From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-oa1-f48.google.com (mail-oa1-f48.google.com [209.85.160.48]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 405E548095A for ; Tue, 5 May 2026 15:39:48 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.160.48 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777995592; cv=none; b=EBxY7AIewfL2NKE3jTdXhvMC4CO44uuTRdCgVb4Bc6LXiUYCcnFbQyTJ6wIdcTs9yrk2u8h7vXhSkhGyLnxvadG7nkc0KSL/HmxJC6NQhBP6AXJtvYE+4R6PFuUB3RCUFCU8wrNib9/hfBQOAcmBTBh6RTgvpBOa+MaWrxmGIn8= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777995592; c=relaxed/simple; bh=Hd+gorxnpksj2SeHzQNvJk6R01VLS+FyxGb3FAaP1tw=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=NDfFjm3CSr4CxuQufGhGsWxPOELdXVzVhS1rAV/WFlTRcGS/+cD/byM97kk3mdp6RihQf1b1+DFq7J4nuJADEnOmF9KwN5ELHPnGy23GVppU9LqH5iyf48EraT3ea4/IhMDcCIVbsByPXS+qSlucbdUMaJtlZTxTf5WC8z9iF4o= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=qGmeBTPs; arc=none smtp.client-ip=209.85.160.48 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="qGmeBTPs" Received: by mail-oa1-f48.google.com with SMTP id 586e51a60fabf-40974bf7781so7233fac.0 for ; Tue, 05 May 2026 08:39:47 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1777995586; x=1778600386; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=5/yyOr/4pj99tbSJw4a+KNInaif0dFyRreHuOMw4Fwo=; b=qGmeBTPs7/1NfMIWORwmMkkfNLkaYdkfgi6w4xCxZgwP10oiKf4+DsICDzS4n+nxHP 6s/Csu7XJ9Lu5N18W0efNz9eA6DOjisIq2GiWAsP1eXMFXtAV3BuZ5yj5R0cSsiKDemC D5TSLPUNooFHlapFQLVFvWqKuvbR2rtbW8cvTKtj0Yh15pCI8epliliq9HEInCURGe26 /hw5Lhajy7x0ybxBBURNMYoiNYuhFc/GATQ/qeRT4oVNYA2Y6I/GreNkemzRP+YlV3XW +t4JdeGLDc+6awDqSXlWSBZzKvrJoNReYHpFFulH6ravER8gKrFS2pSVNcfkVgjkJA1n y0sQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1777995586; x=1778600386; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-gg:x-gm-message-state:from :to:cc:subject:date:message-id:reply-to; bh=5/yyOr/4pj99tbSJw4a+KNInaif0dFyRreHuOMw4Fwo=; b=a21B/M4U9/FojZVm2k2ofF0/2zscvCxpunmS8G8UGY3jNit2GufE9VPuflqItJ6M6q FYiRHK0r9qthd6GZDAf019IDj6Ehflh6TONmeHdDkO96AEm2troY+n4En8CRLQ1o22XF egFoPsXHcaGMXc7YC0GO1atPh1oYmmLjfoIwyx8u41P5KA2zNc76DIA/q2grhNkPla4M /Ft7MCqXFT6yuezH1HHudEe6olvtrEhplNhzNdbmO2rVvsSp75Kxa5ttL6Tod8Kqm6mh 4aliDp6GEpcdl+Iss4/gqLfWe8re7YAojwqxXcuHh451HsIr2bSjfJDNs4Etfa84XiI6 KAqQ== X-Forwarded-Encrypted: i=1; AFNElJ/Ic7QouPCX+VynN3UMYeUexd4gwDQChcnsmcHtEGzQjPm+yt5qBW971Co3+4owpGNVbTtN4eI16xB04hA=@vger.kernel.org X-Gm-Message-State: AOJu0YwNrv7FDsjzMvg9GBM+2mG5GkW8ih++oWZB7cUMK/FYBxlIOAC2 VZQTsAWmAyFYdF+Sik2/PKnZ2zgOhVcr0787CSLkqpa32KxI/Dn7oTFJ X-Gm-Gg: AeBDievKbX49MaM01SOoqbGCAZPvVF57c66XdnOHd5v+S1v2cfSGNc4LufesdCMAx6k Q9mrimizY2jJvHomzBqz4jcnx3+wn8rALmaA+6tYr9f71Zw/cXg/9nHVTqH1Z1csTGuP4Lchz7H r1ZltZnGlmaIvXNU/h8xiVzq+yAi7g3v3ZVejGfiP9B+k2p8q9kl9Kf4QcsREmSmovf5prwPovQ atCma+6nFl3LMSMMFnNl6MXAR8FMnEHveMtKQiYqCgEOzmCgUNteruyOHVVJAw42WyRkces6SXb