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 Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id D34E6E9129F for ; Fri, 6 Feb 2026 04:11:21 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id C3BE76B008A; Thu, 5 Feb 2026 23:11:20 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id C1D726B0092; Thu, 5 Feb 2026 23:11:20 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id AF19B6B0093; Thu, 5 Feb 2026 23:11:20 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0013.hostedemail.com [216.40.44.13]) by kanga.kvack.org (Postfix) with ESMTP id 9F5816B008A for ; Thu, 5 Feb 2026 23:11:20 -0500 (EST) Received: from smtpin17.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay01.hostedemail.com (Postfix) with ESMTP id EFC801C613 for ; Fri, 6 Feb 2026 04:11:19 +0000 (UTC) X-FDA: 84412707078.17.038D5D7 Received: from out-182.mta0.migadu.com (out-182.mta0.migadu.com [91.218.175.182]) by imf03.hostedemail.com (Postfix) with ESMTP id DB87E20010 for ; Fri, 6 Feb 2026 04:11:17 +0000 (UTC) Authentication-Results: imf03.hostedemail.com; dkim=pass header.d=linux.dev header.s=key1 header.b=SAYmCD3c; spf=pass (imf03.hostedemail.com: domain of yosry.ahmed@linux.dev designates 91.218.175.182 as permitted sender) smtp.mailfrom=yosry.ahmed@linux.dev; dmarc=pass (policy=none) header.from=linux.dev ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1770351078; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=D4OW4s3gaXFmm9q/2xsnJJrtYlj7SsdGtTDJwGZIMW4=; b=FhuJ9sdgtHNDG69x1WTMJUmAljs2xbrunSq4QYQvP+JxbxAp4vSKn7EWJrRO+yjp65hPkx 7Ux7lOWoCdiSENZcXXadS8TjSCNED8LO5nCP9T6lhiyo2ckWCEjwU6nd2KKo5aBeWnbu0p SneOCUdAAzcDne6kEaK1FswscHpY4Kc= ARC-Authentication-Results: i=1; imf03.hostedemail.com; dkim=pass header.d=linux.dev header.s=key1 header.b=SAYmCD3c; spf=pass (imf03.hostedemail.com: domain of yosry.ahmed@linux.dev designates 91.218.175.182 as permitted sender) smtp.mailfrom=yosry.ahmed@linux.dev; dmarc=pass (policy=none) header.from=linux.dev ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1770351078; a=rsa-sha256; cv=none; b=ILyGNJf8g9HkxAiCWXYt/k1kpnhQhkbJoa45eG8ziKF2q7Vj/YeZrF5b7K9pt7vDr8sNyI dhFPpSUlCX7P2RupYMRDRpRBdihfAwA4qGY9I272ueQCJRYUj7+Y9a4OlTQ7iFTdFPH+rn 5q2QM6iaoi0ZF47eJ2bR2hQXdHmdI5E= MIME-Version: 1.0 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1770351075; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=D4OW4s3gaXFmm9q/2xsnJJrtYlj7SsdGtTDJwGZIMW4=; b=SAYmCD3ck4yK/hM3itvbD9mKLpxMrdxT7KMxmCG+XVOUfQqmmfWUuy/3IXH3+BGCaiBm+V WhkktKZ4S+fh8PHguT3q0WI69nVlHNFCW7haNtCqkmqAeLIkqzScx/ZKjhGsArRtZ0iwcR ujfDG5W9eyXs/3fFOA+Q6sXz8kDiFTE= Date: Fri, 06 Feb 2026 04:11:13 +0000 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. From: "Yosry Ahmed" Message-ID: <75c6f71f794c04e6c6fc90e6ac0a3d4656651362@linux.dev> TLS-Required: No Subject: Re: [PATCH v1] mm: zswap: add per-memcg stat for incompressible pages To: "Jiayuan Chen" , "SeongJae Park" Cc: "SeongJae Park" , linux-mm@kvack.org, "Jiayuan Chen" , "Johannes Weiner" , "Michal Hocko" , "Roman Gushchin" , "Shakeel Butt" , "Muchun Song" , "Nhat Pham" , "Chengming Zhou" , "Andrew