[dm-devel] misc struct block_device related driver cleanups

[dm-devel] misc struct block_device related driver cleanups

[Christoph Hellwig]

Hi Jens, Minchan and Mike,

this series cleans up a few interactions of driver with struct
block_device, in preparation for big changes to struct block_device
that I plan to send soon.

Diffstat:
 block/loop.c              |    3 -
 block/mtip32xx/mtip32xx.c |   15 --------
 block/mtip32xx/mtip32xx.h |    2 -
 block/zram/zram_drv.c     |   83 ++++++++--------------------------------------
 block/zram/zram_drv.h     |    1 
 md/dm-core.h              |    2 -
 md/dm.c                   |   31 ++++++-----------
 7 files changed, 27 insertions(+), 110 deletions(-)

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

misc struct block_device related driver cleanups

[Christoph Hellwig]

Hi Jens, Minchan and Mike,

this series cleans up a few interactions of driver with struct
block_device, in preparation for big changes to struct block_device
that I plan to send soon.

Diffstat:
 block/loop.c              |    3 -
 block/mtip32xx/mtip32xx.c |   15 --------
 block/mtip32xx/mtip32xx.h |    2 -
 block/zram/zram_drv.c     |   83 ++++++++--------------------------------------
 block/zram/zram_drv.h     |    1 
 md/dm-core.h              |    2 -
 md/dm.c                   |   31 ++++++-----------
 7 files changed, 27 insertions(+), 110 deletions(-)

[dm-devel] [PATCH 1/6] mtip32xx: remove the call to fsync_bdev on removal

[Christoph Hellwig]

del_gendisk already calls fsync_bdev for every partition, no need
to do this twice.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/mtip32xx/mtip32xx.c | 15 ---------------
 drivers/block/mtip32xx/mtip32xx.h |  2 --
 2 files changed, 17 deletions(-)

diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index 153e2cdecb4d40..53ac59d19ae530 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -3687,7 +3687,6 @@ static int mtip_block_initialize(struct driver_data *dd)
 	/* Enable the block device and add it to /dev */
 	device_add_disk(&dd->pdev->dev, dd->disk, NULL);
 
-	dd->bdev = bdget_disk(dd->disk, 0);
 	/*
 	 * Now that the disk is active, initialize any sysfs attributes
 	 * managed by the protocol layer.
@@ -3721,9 +3720,6 @@ static int mtip_block_initialize(struct driver_data *dd)
 	return rv;
 
 kthread_run_error:
-	bdput(dd->bdev);
-	dd->bdev = NULL;
-
 	/* Delete our gendisk. This also removes the device from /dev */
 	del_gendisk(dd->disk);
 
@@ -3804,14 +3800,6 @@ static int mtip_block_remove(struct driver_data *dd)
 	blk_mq_tagset_busy_iter(&dd->tags, mtip_no_dev_cleanup, dd);
 	blk_mq_unquiesce_queue(dd->queue);
 
-	/*
-	 * Delete our gendisk structure. This also removes the device
-	 * from /dev
-	 */
-	if (dd->bdev) {
-		bdput(dd->bdev);
-		dd->bdev = NULL;
-	}
 	if (dd->disk) {
 		if (test_bit(MTIP_DDF_INIT_DONE_BIT, &dd->dd_flag))
 			del_gendisk(dd->disk);
@@ -4206,9 +4194,6 @@ static void mtip_pci_remove(struct pci_dev *pdev)
 	} while (atomic_read(&dd->irq_workers_active) != 0 &&
 		time_before(jiffies, to));
 
-	if (!dd->sr)
-		fsync_bdev(dd->bdev);
-
 	if (atomic_read(&dd->irq_workers_active) != 0) {
 		dev_warn(&dd->pdev->dev,
 			"Completion workers still active!\n");
diff --git a/drivers/block/mtip32xx/mtip32xx.h b/drivers/block/mtip32xx/mtip32xx.h
index e22a7f0523bf30..88f4206310e4c8 100644
--- a/drivers/block/mtip32xx/mtip32xx.h
+++ b/drivers/block/mtip32xx/mtip32xx.h
@@ -463,8 +463,6 @@ struct driver_data {
 
 	int isr_binding;
 
-	struct block_device *bdev;
-
 	struct list_head online_list; /* linkage for online list */
 
 	struct list_head remove_list; /* linkage for removing list */
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 1/6] mtip32xx: remove the call to fsync_bdev on removal