p7KT/ttzzsbsiUSmnOR3GYxaVDBnaxMKgUso83OAhR2UeYpnP7UdnqGoetwy0cTMgwe1AYDHxCy RfhZBLR/RBWnnkApTnmncXMlAkHO7oVJ4RuMQDnVBQCFkD2Gz8Yii1mAhIVhdMYO2ji/fCToJ37 h3DKnoQPOIqxInuu/7gJXDW28+yK2J4m2JivaBB4+JR0wfkUuYQlnw X-Received: by 2002:a05:6808:e64b:b0:467:da0d:537e with SMTP id 5614622812f47-47d260dcc6cmr1641681b6e.7.1777995585821; Tue, 05 May 2026 08:39:45 -0700 (PDT) Received: from localhost ([2a03:2880:10ff::]) by smtp.gmail.com with ESMTPSA id 5614622812f47-47c763b33c1sm8866816b6e.1.2026.05.05.08.39.43 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 05 May 2026 08:39:44 -0700 (PDT) From: Nhat Pham To: kasong@tencent.com Cc: Liam.Howlett@oracle.com, akpm@linux-foundation.org, apopple@nvidia.com, axelrasmussen@google.com, baohua@kernel.org, baolin.wang@linux.alibaba.com, bhe@redhat.com, byungchul@sk.com, cgroups@vger.kernel.org, chengming.zhou@linux.dev, chrisl@kernel.org, corbet@lwn.net, david@kernel.org, dev.jain@arm.com, gourry@gourry.net, hannes@cmpxchg.org, hughd@google.com, jannh@google.com, joshua.hahnjy@gmail.com, lance.yang@linux.dev, lenb@kernel.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, linux-pm@vger.kernel.org, lorenzo.stoakes@oracle.com, matthew.brost@intel.com, mhocko@suse.com, muchun.song@linux.dev, npache@redhat.com, nphamcs@gmail.com, pavel@kernel.org, peterx@redhat.com, peterz@infradead.org, pfalcato@suse.de, rafael@kernel.org, rakie.kim@sk.com, roman.gushchin@linux.dev, rppt@kernel.org, ryan.roberts@arm.com, shakeel.butt@linux.dev, shikemeng@huaweicloud.com, surenb@google.com, tglx@kernel.org, vbabka@suse.cz, weixugc@google.com, ying.huang@linux.alibaba.com, yosry.ahmed@linux.dev, yuanchu@google.com, zhengqi.arch@bytedance.com, ziy@nvidia.com, kernel-team@meta.com, riel@surriel.com, haowenchao22@gmail.com Subject: [PATCH v6 21/22] vswap: batch contiguous vswap free calls Date: Tue, 5 May 2026 08:38:50 -0700 Message-ID: <20260505153854.1612033-22-nphamcs@gmail.com> X-Mailer: git-send-email 2.52.0 In-Reply-To: <20260505153854.1612033-1-nphamcs@gmail.com> References: <20260505153854.1612033-1-nphamcs@gmail.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit In vswap_free(), we release and reacquire the cluster lock for every single entry, even for non-disk-swap backends where the lock drop is unnecessary. Batch consecutive free operations to avoid this overhead. Signed-off-by: Nhat Pham --- mm/vswap.c | 97 ++++++++++++++++++++++++++++++++++-------------------- 1 file changed, 61 insertions(+), 36 deletions(-) diff --git a/mm/vswap.c b/mm/vswap.c index 3f86bbb3a5ea..f07e6d9ec1df 100644 --- a/mm/vswap.c +++ b/mm/vswap.c @@ -529,18 +529,18 @@ static void vswap_cluster_free(struct vswap_cluster *cluster) call_rcu(&cluster->rcu, vswap_cluster_free_rcu); } -static inline void release_vswap_slot(struct vswap_cluster *cluster, - unsigned long index) +static inline void release_vswap_slot_nr(struct vswap_cluster *cluster, + unsigned long index, int