When changing a file's acl mask, __gfs2_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. Signed-off-by: Ernesto A. Fernández <[email protected]> --- Changes in v2: - Rebased on top of the for-next branch. The first part of the patchset does not need to be fixed. fs/gfs2/acl.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index a6d9623..9d5eecb 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -116,6 +116,7 @@ int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) struct gfs2_holder gh; bool need_unlock = false; int ret; + umode_t mode; if (acl && acl->a_count > GFS2_ACL_MAX_ENTRIES(GFS2_SB(inode))) return -E2BIG; @@ -130,17 +131,19 @@ int gfs2_set_acl(struct inode *inode, struct posix_acl *acl, int type) return ret; need_unlock = true; } - if (type == ACL_TYPE_ACCESS && acl) { - umode_t mode = inode->i_mode; - ret = posix_acl_update_mode(inode, &inode->i_mode, &acl); + mode = inode->i_mode; + if (type == ACL_TYPE_ACCESS && acl) { + ret = posix_acl_update_mode(inode, &mode, &acl); if (ret) goto unlock; - if (mode != inode->i_mode) - mark_inode_dirty(inode); } ret = __gfs2_set_acl(inode, acl, type); + if (!ret && mode != inode->i_mode) { + inode->i_mode = mode; + mark_inode_dirty(inode); + } unlock: if (need_unlock) gfs2_glock_dq_uninit(&gh); -- 2.1.4
