From nobody Sat Feb 7 21:24:11 2026 Received: from mail-dl1-f66.google.com (mail-dl1-f66.google.com [74.125.82.66]) (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 BF81D335559 for ; Fri, 23 Jan 2026 06:42:48 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=74.125.82.66 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1769150570; cv=none; b=I8+DNmRk4cG7TvE5Skr4pT9qlSJZTP6mtuGUphYEhGz0m0tU1bq7MEckg0ff8Kz1HAxhc9Hoq/3lJNjBCW32waJhNvxvrgzkUKdDfCdg4GGyyHCwgidFlDWcrWrzhrGTbLOJoHHPV7a39ZPsrKQRvH2eNODQZf8gXubsOZehsYg= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1769150570; c=relaxed/simple; bh=anSJldvMGDdKtmD4mqjHGwl+/ENgn5/LFgwWaAXUyCU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=gnu0BH7TffQABhusjHDt0rsu8p8ehrwhcgKyjoe+MtiGlCSXyRFmF1rMbEDciy5JWYFZN9FxCjWEzuIg1OsWsDgAiOpXgUkz71W8SZlBAzY8ADlpcJzdPM/W/OgipkLSZ3CoVf8QtBQ7mmmaByegeHgbJYajbACnxrgK0O89X+E= 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=ayVwVzzD; arc=none smtp.client-ip=74.125.82.66 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="ayVwVzzD" Received: by mail-dl1-f66.google.com with SMTP id a92af1059eb24-121a0bcd376so6535138c88.0 for ; Thu, 22 Jan 2026 22:42:48 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1769150568; x=1769755368; 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=RurLsspuwixyatIbkQyVE9ShmUOAQwG6MPVdwLrORJU=; b=ayVwVzzD7qVwDmhoKZagt5cXSF8ZVt2WMlWAYJbRXRoCZ3nyTkfrGsObSR7ivipIKp GiJs5KnV5js20Z2GNJGu10/csDzIsLORzLnHom/zTJgyMcwnMB7d5K3iwsfmz85BDtua EeMHe/HK3IdHSkI62LfIsuo/gGuZ7r8FYLtkzyDYZzcKi6ueLE2GadS86orTJnh4H3XA 9R+iE7QwlCkGy2hHyAKOppwWf/G+skuYd/KFjVYM+NsxiL+JCJgVCgtw79YL571xokSN Oy6LUbDXmDnVp/jT+wRT7jxMijcIg+bXPlzIerKYoj3eoiz9UfgEt3d4C/uMA0jHe/X0 TgoA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1769150568; x=1769755368; 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=RurLsspuwixyatIbkQyVE9ShmUOAQwG6MPVdwLrORJU=; b=sHtrnAaEVxYmVHiEXcYtgAuJrYakIPxU9EwuOVxvGiMQRt/qKLUvbbsOit9QtINXEQ HjB9ioPKRs5suip/LPYIDDfTllfe+PpvUrcWp65i7PvQ+6lIRMrJLSC9/y4zlUEg8xFR 3XrXp6HmtoJGXvF0deQae75Jz+URgylAIdW+L0m8iPHa6mOUtHJju+eUAfdhiOAIpybU y5VeCEVlH8sNYzNfeEwtOzASxRWjthLhn6f/nB+IkV3S/sUGkbbauT+vFLWtUsip5qBv gMvSg2zKxDqmDhUIq1hhj8iq3SbPI4ACxI+Gxn/GVf2A+vLAVIMVCeNbal/HsLgi9fjP fNow== X-Forwarded-Encrypted: i=1; AJvYcCWIlfZvvHlQ2FFZdO4+0BcMRjsxtu04DuSlIxigU/ojsieVkOx57pDzSZXzPdIclDu685I9/BvBb4EPGIk=@vger.kernel.org X-Gm-Message-State: AOJu0YyjnEP7nmEc5l5mtRj/VpnJ30MXKCNPZT7/5VTcswc4bqaZAmP9 VzNg7vlgMqE8DMI+GAqI/lquY1PPiT1wZ1U4g5hwWRV9H9dSQAMDTqPJ X-Gm-Gg: AZuq6aJs8JbkbGPK/YvrU95cpdLwTXZtn3Sajkk0AI5wdYlSMQ2LRB22pxlpqTKKKcq J5FEenmRm9tef6g3O+MQfJL2TiHPRi1wb+UEuC2W7vOdS+DuJR+Iko7J2Q092PTY1pjoxt6PfPi dtTGUDYZ/j94JMIDCG+22WpnPQCK5Pj1r7/6UMaErQTbpYm2uX2+ae43ltU3PwUn1ADQQOrh4iq uKp0iI/2UYQLVeQxe21jdU/fi5MNdGKN1BYwkHZuT3634UWq8X7Jcrtv4WeEbPkLKysFGis0iMe nclBXoYHvSD/IBM3A7B1GxamNYCr2rdPVc9bHHSfvkQWwIYnu0c63Ll3SlI+brPf7A2zV3O3Qmf l2TAELh8V8iWoSZbL7ormvhOjlXBfPmPX12yxPbIs6EdE2JSBwrsH2iTPPonCx9HjmvR7daTxkh ANPII= X-Received: by 2002:a05:7022:102:b0:119:e56b:c74b with SMTP id a92af1059eb24-1247dbb463emr967983c88.16.1769150567825; Thu, 22 Jan 2026 22:42:47 -0800 (PST) Received: from debian ([74.48.213.230]) by smtp.gmail.com with ESMTPSA id a92af1059eb24-1247dba0135sm3130527c88.17.2026.01.22.22.42.40 