nr) { unsigned long slot_index = VSWAP_IDX_WITHIN_CLUSTER_VAL(index); lockdep_assert_held(&cluster->lock); - cluster->count--; + cluster->count -= nr; - bitmap_clear(cluster->bitmap, slot_index, 1); + bitmap_clear(cluster->bitmap, slot_index, nr); /* we only free uncached empty clusters */ - if (refcount_dec_and_test(&cluster->refcnt)) + if (refcount_sub_and_test(nr, &cluster->refcnt)) vswap_cluster_free(cluster); else if (cluster->full && cluster_is_alloc_candidate(cluster)) { cluster->full = false; @@ -553,7 +553,7 @@ static inline void release_vswap_slot(struct vswap_cluster *cluster, } } - atomic_dec(&vswap_used); + atomic_sub(nr, &vswap_used); } /* @@ -585,7 +585,7 @@ static unsigned short swp_desc_memcgid(struct swp_desc *desc); * * 1. Callers ensure no concurrent modification of the swap entry's internal * state can occur. This is guaranteed by one of the following: - * - For vswap_free() callers: the swap entry's refcnt (swap count and + * - For vswap_free_nr() callers: the swap entry's refcnt (swap count and * swapcache pin) is down to 0. * - For vswap_store_folio(), swap_zeromap_folio_set(), and zswap_entry_store() * callers: the folio is locked and in the swap cache. @@ -706,26 +706,17 @@ static void __vswap_swap_cgroup_clear(struct vswap_cluster *cluster, /* * Entered with the cluster locked. The cluster lock is held throughout. - * - * This is safe, because: - * - * 1. The swap entry to be freed has refcnt (swap count and swapcache pin) - * down to 0, so no one can change its internal state. - * - * 2. The swap entry to be freed still holds a refcnt to the cluster, keeping - * the cluster itself valid. - * - * 3. swap_slot_free_nr() takes the physical swap cluster lock (ci->lock), - * but the only vswap function called under ci->lock is vswap_rmap_set(), - * which uses atomic ops and does not take cluster->lock. So there is no - * ABBA deadlock risk. */ -static void vswap_free(struct vswap_cluster *cluster, struct swp_desc *desc, - swp_entry_t entry) +static void vswap_free_nr(struct vswap_cluster *cluster, swp_entry_t entry, + int nr) { - unsigned short id = swp_desc_memcgid(desc); + struct swp_desc *desc = __vswap_iter(cluster, entry.val); + unsigned short id; struct mem_cgroup *memcg; + VM_WARN_ON(!desc); + id = swp_desc_memcgid(desc); + /* * The swap_cgroup id reference taken at swapout time pins this * memcg until swap_cgroup_clear() runs below, so we can resolve @@ -733,11 +724,11 @@ static void vswap_free(struct vswap_cluster *cluster, struct swp_desc *desc, */ memcg = id ? mem_cgroup_from_id(id) : NULL; - release_backing(cluster, entry, 1, memcg); - __vswap_swap_cgroup_clear(cluster, entry, 1, memcg); + release_backing(cluster, entry, nr, memcg); + __vswap_swap_cgroup_clear(cluster, entry, nr, memcg); - /* erase forward mapping and release the virtual slot for reallocation */ - release_vswap_slot(cluster, entry.val); + /* erase