linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Sergey Senozhatsky <senozhatsky@chromium.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Yosry Ahmed <yosry.ahmed@linux.dev>,
	Kairui Song <ryncsn@gmail.com>, Minchan Kim <minchan@kernel.org>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Sergey Senozhatsky <senozhatsky@chromium.org>
Subject: [PATCH v5 09/18] zram: rework recompression loop
Date: Wed, 12 Feb 2025 15:27:07 +0900	[thread overview]
Message-ID: <20250212063153.179231-10-senozhatsky@chromium.org> (raw)
In-Reply-To: <20250212063153.179231-1-senozhatsky@chromium.org>

This reworks recompression loop handling:

- set a rule that stream-put NULLs the stream pointer
  If the loop returns with a non-NULL stream then it's a
  successfull recompression, otherwise the stream should
  always be NULL.

- do not count the number of recompressions
  Mark object as incompressible as soon as the algorithm
  with the highest priority failed to compress that object.

- count compression errors as resource usage
  Even if compression has failed, we still need to bump
  num_recomp_pages counter.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 53 +++++++++++++----------------------
 1 file changed, 19 insertions(+), 34 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 31bdf5e0ff74..7c4c296181a8 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -1887,9 +1887,8 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 	unsigned int comp_len_new;
 	unsigned int class_index_old;
 	unsigned int class_index_new;
-	u32 num_recomps = 0;
 	void *src, *dst;
-	int ret;
+	int ret = 0;
 
 	handle_old = zram_get_handle(zram, index);
 	if (!handle_old)
@@ -1932,7 +1931,6 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 		if (!zram->comps[prio])
 			continue;
 
-		num_recomps++;
 		zstrm = zcomp_stream_get(zram->comps[prio]);
 		src = kmap_local_page(page);
 		ret = zcomp_compress(zram->comps[prio], zstrm,
@@ -1941,7 +1939,8 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 
 		if (ret) {
 			zcomp_stream_put(zstrm);
-			return ret;
+			zstrm = NULL;
+			break;
 		}
 
 		class_index_new = zs_lookup_class_index(zram->mem_pool,
@@ -1951,6 +1950,7 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 		if (class_index_new >= class_index_old ||
 		    (threshold && comp_len_new >= threshold)) {
 			zcomp_stream_put(zstrm);
+			zstrm = NULL;
 			continue;
 		}
 
@@ -1958,14 +1958,6 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 		break;
 	}
 
-	/*
-	 * We did not try to recompress, e.g. when we have only one
-	 * secondary algorithm and the page is already recompressed
-	 * using that algorithm
-	 */
-	if (!zstrm)
-		return 0;
-
 	/*
 	 * Decrement the limit (if set) on pages we can recompress, even
 	 * when current recompression was unsuccessful or did not compress
@@ -1975,38 +1967,31 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page,
 	if (*num_recomp_pages)
 		*num_recomp_pages -= 1;
 
-	if (class_index_new >= class_index_old) {
+	/* Compression error */
+	if (ret)
+		return ret;
+
+	if (!zstrm) {
 		/*
 		 * Secondary algorithms failed to re-compress the page
-		 * in a way that would save memory, mark the object as
-		 * incompressible so that we will not try to compress
-		 * it again.
+		 * in a way that would save memory.
 		 *
-		 * We need to make sure that all secondary algorithms have
-		 * failed, so we test if the number of recompressions matches
-		 * the number of active secondary algorithms.
+		 * Mark the object incompressible if the max-priority
+		 * algorithm couldn't re-compress it.
 		 */
-		if (num_recomps == zram->num_active_comps - 1)
-			zram_set_flag(zram, index, ZRAM_INCOMPRESSIBLE);
+		if (prio < zram->num_active_comps)
+			return 0;
+		zram_set_flag(zram, index, ZRAM_INCOMPRESSIBLE);
 		return 0;
 	}
 
