+ mm-zswap-use-separate-compression-and-decompression-requests.patch added to mm-new branch

[Andrew Morton <akpm@linux-foundation.org>]

The patch titled
     Subject: mm: zswap: use separate compression and decompression requests
has been added to the -mm mm-new branch.  Its filename is
     mm-zswap-use-separate-compression-and-decompression-requests.patch

This patch will shortly appear at
     https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-zswap-use-separate-compression-and-decompression-requests.patch

This patch will later appear in the mm-new branch at
    git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

Note, mm-new is a provisional staging ground for work-in-progress
patches, and acceptance into mm-new is a notification for others take
notice and to finish up reviews.  Please do not hesitate to respond to
review feedback and post updated versions to replace or incrementally
fixup patches in mm-new.

The mm-new branch of mm.git is not included in linux-next

If a few days of testing in mm-new is successful, the patch will me moved
into mm.git's mm-unstable branch, which is included in linux-next

Before you just go and hit "reply", please:
   a) Consider who else should be cc'ed
   b) Prefer to cc a suitable mailing list as well
   c) Ideally: find the original patch on the mailing list and do a
      reply-to-all to that, adding suitable additional cc's

*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***

The -mm tree is included into linux-next via various
branches at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there most days

------------------------------------------------------
From: Usama Arif <usama.arif@linux.dev>
Subject: mm: zswap: use separate compression and decompression requests
Date: Fri, 09 Oct 2026 09:30:30 -0700

Patch series "mm: zswap: reduce request contention on loads", v2.

Stores and loads share a per-CPU acomp request and mutex.  A low-priority
store can be preempted right after the compressor drops its stream lock,
while it still holds the zswap mutex, and a higher-priority load on that
CPU then waits for the store to run again.  This follows the work from
Sergey Senozhatsky's zram series which splits it for the same reason [1].

Patch 1 gives compression and decompression separate requests, waits and
mutexes, so loads no longer wait for stores, though they can still wait
for each other.  Patch 2 decompresses with an on-stack request when the
algorithm is synchronous and needs no request context, which covers all
in-tree software compressors, so those loads take no zswap lock. 
Asynchronous algorithms keep the per-CPU request and mutex.  For software
compressors the series allocates the same number of requests as before;
each per-CPU context grows by 72 bytes, and the load path is about 270
bytes deeper on x86-64.

The series does not fix two related cases:
- Stores still serialize on the compression mutex, so a high-priority
  task that reclaims (direct reclaim, MADV_PAGEOUT) can still wait for
  a preempted store.
- On PREEMPT_RT the codec stream locks are preemptible, so a load can
  still wait for a preempted store inside the codec.

The numbers below are the slowest read per run, as a median (min-max) of 5
runs.  Each run is 12 seconds in a zstd VM with lazy preemption,
vm.page-cluster=0 and swap on /dev/ram0.  With 1 vCPU, four nice +10
workers page memory out and read it back while a nice 0 task spins.  A
nice -19 reader pages out its own buffer and measures how long each read
of it takes.  With 8 vCPUs there are 16 workers, 8 spinning tasks and 8
readers.

              Before series (ms)      With series (ms)
  1 vCPU      22.3 (21.6-22.6)        0.97 (0.72-1.4)
  8 vCPUs     314 (97-2542)           7.0 (5.0-98)

Reads over 10 ms fell from 26-35 per run to none with 1 vCPU, and from
3-18 per run to at most one with 8 vCPUs. The benchmark and test programs
were written with the help of an LLM.


This patch (of 2):

Stores and loads serialize on the same per-CPU acomp request and mutex.
A low-priority store can be preempted as soon as the compressor drops
its stream lock, while it still holds the mutex. A higher-priority load
on that CPU then waits until the store runs again, which can take a
long time when other tasks are runnable.

Give compression and decompression their own request, completion wait
and mutex. Since commit e2c3b6b21c77f ("mm: zswap: use SG list
decompression APIs from zsmalloc"), the per-CPU buffer is only used for
compression, so group it with the compression request. Independent
requests can share the per-CPU transform; the codecs and drivers
synchronize their shared state internally. Loads can still wait for
each other on the decompression mutex, and stores still serialize on
the compression mutex.

acomp_request_free() ignores NULL requests, so drop the check in
acomp_ctx_free().

This follows the proposal from Sergey Senozhatsky for the same split
for zram [1].

