io_uring/af_unix: defer registered files gc to io_uring release
authorPavel Begunkov <asml.silence@gmail.com>
Sun, 16 Oct 2022 22:31:26 +0000 (23:31 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 26 Oct 2022 11:25:55 +0000 (13:25 +0200)
commitc378c479c5175833bb22ff71974cda47d7b05401
treed64c25639bfb238e1542c57b175d67dc370148cb
parent67cbc8865a66533fa08c1c13fe9acbaaae63c403
io_uring/af_unix: defer registered files gc to io_uring release

[ upstream commit 0091bfc81741b8d3aeb3b7ab8636f911b2de6e80 ]

Instead of putting io_uring's registered files in unix_gc() we want it
to be done by io_uring itself. The trick here is to consider io_uring
registered files for cycle detection but not actually putting them down.
Because io_uring can't register other ring instances, this will remove
all refs to the ring file triggering the ->release path and clean up
with io_ring_ctx_free().

Cc: stable@vger.kernel.org
Fixes: 6b06314c47e1 ("io_uring: add file set registration")
Reported-and-tested-by: David Bouman <dbouman03@gmail.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
[axboe: add kerneldoc comment to skb, fold in skb leak fix]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/io_uring.c
include/linux/skbuff.h
net/unix/garbage.c