[PATCH v2 1/2] btrfs: make code easy to read in btrfs_open_one_device()

[PATCH v2 1/2] btrfs: make code easy to read in btrfs_open_one_device()

[Anand Jain]

From: Anand Jain <Anand.Jain@oracle.com>

First set the usual case that is writeable, then check
for any special case like a seed device or a rdonly device
and set the state appropriately.

Signed-off-by: Anand Jain <anand.jain@oracle.com>
---
v1->v2:
 Add comment and commit log to explain its safe to set writeable
 first and then reset to the special case such as seed or readonly
 device. (no code change).

 fs/btrfs/volumes.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 5750cd9df417..001894782d84 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -689,14 +689,18 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 	device->generation = btrfs_super_generation(disk_super);
 
+	/*
+	 * Set the device as writeable and then check if its
+	 * seed or readonly device. Both SUPER_FLAG_SEEDING and
+	 * bdev_read_only(bdev) can be set by the user when device
+	 * is unmounted. And this fn is called at the time of mount.
+	 */
+	set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
 		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		fs_devices->seeding = 1;
-	} else {
-		if (bdev_read_only(bdev))
-			clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
-		else
-			set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
+	} else if (bdev_read_only(bdev)) {
+		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	}
 
 	q = bdev_get_queue(bdev);
-- 
2.15.0

[PATCH v2 1/2] btrfs: make code easy to read in btrfs_open_one_device()

[Anand Jain]

From: Anand Jain <Anand.Jain@oracle.com>

First set the usual case that is writeable, then check
for any special case like a seed device or a rdonly device
and set the state appropriately.

Signed-off-by: Anand Jain <anand.jain@oracle.com>
---
v1->v2:
 Add comments, and update change log.

 fs/btrfs/volumes.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 5750cd9df417..001894782d84 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -689,14 +689,18 @@ static int btrfs_open_one_device(struct btrfs_fs_devices *fs_devices,
 
 	device->generation = btrfs_super_generation(disk_super);
 
+	/*
+	 * Set the device as writeable and then check if its
+	 * seed or readonly device. Both SUPER_FLAG_SEEDING and
+	 * bdev_read_only(bdev) can be set by the user when device
+	 * is unmounted. And this fn is called at the time of mount.
+	 */
+	set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
 		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 		fs_devices->seeding = 1;
-	} else {
-		if (bdev_read_only(bdev))
-			clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
-		else
-			set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
+	} else if (bdev_read_only(bdev)) {
+		clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	}
 
 	q = bdev_get_queue(bdev);
-- 
2.7.0

[PATCH v2.1 2/2] btrfs: optimize move uuid_mutex closer to the critical section

[Anand Jain]

Move uuid_mutex closer to the exclusion section.

Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
---
v2->v2.1: updated the v2 reroll log below.
v1->v2:
 Add received reviewed-by. Thxs.
 Fix conflict on top of kdave misc-next.

 fs/btrfs/volumes.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 001894782d84..d9166737c747 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1197,13 +1197,10 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 	 */
 	bytenr = btrfs_sb_offset(0);
 	flags |= FMODE_EXCL;
-	mutex_lock(&uuid_mutex);
 
 	bdev = blkdev_get_by_path(path, flags, holder);
-	if (IS_ERR(bdev)) {
-		ret = PTR_ERR(bdev);
-		goto error;
-	}
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
 
 	if (btrfs_read_disk_super(bdev, bytenr, &page, &disk_super)) {
 		ret = -EINVAL;
@@ -1214,7 +1211,12 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 	transid = btrfs_super_generation(disk_super);
 	total_devices = btrfs_super_num_devices(disk_super);
 
+	mutex_lock(&uuid_mutex);
 	ret = device_list_add(path, disk_super, devid, fs_devices_ret);
+	if (ret >= 0 && fs_devices_ret)
+		(*fs_devices_ret)->total_devices = total_devices;
+	mutex_unlock(&uuid_mutex);
+
 	if (ret > 0) {
 		if (disk_super->label[0])
 			pr_info("BTRFS: device label %s ", disk_super->label);
@@ -1224,15 +1226,12 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 		pr_cont("devid %llu transid %llu %s\n", devid, transid, path);
 		ret = 0;
 	}
-	if (!ret && fs_devices_ret)
-		(*fs_devices_ret)->total_devices = total_devices;
 
 	btrfs_release_disk_super(page);
 
 error_bdev_put:
 	blkdev_put(bdev, flags);
-error:
-	mutex_unlock(&uuid_mutex);
+
 	return ret;
 }
 
-- 
2.7.0