Link: https://lore.kernel.org/20261009163146.83112-1-usama.arif@linux.dev
Link: https://lore.kernel.org/20261009163146.83112-2-usama.arif@linux.dev
Link: https://lore.kernel.org/all/20261005122036.718976-10-senozhatsky@chromium.org/ [1]
Signed-off-by: Usama Arif <usama.arif@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Nhat Pham <nphamcs@gmail.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: <chengming.zhou@linux.dev>
Cc: David Sterba <dsterba@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Nick Terrell <terrelln@fb.com>
Cc: Yosry Ahmed <yosry@kernel.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: <alex@ghiti.fr>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: David S. Miller <davem@davemloft.net>
---

 mm/zswap.c |  108 ++++++++++++++++++++++++++++++---------------------
 1 file changed, 64 insertions(+), 44 deletions(-)

--- a/mm/zswap.c~mm-zswap-use-separate-compression-and-decompression-requests
+++ a/mm/zswap.c
@@ -137,14 +137,25 @@ bool zswap_never_enabled(void)
 * data structures
 **********************************/
 
-struct crypto_acomp_ctx {
-	struct crypto_acomp *acomp;
+struct zswap_acomp_req {
 	struct acomp_req *req;
 	struct crypto_wait wait;
-	u8 *buffer;
 	struct mutex mutex;
 };
 
+/* The compression mutex also protects the output buffer. */
+struct zswap_comp_ctx {
+	struct zswap_acomp_req areq;
+	u8 *buffer;
+};
+
+/* Separate requests, so that decompression does not wait for compression. */
+struct crypto_acomp_ctx {
+	struct crypto_acomp *acomp;
+	struct zswap_comp_ctx comp;
+	struct zswap_acomp_req decomp;
+};
+
 /*
  * The lock ordering is zswap_tree.lock -> zswap_pool.lru_lock.
  * The only case where lru_lock is not acquired while holding tree.lock is
@@ -270,14 +281,10 @@ static void acomp_ctx_free(struct crypto
 	if (!acomp_ctx)
 		return;
 
-	/*
-	 * If there was an error in allocating @acomp_ctx->req, it
-	 * would be set to NULL.
-	 */
-	if (acomp_ctx->req)
-		acomp_request_free(acomp_ctx->req);
-
-	acomp_ctx->req = NULL;
+	acomp_request_free(acomp_ctx->comp.areq.req);
+	acomp_ctx->comp.areq.req = NULL;
+	acomp_request_free(acomp_ctx->decomp.req);
+	acomp_ctx->decomp.req = NULL;
 
 	/*
 	 * We have to handle both cases here: an error pointer return from
@@ -289,8 +296,8 @@ static void acomp_ctx_free(struct crypto
 
 	acomp_ctx->acomp = NULL;
 
-	kfree(acomp_ctx->buffer);
-	acomp_ctx->buffer = NULL;
+	kfree(acomp_ctx->comp.buffer);
+	acomp_ctx->comp.buffer = NULL;
 }
 
 static struct zswap_pool *zswap_pool_create(char *compressor)
@@ -796,6 +803,28 @@ static void zswap_entry_free(struct zswa
 /*********************************
 * compressed storage functions
 **********************************/
+static int zswap_acomp_req_init(struct zswap_acomp_req *areq,
+				struct crypto_acomp *acomp)
+{
+	/* acomp_request_alloc() returns NULL in case of an error. */
+	areq->req = acomp_request_alloc(acomp);
+	if (!areq->req)
+		return -ENOMEM;
+
+	crypto_init_wait(&areq->wait);
+
+	/*
+	 * if the backend of acomp is async zip, crypto_req_done() will wakeup
+	 * crypto_wait_req(); if the backend of acomp is scomp, the callback
+	 * won't be called, crypto_wait_req() will return without blocking.
+	 */
+	acomp_request_set_callback(areq->req, CRYPTO_TFM_REQ_MAY_BACKLOG,
+				   crypto_req_done, &areq->wait);
+
+	mutex_init(&areq->mutex);
+	return 0;
+}
+
 static int zswap_cpu_comp_prepare(unsigned int cpu, struct hlist_node *node)
 {
 	struct zswap_pool *pool = hlist_entry(node, struct zswap_pool, node);
@@ -811,8 +840,8 @@ static int zswap_cpu_comp_prepare(unsign
 		return 0;
 	}
 
-	acomp_ctx->buffer = kmalloc_node(PAGE_SIZE, GFP_KERNEL, cpu_to_node(cpu));
-	if (!acomp_ctx->buffer)
+	acomp_ctx->comp.buffer = kmalloc_node(PAGE_SIZE, GFP_KERNEL, cpu_to_node(cpu));
+	if (!acomp_ctx->comp.buffer)
 		return ret;
 
 	/*
@@ -827,25 +856,13 @@ static int zswap_cpu_comp_prepare(unsign
 		goto fail;
 	}
 
-	/* acomp_request_alloc() returns NULL in case of an error. */
-	acomp_ctx->req = acomp_request_alloc(acomp_ctx->acomp);
-	if (!acomp_ctx->req) {
+	if (zswap_acomp_req_init(&acomp_ctx->comp.areq, acomp_ctx->acomp) ||
+	    zswap_acomp_req_init(&acomp_ctx->decomp, acomp_ctx->acomp)) {
 		pr_err("could not alloc crypto acomp_request %s\n",
 		       pool->tfm_name);
 		goto fail;
 	}
 
