zloop fixes

zloop fixes

[Christoph Hellwig]

Hi all,

this series fixes a bad bug and a few minor issues in the new zloop code
queued up for 7.1.

Diffstat:
 zloop.c |  123 ++++++++++++++++++++++++++++++----------------------------------
 1 file changed, 59 insertions(+), 64 deletions(-)

[PATCH 1/6] zloop: fix write pointer calculation in zloop_forget_cache

[Christoph Hellwig]

The write pointer is absolute and in sector units, so we need to
convert it to a relative byte address first.

Fixes: c505448748f7 ("zloop: forget write cache on force removal")
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index 8baf642037fd..672948d3653e 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -1401,8 +1401,17 @@ static void zloop_forget_cache(struct zloop_device *zlo)
 				zlo->disk->part0, ret);
 			continue;
 		}
-		if (old_wp < zone->wp)
-			zloop_truncate(file, old_wp);
+
+		if (old_wp > zone->wp)
+			continue;
+		/*
+		 * This should not happen, if we recored a full zone, it can't
+		 * be active.
+		 */
+		if (WARN_ON_ONCE(old_wp == ULLONG_MAX))
+			continue;
+
+		zloop_truncate(file, (old_wp - zone->start) << SECTOR_SHIFT);
 	}
 }
 
-- 
2.53.0

Re: [PATCH 1/6] zloop: fix write pointer calculation in zloop_forget_cache

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> The write pointer is absolute and in sector units, so we need to
> convert it to a relative byte address first.
> 
> Fixes: c505448748f7 ("zloop: forget write cache on force removal")
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks good.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

[PATCH 2/6] zloop: use vfs_truncate

[Christoph Hellwig]

While vfs_truncate does various extra checks that we don't really need,
it is always better to use a VFS helper rather than open coding the
logic.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 17 ++---------------
 1 file changed, 2 insertions(+), 15 deletions(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index 672948d3653e..fc54480ed62d 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -1363,20 +1363,6 @@ static int zloop_ctl_add(struct zloop_options *opts)
 	return ret;
 }
 
-static void zloop_truncate(struct file *file, loff_t pos)
-{
-	struct mnt_idmap *idmap = file_mnt_idmap(file);
-	struct dentry *dentry = file_dentry(file);
-	struct iattr newattrs;
-
-	newattrs.ia_size = pos;
-	newattrs.ia_valid = ATTR_SIZE;
-
-	inode_lock(dentry->d_inode);
-	notify_change(idmap, dentry, &newattrs, NULL);
-	inode_unlock(dentry->d_inode);
-}
-
 static void zloop_forget_cache(struct zloop_device *zlo)
 {
 	unsigned int i;
@@ -1411,7 +1397,8 @@ static void zloop_forget_cache(struct zloop_device *zlo)
 		if (WARN_ON_ONCE(old_wp == ULLONG_MAX))
 			continue;
 
-		zloop_truncate(file, (old_wp - zone->start) << SECTOR_SHIFT);
+		vfs_truncate(&file->f_path,
+			(old_wp - zone->start) << SECTOR_SHIFT);
 	}
 }
 
-- 
2.53.0

Re: [PATCH 2/6] zloop: use vfs_truncate

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> While vfs_truncate does various extra checks that we don't really need,
> it is always better to use a VFS helper rather than open coding the
> logic.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks good.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

[PATCH 3/6] zloop: improve the unaligned write pointer warning

[Christoph Hellwig]

Use the IS_ALIGNED helper and avoid extra conversions, and tell the
user what the unaligned size is.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index fc54480ed62d..a738f8629062 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -313,9 +313,9 @@ static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 		return -EINVAL;
 	}
 
-	if (file_sectors & ((zlo->block_size >> SECTOR_SHIFT) - 1)) {
-		pr_err("Zone %u file size not aligned to block size %u\n",
-		       zone_no, zlo->block_size);
+	if (!IS_ALIGNED(stat.size, zlo->block_size)) {
+		pr_err("Zone %u file size (%llu) not aligned to block size %u\n",
+		       zone_no, stat.size, zlo->block_size);
 		return -EINVAL;
 	}
 
