[PATCHv4 0/6] zram: introduce writeback bio batching

[PATCHv4 0/6] zram: introduce writeback bio batching

[Sergey Senozhatsky]

This is a rework of Yuwen's patch [1] that adds writeback bio
batching support to zram, which can improve throughput of
writeback operation.

[1] https://lore.kernel.org/linux-block/tencent_0FBBFC8AE0B97BC63B5D47CE1FF2BABFDA09@qq.com/

v3 -> v4:
- fixed bdev block id leak (Yuwen)
- fixed missing pps index assignment (Yuwen)
- fixed preliminary wb IO completion slot release
- we now use bit 0 for writeback as well
- access bdev read slot block id under slot lock
- some other cleanups

Sergey Senozhatsky (5):
  zram: add writeback batch size device attr
  zram: take write lock in wb limit store handlers
  zram: drop wb_limit_lock
  zram: rework bdev block allocation
  zram: read slot block idx under slot lock

Yuwen Chen (1):
  zram: introduce writeback bio batching support

 drivers/block/zram/zram_drv.c | 452 ++++++++++++++++++++++++++--------
 drivers/block/zram/zram_drv.h |   2 +-
 2 files changed, 348 insertions(+), 106 deletions(-)

--
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 1/6] zram: introduce writeback bio batching support

[Sergey Senozhatsky]

From: Yuwen Chen <ywen.chen@foxmail.com>

Currently, zram writeback supports only a single bio writeback
operation, waiting for bio completion before post-processing
next pp-slot.  This works, in general, but has certain throughput
limitations.  Implement batched (multiple) bio writeback support
to take advantage of parallel requests processing and better
requests scheduling.

For the time being the writeback batch size (maximum number of
in-flight bio requests) is set to 32 for all devices.  A follow
up patch adds a writeback_batch_size device attribute, so the
batch size becomes run-time configurable.

Please refer to [1] and [2] for benchmarks.

[1] https://lore.kernel.org/linux-block/tencent_B2DC37E3A2AED0E7F179365FCB5D82455B08@qq.com
[2] https://lore.kernel.org/linux-block/tencent_0FBBFC8AE0B97BC63B5D47CE1FF2BABFDA09@qq.com

[senozhatsky: significantly reworked the initial patch so that the
approach and implementation resemble current zram post-processing
code]

Signed-off-by: Yuwen Chen <ywen.chen@foxmail.com>
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Co-developed-by: Richard Chang <richardycc@google.com>
Suggested-by: Minchan Kim <minchan@google.com>
---
 drivers/block/zram/zram_drv.c | 348 +++++++++++++++++++++++++++-------
 1 file changed, 282 insertions(+), 66 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index a43074657531..ea06f4d7b623 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -500,6 +500,24 @@ static ssize_t idle_store(struct device *dev,
 }
 
 #ifdef CONFIG_ZRAM_WRITEBACK
+struct zram_wb_ctl {
+	struct list_head idle_reqs;
+	struct list_head inflight_reqs;
+
+	atomic_t num_inflight;
+	struct completion done;
+};
+
+struct zram_wb_req {
+	unsigned long blk_idx;
+	struct page *page;
+	struct zram_pp_slot *pps;
+	struct bio_vec bio_vec;
+	struct bio bio;
+
+	struct list_head entry;
+};
+
 static ssize_t writeback_limit_enable_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t len)
 {
@@ -734,20 +752,209 @@ static void read_from_bdev_async(struct zram *zram, struct page *page,
 	submit_bio(bio);
 }
 