[Christoph Hellwig]

del_gendisk already calls fsync_bdev for every partition, no need
to do this twice.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/mtip32xx/mtip32xx.c | 15 ---------------
 drivers/block/mtip32xx/mtip32xx.h |  2 --
 2 files changed, 17 deletions(-)

diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index 153e2cdecb4d40..53ac59d19ae530 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -3687,7 +3687,6 @@ static int mtip_block_initialize(struct driver_data *dd)
 	/* Enable the block device and add it to /dev */
 	device_add_disk(&dd->pdev->dev, dd->disk, NULL);
 
-	dd->bdev = bdget_disk(dd->disk, 0);
 	/*
 	 * Now that the disk is active, initialize any sysfs attributes
 	 * managed by the protocol layer.
@@ -3721,9 +3720,6 @@ static int mtip_block_initialize(struct driver_data *dd)
 	return rv;
 
 kthread_run_error:
-	bdput(dd->bdev);
-	dd->bdev = NULL;
-
 	/* Delete our gendisk. This also removes the device from /dev */
 	del_gendisk(dd->disk);
 
@@ -3804,14 +3800,6 @@ static int mtip_block_remove(struct driver_data *dd)
 	blk_mq_tagset_busy_iter(&dd->tags, mtip_no_dev_cleanup, dd);
 	blk_mq_unquiesce_queue(dd->queue);
 
-	/*
-	 * Delete our gendisk structure. This also removes the device
-	 * from /dev
-	 */
-	if (dd->bdev) {
-		bdput(dd->bdev);
-		dd->bdev = NULL;
-	}
 	if (dd->disk) {
 		if (test_bit(MTIP_DDF_INIT_DONE_BIT, &dd->dd_flag))
 			del_gendisk(dd->disk);
@@ -4206,9 +4194,6 @@ static void mtip_pci_remove(struct pci_dev *pdev)
 	} while (atomic_read(&dd->irq_workers_active) != 0 &&
 		time_before(jiffies, to));
 
-	if (!dd->sr)
-		fsync_bdev(dd->bdev);
-
 	if (atomic_read(&dd->irq_workers_active) != 0) {
 		dev_warn(&dd->pdev->dev,
 			"Completion workers still active!\n");
diff --git a/drivers/block/mtip32xx/mtip32xx.h b/drivers/block/mtip32xx/mtip32xx.h
index e22a7f0523bf30..88f4206310e4c8 100644
--- a/drivers/block/mtip32xx/mtip32xx.h
+++ b/drivers/block/mtip32xx/mtip32xx.h
@@ -463,8 +463,6 @@ struct driver_data {
 
 	int isr_binding;
 
-	struct block_device *bdev;
-
 	struct list_head online_list; /* linkage for online list */
 
 	struct list_head remove_list; /* linkage for removing list */
-- 
2.29.2

[dm-devel] [PATCH 2/6] zram: remove the claim mechanism

[Christoph Hellwig]

The zram claim mechanism was added to ensure no new opens come in
during teardown.  But the proper way to archive that is to call
del_gendisk first, which takes care of all that.  Once del_gendisk
is called in the right place, the reset side can also be simplified
as no I/O can be outstanding on a block device that is not open.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zram/zram_drv.c | 72 ++++++++---------------------------
 1 file changed, 15 insertions(+), 57 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 6d15d51cee2b7e..2e6d75ec1afddb 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -1756,64 +1756,33 @@ static ssize_t disksize_store(struct device *dev,
 static ssize_t reset_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t len)
 {
-	int ret;
-	unsigned short do_reset;
-	struct zram *zram;
+	struct zram *zram = dev_to_zram(dev);
 	struct block_device *bdev;
+	unsigned short do_reset;
+	int ret = 0;
 
 	ret = kstrtou16(buf, 10, &do_reset);
 	if (ret)
 		return ret;
-
 	if (!do_reset)
 		return -EINVAL;
 
-	zram = dev_to_zram(dev);
 	bdev = bdget_disk(zram->disk, 0);
 	if (!bdev)
 		return -ENOMEM;
 
 	mutex_lock(&bdev->bd_mutex);
-	/* Do not reset an active device or claimed device */
-	if (bdev->bd_openers || zram->claim) {
-		mutex_unlock(&bdev->bd_mutex);
-		bdput(bdev);
-		return -EBUSY;
-	}
-
-	/* From now on, anyone can't open /dev/zram[0-9] */
-	zram->claim = true;
+	if (bdev->bd_openers)
+		ret = -EBUSY;
+	else
+		zram_reset_device(zram);
 	mutex_unlock(&bdev->bd_mutex);
-
-	/* Make sure all the pending I/O are finished */
-	fsync_bdev(bdev);
-	zram_reset_device(zram);
 	bdput(bdev);
 