-- 
2.53.0

Re: [PATCH 3/6] zloop: improve the unaligned write pointer warning

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> Use the IS_ALIGNED helper and avoid extra conversions, and tell the
> user what the unaligned size is.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks good.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

[PATCH 4/6] zloop: set RQF_QUIET when completing requests on deleted devices

[Christoph Hellwig]

Reduce the dmesg spam for tests that involve device deletion.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index a738f8629062..7257188dd3a8 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -891,8 +891,10 @@ static blk_status_t zloop_queue_rq(struct blk_mq_hw_ctx *hctx,
 	struct zloop_cmd *cmd = blk_mq_rq_to_pdu(rq);
 	struct zloop_device *zlo = rq->q->queuedata;
 
-	if (data_race(READ_ONCE(zlo->state)) == Zlo_deleting)
+	if (data_race(READ_ONCE(zlo->state)) == Zlo_deleting) {
+		rq->rq_flags |= RQF_QUIET;
 		return BLK_STS_IOERR;
+	}
 
 	/*
 	 * If we need to strongly order zone append operations, set the request
-- 
2.53.0

Re: [PATCH 4/6] zloop: set RQF_QUIET when completing requests on deleted devices

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> Reduce the dmesg spam for tests that involve device deletion.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks good.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

[PATCH 5/6] zloop: factor out zloop_mark_{full,empty} helpers

[Christoph Hellwig]

Move a few chunks of duplicated code into helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 48 +++++++++++++++++++++++--------------------
 1 file changed, 26 insertions(+), 22 deletions(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index 7257188dd3a8..3f69206d674a 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -288,6 +288,24 @@ static bool zloop_do_open_zone(struct zloop_device *zlo,
 	}
 }
 
+static void zloop_mark_full(struct zloop_device *zlo, struct zloop_zone *zone)
+{
+	lockdep_assert_held(&zone->wp_lock);
+
+	zloop_lru_remove_open_zone(zlo, zone);
+	zone->cond = BLK_ZONE_COND_FULL;
+	zone->wp = ULLONG_MAX;
+}
+
+static void zloop_mark_empty(struct zloop_device *zlo, struct zloop_zone *zone)
+{
+	lockdep_assert_held(&zone->wp_lock);
+
+	zloop_lru_remove_open_zone(zlo, zone);
+	zone->cond = BLK_ZONE_COND_EMPTY;
+	zone->wp = zone->start;
+}
+
 static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 {
 	struct zloop_zone *zone = &zlo->zones[zone_no];
@@ -321,13 +339,9 @@ static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 
 	spin_lock_irqsave(&zone->wp_lock, flags);
 	if (!file_sectors) {
-		zloop_lru_remove_open_zone(zlo, zone);
-		zone->cond = BLK_ZONE_COND_EMPTY;
-		zone->wp = zone->start;
+		zloop_mark_empty(zlo, zone);
 	} else if (file_sectors == zlo->zone_capacity) {
-		zloop_lru_remove_open_zone(zlo, zone);
-		zone->cond = BLK_ZONE_COND_FULL;
-		zone->wp = ULLONG_MAX;
+		zloop_mark_full(zlo, zone);
 	} else {
 		if (zone->cond != BLK_ZONE_COND_IMP_OPEN &&
 		    zone->cond != BLK_ZONE_COND_EXP_OPEN)
@@ -429,9 +443,7 @@ static int zloop_reset_zone(struct zloop_device *zlo, unsigned int zone_no)
 	}
 
 	spin_lock_irqsave(&zone->wp_lock, flags);
-	zloop_lru_remove_open_zone(zlo, zone);
-	zone->cond = BLK_ZONE_COND_EMPTY;
-	zone->wp = zone->start;
+	zloop_mark_empty(zlo, zone);
 	clear_bit(ZLOOP_ZONE_SEQ_ERROR, &zone->flags);
 	spin_unlock_irqrestore(&zone->wp_lock, flags);
 