-static int zram_writeback_slots(struct zram *zram, struct zram_pp_ctl *ctl)
+static void release_wb_req(struct zram_wb_req *req)
+{
+	__free_page(req->page);
+	kfree(req);
+}
+
+static void release_wb_ctl(struct zram_wb_ctl *wb_ctl)
+{
+	/* We should never have inflight requests at this point */
+	WARN_ON(!list_empty(&wb_ctl->inflight_reqs));
+
+	while (!list_empty(&wb_ctl->idle_reqs)) {
+		struct zram_wb_req *req;
+
+		req = list_first_entry(&wb_ctl->idle_reqs,
+				       struct zram_wb_req, entry);
+		list_del(&req->entry);
+		release_wb_req(req);
+	}
+
+	kfree(wb_ctl);
+}
+
+/* XXX: should be a per-device sysfs attr */
+#define ZRAM_WB_REQ_CNT 32
+
+static struct zram_wb_ctl *init_wb_ctl(void)
+{
+	struct zram_wb_ctl *wb_ctl;
+	int i;
+
+	wb_ctl = kmalloc(sizeof(*wb_ctl), GFP_KERNEL);
+	if (!wb_ctl)
+		return NULL;
+
+	INIT_LIST_HEAD(&wb_ctl->idle_reqs);
+	INIT_LIST_HEAD(&wb_ctl->inflight_reqs);
+	atomic_set(&wb_ctl->num_inflight, 0);
+	init_completion(&wb_ctl->done);
+
+	for (i = 0; i < ZRAM_WB_REQ_CNT; i++) {
+		struct zram_wb_req *req;
+
+		/*
+		 * This is fatal condition only if we couldn't allocate
+		 * any requests at all.  Otherwise we just work with the
+		 * requests that we have successfully allocated, so that
+		 * writeback can still proceed, even if there is only one
+		 * request on the idle list.
+		 */
+		req = kzalloc(sizeof(*req), GFP_KERNEL | __GFP_NOWARN);
+		if (!req)
+			break;
+
+		req->page = alloc_page(GFP_KERNEL | __GFP_NOWARN);
+		if (!req->page) {
+			kfree(req);
+			break;
+		}
+
+		list_add(&req->entry, &wb_ctl->idle_reqs);
+	}
+
+	/* We couldn't allocate any requests, so writeabck is not possible */
+	if (list_empty(&wb_ctl->idle_reqs))
+		goto release_wb_ctl;
+
+	return wb_ctl;
+
+release_wb_ctl:
+	release_wb_ctl(wb_ctl);
+	return NULL;
+}
+
+static void zram_account_writeback_rollback(struct zram *zram)
+{
+	spin_lock(&zram->wb_limit_lock);
+	if (zram->wb_limit_enable)
+		zram->bd_wb_limit +=  1UL << (PAGE_SHIFT - 12);
+	spin_unlock(&zram->wb_limit_lock);
+}
+
+static void zram_account_writeback_submit(struct zram *zram)
+{
+	spin_lock(&zram->wb_limit_lock);
+	if (zram->wb_limit_enable && zram->bd_wb_limit > 0)
+		zram->bd_wb_limit -=  1UL << (PAGE_SHIFT - 12);
+	spin_unlock(&zram->wb_limit_lock);
+}
+
+static int zram_writeback_complete(struct zram *zram, struct zram_wb_req *req)
+{
+	u32 index = req->pps->index;
+	int err;
+
+	err = blk_status_to_errno(req->bio.bi_status);
+	if (err) {
+		/*
+		 * Failed wb requests should not be accounted in wb_limit
+		 * (if enabled).
+		 */
+		zram_account_writeback_rollback(zram);
+		free_block_bdev(zram, req->blk_idx);
+		return err;
+	}
+
+	atomic64_inc(&zram->stats.bd_writes);
+	zram_slot_lock(zram, index);
+	/*
+	 * We release slot lock during writeback so slot can change under us:
+	 * slot_free() or slot_free() and zram_write_page(). In both cases
+	 * slot loses ZRAM_PP_SLOT flag. No concurrent post-processing can
+	 * set ZRAM_PP_SLOT on such slots until current post-processing
+	 * finishes.
+	 */
+	if (!zram_test_flag(zram, index, ZRAM_PP_SLOT)) {
+		free_block_bdev(zram, req->blk_idx);
+		goto out;
+	}
+
+	zram_free_page(zram, index);
+	zram_set_flag(zram, index, ZRAM_WB);
+	zram_set_handle(zram, index, req->blk_idx);
+	atomic64_inc(&zram->stats.pages_stored);
+
+out:
+	zram_slot_unlock(zram, index);
+	return 0;
+}
+
+static void zram_writeback_endio(struct bio *bio)
+{
+	struct zram_wb_ctl *wb_ctl = bio->bi_private;
+
+	if (atomic_dec_return(&wb_ctl->num_inflight) == 0)
+		complete(&wb_ctl->done);
+}
+
+static void zram_submit_wb_request(struct zram *zram,
+				   struct zram_wb_ctl *wb_ctl,
+				   struct zram_wb_req *req)
+{
+	/*
+	 * wb_limit (if enabled) should be adjusted before submission,
+	 * so that we don't over-submit.
+	 */
+	zram_account_writeback_submit(zram);
+	atomic_inc(&wb_ctl->num_inflight);
+	list_add_tail(&req->entry, &wb_ctl->inflight_reqs);
+	submit_bio(&req->bio);
+}
+
+static struct zram_wb_req *select_idle_req(struct zram_wb_ctl *wb_ctl)
 {
+	struct zram_wb_req *req;
+
+	req = list_first_entry_or_null(&wb_ctl->idle_reqs,
+				       struct zram_wb_req, entry);
+	if (req)
+		list_del(&req->entry);
+	return req;
+}
+
+static int zram_wb_wait_for_completion(struct zram *zram,
+				       struct zram_wb_ctl *wb_ctl)
+{
+	int ret = 0;
+
+	if (atomic_read(&wb_ctl->num_inflight))
+		wait_for_completion_io(&wb_ctl->done);
+
+	reinit_completion(&wb_ctl->done);
+	while (!list_empty(&wb_ctl->inflight_reqs)) {
+		struct zram_wb_req *req;
+		int err;
+
+		req = list_first_entry(&wb_ctl->inflight_reqs,
+				       struct zram_wb_req, entry);
+		list_move(&req->entry, &wb_ctl->idle_reqs);
+
+		err = zram_writeback_complete(zram, req);
+		if (err)
+			ret = err;
+
+		release_pp_slot(zram, req->pps);
+		req->pps = NULL;
+	}
+
+	return ret;
+}
+
+static int zram_writeback_slots(struct zram *zram,
+				struct zram_pp_ctl *ctl,
+				struct zram_wb_ctl *wb_ctl)
+{
+	struct zram_wb_req *req = NULL;
 	unsigned long blk_idx = 0;
-	struct page *page = NULL;
 	struct zram_pp_slot *pps;
-	struct bio_vec bio_vec;
-	struct bio bio;
+	struct blk_plug io_plug;
 	int ret = 0, err;
-	u32 index;
-
-	page = alloc_page(GFP_KERNEL);
-	if (!page)
-		return -ENOMEM;
+	u32 index = 0;
 