-	mutex_lock(&bdev->bd_mutex);
-	zram->claim = false;
-	mutex_unlock(&bdev->bd_mutex);
-
-	return len;
-}
-
-static int zram_open(struct block_device *bdev, fmode_t mode)
-{
-	int ret = 0;
-	struct zram *zram;
-
-	WARN_ON(!mutex_is_locked(&bdev->bd_mutex));
-
-	zram = bdev->bd_disk->private_data;
-	/* zram was claimed to reset so open request fails */
-	if (zram->claim)
-		ret = -EBUSY;
-
-	return ret;
+	return ret ? ret : len;
 }
 
 static const struct block_device_operations zram_devops = {
-	.open = zram_open,
 	.submit_bio = zram_submit_bio,
 	.swap_slot_free_notify = zram_slot_free_notify,
 	.rw_page = zram_rw_page,
@@ -1821,7 +1790,6 @@ static const struct block_device_operations zram_devops = {
 };
 
 static const struct block_device_operations zram_wb_devops = {
-	.open = zram_open,
 	.submit_bio = zram_submit_bio,
 	.swap_slot_free_notify = zram_slot_free_notify,
 	.owner = THIS_MODULE
@@ -1974,32 +1942,22 @@ static int zram_add(void)
 
 static int zram_remove(struct zram *zram)
 {
-	struct block_device *bdev;
-
-	bdev = bdget_disk(zram->disk, 0);
-	if (!bdev)
-		return -ENOMEM;
+	struct block_device *bdev = bdget_disk(zram->disk, 0);
 
-	mutex_lock(&bdev->bd_mutex);
-	if (bdev->bd_openers || zram->claim) {
-		mutex_unlock(&bdev->bd_mutex);
+	if (bdev) {
+		if (bdev->bd_openers) {
+			bdput(bdev);
+			return -EBUSY;
+		}
 		bdput(bdev);
-		return -EBUSY;
 	}
 
-	zram->claim = true;
-	mutex_unlock(&bdev->bd_mutex);
-
+	del_gendisk(zram->disk);
 	zram_debugfs_unregister(zram);
-
-	/* Make sure all the pending I/O are finished */
-	fsync_bdev(bdev);
 	zram_reset_device(zram);
-	bdput(bdev);
 
 	pr_info("Removed device: %s\n", zram->disk->disk_name);
 
-	del_gendisk(zram->disk);
 	blk_cleanup_queue(zram->disk->queue);
 	put_disk(zram->disk);
 	kfree(zram);
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 2/6] zram: remove the claim mechanism

[Christoph Hellwig]

The zram claim mechanism was added to ensure no new opens come in
during teardown.  But the proper way to archive that is to call
del_gendisk first, which takes care of all that.  Once del_gendisk
is called in the right place, the reset side can also be simplified
as no I/O can be outstanding on a block device that is not open.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zram/zram_drv.c | 72 ++++++++---------------------------
 1 file changed, 15 insertions(+), 57 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 6d15d51cee2b7e..2e6d75ec1afddb 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -1756,64 +1756,33 @@ static ssize_t disksize_store(struct device *dev,
 static ssize_t reset_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t len)
 {
-	int ret;
-	unsigned short do_reset;
-	struct zram *zram;
+	struct zram *zram = dev_to_zram(dev);
 	struct block_device *bdev;
+	unsigned short do_reset;
+	int ret = 0;
 
 	ret = kstrtou16(buf, 10, &do_reset);
 	if (ret)
 		return ret;
-
 	if (!do_reset)
 		return -EINVAL;
 
-	zram = dev_to_zram(dev);
 	bdev = bdget_disk(zram->disk, 0);
 	if (!bdev)
 		return -ENOMEM;
 
 	mutex_lock(&bdev->bd_mutex);
-	/* Do not reset an active device or claimed device */
-	if (bdev->bd_openers || zram->claim) {
-		mutex_unlock(&bdev->bd_mutex);
-		bdput(bdev);
-		return -EBUSY;
-	}
-
-	/* From now on, anyone can't open /dev/zram[0-9] */
-	zram->claim = true;
+	if (bdev->bd_openers)
+		ret = -EBUSY;
+	else
+		zram_reset_device(zram);
 	mutex_unlock(&bdev->bd_mutex);
-
-	/* Make sure all the pending I/O are finished */
-	fsync_bdev(bdev);
-	zram_reset_device(zram);
 	bdput(bdev);
 
