io-wq: ensure we exit if thread group is exiting

Dave reports that a coredumping workload gets stuck in 5.15-rc2, and
identified the culprit in the Fixes line below. The problem is that
relying solely on fatal_signal_pending() to gate whether to exit or not
fails miserably if a process gets eg SIGILL sent. Don't exclusively
rely on fatal signals, also check if the thread group is exiting.

Fixes: 15e20db2e0ce ("io-wq: only exit on fatal signals")
Reported-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/fs/io-wq.c b/fs/io-wq.c
index c2e0e8e80949a7089268e2854a6b48c5802825f3..c2360cdc403dd42bbe395b41e4b7de3c9f876cef 100644 (file)
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -584,7 +584,8 @@ loop:
 
                        if (!get_signal(&ksig))
                                continue;
-                       if (fatal_signal_pending(current))
+                       if (fatal_signal_pending(current) ||
+                           signal_group_exit(current->signal))
                                break;
                        continue;
                }