@@ -477,9 +489,7 @@ static int zloop_finish_zone(struct zloop_device *zlo, unsigned int zone_no)
 	}
 
 	spin_lock_irqsave(&zone->wp_lock, flags);
-	zloop_lru_remove_open_zone(zlo, zone);
-	zone->cond = BLK_ZONE_COND_FULL;
-	zone->wp = ULLONG_MAX;
+	zloop_mark_full(zlo, zone);
 	clear_bit(ZLOOP_ZONE_SEQ_ERROR, &zone->flags);
 	spin_unlock_irqrestore(&zone->wp_lock, flags);
 
@@ -616,11 +626,8 @@ static int zloop_seq_write_prep(struct zloop_cmd *cmd)
 	 */
 	if (!is_append || !zlo->ordered_zone_append) {
 		zone->wp += nr_sectors;
-		if (zone->wp == zone_end) {
-			zloop_lru_remove_open_zone(zlo, zone);
-			zone->cond = BLK_ZONE_COND_FULL;
-			zone->wp = ULLONG_MAX;
-		}
+		if (zone->wp == zone_end)
+			zloop_mark_full(zlo, zone);
 	}
 out_unlock:
 	spin_unlock_irqrestore(&zone->wp_lock, flags);
@@ -873,11 +880,8 @@ static bool zloop_set_zone_append_sector(struct request *rq)
 
 	rq->__sector = zone->wp;
 	zone->wp += blk_rq_sectors(rq);
-	if (zone->wp >= zone_end) {
-		zloop_lru_remove_open_zone(zlo, zone);
-		zone->cond = BLK_ZONE_COND_FULL;
-		zone->wp = ULLONG_MAX;
-	}
+	if (zone->wp >= zone_end)
+		zloop_mark_full(zlo, zone);
 
 	spin_unlock_irqrestore(&zone->wp_lock, flags);
 
-- 
2.53.0

Re: [PATCH 5/6] zloop: factor out zloop_mark_{full,empty} helpers

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> Move a few chunks of duplicated code into helpers.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Nice cleanup !

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

[PATCH 6/6] zloop: remove irq-safe locking

[Christoph Hellwig]

All of zloop runs in user context, so drop the irq-safe locking.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zloop.c | 37 +++++++++++++++----------------------
 1 file changed, 15 insertions(+), 22 deletions(-)

diff --git a/drivers/block/zloop.c b/drivers/block/zloop.c
index 3f69206d674a..55eeb6aac0ea 100644
--- a/drivers/block/zloop.c
+++ b/drivers/block/zloop.c
@@ -311,7 +311,6 @@ static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 	struct zloop_zone *zone = &zlo->zones[zone_no];
 	struct kstat stat;
 	sector_t file_sectors;
-	unsigned long flags;
 	int ret;
 
 	lockdep_assert_held(&zone->lock);
@@ -337,7 +336,7 @@ static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 		return -EINVAL;
 	}
 
-	spin_lock_irqsave(&zone->wp_lock, flags);
+	spin_lock(&zone->wp_lock);
 	if (!file_sectors) {
 		zloop_mark_empty(zlo, zone);
 	} else if (file_sectors == zlo->zone_capacity) {
@@ -348,7 +347,7 @@ static int zloop_update_seq_zone(struct zloop_device *zlo, unsigned int zone_no)
 			zone->cond = BLK_ZONE_COND_CLOSED;
 		zone->wp = zone->start + file_sectors;
 	}
-	spin_unlock_irqrestore(&zone->wp_lock, flags);
+	spin_unlock(&zone->wp_lock);
 
 	return 0;
 }
@@ -381,7 +380,6 @@ static int zloop_open_zone(struct zloop_device *zlo, unsigned int zone_no)
 static int zloop_close_zone(struct zloop_device *zlo, unsigned int zone_no)
 {
 	struct zloop_zone *zone = &zlo->zones[zone_no];
-	unsigned long flags;
 	int ret = 0;
 
