3.16.49-rc1 review patch. If anyone has any objections, please let me know. ------------------ From: Ernesto A. Fernández <ernesto.mnd.fernandez@xxxxxxxxx> commit fe26569eb9197d845d73abe7dd20f603d79eb031 upstream. When changing a file's acl mask, ext2_set_acl() will first set the group bits of i_mode to the value of the mask, and only then set the actual extended attribute representing the new acl. If the second part fails (due to lack of space, for example) and the file had no acl attribute to begin with, the system will from now on assume that the mask permission bits are actual group permission bits, potentially granting access to the wrong users. Prevent this by only changing the inode mode after the acl has been set. [JK: Rebased on top of "ext2: Don't clear SGID when inheriting ACLs"] Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@xxxxxxxxx> Signed-off-by: Jan Kara <jack@xxxxxxx> [bwh: Backported to 3.16: adjust context] Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx> --- fs/ext2/acl.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) --- a/fs/ext2/acl.c +++ b/fs/ext2/acl.c @@ -221,15 +221,22 @@ int ext2_set_acl(struct inode *inode, struct posix_acl *acl, int type) { int error; + int update_mode = 0; + umode_t mode = inode->i_mode; if (type == ACL_TYPE_ACCESS && acl) { - error = posix_acl_update_mode(inode, &inode->i_mode, &acl); + error = posix_acl_update_mode(inode, &mode, &acl); if (error) return error; + update_mode = 1; + } + error = __ext2_set_acl(inode, acl, type); + if (!error && update_mode) { + inode->i_mode = mode; inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); } - return __ext2_set_acl(inode, acl, type); + return error; } /*