-	mutex_lock(&bdev->bd_mutex);
-	zram->claim = false;
-	mutex_unlock(&bdev->bd_mutex);
-
-	return len;
-}
-
-static int zram_open(struct block_device *bdev, fmode_t mode)
-{
-	int ret = 0;
-	struct zram *zram;
-
-	WARN_ON(!mutex_is_locked(&bdev->bd_mutex));
-
-	zram = bdev->bd_disk->private_data;
-	/* zram was claimed to reset so open request fails */
-	if (zram->claim)
-		ret = -EBUSY;
-
-	return ret;
+	return ret ? ret : len;
 }
 
 static const struct block_device_operations zram_devops = {
-	.open = zram_open,
 	.submit_bio = zram_submit_bio,
 	.swap_slot_free_notify = zram_slot_free_notify,
 	.rw_page = zram_rw_page,
@@ -1821,7 +1790,6 @@ static const struct block_device_operations zram_devops = {
 };
 
 static const struct block_device_operations zram_wb_devops = {
-	.open = zram_open,
 	.submit_bio = zram_submit_bio,
 	.swap_slot_free_notify = zram_slot_free_notify,
 	.owner = THIS_MODULE
@@ -1974,32 +1942,22 @@ static int zram_add(void)
 
 static int zram_remove(struct zram *zram)
 {
-	struct block_device *bdev;
-
-	bdev = bdget_disk(zram->disk, 0);
-	if (!bdev)
-		return -ENOMEM;
+	struct block_device *bdev = bdget_disk(zram->disk, 0);
 
-	mutex_lock(&bdev->bd_mutex);
-	if (bdev->bd_openers || zram->claim) {
-		mutex_unlock(&bdev->bd_mutex);
+	if (bdev) {
+		if (bdev->bd_openers) {
+			bdput(bdev);
+			return -EBUSY;
+		}
 		bdput(bdev);
-		return -EBUSY;
 	}
 
-	zram->claim = true;
-	mutex_unlock(&bdev->bd_mutex);
-
+	del_gendisk(zram->disk);
 	zram_debugfs_unregister(zram);
-
-	/* Make sure all the pending I/O are finished */
-	fsync_bdev(bdev);
 	zram_reset_device(zram);
-	bdput(bdev);
 
 	pr_info("Removed device: %s\n", zram->disk->disk_name);
 
-	del_gendisk(zram->disk);
 	blk_cleanup_queue(zram->disk->queue);
 	put_disk(zram->disk);
 	kfree(zram);
-- 
2.29.2

[dm-devel] [PATCH 3/6] zram: do not call set_blocksize

[Christoph Hellwig]

set_blocksize is used by file systems to use their preferred buffer cache
block size.  Block drivers should not set it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zram/zram_drv.c | 11 +----------
 drivers/block/zram/zram_drv.h |  1 -
 2 files changed, 1 insertion(+), 11 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 2e6d75ec1afddb..88baa6158eaee1 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -403,13 +403,10 @@ static void reset_bdev(struct zram *zram)
 		return;
 
 	bdev = zram->bdev;
-	if (zram->old_block_size)
-		set_blocksize(bdev, zram->old_block_size);
 	blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
 	/* hope filp_close flush all of IO */
 	filp_close(zram->backing_dev, NULL);
 	zram->backing_dev = NULL;
-	zram->old_block_size = 0;
 	zram->bdev = NULL;
 	zram->disk->fops = &zram_devops;
 	kvfree(zram->bitmap);
@@ -454,7 +451,7 @@ static ssize_t backing_dev_store(struct device *dev,
 	struct file *backing_dev = NULL;
 	struct inode *inode;
 	struct address_space *mapping;
-	unsigned int bitmap_sz, old_block_size = 0;
+	unsigned int bitmap_sz;
 	unsigned long nr_pages, *bitmap = NULL;
 	struct block_device *bdev = NULL;
 	int err;
@@ -509,14 +506,8 @@ static ssize_t backing_dev_store(struct device *dev,
 		goto out;
 	}
 
-	old_block_size = block_size(bdev);
-	err = set_blocksize(bdev, PAGE_SIZE);
-	if (err)
-		goto out;
-
 	reset_bdev(zram);
 
-	zram->old_block_size = old_block_size;
 	zram->bdev = bdev;
 	zram->backing_dev = backing_dev;
 	zram->bitmap = bitmap;
diff --git a/drivers/block/zram/zram_drv.h b/drivers/block/zram/zram_drv.h
index f2fd46daa76045..712354a4207c77 100644
--- a/drivers/block/zram/zram_drv.h
+++ b/drivers/block/zram/zram_drv.h
@@ -118,7 +118,6 @@ struct zram {
 	bool wb_limit_enable;
 	u64 bd_wb_limit;
 	struct block_device *bdev;
-	unsigned int old_block_size;
 	unsigned long *bitmap;
 	unsigned long nr_pages;
 #endif
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 3/6] zram: do not call set_blocksize