+	blk_start_plug(&io_plug);
 	while ((pps = select_pp_slot(ctl))) {
 		spin_lock(&zram->wb_limit_lock);
 		if (zram->wb_limit_enable && !zram->bd_wb_limit) {
@@ -757,6 +964,26 @@ static int zram_writeback_slots(struct zram *zram, struct zram_pp_ctl *ctl)
 		}
 		spin_unlock(&zram->wb_limit_lock);
 
+		while (!req) {
+			req = select_idle_req(wb_ctl);
+			if (req)
+				break;
+
+			blk_finish_plug(&io_plug);
+			err = zram_wb_wait_for_completion(zram, wb_ctl);
+			blk_start_plug(&io_plug);
+			/*
+			 * BIO errors are not fatal, we continue and simply
+			 * attempt to writeback the remaining objects (pages).
+			 * At the same time we need to signal user-space that
+			 * some writes (at least one, but also could be all of
+			 * them) were not successful and we do so by returning
+			 * the most recent BIO error.
+			 */
+			if (err)
+				ret = err;
+		}
+
 		if (!blk_idx) {
 			blk_idx = alloc_block_bdev(zram);
 			if (!blk_idx) {
@@ -775,67 +1002,46 @@ static int zram_writeback_slots(struct zram *zram, struct zram_pp_ctl *ctl)
 		 */
 		if (!zram_test_flag(zram, index, ZRAM_PP_SLOT))
 			goto next;
-		if (zram_read_from_zspool(zram, page, index))
+		if (zram_read_from_zspool(zram, req->page, index))
 			goto next;
 		zram_slot_unlock(zram, index);
 
-		bio_init(&bio, zram->bdev, &bio_vec, 1,
-			 REQ_OP_WRITE | REQ_SYNC);
-		bio.bi_iter.bi_sector = blk_idx * (PAGE_SIZE >> 9);
-		__bio_add_page(&bio, page, PAGE_SIZE, 0);
-
 		/*
-		 * XXX: A single page IO would be inefficient for write
-		 * but it would be not bad as starter.
+		 * From now on pp-slot is owned by the req, remove it from
+		 * its pp bucket.
 		 */
-		err = submit_bio_wait(&bio);
-		if (err) {
-			release_pp_slot(zram, pps);
-			/*
-			 * BIO errors are not fatal, we continue and simply
-			 * attempt to writeback the remaining objects (pages).
-			 * At the same time we need to signal user-space that
-			 * some writes (at least one, but also could be all of
-			 * them) were not successful and we do so by returning
-			 * the most recent BIO error.
-			 */
-			ret = err;
-			continue;
-		}
+		list_del_init(&pps->entry);
 
-		atomic64_inc(&zram->stats.bd_writes);
-		zram_slot_lock(zram, index);
-		/*
-		 * Same as above, we release slot lock during writeback so
-		 * slot can change under us: slot_free() or slot_free() and
-		 * reallocation (zram_write_page()). In both cases slot loses
-		 * ZRAM_PP_SLOT flag. No concurrent post-processing can set
-		 * ZRAM_PP_SLOT on such slots until current post-processing
-		 * finishes.
-		 */
-		if (!zram_test_flag(zram, index, ZRAM_PP_SLOT))
-			goto next;
+		req->blk_idx = blk_idx;
+		req->pps = pps;
+		bio_init(&req->bio, zram->bdev, &req->bio_vec, 1, REQ_OP_WRITE);
+		req->bio.bi_iter.bi_sector = req->blk_idx * (PAGE_SIZE >> 9);
+		req->bio.bi_end_io = zram_writeback_endio;
+		req->bio.bi_private = wb_ctl;
+		__bio_add_page(&req->bio, req->page, PAGE_SIZE, 0);
 
