diff mbox

[v2] btrfs: Change qgroup_meta_rsv to 64bit

Message ID 20170314102509.10648-1-rgoldwyn@suse.de
State Accepted
Headers show

Commit Message

Goldwyn Rodrigues March 14, 2017, 10:25 a.m. UTC
From: Goldwyn Rodrigues <rgoldwyn@suse.com>

Using an int value is causing qg->reserved to become negative and
exclusive -EDQUOT to be reached prematurely.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
---

Changes since v1:
 - changed reserved from long to u64

 fs/btrfs/ctree.h   |  2 +-
 fs/btrfs/disk-io.c |  2 +-
 fs/btrfs/qgroup.c  | 10 +++++-----
 3 files changed, 7 insertions(+), 7 deletions(-)

Comments

David Sterba March 14, 2017, 12:30 p.m. UTC | #1
On Tue, Mar 14, 2017 at 05:25:09AM -0500, Goldwyn Rodrigues wrote:
> From: Goldwyn Rodrigues <rgoldwyn@suse.com>
> 
> Using an int value is causing qg->reserved to become negative and
> exclusive -EDQUOT to be reached prematurely.

V1 had a reproducer, I think it's fine to keep it here so I'll copy it.

Reviewed-by: David Sterba <dsterba@suse.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Goldwyn Rodrigues March 14, 2017, 1:31 p.m. UTC | #2
On 03/14/2017 07:30 AM, David Sterba wrote:
> On Tue, Mar 14, 2017 at 05:25:09AM -0500, Goldwyn Rodrigues wrote:
>> From: Goldwyn Rodrigues <rgoldwyn@suse.com>
>>
>> Using an int value is causing qg->reserved to become negative and
>> exclusive -EDQUOT to be reached prematurely.
> 
> V1 had a reproducer, I think it's fine to keep it here so I'll copy it.
> 
> Reviewed-by: David Sterba <dsterba@suse.com>
> 

I am adding a testcase in (x)fstests which covers both qgroup issues for
which I have sent the patches. I am fine either ways.
David Sterba March 14, 2017, 2:04 p.m. UTC | #3
On Tue, Mar 14, 2017 at 08:31:26AM -0500, Goldwyn Rodrigues wrote:
> 
> 
> On 03/14/2017 07:30 AM, David Sterba wrote:
> > On Tue, Mar 14, 2017 at 05:25:09AM -0500, Goldwyn Rodrigues wrote:
> >> From: Goldwyn Rodrigues <rgoldwyn@suse.com>
> >>
> >> Using an int value is causing qg->reserved to become negative and
> >> exclusive -EDQUOT to be reached prematurely.
> > 
> > V1 had a reproducer, I think it's fine to keep it here so I'll copy it.
> > 
> > Reviewed-by: David Sterba <dsterba@suse.com>
> > 
> 
> I am adding a testcase in (x)fstests which covers both qgroup issues for
> which I have sent the patches. I am fine either ways.

The fstests testcase is orthogonal to the changelog describing the steps
to reproduce the bug (it could just roughly describe the idea), as it
helps to understand what's the patch going to fix.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 29b7fc2..c411590 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -1259,7 +1259,7 @@  struct btrfs_root {
 	atomic_t will_be_snapshoted;
 
 	/* For qgroup metadata space reserve */
-	atomic_t qgroup_meta_rsv;
+	atomic64_t qgroup_meta_rsv;
 };
 static inline u32 btrfs_inode_sectorsize(const struct inode *inode)
 {
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 08b74da..eb1ee7b 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1342,7 +1342,7 @@  static void __setup_root(struct btrfs_root *root, struct btrfs_fs_info *fs_info,
 	atomic_set(&root->orphan_inodes, 0);
 	atomic_set(&root->refs, 1);
 	atomic_set(&root->will_be_snapshoted, 0);
-	atomic_set(&root->qgroup_meta_rsv, 0);
+	atomic64_set(&root->qgroup_meta_rsv, 0);
 	root->log_transid = 0;
 	root->log_transid_committed = -1;
 	root->last_log_commit = 0;
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index 6215264..3757902 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -2963,20 +2963,20 @@  int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
 	ret = qgroup_reserve(root, num_bytes, enforce);
 	if (ret < 0)
 		return ret;
-	atomic_add(num_bytes, &root->qgroup_meta_rsv);
+	atomic64_add(num_bytes, &root->qgroup_meta_rsv);
 	return ret;
 }
 
 void btrfs_qgroup_free_meta_all(struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	int reserved;
+	u64 reserved;
 
 	if (!test_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags) ||
 	    !is_fstree(root->objectid))
 		return;
 
-	reserved = atomic_xchg(&root->qgroup_meta_rsv, 0);
+	reserved = atomic64_xchg(&root->qgroup_meta_rsv, 0);
 	if (reserved == 0)
 		return;
 	btrfs_qgroup_free_refroot(fs_info, root->objectid, reserved);
@@ -2991,8 +2991,8 @@  void btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes)
 		return;
 
 	BUG_ON(num_bytes != round_down(num_bytes, fs_info->nodesize));
-	WARN_ON(atomic_read(&root->qgroup_meta_rsv) < num_bytes);
-	atomic_sub(num_bytes, &root->qgroup_meta_rsv);
+	WARN_ON(atomic64_read(&root->qgroup_meta_rsv) < num_bytes);
+	atomic64_sub(num_bytes, &root->qgroup_meta_rsv);
 	btrfs_qgroup_free_refroot(fs_info, root->objectid, num_bytes);
 }