block: an exiting task should be allowed to create io_context

While fixing io_context creation / task exit race condition,
6e736be7f2 "block: make ioc get/put interface more conventional and
fix race on alloction" also prevented an exiting (%PF_EXITING) task
from creating its own io_context.  This is incorrect as exit path may
issue IOs, e.g. from exit_files(), and if those IOs are the first ones
issued by the task, io_context needs to be created to process the IOs.

Combined with the existing problem of io_context / io_cq creation
failure having the possibility of stalling IO, this problem results in
deterministic full IO lockup with certain workloads.

Fix it by allowing io_context creation regardless of %PF_EXITING for
%current.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Reported-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/block/blk-ioc.c b/block/blk-ioc.c
index ce9b35a9468eb3816dc9bbfbf66e9d20b17d7cca..33fae7df16a016d7ccbb0ccb6b3f78e47d9f0765 100644 (file)
--- a/block/blk-ioc.c
+++ b/block/blk-ioc.c
@@ -281,9 +281,16 @@ void create_io_context_slowpath(struct task_struct *task, gfp_t gfp_flags,
        INIT_HLIST_HEAD(&ioc->icq_list);
        INIT_WORK(&ioc->release_work, ioc_release_fn);
 
-       /* try to install, somebody might already have beaten us to it */
+       /*
+        * Try to install.  ioc shouldn't be installed if someone else
+        * already did or @task, which isn't %current, is exiting.  Note
+        * that we need to allow ioc creation on exiting %current as exit
+        * path may issue IOs from e.g. exit_files().  The exit path is
+        * responsible for not issuing IO after exit_io_context().
+        */
        task_lock(task);
-       if (!task->io_context && !(task->flags & PF_EXITING))
+       if (!task->io_context &&
+           (task == current || !(task->flags & PF_EXITING)))
                task->io_context = ioc;
        else
                kmem_cache_free(iocontext_cachep, ioc);