-		zram_free_page(zram, index);
-		zram_set_flag(zram, index, ZRAM_WB);
-		zram_set_handle(zram, index, blk_idx);
+		zram_submit_wb_request(zram, wb_ctl, req);
 		blk_idx = 0;
-		atomic64_inc(&zram->stats.pages_stored);
-		spin_lock(&zram->wb_limit_lock);
-		if (zram->wb_limit_enable && zram->bd_wb_limit > 0)
-			zram->bd_wb_limit -=  1UL << (PAGE_SHIFT - 12);
-		spin_unlock(&zram->wb_limit_lock);
+		req = NULL;
+		continue;
+
 next:
 		zram_slot_unlock(zram, index);
 		release_pp_slot(zram, pps);
-
 		cond_resched();
 	}
 
-	if (blk_idx)
-		free_block_bdev(zram, blk_idx);
-	if (page)
-		__free_page(page);
+	/*
+	 * Selected idle req, but never submitted it due to some error or
+	 * wb limit.
+	 */
+	if (req)
+		release_wb_req(req);
+
+	blk_finish_plug(&io_plug);
+	err = zram_wb_wait_for_completion(zram, wb_ctl);
+	if (err)
+		ret = err;
 
 	return ret;
 }
@@ -948,7 +1154,8 @@ static ssize_t writeback_store(struct device *dev,
 	struct zram *zram = dev_to_zram(dev);
 	u64 nr_pages = zram->disksize >> PAGE_SHIFT;
 	unsigned long lo = 0, hi = nr_pages;
-	struct zram_pp_ctl *ctl = NULL;
+	struct zram_pp_ctl *pp_ctl = NULL;
+	struct zram_wb_ctl *wb_ctl = NULL;
 	char *args, *param, *val;
 	ssize_t ret = len;
 	int err, mode = 0;
@@ -970,8 +1177,14 @@ static ssize_t writeback_store(struct device *dev,
 		goto release_init_lock;
 	}
 
-	ctl = init_pp_ctl();
-	if (!ctl) {
+	pp_ctl = init_pp_ctl();
+	if (!pp_ctl) {
+		ret = -ENOMEM;
+		goto release_init_lock;
+	}
+
+	wb_ctl = init_wb_ctl();
+	if (!wb_ctl) {
 		ret = -ENOMEM;
 		goto release_init_lock;
 	}
@@ -1000,7 +1213,7 @@ static ssize_t writeback_store(struct device *dev,
 				goto release_init_lock;
 			}
 
-			scan_slots_for_writeback(zram, mode, lo, hi, ctl);
+			scan_slots_for_writeback(zram, mode, lo, hi, pp_ctl);
 			break;
 		}
 
@@ -1011,7 +1224,7 @@ static ssize_t writeback_store(struct device *dev,
 				goto release_init_lock;
 			}
 
-			scan_slots_for_writeback(zram, mode, lo, hi, ctl);
+			scan_slots_for_writeback(zram, mode, lo, hi, pp_ctl);
 			break;
 		}
 
@@ -1022,7 +1235,7 @@ static ssize_t writeback_store(struct device *dev,
 				goto release_init_lock;
 			}
 
-			scan_slots_for_writeback(zram, mode, lo, hi, ctl);
+			scan_slots_for_writeback(zram, mode, lo, hi, pp_ctl);
 			continue;
 		}
 
@@ -1033,17 +1246,18 @@ static ssize_t writeback_store(struct device *dev,
 				goto release_init_lock;
 			}
 
-			scan_slots_for_writeback(zram, mode, lo, hi, ctl);
+			scan_slots_for_writeback(zram, mode, lo, hi, pp_ctl);
 			continue;
 		}
 	}
 
-	err = zram_writeback_slots(zram, ctl);
+	err = zram_writeback_slots(zram, pp_ctl, wb_ctl);
 	if (err)
 		ret = err;
 
 release_init_lock:
-	release_pp_ctl(zram, ctl);
+	release_pp_ctl(zram, pp_ctl);
+	release_wb_ctl(wb_ctl);
 	atomic_set(&zram->pp_in_progress, 0);
 	up_read(&zram->init_lock);
 
@@ -1112,7 +1326,9 @@ static int read_from_bdev(struct zram *zram, struct page *page,
 	return -EIO;
 }
 
-static void free_block_bdev(struct zram *zram, unsigned long blk_idx) {};
+static void free_block_bdev(struct zram *zram, unsigned long blk_idx)
+{
+}
 #endif
 
 #ifdef CONFIG_ZRAM_MEMORY_TRACKING
-- 
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 2/6] zram: add writeback batch size device attr

[Sergey Senozhatsky]

Introduce writeback_batch_size device attribute so that
the maximum number of in-flight writeback bio requests
can be configured at run-time per-device.  This essentially
enables batched bio writeback.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 48 ++++++++++++++++++++++++++++++-----
 drivers/block/zram/zram_drv.h |  1 +
 2 files changed, 43 insertions(+), 6 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index ea06f4d7b623..be39fe04b9b1 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -588,6 +588,42 @@ static ssize_t writeback_limit_show(struct device *dev,
 	return sysfs_emit(buf, "%llu\n", val);
 }
 