[Christoph Hellwig]

set_blocksize is used by file systems to use their preferred buffer cache
block size.  Block drivers should not set it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/zram/zram_drv.c | 11 +----------
 drivers/block/zram/zram_drv.h |  1 -
 2 files changed, 1 insertion(+), 11 deletions(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 2e6d75ec1afddb..88baa6158eaee1 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -403,13 +403,10 @@ static void reset_bdev(struct zram *zram)
 		return;
 
 	bdev = zram->bdev;
-	if (zram->old_block_size)
-		set_blocksize(bdev, zram->old_block_size);
 	blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
 	/* hope filp_close flush all of IO */
 	filp_close(zram->backing_dev, NULL);
 	zram->backing_dev = NULL;
-	zram->old_block_size = 0;
 	zram->bdev = NULL;
 	zram->disk->fops = &zram_devops;
 	kvfree(zram->bitmap);
@@ -454,7 +451,7 @@ static ssize_t backing_dev_store(struct device *dev,
 	struct file *backing_dev = NULL;
 	struct inode *inode;
 	struct address_space *mapping;
-	unsigned int bitmap_sz, old_block_size = 0;
+	unsigned int bitmap_sz;
 	unsigned long nr_pages, *bitmap = NULL;
 	struct block_device *bdev = NULL;
 	int err;
@@ -509,14 +506,8 @@ static ssize_t backing_dev_store(struct device *dev,
 		goto out;
 	}
 
-	old_block_size = block_size(bdev);
-	err = set_blocksize(bdev, PAGE_SIZE);
-	if (err)
-		goto out;
-
 	reset_bdev(zram);
 
-	zram->old_block_size = old_block_size;
 	zram->bdev = bdev;
 	zram->backing_dev = backing_dev;
 	zram->bitmap = bitmap;
diff --git a/drivers/block/zram/zram_drv.h b/drivers/block/zram/zram_drv.h
index f2fd46daa76045..712354a4207c77 100644
--- a/drivers/block/zram/zram_drv.h
+++ b/drivers/block/zram/zram_drv.h
@@ -118,7 +118,6 @@ struct zram {
 	bool wb_limit_enable;
 	u64 bd_wb_limit;
 	struct block_device *bdev;
-	unsigned int old_block_size;
 	unsigned long *bitmap;
 	unsigned long nr_pages;
 #endif
-- 
2.29.2

[dm-devel] [PATCH 4/6] loop: do not call set_blocksize

[Christoph Hellwig]

set_blocksize is used by file systems to use their preferred buffer cache
block size.  Block drivers should not set it.

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

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 9a27d4f1c08aac..b42c728620c9e4 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1164,9 +1164,6 @@ static int loop_configure(struct loop_device *lo, fmode_t mode,
 	size = get_loop_size(lo, file);
 	loop_set_size(lo, size);
 
-	set_blocksize(bdev, S_ISBLK(inode->i_mode) ?
-		      block_size(inode->i_bdev) : PAGE_SIZE);
-
 	lo->lo_state = Lo_bound;
 	if (part_shift)
 		lo->lo_flags |= LO_FLAGS_PARTSCAN;
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 4/6] loop: do not call set_blocksize

[Christoph Hellwig]

set_blocksize is used by file systems to use their preferred buffer cache
block size.  Block drivers should not set it.

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

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 9a27d4f1c08aac..b42c728620c9e4 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1164,9 +1164,6 @@ static int loop_configure(struct loop_device *lo, fmode_t mode,
 	size = get_loop_size(lo, file);
 	loop_set_size(lo, size);
 
-	set_blocksize(bdev, S_ISBLK(inode->i_mode) ?
-		      block_size(inode->i_bdev) : PAGE_SIZE);
-
 	lo->lo_state = Lo_bound;
 	if (part_shift)
 		lo->lo_flags |= LO_FLAGS_PARTSCAN;
-- 
2.29.2

[dm-devel] [PATCH 5/6] dm: simplify flush_bio initialization in __send_empty_flush

[Christoph Hellwig]

We don't really need the struct block_device to initialize a bio.  So
switch from using bio_set_dev to manually setting up bi_disk (bi_partno
will always be zero and has been cleared by bio_init already).