 	if (test_bit(ZLOOP_ZONE_CONV, &zone->flags))
@@ -400,13 +398,13 @@ static int zloop_close_zone(struct zloop_device *zlo, unsigned int zone_no)
 		break;
 	case BLK_ZONE_COND_IMP_OPEN:
 	case BLK_ZONE_COND_EXP_OPEN:
-		spin_lock_irqsave(&zone->wp_lock, flags);
+		spin_lock(&zone->wp_lock);
 		zloop_lru_remove_open_zone(zlo, zone);
 		if (zone->wp == zone->start)
 			zone->cond = BLK_ZONE_COND_EMPTY;
 		else
 			zone->cond = BLK_ZONE_COND_CLOSED;
-		spin_unlock_irqrestore(&zone->wp_lock, flags);
+		spin_unlock(&zone->wp_lock);
 		break;
 	case BLK_ZONE_COND_EMPTY:
 	case BLK_ZONE_COND_FULL:
@@ -424,7 +422,6 @@ static int zloop_close_zone(struct zloop_device *zlo, unsigned int zone_no)
 static int zloop_reset_zone(struct zloop_device *zlo, unsigned int zone_no)
 {
 	struct zloop_zone *zone = &zlo->zones[zone_no];
-	unsigned long flags;
 	int ret = 0;
 
 	if (test_bit(ZLOOP_ZONE_CONV, &zone->flags))
@@ -442,10 +439,10 @@ static int zloop_reset_zone(struct zloop_device *zlo, unsigned int zone_no)
 		goto unlock;
 	}
 
-	spin_lock_irqsave(&zone->wp_lock, flags);
+	spin_lock(&zone->wp_lock);
 	zloop_mark_empty(zlo, zone);
 	clear_bit(ZLOOP_ZONE_SEQ_ERROR, &zone->flags);
-	spin_unlock_irqrestore(&zone->wp_lock, flags);
+	spin_unlock(&zone->wp_lock);
 
 unlock:
 	mutex_unlock(&zone->lock);
@@ -470,7 +467,6 @@ static int zloop_reset_all_zones(struct zloop_device *zlo)
 static int zloop_finish_zone(struct zloop_device *zlo, unsigned int zone_no)
 {
 	struct zloop_zone *zone = &zlo->zones[zone_no];
-	unsigned long flags;
 	int ret = 0;
 
 	if (test_bit(ZLOOP_ZONE_CONV, &zone->flags))
@@ -488,10 +484,10 @@ static int zloop_finish_zone(struct zloop_device *zlo, unsigned int zone_no)
 		goto unlock;
 	}
 
-	spin_lock_irqsave(&zone->wp_lock, flags);
+	spin_lock(&zone->wp_lock);
 	zloop_mark_full(zlo, zone);
 	clear_bit(ZLOOP_ZONE_SEQ_ERROR, &zone->flags);
-	spin_unlock_irqrestore(&zone->wp_lock, flags);
+	spin_unlock(&zone->wp_lock);
 
  unlock:
 	mutex_unlock(&zone->lock);
@@ -581,10 +577,9 @@ static int zloop_seq_write_prep(struct zloop_cmd *cmd)
 	bool is_append = req_op(rq) == REQ_OP_ZONE_APPEND;
 	struct zloop_zone *zone = &zlo->zones[zone_no];
 	sector_t zone_end = zone->start + zlo->zone_capacity;
-	unsigned long flags;
 	int ret = 0;
 
-	spin_lock_irqsave(&zone->wp_lock, flags);
+	spin_lock(&zone->wp_lock);
 
 	/*
 	 * Zone append operations always go at the current write pointer, but
@@ -630,7 +625,7 @@ static int zloop_seq_write_prep(struct zloop_cmd *cmd)
 			zloop_mark_full(zlo, zone);
 	}
 out_unlock:
-	spin_unlock_irqrestore(&zone->wp_lock, flags);
+	spin_unlock(&zone->wp_lock);
 	return ret;
 }
 
@@ -868,13 +863,12 @@ static bool zloop_set_zone_append_sector(struct request *rq)
 	struct zloop_zone *zone = &zlo->zones[zone_no];
 	sector_t zone_end = zone->start + zlo->zone_capacity;
 	sector_t nr_sectors = blk_rq_sectors(rq);
-	unsigned long flags;
 