+static ssize_t writeback_batch_size_store(struct device *dev,
+					  struct device_attribute *attr,
+					  const char *buf, size_t len)
+{
+	struct zram *zram = dev_to_zram(dev);
+	u32 val;
+	ssize_t ret = -EINVAL;
+
+	if (kstrtouint(buf, 10, &val))
+		return ret;
+
+	if (!val)
+		val = 1;
+
+	down_read(&zram->init_lock);
+	zram->wb_batch_size = val;
+	up_read(&zram->init_lock);
+	ret = len;
+
+	return ret;
+}
+
+static ssize_t writeback_batch_size_show(struct device *dev,
+					 struct device_attribute *attr,
+					 char *buf)
+{
+	u32 val;
+	struct zram *zram = dev_to_zram(dev);
+
+	down_read(&zram->init_lock);
+	val = zram->wb_batch_size;
+	up_read(&zram->init_lock);
+
+	return sysfs_emit(buf, "%u\n", val);
+}
+
 static void reset_bdev(struct zram *zram)
 {
 	if (!zram->backing_dev)
@@ -775,10 +811,7 @@ static void release_wb_ctl(struct zram_wb_ctl *wb_ctl)
 	kfree(wb_ctl);
 }
 
-/* XXX: should be a per-device sysfs attr */
-#define ZRAM_WB_REQ_CNT 32
-
-static struct zram_wb_ctl *init_wb_ctl(void)
+static struct zram_wb_ctl *init_wb_ctl(struct zram *zram)
 {
 	struct zram_wb_ctl *wb_ctl;
 	int i;
@@ -792,7 +825,7 @@ static struct zram_wb_ctl *init_wb_ctl(void)
 	atomic_set(&wb_ctl->num_inflight, 0);
 	init_completion(&wb_ctl->done);
 
-	for (i = 0; i < ZRAM_WB_REQ_CNT; i++) {
+	for (i = 0; i < zram->wb_batch_size; i++) {
 		struct zram_wb_req *req;
 
 		/*
@@ -1183,7 +1216,7 @@ static ssize_t writeback_store(struct device *dev,
 		goto release_init_lock;
 	}
 
-	wb_ctl = init_wb_ctl();
+	wb_ctl = init_wb_ctl(zram);
 	if (!wb_ctl) {
 		ret = -ENOMEM;
 		goto release_init_lock;
@@ -2826,6 +2859,7 @@ static DEVICE_ATTR_RW(backing_dev);
 static DEVICE_ATTR_WO(writeback);
 static DEVICE_ATTR_RW(writeback_limit);
 static DEVICE_ATTR_RW(writeback_limit_enable);
+static DEVICE_ATTR_RW(writeback_batch_size);
 #endif
 #ifdef CONFIG_ZRAM_MULTI_COMP
 static DEVICE_ATTR_RW(recomp_algorithm);
@@ -2847,6 +2881,7 @@ static struct attribute *zram_disk_attrs[] = {
 	&dev_attr_writeback.attr,
 	&dev_attr_writeback_limit.attr,
 	&dev_attr_writeback_limit_enable.attr,
+	&dev_attr_writeback_batch_size.attr,
 #endif
 	&dev_attr_io_stat.attr,
 	&dev_attr_mm_stat.attr,
@@ -2908,6 +2943,7 @@ static int zram_add(void)
 
 	init_rwsem(&zram->init_lock);
 #ifdef CONFIG_ZRAM_WRITEBACK
+	zram->wb_batch_size = 32;
 	spin_lock_init(&zram->wb_limit_lock);
 #endif
 
diff --git a/drivers/block/zram/zram_drv.h b/drivers/block/zram/zram_drv.h
index 6cee93f9c0d0..1a647f42c1a4 100644
--- a/drivers/block/zram/zram_drv.h
+++ b/drivers/block/zram/zram_drv.h
@@ -129,6 +129,7 @@ struct zram {
 	struct file *backing_dev;
 	spinlock_t wb_limit_lock;
 	bool wb_limit_enable;
+	u32 wb_batch_size;
 	u64 bd_wb_limit;
 	struct block_device *bdev;
 	unsigned long *bitmap;
-- 
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 3/6] zram: take write lock in wb limit store handlers

[Sergey Senozhatsky]

Write device attrs handlers should take write zram init_lock.
While at it, fixup coding styles.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index be39fe04b9b1..073a12132cb3 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -519,7 +519,8 @@ struct zram_wb_req {
 };
 
 static ssize_t writeback_limit_enable_store(struct device *dev,
-		struct device_attribute *attr, const char *buf, size_t len)
+					    struct device_attribute *attr,
+					    const char *buf, size_t len)
 {
 	struct zram *zram = dev_to_zram(dev);
 	u64 val;
@@ -528,18 +529,19 @@ static ssize_t writeback_limit_enable_store(struct device *dev,
 	if (kstrtoull(buf, 10, &val))
 		return ret;
 