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/dm.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 54739f1b579bc8..6d7eb72d41f9ea 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1422,18 +1422,12 @@ static int __send_empty_flush(struct clone_info *ci)
 	 */
 	bio_init(&flush_bio, NULL, 0);
 	flush_bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC;
+	flush_bio.bi_disk = ci->io->md->disk;
+	bio_associate_blkg(&flush_bio);
+
 	ci->bio = &flush_bio;
 	ci->sector_count = 0;
 
-	/*
-	 * Empty flush uses a statically initialized bio, as the base for
-	 * cloning.  However, blkg association requires that a bdev is
-	 * associated with a gendisk, which doesn't happen until the bdev is
-	 * opened.  So, blkg association is done at issue time of the flush
-	 * rather than when the device is created in alloc_dev().
-	 */
-	bio_set_dev(ci->bio, ci->io->md->bdev);
-
 	BUG_ON(bio_has_data(ci->bio));
 	while ((ti = dm_table_get_target(ci->map, target_nr++)))
 		__send_duplicate_bios(ci, ti, ti->num_flush_bios, NULL);
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 5/6] dm: simplify flush_bio initialization in __send_empty_flush

[Christoph Hellwig]

We don't really need the struct block_device to initialize a bio.  So
switch from using bio_set_dev to manually setting up bi_disk (bi_partno
will always be zero and has been cleared by bio_init already).

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/dm.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 54739f1b579bc8..6d7eb72d41f9ea 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1422,18 +1422,12 @@ static int __send_empty_flush(struct clone_info *ci)
 	 */
 	bio_init(&flush_bio, NULL, 0);
 	flush_bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC;
+	flush_bio.bi_disk = ci->io->md->disk;
+	bio_associate_blkg(&flush_bio);
+
 	ci->bio = &flush_bio;
 	ci->sector_count = 0;
 
-	/*
-	 * Empty flush uses a statically initialized bio, as the base for
-	 * cloning.  However, blkg association requires that a bdev is
-	 * associated with a gendisk, which doesn't happen until the bdev is
-	 * opened.  So, blkg association is done at issue time of the flush
-	 * rather than when the device is created in alloc_dev().
-	 */
-	bio_set_dev(ci->bio, ci->io->md->bdev);
-
 	BUG_ON(bio_has_data(ci->bio));
 	while ((ti = dm_table_get_target(ci->map, target_nr++)))
 		__send_duplicate_bios(ci, ti, ti->num_flush_bios, NULL);
-- 
2.29.2

Re: [dm-devel] [PATCH 5/6] dm: simplify flush_bio initialization in __send_empty_flush

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> We don't really need the struct block_device to initialize a bio.  So
> switch from using bio_set_dev to manually setting up bi_disk (bi_partno
> will always be zero and has been cleared by bio_init already).
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Mike Snitzer <snitzer@redhat.com>

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Re: [PATCH 5/6] dm: simplify flush_bio initialization in __send_empty_flush

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> We don't really need the struct block_device to initialize a bio.  So
> switch from using bio_set_dev to manually setting up bi_disk (bi_partno
> will always be zero and has been cleared by bio_init already).
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Mike Snitzer <snitzer@redhat.com>

[dm-devel] [PATCH 6/6] dm: remove the block_device reference in struct mapped_device

[Christoph Hellwig]

Get rid of the long-lasting struct block_device reference in
struct mapped_device.  The only remaining user is the freeze code,
where we can trivially look up the block device at freeze time
and release the reference at thaw time.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/dm-core.h |  2 --
 drivers/md/dm.c      | 19 ++++++++-----------
 2 files changed, 8 insertions(+), 13 deletions(-)

diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h
index d522093cb39dda..b1b400ed76fe90 100644
--- a/drivers/md/dm-core.h
+++ b/drivers/md/dm-core.h
@@ -107,8 +107,6 @@ struct mapped_device {
 	/* kobject and completion */
 	struct dm_kobject_holder kobj_holder;
 
-	struct block_device *bdev;
-
 	struct dm_stats stats;
 
 	/* for blk-mq request-based DM support */
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 6d7eb72d41f9ea..83fe1e7f13e6b0 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1744,11 +1744,6 @@ static void cleanup_mapped_device(struct mapped_device *md)
 
 	cleanup_srcu_struct(&md->io_barrier);
 
-	if (md->bdev) {
-		bdput(md->bdev);
-		md->bdev = NULL;
-	}
-
 	mutex_destroy(&md->suspend_lock);
 	mutex_destroy(&md->type_lock);
 	mutex_destroy(&md->table_devices_lock);
@@ -1840,10 +1835,6 @@ static struct mapped_device *alloc_dev(int minor)
 	if (!md->wq)
 		goto bad;
 