-	crypto_init_wait(&acomp_ctx->wait);
-
-	/*
-	 * if the backend of acomp is async zip, crypto_req_done() will wakeup
-	 * crypto_wait_req(); if the backend of acomp is scomp, the callback
-	 * won't be called, crypto_wait_req() will return without blocking.
-	 */
-	acomp_request_set_callback(acomp_ctx->req, CRYPTO_TFM_REQ_MAY_BACKLOG,
-				   crypto_req_done, &acomp_ctx->wait);
-
-	mutex_init(&acomp_ctx->mutex);
 	return 0;
 
 fail:
@@ -856,7 +873,7 @@ fail:
 static bool zswap_compress(struct folio *folio, long index,
 			   struct zswap_entry *entry, struct zswap_pool *pool)
 {
-	struct crypto_acomp_ctx *acomp_ctx;
+	struct zswap_comp_ctx *comp_ctx;
 	struct scatterlist input, output;
 	int comp_ret = 0, alloc_ret = 0;
 	unsigned int dlen = PAGE_SIZE;
@@ -865,15 +882,16 @@ static bool zswap_compress(struct folio
 	u8 *dst;
 	bool mapped = false;
 
-	acomp_ctx = raw_cpu_ptr(pool->acomp_ctx);
-	mutex_lock(&acomp_ctx->mutex);
+	comp_ctx = &raw_cpu_ptr(pool->acomp_ctx)->comp;
+	mutex_lock(&comp_ctx->areq.mutex);
 
-	dst = acomp_ctx->buffer;
+	dst = comp_ctx->buffer;
 	sg_init_table(&input, 1);
 	sg_set_folio(&input, folio, PAGE_SIZE, index * PAGE_SIZE);
 
 	sg_init_one(&output, dst, PAGE_SIZE);
-	acomp_request_set_params(acomp_ctx->req, &input, &output, PAGE_SIZE, dlen);
+	acomp_request_set_params(comp_ctx->areq.req, &input, &output,
+				 PAGE_SIZE, dlen);
 
 	/*
 	 * it maybe looks a little bit silly that we send an asynchronous request,
@@ -885,10 +903,12 @@ static bool zswap_compress(struct folio
 	 * existing method to send the second page before the first page is done
 	 * in one thread doing zswap.
 	 * but in different threads running on different cpu, we have different
-	 * acomp instance, so multiple threads can do (de)compression in parallel.
+	 * acomp instance, and compression and decompression use separate
+	 * requests, so multiple threads can do (de)compression in parallel.
 	 */
-	comp_ret = crypto_wait_req(crypto_acomp_compress(acomp_ctx->req), &acomp_ctx->wait);
-	dlen = acomp_ctx->req->dlen;
+	comp_ret = crypto_wait_req(crypto_acomp_compress(comp_ctx->areq.req),
+				   &comp_ctx->areq.wait);
+	dlen = comp_ctx->areq.req->dlen;
 
 	/*
 	 * If a page cannot be compressed into a size smaller than PAGE_SIZE,
@@ -932,7 +952,7 @@ unlock:
 	else if (alloc_ret)
 		zswap_reject_alloc_fail++;
 
-	mutex_unlock(&acomp_ctx->mutex);
+	mutex_unlock(&comp_ctx->areq.mutex);
 	return comp_ret == 0 && alloc_ret == 0;
 }
 
@@ -948,7 +968,7 @@ static bool zswap_decompress(struct zswa
 		return false;
 
 	acomp_ctx = raw_cpu_ptr(pool->acomp_ctx);
-	mutex_lock(&acomp_ctx->mutex);
+	mutex_lock(&acomp_ctx->decomp.mutex);
 	zs_obj_read_sg_begin(pool->zs_pool, entry->handle, input, entry->length);
 
 	/* zswap entries of length PAGE_SIZE are not compressed. */
@@ -965,15 +985,15 @@ static bool zswap_decompress(struct zswa
 	} else {
 		sg_init_table(&output, 1);
 		sg_set_folio(&output, folio, PAGE_SIZE, 0);
-		acomp_request_set_params(acomp_ctx->req, input, &output,
+		acomp_request_set_params(acomp_ctx->decomp.req, input, &output,
 					 entry->length, PAGE_SIZE);
-		ret = crypto_acomp_decompress(acomp_ctx->req);
-		ret = crypto_wait_req(ret, &acomp_ctx->wait);
-		dlen = acomp_ctx->req->dlen;
+		ret = crypto_acomp_decompress(acomp_ctx->decomp.req);
+		ret = crypto_wait_req(ret, &acomp_ctx->decomp.wait);
+		dlen = acomp_ctx->decomp.req->dlen;
 	}
 
 	zs_obj_read_sg_end(pool->zs_pool, entry->handle);
-	mutex_unlock(&acomp_ctx->mutex);
+	mutex_unlock(&acomp_ctx->decomp.mutex);
 
 	if (!ret && dlen == PAGE_SIZE)
 		return true;
_

Patches currently in -mm which might be from usama.arif@linux.dev are

mm-filemap-move-lruvec-accounting-outside-the-xarray-lock.patch
mm-remove-the-unused-zone-unaccepted_cleanup.patch
mm-zswap-use-separate-compression-and-decompression-requests.patch
mm-zswap-use-stack-requests-for-synchronous-decompression.patch