-	down_read(&zram->init_lock);
+	down_write(&zram->init_lock);
 	spin_lock(&zram->wb_limit_lock);
 	zram->wb_limit_enable = val;
 	spin_unlock(&zram->wb_limit_lock);
-	up_read(&zram->init_lock);
+	up_write(&zram->init_lock);
 	ret = len;
 
 	return ret;
 }
 
 static ssize_t writeback_limit_enable_show(struct device *dev,
-		struct device_attribute *attr, char *buf)
+					   struct device_attribute *attr,
+					   char *buf)
 {
 	bool val;
 	struct zram *zram = dev_to_zram(dev);
@@ -554,7 +556,8 @@ static ssize_t writeback_limit_enable_show(struct device *dev,
 }
 
 static ssize_t writeback_limit_store(struct device *dev,
-		struct device_attribute *attr, const char *buf, size_t len)
+				     struct device_attribute *attr,
+				     const char *buf, size_t len)
 {
 	struct zram *zram = dev_to_zram(dev);
 	u64 val;
@@ -563,11 +566,11 @@ static ssize_t writeback_limit_store(struct device *dev,
 	if (kstrtoull(buf, 10, &val))
 		return ret;
 
-	down_read(&zram->init_lock);
+	down_write(&zram->init_lock);
 	spin_lock(&zram->wb_limit_lock);
 	zram->bd_wb_limit = val;
 	spin_unlock(&zram->wb_limit_lock);
-	up_read(&zram->init_lock);
+	up_write(&zram->init_lock);
 	ret = len;
 
 	return ret;
-- 
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 4/6] zram: drop wb_limit_lock

[Sergey Senozhatsky]

We don't need wb_limit_lock.  Writeback limit setters take an
exclusive write zram init_lock, while wb_limit modifications
happen only from a single task and under zram read init_lock.
No concurrent wb_limit modifications are possible (we permit
only one post-processing task at a time).  Add lockdep
assertions to wb_limit mutators.

While at it, fixup coding styles.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 22 +++++-----------------
 drivers/block/zram/zram_drv.h |  1 -
 2 files changed, 5 insertions(+), 18 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 073a12132cb3..1cfb58516a8e 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -530,9 +530,7 @@ static ssize_t writeback_limit_enable_store(struct device *dev,
 		return ret;
 
 	down_write(&zram->init_lock);
-	spin_lock(&zram->wb_limit_lock);
 	zram->wb_limit_enable = val;
-	spin_unlock(&zram->wb_limit_lock);
 	up_write(&zram->init_lock);
 	ret = len;
 
@@ -547,9 +545,7 @@ static ssize_t writeback_limit_enable_show(struct device *dev,
 	struct zram *zram = dev_to_zram(dev);
 
 	down_read(&zram->init_lock);
-	spin_lock(&zram->wb_limit_lock);
 	val = zram->wb_limit_enable;
-	spin_unlock(&zram->wb_limit_lock);
 	up_read(&zram->init_lock);
 
 	return sysfs_emit(buf, "%d\n", val);
@@ -567,9 +563,7 @@ static ssize_t writeback_limit_store(struct device *dev,
 		return ret;
 
 	down_write(&zram->init_lock);
-	spin_lock(&zram->wb_limit_lock);
 	zram->bd_wb_limit = val;
-	spin_unlock(&zram->wb_limit_lock);
 	up_write(&zram->init_lock);
 	ret = len;
 
@@ -577,15 +571,13 @@ static ssize_t writeback_limit_store(struct device *dev,
 }
 
 static ssize_t writeback_limit_show(struct device *dev,
-		struct device_attribute *attr, char *buf)
+				    struct device_attribute *attr, char *buf)
 {
 	u64 val;
 	struct zram *zram = dev_to_zram(dev);
 
 	down_read(&zram->init_lock);
-	spin_lock(&zram->wb_limit_lock);
 	val = zram->bd_wb_limit;
-	spin_unlock(&zram->wb_limit_lock);
 	up_read(&zram->init_lock);
 
 	return sysfs_emit(buf, "%llu\n", val);
@@ -864,18 +856,18 @@ static struct zram_wb_ctl *init_wb_ctl(struct zram *zram)
 
 static void zram_account_writeback_rollback(struct zram *zram)
 {
-	spin_lock(&zram->wb_limit_lock);
+	lockdep_assert_held_read(&zram->init_lock);
+
 	if (zram->wb_limit_enable)
 		zram->bd_wb_limit +=  1UL << (PAGE_SHIFT - 12);
-	spin_unlock(&zram->wb_limit_lock);
 }
 
 static void zram_account_writeback_submit(struct zram *zram)
 {
-	spin_lock(&zram->wb_limit_lock);
+	lockdep_assert_held_read(&zram->init_lock);
+
 	if (zram->wb_limit_enable && zram->bd_wb_limit > 0)
 		zram->bd_wb_limit -=  1UL << (PAGE_SHIFT - 12);
-	spin_unlock(&zram->wb_limit_lock);
 }
 
 static int zram_writeback_complete(struct zram *zram, struct zram_wb_req *req)
@@ -992,13 +984,10 @@ static int zram_writeback_slots(struct zram *zram,
 