-	spin_lock_irqsave(&zone->wp_lock, flags);
+	spin_lock(&zone->wp_lock);
 
 	if (zone->cond == BLK_ZONE_COND_FULL ||
 	    zone->wp + nr_sectors > zone_end) {
-		spin_unlock_irqrestore(&zone->wp_lock, flags);
+		spin_unlock(&zone->wp_lock);
 		return false;
 	}
 
@@ -883,7 +877,7 @@ static bool zloop_set_zone_append_sector(struct request *rq)
 	if (zone->wp >= zone_end)
 		zloop_mark_full(zlo, zone);
 
-	spin_unlock_irqrestore(&zone->wp_lock, flags);
+	spin_unlock(&zone->wp_lock);
 
 	return true;
 }
@@ -944,7 +938,6 @@ static int zloop_report_zones(struct gendisk *disk, sector_t sector,
 	struct zloop_device *zlo = disk->private_data;
 	struct blk_zone blkz = {};
 	unsigned int first, i;
-	unsigned long flags;
 	int ret;
 
 	first = disk_zone_no(disk, sector);
@@ -968,9 +961,9 @@ static int zloop_report_zones(struct gendisk *disk, sector_t sector,
 
 		blkz.start = zone->start;
 		blkz.len = zlo->zone_size;
-		spin_lock_irqsave(&zone->wp_lock, flags);
+		spin_lock(&zone->wp_lock);
 		blkz.wp = zone->wp;
-		spin_unlock_irqrestore(&zone->wp_lock, flags);
+		spin_unlock(&zone->wp_lock);
 		blkz.cond = zone->cond;
 		if (test_bit(ZLOOP_ZONE_CONV, &zone->flags)) {
 			blkz.type = BLK_ZONE_TYPE_CONVENTIONAL;
-- 
2.53.0

Re: [PATCH 6/6] zloop: remove irq-safe locking

[Damien Le Moal]

On 4/14/26 10:17, Christoph Hellwig wrote:
> All of zloop runs in user context, so drop the irq-safe locking.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks good.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

Re: zloop fixes

[Johannes Thumshirn]

On 4/14/26 10:17 AM, Christoph Hellwig wrote:
> Hi all,
>
> this series fixes a bad bug and a few minor issues in the new zloop code
> queued up for 7.1.
>
> Diffstat:
>   zloop.c |  123 ++++++++++++++++++++++++++++++----------------------------------
>   1 file changed, 59 insertions(+), 64 deletions(-)
>
Looks good,

Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Re: zloop fixes

[Jens Axboe]

On Tue, 14 Apr 2026 10:17:45 +0200, Christoph Hellwig wrote:
> this series fixes a bad bug and a few minor issues in the new zloop code
> queued up for 7.1.
> 
> Diffstat:
>  zloop.c |  123 ++++++++++++++++++++++++++++++----------------------------------
>  1 file changed, 59 insertions(+), 64 deletions(-)
> 
> [...]

Applied, thanks!

[1/6] zloop: fix write pointer calculation in zloop_forget_cache
      commit: 32be3c01c3b8e948a4326ab7e76c1c63dd3e27bc
[2/6] zloop: use vfs_truncate
      commit: 14e0077911e3d5e11e94417861e700cbb521a107
[3/6] zloop: improve the unaligned write pointer warning
      commit: 6466b211f797ae88073b5826dd764a6a98b67edb
[4/6] zloop: set RQF_QUIET when completing requests on deleted devices
      commit: 5b680d7afc4a2fefa0b4f584462c7540de56e2e4
[5/6] zloop: factor out zloop_mark_{full,empty} helpers
      commit: ec5c045f6cc879637cb52c9902d5fb7d419bdf47
[6/6] zloop: remove irq-safe locking
      commit: 64b437c4a96ae088d46c7d9930c35e77ee1b5b21

Best regards,
-- 
Jens Axboe