forward mapping and release the virtual slots for reallocation */ + release_vswap_slot_nr(cluster, entry.val, nr); } @@ -908,10 +899,18 @@ static bool vswap_free_nr_any_cache_only(swp_entry_t entry, int nr) struct vswap_cluster *cluster = NULL; struct swp_desc *desc; bool ret = false; - int i; + swp_entry_t free_start; + unsigned short batch_memcgid = 0; + int i, free_nr = 0; + free_start.val = 0; rcu_read_lock(); for (i = 0; i < nr; i++) { + /* flush pending free batch at cluster boundary */ + if (free_nr && !VSWAP_IDX_WITHIN_CLUSTER_VAL(entry.val)) { + vswap_free_nr(cluster, free_start, free_nr); + free_nr = 0; + } desc = vswap_iter(&cluster, entry.val); VM_WARN_ON(!desc); ret |= (desc->swap_count == 1 && desc->in_swapcache); @@ -919,18 +918,34 @@ static bool vswap_free_nr_any_cache_only(swp_entry_t entry, int nr) if (!desc->swap_count && !desc->in_swapcache) { if (xa_is_value(desc->shadow)) desc->shadow = NULL; - vswap_free(cluster, desc, entry); - } else if (!desc->swap_count && desc->in_swapcache && - desc->type == VSWAP_SWAPFILE) { + /* flush at cgroup boundary */ + if (free_nr && + swp_desc_memcgid(desc) != batch_memcgid) { + vswap_free_nr(cluster, free_start, free_nr); + free_nr = 0; + } + if (!free_nr) + batch_memcgid = swp_desc_memcgid(desc); + if (!free_nr++) + free_start = entry; + } else { + if (free_nr) { + vswap_free_nr(cluster, free_start, free_nr); + free_nr = 0; + } /* * swap_count just dropped to 0, but still in swap * cache. If backed by a physical swap slot, mark it * so the physical swap allocator can check cheaply. */ - swap_rmap_mark_cache_only(desc->slot); + if (!desc->swap_count && desc->in_swapcache && + desc->type == VSWAP_SWAPFILE) + swap_rmap_mark_cache_only(desc->slot); } entry.val++; } + if (free_nr) + vswap_free_nr(cluster, free_start, free_nr); if (cluster) spin_unlock(&cluster->lock); rcu_read_unlock(); @@ -1032,8 +1047,9 @@ bool folio_free_swap(struct folio *folio) VM_WARN_ON_FOLIO(!desc || desc->swap_cache != folio, folio); desc->swap_cache = NULL; desc->in_swapcache = false; - vswap_free(cluster, desc, (swp_entry_t){ entry.val + i }); } + + vswap_free_nr(cluster, entry, nr); spin_unlock_irq(&cluster->lock); rcu_read_unlock(); @@ -1095,14 +1111,23 @@ static void __swapcache_clear(struct vswap_cluster *cluster, swp_entry_t entry, int nr) { struct swp_desc *desc; - int i; + swp_entry_t free_start; + int i, free_nr = 0; + free_start = entry; for (i = 0; i < nr; i++) { desc = __vswap_iter(cluster, entry.val + i); desc->in_swapcache = false; - if (!desc->swap_count) - vswap_free(cluster, desc, (swp_entry_t){ entry.val + i }); + if (!desc->swap_count) { + if (!free_nr++) + free_start.val = entry.val + i; + } else if (free_nr) { + vswap_free_nr(cluster, free_start, free_nr); + free_nr = 0; + } } + if (free_nr) + vswap_free_nr(cluster, free_start, free_nr); } void swapcache_clear(swp_entry_t entry, int nr) -- 2.52.0