 	blk_start_plug(&io_plug);
 	while ((pps = select_pp_slot(ctl))) {
-		spin_lock(&zram->wb_limit_lock);
 		if (zram->wb_limit_enable && !zram->bd_wb_limit) {
-			spin_unlock(&zram->wb_limit_lock);
 			ret = -EIO;
 			break;
 		}
-		spin_unlock(&zram->wb_limit_lock);
 
 		while (!req) {
 			req = select_idle_req(wb_ctl);
@@ -2947,7 +2936,6 @@ static int zram_add(void)
 	init_rwsem(&zram->init_lock);
 #ifdef CONFIG_ZRAM_WRITEBACK
 	zram->wb_batch_size = 32;
-	spin_lock_init(&zram->wb_limit_lock);
 #endif
 
 	/* gendisk structure */
diff --git a/drivers/block/zram/zram_drv.h b/drivers/block/zram/zram_drv.h
index 1a647f42c1a4..c6d94501376c 100644
--- a/drivers/block/zram/zram_drv.h
+++ b/drivers/block/zram/zram_drv.h
@@ -127,7 +127,6 @@ struct zram {
 	bool claim; /* Protected by disk->open_mutex */
 #ifdef CONFIG_ZRAM_WRITEBACK
 	struct file *backing_dev;
-	spinlock_t wb_limit_lock;
 	bool wb_limit_enable;
 	u32 wb_batch_size;
 	u64 bd_wb_limit;
-- 
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 5/6] zram: rework bdev block allocation

[Sergey Senozhatsky]

First, writeback bdev ->bitmap bits are set only from one
context, as we can have only one single task performing
writeback, so we cannot race with anything else.  Remove
retry path.

Second, we always check ZRAM_WB flag to distinguish writtenback
slots, so we should not confuse 0 bdev block index and 0 handle.
We can use first bdev block (0 bit) for writeback as well.

While at it, give functions slightly more accurate names, as
we don't alloc/free anything there, we reserve a block for
async writeback or release the block.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 37 +++++++++++++++++------------------
 1 file changed, 18 insertions(+), 19 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 1cfb58516a8e..93365811781b 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -500,6 +500,8 @@ static ssize_t idle_store(struct device *dev,
 }
 
 #ifdef CONFIG_ZRAM_WRITEBACK
+#define INVALID_BDEV_BLOCK		(~0UL)
+
 struct zram_wb_ctl {
 	struct list_head idle_reqs;
 	struct list_head inflight_reqs;
@@ -746,23 +748,20 @@ static ssize_t backing_dev_store(struct device *dev,
 	return err;
 }
 
-static unsigned long alloc_block_bdev(struct zram *zram)
+static unsigned long zram_reserve_bdev_block(struct zram *zram)
 {
-	unsigned long blk_idx = 1;
-retry:
-	/* skip 0 bit to confuse zram.handle = 0 */
-	blk_idx = find_next_zero_bit(zram->bitmap, zram->nr_pages, blk_idx);
-	if (blk_idx == zram->nr_pages)
-		return 0;
+	unsigned long blk_idx;
 
-	if (test_and_set_bit(blk_idx, zram->bitmap))
-		goto retry;
+	blk_idx = find_next_zero_bit(zram->bitmap, zram->nr_pages, 0);
+	if (blk_idx == zram->nr_pages)
+		return INVALID_BDEV_BLOCK;
 
+	set_bit(blk_idx, zram->bitmap);
 	atomic64_inc(&zram->stats.bd_count);
 	return blk_idx;
 }
 
-static void free_block_bdev(struct zram *zram, unsigned long blk_idx)
+static void zram_release_bdev_block(struct zram *zram, unsigned long blk_idx)
 {
 	int was_set;
 
@@ -882,7 +881,7 @@ static int zram_writeback_complete(struct zram *zram, struct zram_wb_req *req)
 		 * (if enabled).
 		 */
 		zram_account_writeback_rollback(zram);
-		free_block_bdev(zram, req->blk_idx);
+		zram_release_bdev_block(zram, req->blk_idx);
 		return err;
 	}
 
@@ -896,7 +895,7 @@ static int zram_writeback_complete(struct zram *zram, struct zram_wb_req *req)
 	 * finishes.
 	 */
 	if (!zram_test_flag(zram, index, ZRAM_PP_SLOT)) {
-		free_block_bdev(zram, req->blk_idx);
+		zram_release_bdev_block(zram, req->blk_idx);
 		goto out;
 	}
 
@@ -975,8 +974,8 @@ static int zram_writeback_slots(struct zram *zram,
 				struct zram_pp_ctl *ctl,
 				struct zram_wb_ctl *wb_ctl)
 {
+	unsigned long blk_idx = INVALID_BDEV_BLOCK;
 	struct zram_wb_req *req = NULL;
-	unsigned long blk_idx = 0;
 	struct zram_pp_slot *pps;
 	struct blk_plug io_plug;
 	int ret = 0, err;
@@ -1009,9 +1008,9 @@ static int zram_writeback_slots(struct zram *zram,
 				ret = err;
 		}
 