Morton" , "Nick Terrell" , "David Sterba" , cgroups@vger.kernel.org, linux-kernel@vger.kernel.org In-Reply-To: References: <20260206022152.67992-1-sj@kernel.org> X-Migadu-Flow: FLOW_OUT X-Rspamd-Server: rspam01 X-Rspamd-Queue-Id: DB87E20010 X-Stat-Signature: j8759ddaft9r4ak8p8jt1m59nndfo1z3 X-Rspam-User: X-HE-Tag: 1770351077-228467 X-HE-Meta: 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 RTMKEPF6 u87ChoeuCEax14dHVS+RT06kQwY218jZCXVeEs2Pufx7xVGESpEuHDj0vmH4yljcwOqxY0PJIJ1MVsCMiWkxWZgFfYWnNaCmKz/dgnS77vSQQjQ/9V3n0kHkhDSkpuieOTA98MzpwlcTqQxL/UMlu9icvioptrHbKcOu9c4LLrQW62lGkXdXI132zO/ClKU5cV1PJUdRCbWimZUTAbC1biR9NwfV66FddmVJL7yczif6plsjzXTfe1lFFySfBcbjHFHkskIiW7u7xmj9s10YugztNRclJ1NH7giobSnh4lLuoBBg= X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: On Thu, 5 Feb 2026 13:30:12 +0800 Jiayuan Chen w= rote: > >=20=20 >=20>=20=20 >=20> From: Jiayuan Chen > >=20=20 >=20> The global zswap_stored_incompressible_pages counter was added in = commit > > dca4437a5861 ("mm/zswap: store > to track how many pages are stored in raw (uncompressed) form in zsw= ap. > > However, in containerized environments, knowing which cgroup is > > contributing incompressible pages is essential for effective resourc= e > > management. > >=20=20 >=20> Add a new memcg stat 'zswpraw' to track incompressible pages per c= group. > > This helps administrators and orchestrators to: > >=20=20 >=20> 1. Identify workloads that produce incompressible data (e.g., encr= ypted > > data, already-compressed media, random data) and may not benefit fro= m > > zswap. > >=20=20 >=20> 2. Make informed decisions about workload placement - moving > > incompressible workloads to nodes with larger swap backing devices > > rather than relying on zswap. > >=20=20 >=20> 3. Debug zswap efficiency issues at the cgroup level without needi= ng to > > correlate global stats with individual cgroups. > >=20=20 >=20> While the compression ratio can be estimated from existing stats > > (zswap / zswapped * PAGE_SIZE), this doesn't distinguish between > > "uniformly poor compression" and "a few completely incompressible pa= ges > > mixed with highly compressible ones". The zswpraw stat provides dire= ct > > visibility into the latter case. > >=20=20 >=20> Changes > > ------- > >=20=20 >=20> 1. Add zswap_is_raw() helper (include/linux/zswap.h) > > - Abstract the PAGE_SIZE comparison logic for identifying raw entrie= s > > - Keep the incompressible check in one place for maintainability > >=20=20 >=20> 2. Add MEMCG_ZSWAP_RAW stat definition (include/linux/memcontrol.h= , > > mm/memcontrol.c) > > - Add MEMCG_ZSWAP_RAW to memcg_stat_item enum > > - Register in memcg_stat_items[] and memory_stats[] arrays > > - Export as "zswpraw" in memory.stat > >=20=20 >=20> 3. Update statistics accounting (mm/memcontrol.c, mm/zswap.c) > > - Track MEMCG_ZSWAP_RAW in obj_cgroup_charge/uncharge_zswap() > > - Use zswap_is_raw() helper in zswap.c for consistency > >=20=20 >=20> Test > > ---- > >=20=20 >=20> I wrote a simple test program[1] that allocates memory and compres= ses it > > with zstd, so kernel zswap cannot compress