Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux

exec: Correct comments about "point of no return"

In commit 221af7f87b97 ("Split 'flush_old_exec' into two functions"),
the comment about the point of no return should have stayed in
flush_old_exec() since it refers to "bprm->mm = NULL;" line, but prior
changes in commits c89681ed7d0e ("remove steal_locks()"), and
fd8328be874f ("sanitize handling of shared descriptor tables in failing
execve()") made it look like it meant the current->sas_ss_sp line instead.

The comment was referring to the fact that once bprm->mm is NULL, all
failures from a binfmt load_binary hook (e.g. load_elf_binary), will
get SEGV raised against current. Move this comment and expand the
explanation a bit, putting it above the assignment this time, and add
details about the true nature of "point of no return" being the call
to flush_old_exec() itself.

This also removes an erroneous commet about when credentials are being
installed. That has its own dedicated function, install_exec_creds(),
which carries a similar (and correct) comment, so remove the bogus comment
where installation is not actually happening.

Cc: David Howells <dhowells@redhat.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Serge Hallyn <serge@hallyn.com>

+13 -3
+13 -3
fs/exec.c
··· 1259 1259 perf_event_comm(tsk, exec); 1260 1260 } 1261 1261 1262 + /* 1263 + * Calling this is the point of no return. None of the failures will be 1264 + * seen by userspace since either the process is already taking a fatal 1265 + * signal (via de_thread() or coredump), or will have SEGV raised 1266 + * (after exec_mmap()) by search_binary_handlers (see below). 1267 + */ 1262 1268 int flush_old_exec(struct linux_binprm * bprm) 1263 1269 { 1264 1270 int retval; ··· 1292 1286 if (retval) 1293 1287 goto out; 1294 1288 1295 - bprm->mm = NULL; /* We're using it now */ 1289 + /* 1290 + * After clearing bprm->mm (to mark that current is using the 1291 + * prepared mm now), we have nothing left of the original 1292 + * process. If anything from here on returns an error, the check 1293 + * in search_binary_handler() will SEGV current. 1294 + */ 1295 + bprm->mm = NULL; 1296 1296 1297 1297 set_fs(USER_DS); 1298 1298 current->flags &= ~(PF_RANDOMIZE | PF_FORKNOEXEC | PF_KTHREAD | ··· 1345 1333 { 1346 1334 arch_pick_mmap_layout(current->mm); 1347 1335 1348 - /* This is the point of no return */ 1349 1336 current->sas_ss_sp = current->sas_ss_size = 0; 1350 1337 1351 1338 if (uid_eq(current_euid(), current_uid()) && gid_eq(current_egid(), current_gid())) ··· 1362 1351 */ 1363 1352 current->mm->task_size = TASK_SIZE; 1364 1353 1365 - /* install the new credentials */ 1366 1354 if (!uid_eq(bprm->cred->uid, current_euid()) || 1367 1355 !gid_eq(bprm->cred->gid, current_egid())) { 1368 1356 current->pdeath_signal = 0;