-	md->bdev = bdget_disk(md->disk, 0);
-	if (!md->bdev)
-		goto bad;
-
 	dm_stats_init(&md->stats);
 
 	/* Populate the mapping, nobody knows we exist yet */
@@ -2384,12 +2375,17 @@ struct dm_table *dm_swap_table(struct mapped_device *md, struct dm_table *table)
  */
 static int lock_fs(struct mapped_device *md)
 {
+	struct block_device *bdev;
 	int r;
 
 	WARN_ON(md->frozen_sb);
 
-	md->frozen_sb = freeze_bdev(md->bdev);
+	bdev = bdget_disk(md->disk, 0);
+	if (!bdev)
+		return -ENOMEM;
+	md->frozen_sb = freeze_bdev(bdev);
 	if (IS_ERR(md->frozen_sb)) {
+		bdput(bdev);
 		r = PTR_ERR(md->frozen_sb);
 		md->frozen_sb = NULL;
 		return r;
@@ -2405,7 +2401,8 @@ static void unlock_fs(struct mapped_device *md)
 	if (!test_bit(DMF_FROZEN, &md->flags))
 		return;
 
-	thaw_bdev(md->bdev, md->frozen_sb);
+	thaw_bdev(md->frozen_sb->s_bdev, md->frozen_sb);
+	bdput(md->frozen_sb->s_bdev);
 	md->frozen_sb = NULL;
 	clear_bit(DMF_FROZEN, &md->flags);
 }
-- 
2.29.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

[PATCH 6/6] dm: remove the block_device reference in struct mapped_device

[Christoph Hellwig]

Get rid of the long-lasting struct block_device reference in
struct mapped_device.  The only remaining user is the freeze code,
where we can trivially look up the block device at freeze time
and release the reference at thaw time.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/md/dm-core.h |  2 --
 drivers/md/dm.c      | 19 ++++++++-----------
 2 files changed, 8 insertions(+), 13 deletions(-)

diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h
index d522093cb39dda..b1b400ed76fe90 100644
--- a/drivers/md/dm-core.h
+++ b/drivers/md/dm-core.h
@@ -107,8 +107,6 @@ struct mapped_device {
 	/* kobject and completion */
 	struct dm_kobject_holder kobj_holder;
 
-	struct block_device *bdev;
-
 	struct dm_stats stats;
 
 	/* for blk-mq request-based DM support */
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 6d7eb72d41f9ea..83fe1e7f13e6b0 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -1744,11 +1744,6 @@ static void cleanup_mapped_device(struct mapped_device *md)
 
 	cleanup_srcu_struct(&md->io_barrier);
 
-	if (md->bdev) {
-		bdput(md->bdev);
-		md->bdev = NULL;
-	}
-
 	mutex_destroy(&md->suspend_lock);
 	mutex_destroy(&md->type_lock);
 	mutex_destroy(&md->table_devices_lock);
@@ -1840,10 +1835,6 @@ static struct mapped_device *alloc_dev(int minor)
 	if (!md->wq)
 		goto bad;
 
-	md->bdev = bdget_disk(md->disk, 0);
-	if (!md->bdev)
-		goto bad;
-
 	dm_stats_init(&md->stats);
 
 	/* Populate the mapping, nobody knows we exist yet */
@@ -2384,12 +2375,17 @@ struct dm_table *dm_swap_table(struct mapped_device *md, struct dm_table *table)
  */
 static int lock_fs(struct mapped_device *md)
 {
+	struct block_device *bdev;
 	int r;
 
 	WARN_ON(md->frozen_sb);
 
-	md->frozen_sb = freeze_bdev(md->bdev);
+	bdev = bdget_disk(md->disk, 0);
+	if (!bdev)
+		return -ENOMEM;
+	md->frozen_sb = freeze_bdev(bdev);
 	if (IS_ERR(md->frozen_sb)) {
+		bdput(bdev);
 		r = PTR_ERR(md->frozen_sb);
 		md->frozen_sb = NULL;
 		return r;
@@ -2405,7 +2401,8 @@ static void unlock_fs(struct mapped_device *md)
 	if (!test_bit(DMF_FROZEN, &md->flags))
 		return;
 
-	thaw_bdev(md->bdev, md->frozen_sb);
+	thaw_bdev(md->frozen_sb->s_bdev, md->frozen_sb);
+	bdput(md->frozen_sb->s_bdev);
 	md->frozen_sb = NULL;
 	clear_bit(DMF_FROZEN, &md->flags);
 }
-- 
2.29.2