further. > >=20=20 >=20> $ cgcreate -g memory:test > > $ cgexec -g memory:test ./test_zswpraw & > > $ cat /sys/fs/cgroup/test/memory.stat | grep zswp > > zswpraw 0 > > zswpin 0 > > zswpout 0 > > zswpwb 0 > >=20=20 >=20> $ echo "100M" > /sys/fs/cgroup/test/memory.reclaim > > $ cat /sys/fs/cgroup/test/memory.stat | grep zswp > > zswpraw 104800256 > > zswpin 0 > > zswpout 51222 > > zswpwb 0 > >=20=20 >=20> $ pkill test_zswpraw > > $ cat /sys/fs/cgroup/test/memory.stat | grep zswp > > zswpraw 0 > > zswpin 1 > > zswpout 51222 > > zswpwb 0 > >=20=20 >=20> [1] https://gist.github.com/mrpre/00432c6154250326994fbeaf62e0e6f1 > >=20=20 >=20> Signed-off-by: Jiayuan Chen > > --- > > include/linux/memcontrol.h | 1 + > > include/linux/zswap.h | 9 +++++++++ > > mm/memcontrol.c | 6 ++++++ > > mm/zswap.c | 6 +++--- > > 4 files changed, 19 insertions(+), 3 deletions(-) > >=20=20 >=20> As others also mentioned, the documentation of the new stat would = be needed. > >=20=20 >=20>=20=20 >=20> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol= .h > > index b6c82c8f73e1..83d1328f81d1 100644 > > --- a/include/linux/memcontrol.h > > +++ b/include/linux/memcontrol.h > > @@ -39,6 +39,7 @@ enum memcg_stat_item { > > MEMCG_KMEM, > > MEMCG_ZSWAP_B, > > MEMCG_ZSWAPPED, > > + MEMCG_ZSWAP_RAW, > > MEMCG_NR_STAT, > > }; > >=20=20 >=20> diff --git a/include/linux/zswap.h b/include/linux/zswap.h > > index 30c193a1207e..94f84b154b71 100644 > > --- a/include/linux/zswap.h > > +++ b/include/linux/zswap.h > > @@ -7,6 +7,15 @@ > >=20=20 >=20> struct lruvec; > >=20=20 >=20> +/* > > + * Check if a zswap entry is stored in raw (uncompressed) form. > > + * This happens when compression doesn't reduce the size. > > + */ > > +static inline bool zswap_is_raw(size_t size) > > +{ > > + return size =3D=3D PAGE_SIZE; > > +} > > + > >=20=20 >=20> No strong opinion, but I'm not really sure if the helper is needed= , because it > > feels quite simple logic: > >=20=20 >=20> "If an object is compressed and the size is same to the original o= ne, the > > object is incompressible." > >=20=20 >=20> I also feel the function name bit odd, given the type of the param= eter. Based > > on the function name and the comment, I'd expect it to receive a zsw= ap_entry > > object. I understand it is better to receive a size_t, to be called = from > > obj_cgroup_[un]charge_zswap(), though. Even in the case, I think the= name can > > be better (e.g., zswap_compression_failed() or zswap_was_incompressi= ble() ?), > > or at least the coment can be more kindly explain the fact that the = parameter > > is the size of object after the compression attempt. > >=20=20 >=20>=20=20 >=20> extern atomic_long_t zswap_stored_pages; > >=20=20 >=20> #ifdef CONFIG_ZSWAP > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > > index 007413a53b45..32fb801530a3 100644 > > --- a/mm/memcontrol.c > > +++ b/mm/memcontrol.c > > @@ -341,6 +341,7 @@ static const unsigned int memcg_stat_items[] =3D= { > > MEMCG_KMEM, > > MEMCG_ZSWAP_B, > > MEMCG_ZSWAPPED, > > + MEMCG_ZSWAP_RAW, > > }; > >=20=20 >=20> No strong opinion, but I think Shakeel's