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 22 Jan 2026 22:42:47 -0800 (PST) From: Qiliang Yuan To: vbabka@suse.cz Cc: akpm@linux-foundation.org, david@kernel.org, edumazet@google.com, hannes@cmpxchg.org, jackmanb@google.com, jis1@chinatelecom.cn, lance.yang@linux.dev, linux-kernel@vger.kernel.org, linux-mm@kvack.org, liyi1@chinatelecom.cn, mhocko@suse.com, netdev@vger.kernel.org, realwujing@gmail.com, rppt@kernel.org, sunshx@chinatelecom.cn, surenb@google.com, wangh13@chinatelecom.cn, weixugc@google.com, willy@infradead.org, yuanql9@chinatelecom.cn, zhangjn11@chinatelecom.cn, zhangzq20@chinatelecom.cn, ziy@nvidia.com Subject: [PATCH v7] mm/page_alloc: boost watermarks on atomic allocation failure Date: Fri, 23 Jan 2026 01:42:30 -0500 Message-ID: <20260123064231.250767-1-realwujing@gmail.com> X-Mailer: git-send-email 2.51.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Atomic allocations (GFP_ATOMIC) are prone to failure under heavy memory pressure as they cannot enter direct reclaim. This patch introduces a watermark boost mechanism to mitigate this issue. When a GFP_ATOMIC request enters the slowpath, the preferred zone's watermark_boost is increased under zone->lock protection. This triggers kswapd to proactively reclaim memory, creating a safety buffer for future atomic allocations. A 1-second debounce timer prevents excessive boosts during traffic bursts. This approach reuses existing watermark_boost infrastructure with minimal overhead and proper locking to ensure thread safety. Allocation failure logs: [38535644.718700] node 0: slabs: 1031, objs: 43328, free: 0 [38535644.725059] node 1: slabs: 339, objs: 17616, free: 317 [38535645.428345] SLUB: Unable to allocate memory on node -1, gfp=3D0x48002= 0(GFP_ATOMIC) [38535645.436888] cache: skbuff_head_cache, object size: 232, buffer size: = 256, default order: 2, min order: 0 [38535645.447664] node 0: slabs: 940, objs: 40864, free: 144 [38535645.454026] node 1: slabs: 322, objs: 19168, free: 383 [38535645.556122] SLUB: Unable to allocate memory on node -1, gfp=3D0x48002= 0(GFP_ATOMIC) [38535645.564576] cache: skbuff_head_cache, object size: 232, buffer size: = 256, default order: 2, min order: 0 [38535649.655523] warn_alloc: 59 callbacks suppressed [38535649.655527] swapper/100: page allocation failure: order:0, mode:0x480= 020(GFP_ATOMIC), nodemask=3D(null) [38535649.671692] swapper/100 cpuset=3D/ mems_allowed=3D0-1 Signed-off-by: Qiliang Yuan Signed-off-by: Qiliang Yuan --- v7: - Use local variable for boost_amount to improve code readability - Add zone->lock protection in boost_zones_for_atomic() - Add lockdep assertion in boost_watermark() to prevent locking mistakes - Remove redundant boost call at fail label due to 1-second debounce v6: - Replace magic number ">> 10" with ATOMIC_BOOST_SCALE_SHIFT define - Add documentation explaining 0.1% zone size boost rationale v5: - Simplify to use native boost_watermark() instead of custom logic v4: - Add watermark_scale_boost and gradual decay via balance_pgdat v3: - Move debounce timer to per-zone; optimize zone selection v2: - Add debounce logic and zone-proportional boosting v1: - Initial: boost min_free_kbytes on GFP_ATOMIC failure