-	/* Successful recompression but above threshold */
-	if (threshold && comp_len_new >= threshold)
-		return 0;
-
 	/*
-	 * No direct reclaim (slow path) for handle allocation and no
-	 * re-compression attempt (unlike in zram_write_bvec()) since
-	 * we already have stored that object in zsmalloc. If we cannot
-	 * alloc memory for recompressed object then we bail out and
-	 * simply keep the old (existing) object in zsmalloc.
+	 * We are holding per-CPU stream mutex and entry lock so better
+	 * avoid direct reclaim.  Allocation error is not fatal since
+	 * we still have the old object in the mem_pool.
 	 */
 	handle_new = zs_malloc(zram->mem_pool, comp_len_new,
-			       __GFP_KSWAPD_RECLAIM |
-			       __GFP_NOWARN |
-			       __GFP_HIGHMEM |
-			       __GFP_MOVABLE);
+			       GFP_NOWAIT | __GFP_HIGHMEM | __GFP_MOVABLE);
 	if (IS_ERR_VALUE(handle_new)) {
 		zcomp_stream_put(zstrm);
 		return PTR_ERR((void *)handle_new);
-- 
2.48.1.502.g6dc24dfdaf-goog



  parent reply	other threads:[~2025-02-12  6:32 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-12  6:26 [PATCH v5 00/18] zsmalloc/zram: there be preemption Sergey Senozhatsky
2025-02-12  6:26 ` [PATCH v5 01/18] zram: sleepable entry locking Sergey Senozhatsky
2025-02-13  0:08   ` Andrew Morton
2025-02-13  0:52     ` Sergey Senozhatsky
2025-02-13  1:42       ` Sergey Senozhatsky
2025-02-13  8:49         ` Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 02/18] zram: permit preemption with active compression stream Sergey Senozhatsky
2025-02-12 16:01   ` Yosry Ahmed
2025-02-13  1:04     ` Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 03/18] zram: remove crypto include Sergey Senozhatsky
2025-02-12 16:13   ` Yosry Ahmed
2025-02-13  0:53     ` Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 04/18] zram: remove max_comp_streams device attr Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 05/18] zram: remove two-staged handle allocation Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 06/18] zram: remove writestall zram_stats member Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 07/18] zram: limit max recompress prio to num_active_comps Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 08/18] zram: filter out recomp targets based on priority Sergey Senozhatsky
2025-02-12  6:27 ` Sergey Senozhatsky [this message]
2025-02-12  6:27 ` [PATCH v5 10/18] zsmalloc: factor out pool locking helpers Sergey Senozhatsky
2025-02-12 16:18   ` Yosry Ahmed
2025-02-12 16:19     ` Yosry Ahmed
2025-02-13  0:57     ` Sergey Senozhatsky
2025-02-13  1:12       ` Yosry Ahmed
2025-02-13  2:54         ` Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 11/18] zsmalloc: factor out size-class " Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 12/18] zsmalloc: make zspage lock preemptible Sergey Senozhatsky
2025-02-12 17:14   ` Yosry Ahmed
2025-02-13  1:20     ` Sergey Senozhatsky
2025-02-13  1:31       ` Yosry Ahmed
2025-02-13  1:53         ` Sergey Senozhatsky
2025-02-13 11:32   ` Hillf Danton
2025-02-13 12:29     ` Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 13/18] zsmalloc: introduce new object mapping API Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 14/18] zram: switch to new zsmalloc " Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 15/18] zram: permit reclaim in zstd custom allocator Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 16/18] zram: do not leak page on recompress_store error path Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 17/18] zram: do not leak page on writeback_store " Sergey Senozhatsky
2025-02-12  6:27 ` [PATCH v5 18/18] zram: add might_sleep to zcomp API Sergey Senozhatsky
2025-02-13  0:09 ` [PATCH v5 00/18] zsmalloc/zram: there be preemption Andrew Morton
2025-02-13  0:51   ` Sergey Senozhatsky

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250212063153.179231-10-senozhatsky@chromium.org \
    --to=senozhatsky@chromium.org \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=minchan@kernel.org \
    --cc=ryncsn@gmail.com \
    --cc=yosry.ahmed@linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox