations. However targeted memcg charging does not have such limitation and can work even for allocations by kthreads and for allocations during interrupts. For now due to lack of actual use-case, targeted memcg charging for such allocations is not added. Though this might change in future. Signed-off-by: Shakeel Butt Cc: Michal Hocko Cc: Jan Kara Cc: Amir Goldstein Cc: Christoph Lameter Cc: Pekka Enberg Cc: David Rientjes Cc: Joonsoo Kim Cc: Greg Thelen Cc: Johannes Weiner Cc: Vladimir Davydov Cc: Mel Gorman Cc: Vlastimil Babka Cc: Alexander Viro Cc: Andrew Morton --- Changelog sinve v5: - Added more explanation in commit message. - Added handling of NULL memcg for targeted memcg allocation functions. Changelog since v4: - Removed branch from hot path of memory charging. Changelog since v3: - Added node variant of directed kmem allocation functions. Changelog since v2: - Merge the kmalloc_memcg patch into this patch. - Instead of plumbing memcg throughout, use field in task_struct to pass the target_memcg. Changelog since v1: - Fixed build for SLOB include/linux/sched.h | 3 ++ include/linux/sched/mm.h | 24 ++++++++++++ include/linux/slab.h | 83 ++++++++++++++++++++++++++++++++++++++++ kernel/fork.c | 3 ++ mm/memcontrol.c | 18 ++++++++- 5 files changed, 129 insertions(+), 2 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 87bf02d93a27..cbd0def60fd4 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1149,6 +1149,9 @@ struct task_struct { /* Number of pages to reclaim on returning to userland: */ unsigned int memcg_nr_pages_over_high; + + /* Used by memcontrol for targeted memcg charge: */ + struct mem_cgroup *target_memcg; #endif #ifdef CONFIG_UPROBES diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index 44d356f5e47c..2ffb194f3f32 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -248,6 +248,30 @@ static inline void memalloc_noreclaim_restore(unsigned int flags) current->flags = (current->flags & ~PF_MEMALLOC) | flags; } +#ifdef CONFIG_MEMCG +static inline struct mem_cgroup *memalloc_memcg_save(struct mem_cgroup *memcg) +{ + struct mem_cgroup *old_memcg = current->target_memcg; + + current->target_memcg = memcg; + return old_memcg; +} + +static inline void memalloc_memcg_restore(struct mem_cgroup *memcg) +{ + current->target_memcg = memcg; +} +#else +static inline struct mem_cgroup *memalloc_memcg_save(struct mem_cgroup *memcg) +{ + return NULL; +} + +static inline void memalloc_memcg_restore(struct mem_cgroup *memcg) +{ +} +#endif /* CONFIG_MEMCG */ + #ifdef CONFIG_MEMBARRIER enum { MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0), diff --git a/include/linux/slab.h b/include/linux/slab.h index 14e3fe4bd6a1..2f6319fa0d3d 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -16,6 +16,7 @@ #include #include #include +#include /* @@ -375,6 +376,27 @@ static __always_inline void kfree_bulk(size_t size, void **p) kmem_cache_free_bulk(NULL, size, p); } +/* + * Calling kmem_cache_alloc_memcg implicitly assumes that the caller wants + * a __GFP_ACCOUNT allocation. However if memcg is NULL then + * kmem_cache_alloc_memcg is same as kmem_cache_alloc. + */ +static __always_inline void *kmem_cache_alloc_memcg(struct kmem_cache *cachep, + gfp_t flags, + struct mem_cgroup *memcg) +{ + struct mem_cgroup *old_memcg; + void *ptr; + + if (!memcg) + return kmem_cache_alloc(cachep, flags); + + old_memcg = memalloc_memcg_save(memcg); + ptr = kmem_cache_alloc(cachep, flags | __GFP_ACCOUNT); + memalloc_memcg_restore(old_memcg); + return ptr; +} + #ifdef CONFIG_NUMA void *__kmalloc_node(size_t size, gfp_t flags, int node) __assume_kmalloc_alignment __malloc; void *kmem_cache_alloc_node(struct kmem_cache *, gfp_t flags, int node) __assume_slab_alignment __malloc; @@ -390,6 +412,27 @@ static __always_inline void *kmem_cache_alloc_node(struct kmem_cache *s, gfp_t f } #endif +/* + * Calling kmem_cache_alloc_node_memcg implicitly assumes that the caller + * wants a __GFP_ACCOUNT allocation. However if memcg is NULL then + * kmem_cache_alloc_node_memcg is same as kmem_cache_alloc_node. + */ +static __always_inline void * +kmem_cache_alloc_node_memcg(struct kmem_cache *cachep, gfp_t flags, int node, + struct mem_cgroup *memcg) +{ + struct mem_cgroup *old_memcg; + void *ptr; + + if (!memcg) + return kmem_cache_alloc_node(cachep, flags, node); + + old_memcg = memalloc_memcg_save(memcg); + ptr = kmem_cache_alloc_node(cachep, flags | __GFP_ACCOUNT, node); + memalloc_memcg_restore(old_memcg); + return ptr; +} + #ifdef CONFIG_TRACING extern void *kmem_cache_alloc_trace(struct kmem_cache *, gfp_t, size_t) __assume_slab_alignment __malloc; @@ -518,6 +561,26 @@ static __always_inline void *kmalloc(size_t size, gfp_t flags) return __kmalloc(size, flags); } +/* + * Calling kmalloc_memcg implicitly assumes that the caller wants a + * __GFP_ACCOUNT allocation. However if memcg is NULL then kmalloc_memcg + * is same as kmalloc. + */ +static __always_inline void *kmalloc_memcg(size_t size, gfp_t flags, + struct mem_cgroup *memcg) +{ + struct mem_cgroup *old_memcg; + void *ptr; + + if (!memcg) + return kmalloc(size, flags); + + old_memcg = memalloc_memcg_save(memcg); + ptr = kmalloc(size, flags | __GFP_ACCOUNT); + memalloc_memcg_restore(old_memcg); + return ptr; +} + /* * Determine size used for the nth kmalloc cache. * return size or 0 if a kmalloc cache for that @@ -555,6 +618,26 @@ static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) return __kmalloc_node(size, flags, node); } +/* + * Calling kmalloc_node_memcg implicitly assumes that the caller wants a + * __GFP_ACCOUNT allocation. However if memcg is NULL then kmalloc_node_memcg + * is same as kmalloc_node. + */ +static __always_inline void * +kmalloc_node_memcg(size_t size, gfp_t flags, int node, struct mem_cgroup *memcg) +{ + struct mem_cgroup *old_memcg; + void *ptr; + + if (!memcg) + return kmalloc_node(size, flags, node); + + old_memcg = memalloc_memcg_save(memcg); + ptr = kmalloc_node(size, flags | __GFP_ACCOUNT, node); + memalloc_memcg_restore(old_memcg); + return ptr; +} + struct memcg_cache_array { struct rcu_head rcu; struct kmem_cache *entries[0]; diff --git a/kernel/fork.c b/kernel/fork.c index a64d0a19f174..5bf300015790 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -843,6 +843,9 @@ static struct task_struct *dup_task_struct(struct task_struct *orig, int node) tsk->fail_nth = 0; #endif +#ifdef CONFIG_MEMCG + tsk->target_memcg = NULL; +#endif return tsk; free_stack: diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 02e77c88967a..08bfb8c2411b 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -701,6 +701,20 @@ static struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm) return memcg; } +static __always_inline struct mem_cgroup *get_mem_cgroup( + struct mem_cgroup *memcg, struct mm_struct *mm) +{ + if (unlikely(memcg)) { + rcu_read_lock(); + if (css_tryget_online(&memcg->css)) { + rcu_read_unlock(); + return memcg; + } + rcu_read_unlock(); + } + return get_mem_cgroup_from_mm(mm); +} + /** * mem_cgroup_iter - iterate over memory cgroup hierarchy * @root: hierarchy root @@ -2260,7 +2274,7 @@ struct kmem_cache *memcg_kmem_get_cache(struct kmem_cache *cachep) if (current->memcg_kmem_skip_account) return cachep; - memcg = get_mem_cgroup_from_mm(current->mm); + memcg = get_mem_cgroup(current->target_memcg, current->mm); kmemcg_id = READ_ONCE(memcg->kmemcg_id); if (kmemcg_id < 0) goto out; @@ -2344,7 +2358,7 @@ int memcg_kmem_charge(struct page *page, gfp_t gfp, int order) if (memcg_kmem_bypass()) return 0; - memcg = get_mem_cgroup_from_mm(current->mm); + memcg = get_mem_cgroup(current->target_memcg, current->mm); if (!mem_cgroup_is_root(memcg)) { ret = memcg_kmem_charge_memcg(page, gfp, order, memcg); if (!ret) -- 2.18.0.rc1.244.gcf134e6275-goog