Re: [dm-devel] [PATCH 6/6] dm: remove the block_device reference in struct mapped_device

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> Get rid of the long-lasting struct block_device reference in
> struct mapped_device.  The only remaining user is the freeze code,
> where we can trivially look up the block device at freeze time
> and release the reference at thaw time.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Mike Snitzer <snitzer@redhat.com>

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Re: [PATCH 6/6] dm: remove the block_device reference in struct mapped_device

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> Get rid of the long-lasting struct block_device reference in
> struct mapped_device.  The only remaining user is the freeze code,
> where we can trivially look up the block device at freeze time
> and release the reference at thaw time.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Mike Snitzer <snitzer@redhat.com>

Re: [dm-devel] misc struct block_device related driver cleanups

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> Hi Jens, Minchan and Mike,
> 
> this series cleans up a few interactions of driver with struct
> block_device, in preparation for big changes to struct block_device
> that I plan to send soon.

Thanks, I've picked up 5 and 6 for 5.11.

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Re: misc struct block_device related driver cleanups

[Mike Snitzer]

On Mon, Nov 16 2020 at  4:20pm -0500,
Christoph Hellwig <hch@lst.de> wrote:

> Hi Jens, Minchan and Mike,
> 
> this series cleans up a few interactions of driver with struct
> block_device, in preparation for big changes to struct block_device
> that I plan to send soon.

Thanks, I've picked up 5 and 6 for 5.11.

Re: [dm-devel] misc struct block_device related driver cleanups

[Christoph Hellwig]

On Tue, Nov 17, 2020 at 10:46:29AM -0500, Mike Snitzer wrote:
> On Mon, Nov 16 2020 at  4:20pm -0500,
> Christoph Hellwig <hch@lst.de> wrote:
> 
> > Hi Jens, Minchan and Mike,
> > 
> > this series cleans up a few interactions of driver with struct
> > block_device, in preparation for big changes to struct block_device
> > that I plan to send soon.
> 
> Thanks, I've picked up 5 and 6 for 5.11.

I actually need them in Jens' for-5.11/block tree for my next series..

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Re: misc struct block_device related driver cleanups

[Christoph Hellwig]

On Tue, Nov 17, 2020 at 10:46:29AM -0500, Mike Snitzer wrote:
> On Mon, Nov 16 2020 at  4:20pm -0500,
> Christoph Hellwig <hch@lst.de> wrote:
> 
> > Hi Jens, Minchan and Mike,
> > 
> > this series cleans up a few interactions of driver with struct
> > block_device, in preparation for big changes to struct block_device
> > that I plan to send soon.
> 
> Thanks, I've picked up 5 and 6 for 5.11.

I actually need them in Jens' for-5.11/block tree for my next series..

Re: [dm-devel] misc struct block_device related driver cleanups

[Mike Snitzer]

On Tue, Nov 17 2020 at 10:51am -0500,
Christoph Hellwig <hch@lst.de> wrote:

> On Tue, Nov 17, 2020 at 10:46:29AM -0500, Mike Snitzer wrote:
> > On Mon, Nov 16 2020 at  4:20pm -0500,
> > Christoph Hellwig <hch@lst.de> wrote:
> > 
> > > Hi Jens, Minchan and Mike,
> > > 
> > > this series cleans up a few interactions of driver with struct
> > > block_device, in preparation for big changes to struct block_device
> > > that I plan to send soon.
> > 
> > Thanks, I've picked up 5 and 6 for 5.11.
> 
> I actually need them in Jens' for-5.11/block tree for my next series..

Ah, OK.. no problem.

I'll reply with my Acked-by:s

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Re: misc struct block_device related driver cleanups

[Mike Snitzer]

On Tue, Nov 17 2020 at 10:51am -0500,
Christoph Hellwig <hch@lst.de> wrote:

> On Tue, Nov 17, 2020 at 10:46:29AM -0500, Mike Snitzer wrote:
> > On Mon, Nov 16 2020 at  4:20pm -0500,
> > Christoph Hellwig <hch@lst.de> wrote:
> > 
> > > Hi Jens, Minchan and Mike,
> > > 
> > > this series cleans up a few interactions of driver with struct
> > > block_device, in preparation for big changes to struct block_device
> > > that I plan to send soon.
> > 
> > Thanks, I've picked up 5 and 6 for 5.11.
> 
> I actually need them in Jens' for-5.11/block tree for my next series..

Ah, OK.. no problem.

I'll reply with my Acked-by:s