suggestion of other names= is > > reasonable. > >=20=20 >=20>=20=20 >=20> #define NR_MEMCG_NODE_STAT_ITEMS ARRAY_SIZE(memcg_node_stat_items) > > @@ -1346,6 +1347,7 @@ static const struct memory_stat memory_stats[]= =3D { > > #ifdef CONFIG_ZSWAP > > { "zswap", MEMCG_ZSWAP_B }, > > { "zswapped", MEMCG_ZSWAPPED }, > > + { "zswpraw", MEMCG_ZSWAP_RAW }, > >=20=20 >=20> Ditto. > >=20=20 >=20>=20=20 >=20> #endif > > { "file_mapped", NR_FILE_MAPPED }, > > { "file_dirty", NR_FILE_DIRTY }, > > @@ -5458,6 +5460,8 @@ void obj_cgroup_charge_zswap(struct obj_cgroup= *objcg, size_t size) > > memcg =3D obj_cgroup_memcg(objcg); > > mod_memcg_state(memcg, MEMCG_ZSWAP_B, size); > > mod_memcg_state(memcg, MEMCG_ZSWAPPED, 1); > > + if (zswap_is_raw(size)) > > + mod_memcg_state(memcg, MEMCG_ZSWAP_RAW, 1); > >=20=20 >=20> I understand the helper function is better to receive size_t rathe= r than > > zswap_entry for this. > >=20=20 >=20>=20=20 >=20> rcu_read_unlock(); > > } > >=20=20 >=20> @@ -5481,6 +5485,8 @@ void obj_cgroup_uncharge_zswap(struct obj_cg= roup *objcg, size_t size) > > memcg =3D obj_cgroup_memcg(objcg); > > mod_memcg_state(memcg, MEMCG_ZSWAP_B, -size); > > mod_memcg_state(memcg, MEMCG_ZSWAPPED, -1); > > + if (zswap_is_raw(size)) > > + mod_memcg_state(memcg, MEMCG_ZSWAP_RAW, -1); > > rcu_read_unlock(); > > } > >=20=20 >=20> diff --git a/mm/zswap.c b/mm/zswap.c > > index 3d2d59ac3f9c..54ab4d126f64 100644 > > --- a/mm/zswap.c > > +++ b/mm/zswap.c > > @@ -723,7 +723,7 @@ static void zswap_entry_free(struct zswap_entry = *entry) > > obj_cgroup_uncharge_zswap(entry->objcg, entry->length); > > obj_cgroup_put(entry->objcg); > > } > > - if (entry->length =3D=3D PAGE_SIZE) > > + if (zswap_is_raw(entry->length)) > > atomic_long_dec(&zswap_stored_incompressible_pages); > > zswap_entry_cache_free(entry); > > atomic_long_dec(&zswap_stored_pages); > > @@ -941,7 +941,7 @@ static bool zswap_decompress(struct zswap_entry = *entry, struct folio *folio) > > zs_obj_read_sg_begin(pool->zs_pool, entry->handle, input, entry->len= gth); > >=20=20 >=20> /* zswap entries of length PAGE_SIZE are not compressed. */ > > - if (entry->length =3D=3D PAGE_SIZE) { > > + if (zswap_is_raw(entry->length)) { > > WARN_ON_ONCE(input->length !=3D PAGE_SIZE); > > memcpy_from_sglist(kmap_local_folio(folio, 0), input, 0, PAGE_SIZE); > > dlen =3D PAGE_SIZE; > >=20=20 >=20> Below this part, I show 'dlen =3D=3D PAGE_SIZE'. Should it also be= converted to > > use the helper function? > >=20 >=20The dlen variable represents the decompressed (plaintext) size. > Since we compress individual pages, the decompressed output should > always be PAGE_SIZE in normal cases. >=20 >=20This check validates whether decompression produced the expected resu= lt, not whether the entry is incompressible. >=20 >=20Using zswap_is_incomp() here would be semantically incorrect - the he= lper is meant to check if an > entry was stored without compression (i.e., compression failed to reduc= e size), while dlen =3D=3D PAGE_SIZE > verifies the output of decompression is valid. This confusion is exactly why we should drop the helper.