OSDN Git Service

io_uring: io_allocate_scq_urings() should return a sane state
authorJens Axboe <axboe@kernel.dk>
Wed, 20 Nov 2019 16:26:29 +0000 (09:26 -0700)
committerJens Axboe <axboe@kernel.dk>
Tue, 26 Nov 2019 02:56:06 +0000 (19:56 -0700)
commiteb065d301e8c83643367bdb0898becc364046bda
treedaa735914c94651012eaf4ef7776d23e2417fc5c
parentbbad27b2f622fa26d107f8a72c0cd5cc102dc56e
io_uring: io_allocate_scq_urings() should return a sane state

We currently rely on the ring destroy on cleaning things up in case of
failure, but io_allocate_scq_urings() can leave things half initialized
if only parts of it fails.

Be nice and return with either everything setup in success, or return an
error with things nicely cleaned up.

Reported-by: syzbot+0d818c0d39399188f393@syzkaller.appspotmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
fs/io_uring.c