Another attempt to fix bug #4308 - Excel save operation corrupts file ACLs.

Simo is completely correct. We should be doing the chown *first*, and fail the
ACL set if this fails. The long standing assumption I made when writing the
initial POSIX ACL code was that Windows didn't control who could chown a file
in the same was as POSIX. In POSIX only root can do this whereas I wasn't sure
who could do this in Windows at the time (I didn't understand the privilege
model). So the assumption was that setting the ACL was more important (early
tests showed many failed ACL set's due to inability to chown). But now we have
privileges in smbd, and we must always fail an ACL set when we can't chown
first. The key that Simo noticed is that the CREATOR_OWNER bits in the ACL
incoming are relative to the *new* owner, not the old one. This is why the old
user owner disappears on ACL set - their access was set via the USER_OBJ in the
creator POSIX ACL and when the ownership changes they lose their access.

Patch is simple - just ensure we do the chown first before evaluating the
incoming ACL re-read the owners. We already have code to do this it just wasn't
rigorously being applied.
Jeremy.
(cherry picked from commit 96b819e04cd71a6c899801ae68031bf55b54ea46)

1 files changed, 4 insertions, 25 deletions

diff --git a/source/smbd/posix_acls.c b/source/smbd/posix_acls.c
index 75cca51648d..5ccfb26f595 100644
--- a/source/smbd/posix_acls.c
+++ b/source/smbd/posix_acls.c
@@ -3424,7 +3424,6 @@ NTSTATUS set_nt_acl(files_struct *fsp, uint32 security_info_sent, const SEC_DESC
 	NTSTATUS status;
 	uid_t orig_uid;
 	gid_t orig_gid;
-	bool need_chown = False;
 
 	DEBUG(10,("set_nt_acl: called for file %s\n", fsp->fsp_name ));
 
@@ -3460,14 +3459,12 @@ NTSTATUS set_nt_acl(files_struct *fsp, uint32 security_info_sent, const SEC_DESC
 	}
 
 	/*
-	 * Do we need to chown ?
+	 * Do we need to chown ? If so this must be done first as the incoming
+	 * CREATOR_OWNER acl will be relative to the *new* owner, not the old.
+	 * Noticed by Simo.
 	 */
 
 	if (((user != (uid_t)-1) && (orig_uid != user)) || (( grp != (gid_t)-1) && (orig_gid != grp))) {
-		need_chown = True;
-	}
-
-	if (need_chown && (user == (uid_t)-1 || user == current_user.ut.uid)) {
 
 		DEBUG(3,("set_nt_acl: chown %s. uid = %u, gid = %u.\n",
 				fsp->fsp_name, (unsigned int)user, (unsigned int)grp ));
@@ -3507,9 +3504,6 @@ NTSTATUS set_nt_acl(files_struct *fsp, uint32 security_info_sent, const SEC_DESC
 		orig_mode = sbuf.st_mode;
 		orig_uid = sbuf.st_uid;
 		orig_gid = sbuf.st_gid;
-
-		/* We did chown already, drop the flag */
-		need_chown = False;
 	}
 
 	create_file_sids(&sbuf, &file_owner_sid, &file_grp_sid);
@@ -3660,24 +3654,9 @@ NTSTATUS set_nt_acl(files_struct *fsp, uint32 security_info_sent, const SEC_DESC
 		}
 
 		free_canon_ace_list(file_ace_list);
-		free_canon_ace_list(dir_ace_list); 
+		free_canon_ace_list(dir_ace_list);
 	}
 
-	/* Any chown pending? */
-	if (need_chown) {
-		DEBUG(3,("set_nt_acl: chown %s. uid = %u, gid = %u.\n",
-			 fsp->fsp_name, (unsigned int)user, (unsigned int)grp ));
-		
-		if(try_chown( fsp->conn, fsp->fsp_name, user, grp) == -1) {
-			DEBUG(3,("set_nt_acl: chown %s, %u, %u failed. Error = %s.\n",
-				 fsp->fsp_name, (unsigned int)user, (unsigned int)grp, strerror(errno) ));
-			if (errno == EPERM) {
-				return NT_STATUS_INVALID_OWNER;
-			}
-			return map_nt_error_from_unix(errno);
-		}
-	}
-	
 	return NT_STATUS_OK;
 }