sin/samba-mirror
1
0
Fork 0
mirror of https://github.com/samba-team/samba.git synced 2024-12-22 13:34:15 +03:00
Code

smbd: improve reinit_after_fork error handling

Browse Source 
reinit_after_fork() may panic or return an error on failure. When smbd
is started in interactive mode, it ignores the reinit_after_fork()
return status and unconditionally proceeds to smbd_process().

Similarly, if messaging_reinit() fails within reinit_after_fork() then
it will subsequently call ctdb_async_ctx_reinit() if clustering is
enabled.

There's no reason why these errors shouldn't be handled immediately, so
add appropriate error handling.

Found by code inspection; not seen in the wild.

Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Wed Sep  4 09:53:01 UTC 2024 on atb-devel-224
This commit is contained in:
David Disseldorp 2024-09-04 02:54:09 +00:00
parent ecb8a99a2c
commit 22edd69503
2 changed files with 8 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
source3/lib/util.c
View File

@ -482,6 +482,7 @@ NTSTATUS reinit_after_fork(struct messaging_context *msg_ctx,
if (!NT_STATUS_IS_OK(status)) {
DEBUG(0,("messaging_reinit() failed: %s\n",
nt_errstr(status)));
goto done;
}
if (lp_clustering()) {

8
source3/smbd/server.c
View File

@ -976,7 +976,13 @@ static void smbd_accept_connection(struct tevent_context *ev,
smb_set_close_on_exec(fd);
if (s->parent->interactive) {
reinit_after_fork(msg_ctx, ev, true);
NTSTATUS status;
status = reinit_after_fork(msg_ctx, ev, true);
if (!NT_STATUS_IS_OK(status)) {
exit_server("reinit_after_fork() failed");
return;
}
smbd_process(ev, msg_ctx, fd, true);
exit_server_cleanly("end of interactive mode");
return;