include/linux/mmzone.h | 1 + mm/page_alloc.c | 46 ++++++++++++++++++++++++++++++++++++++++-- 2 files changed, 45 insertions(+), 2 deletions(-) diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 75ef7c9f9307..8e37e4e6765b 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -882,6 +882,7 @@ struct zone { /* zone watermarks, access with *_wmark_pages(zone) macros */ unsigned long _watermark[NR_WMARK]; unsigned long watermark_boost; + unsigned long last_boost_jiffies; =20 unsigned long nr_reserved_highatomic; unsigned long nr_free_highatomic; diff --git a/mm/page_alloc.c b/mm/page_alloc.c index c380f063e8b7..94168571cc38 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -218,6 +218,13 @@ unsigned int pageblock_order __read_mostly; static void __free_pages_ok(struct page *page, unsigned int order, fpi_t fpi_flags); =20 +/* + * Boost watermarks by ~0.1% of zone size on atomic allocation pressure. + * This provides zone-proportional safety buffers: ~1MB per 1GB of zone si= ze. + * Larger zones under GFP_ATOMIC pressure need proportionally larger reser= ves. + */ +#define ATOMIC_BOOST_SCALE_SHIFT 10 + /* * results with 256, 32 in the lowmem_reserve sysctl: * 1G machine -> (16M dma, 800M-16M normal, 1G-800M high) @@ -2161,6 +2168,9 @@ bool pageblock_unisolate_and_move_free_pages(struct z= one *zone, struct page *pag static inline bool boost_watermark(struct zone *zone) { unsigned long max_boost; + unsigned long boost_amount; + + lockdep_assert_held(&zone->lock); =20 if (!watermark_boost_factor) return false; @@ -2189,12 +2199,40 @@ static inline bool boost_watermark(struct zone *zon= e) =20 max_boost =3D max(pageblock_nr_pages, max_boost); =20 - zone->watermark_boost =3D min(zone->watermark_boost + pageblock_nr_pages, - max_boost); + boost_amount =3D max(pageblock_nr_pages, + zone_managed_pages(zone) >> ATOMIC_BOOST_SCALE_SHIFT); + zone->watermark_boost =3D min(zone->watermark_boost + boost_amount, + max_boost); =20 return true; } =20 +static void boost_zones_for_atomic(struct alloc_context *ac, gfp_t gfp_mas= k) +{ + struct zoneref *z; + struct zone *zone; + unsigned long now =3D jiffies; + bool should_wake; + + for_each_zone_zonelist(zone, z, ac->zonelist, ac->highest_zoneidx) { + /* Rate-limit boosts to once per second per zone */ + if (time_after(now, zone->last_boost_jiffies + HZ)) { + zone->last_boost_jiffies =3D now; + + /* Modify watermark under lock, wake kswapd outside */ + spin_lock(&zone->lock); + should_wake =3D boost_watermark(zone); + spin_unlock(&zone->lock); + + if (should_wake) + wakeup_kswapd(zone, gfp_mask, 0, ac->highest_zoneidx); + + /* Boost only the preferred zone */ + break; + } + } +} + /* * When we are falling back to another migratetype during allocation, shou= ld we * try to claim an entire block to satisfy further allocations, instead of @@ -4742,6 +4780,10 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int = order, if (page) goto got_pg; =20 + /* Boost watermarks for atomic requests entering slowpath */ + if ((gfp_mask & GFP_ATOMIC) && order =3D=3D 0) + boost_zones_for_atomic(ac, gfp_mask); + /* * For costly allocations, try direct compaction first, as it's likely * that we have enough base pages and don't need to reclaim. For non- --=20 2.51.0