Skip to contentSkip to conversation

xfs: don't allow SWAPEXT if we'd screw up quota accounting

PATCH In review Darrick J. Wong 3 participants 4 messages May 14, 2020 – May 17, 2020 XFS filesystem

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xfs: don't allow SWAPEXT if we'd screw up quota accounting

From: Darrick J. Wong <[email protected]>

Since the old SWAPEXT ioctl doesn't know how to adjust quota ids, bail out of the ids don't match and quotas are enabled.

---
 fs/xfs/xfs_bmap_util.c |    6 ++++++
 1 file changed, 6 insertions(+)
fs/xfs/xfs_bmap_util.c +6 -0
index cc23a3e23e2d..5e7da27c6e98 100644
@@ -1210,6 +1210,12 @@ xfs_swap_extents_check_format(
 	struct xfs_inode	*ip,	/* target inode */
 	struct xfs_inode	*tip)	/* tmp inode */
 {
+	/* User/group/project quota ids must match if quotas are enforced. */
+	if (XFS_IS_QUOTA_ON(ip->i_mount) &&
+	    (!uid_eq(VFS_I(ip)->i_uid, VFS_I(tip)->i_uid) ||
+	     !gid_eq(VFS_I(ip)->i_gid, VFS_I(tip)->i_gid) ||
+	     ip->i_d.di_projid != tip->i_d.di_projid))
+		return -EINVAL;
 
 	/* Should never get a local format */
 	if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||
Reviewed-by on v1 1/1
#1 lore raw
Quoted 6 lines · Darrick J. Wong

From: Darrick J. Wong <[email protected]>

Since the old SWAPEXT ioctl doesn't know how to adjust quota ids, bail out of the ids don't match and quotas are enabled.

makes sense, I probably missed the discussion that presumably arrived at "ye gods trying to fix up the quota allocations is nigh impossible?"

And in the end, what's yet another -EINVAL return here, anyway? ;)

Quoted 22 lines
---
 fs/xfs/xfs_bmap_util.c |    6 ++++++
 1 file changed, 6 insertions(+)
fs/xfs/xfs_bmap_util.c +6 -0
index cc23a3e23e2d..5e7da27c6e98 100644
@@ -1210,6 +1210,12 @@ xfs_swap_extents_check_format(
 	struct xfs_inode	*ip,	/* target inode */
 	struct xfs_inode	*tip)	/* tmp inode */
 {
+	/* User/group/project quota ids must match if quotas are enforced. */
+	if (XFS_IS_QUOTA_ON(ip->i_mount) &&
+	    (!uid_eq(VFS_I(ip)->i_uid, VFS_I(tip)->i_uid) ||
+	     !gid_eq(VFS_I(ip)->i_gid, VFS_I(tip)->i_gid) ||
+	     ip->i_d.di_projid != tip->i_d.di_projid))
+		return -EINVAL;
 
 	/* Should never get a local format */
 	if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||
#2 lore raw
Quoted 10 lines · Eric Sandeen
Quoted 6 lines · Darrick J. Wong

From: Darrick J. Wong <[email protected]>

Since the old SWAPEXT ioctl doesn't know how to adjust quota ids, bail out of the ids don't match and quotas are enabled.

makes sense, I probably missed the discussion that presumably arrived at "ye gods trying to fix up the quota allocations is nigh impossible?"

No, it's not impossible[1], it's just the one person who replied also declined to review it, so now I'm racing patches. :P

--D

[1] https://lore.kernel.org/linux-xfs/158864102885.182577.15936710415441871446.stgit@magnolia/

Quoted 26 lines

And in the end, what's yet another -EINVAL return here, anyway? ;)

Quoted 22 lines
---
 fs/xfs/xfs_bmap_util.c |    6 ++++++
 1 file changed, 6 insertions(+)
fs/xfs/xfs_bmap_util.c +6 -0
index cc23a3e23e2d..5e7da27c6e98 100644
@@ -1210,6 +1210,12 @@ xfs_swap_extents_check_format(
 	struct xfs_inode	*ip,	/* target inode */
 	struct xfs_inode	*tip)	/* tmp inode */
 {
+	/* User/group/project quota ids must match if quotas are enforced. */
+	if (XFS_IS_QUOTA_ON(ip->i_mount) &&
+	    (!uid_eq(VFS_I(ip)->i_uid, VFS_I(tip)->i_uid) ||
+	     !gid_eq(VFS_I(ip)->i_gid, VFS_I(tip)->i_gid) ||
+	     ip->i_d.di_projid != tip->i_d.di_projid))
+		return -EINVAL;
 
 	/* Should never get a local format */
 	if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||