Hi Linus,
A set of fixes that should go into this release. This pull request
contains:
- Ensure ->result is always set when IO is retried (Bijan)
- In conjunction with the above, fix a regression in polled IO issue
when retried (me/Bijan)
- Don't setup async context for read/write fixed, otherwise we may
wrongly map the iovec on retry (me)
- Cancel io-wq work if we fail getting mm reference (me)
- Ensure dependent work is always initialized correctly (me)
- Only allow original task to submit IO, don't allow it from a passed
ring fd (me)
Please pull!
git://git.kernel.dk/linux-block.git tags/io_uring-5.5-2020-01-16
----------------------------------------------------------------
Bijan Mottahedeh (1):
io_uring: clear req->result always before issuing a read/write request
Jens Axboe (5):
io_uring: don't setup async context for read/write fixed
io-wq: cancel work if we fail getting a mm reference
io_uring: be consistent in assigning next work from handler
io_uring: ensure workqueue offload grabs ring mutex for poll list
io_uring: only allow submit from owning task
fs/io-wq.c | 12 ++++++----
fs/io_uring.c | 72 +++++++++++++++++++++++++++++++++++++++--------------------
2 files changed, 56 insertions(+), 28 deletions(-)
--
Jens Axboe