-		if (!blk_idx) {
-			blk_idx = alloc_block_bdev(zram);
-			if (!blk_idx) {
+		if (blk_idx == INVALID_BDEV_BLOCK) {
+			blk_idx = zram_reserve_bdev_block(zram);
+			if (blk_idx == INVALID_BDEV_BLOCK) {
 				ret = -ENOSPC;
 				break;
 			}
@@ -1046,7 +1045,7 @@ static int zram_writeback_slots(struct zram *zram,
 		__bio_add_page(&req->bio, req->page, PAGE_SIZE, 0);
 
 		zram_submit_wb_request(zram, wb_ctl, req);
-		blk_idx = 0;
+		blk_idx = INVALID_BDEV_BLOCK;
 		req = NULL;
 		continue;
 
@@ -1351,7 +1350,7 @@ static int read_from_bdev(struct zram *zram, struct page *page,
 	return -EIO;
 }
 
-static void free_block_bdev(struct zram *zram, unsigned long blk_idx)
+static void zram_release_bdev_block(struct zram *zram, unsigned long blk_idx)
 {
 }
 #endif
@@ -1875,7 +1874,7 @@ static void zram_free_page(struct zram *zram, size_t index)
 
 	if (zram_test_flag(zram, index, ZRAM_WB)) {
 		zram_clear_flag(zram, index, ZRAM_WB);
-		free_block_bdev(zram, zram_get_handle(zram, index));
+		zram_release_bdev_block(zram, zram_get_handle(zram, index));
 		goto out;
 	}
 
-- 
2.52.0.rc1.455.g30608eb744-goog

[PATCHv4 6/6] zram: read slot block idx under slot lock

[Sergey Senozhatsky]

Read slot's block id under slot-lock.  We release the
slot-lock for bdev read so, technically, slot still can
get freed in the meantime, but at least we will read
bdev block (page) that holds previous know slot data,
not from slot->handle bdev block, which can be anything
at that point.

Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
 drivers/block/zram/zram_drv.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 93365811781b..77d0e636e053 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -1980,14 +1980,14 @@ static int zram_read_page(struct zram *zram, struct page *page, u32 index,
 		ret = zram_read_from_zspool(zram, page, index);
 		zram_slot_unlock(zram, index);
 	} else {
+		unsigned long blk_idx = zram_get_handle(zram, index);
+
 		/*
 		 * The slot should be unlocked before reading from the backing
 		 * device.
 		 */
 		zram_slot_unlock(zram, index);
-
-		ret = read_from_bdev(zram, page, zram_get_handle(zram, index),
-				     parent);
+		ret = read_from_bdev(zram, page, blk_idx, parent);
 	}
 
 	/* Should NEVER happen. Return bio error if it does. */
-- 
2.52.0.rc1.455.g30608eb744-goog

Re: [PATCHv4 2/6] zram: add writeback batch size device attr

[Andrew Morton]

On Tue, 18 Nov 2025 16:29:56 +0900 Sergey Senozhatsky <senozhatsky@chromium.org> wrote:

> Introduce writeback_batch_size device attribute so that
> the maximum number of in-flight writeback bio requests
> can be configured at run-time per-device.  This essentially
> enables batched bio writeback.
> 
> ...
>
> --- a/drivers/block/zram/zram_drv.c
> +++ b/drivers/block/zram/zram_drv.c
> @@ -588,6 +588,42 @@ static ssize_t writeback_limit_show(struct device *dev,
>  	return sysfs_emit(buf, "%llu\n", val);
>  }
>  
> +static ssize_t writeback_batch_size_store(struct device *dev,
> +					  struct device_attribute *attr,
> +					  const char *buf, size_t len)
> +{
> +	struct zram *zram = dev_to_zram(dev);
> +	u32 val;
> +	ssize_t ret = -EINVAL;
> +
> +	if (kstrtouint(buf, 10, &val))
> +		return ret;
> +
> +	if (!val)
> +		val = 1;
> +
> +	down_read(&zram->init_lock);

down_write()?

> +	zram->wb_batch_size = val;
> +	up_read(&zram->init_lock);
> +	ret = len;
> +
> +	return ret;
> +}

Re: [PATCHv4 2/6] zram: add writeback batch size device attr

[Sergey Senozhatsky]

On (25/11/20 17:17), Andrew Morton wrote:
> > +static ssize_t writeback_batch_size_store(struct device *dev,
> > +					  struct device_attribute *attr,
> > +					  const char *buf, size_t len)
> > +{
> > +	struct zram *zram = dev_to_zram(dev);
> > +	u32 val;
> > +	ssize_t ret = -EINVAL;
> > +
> > +	if (kstrtouint(buf, 10, &val))
> > +		return ret;
> > +
> > +	if (!val)
> > +		val = 1;
> > +
> > +	down_read(&zram->init_lock);
> 
> down_write()?

Indeed.  Thanks!