Re: [PATCH for-next v2 10/25] io_uring: kill REQ_F_COMPLETE_INLINE

[Pavel Begunkov <asml.silence@gmail.com>]

On 6/15/22 11:18, Pavel Begunkov wrote:
> On 6/15/22 09:20, Hao Xu wrote:
>> On 6/14/22 22:37, Pavel Begunkov wrote:
>>> REQ_F_COMPLETE_INLINE is only needed to delay queueing into the
>>> completion list to io_queue_sqe() as __io_req_complete() is inlined and
>>> we don't want to bloat the kernel.
>>>
>>> As now we complete in a more centralised fashion in io_issue_sqe() we
>>> can get rid of the flag and queue to the list directly.
>>>
>>> Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
>>> ---
>>>   io_uring/io_uring.c       | 20 ++++++++------------
>>>   io_uring/io_uring.h       |  5 -----
>>>   io_uring/io_uring_types.h |  3 ---
>>>   3 files changed, 8 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
>>> index 1fb93fdcfbab..fcee58c6c35e 100644
>>> --- a/io_uring/io_uring.c
>>> +++ b/io_uring/io_uring.c
>>> @@ -1278,17 +1278,14 @@ static void io_req_complete_post32(struct io_kiocb *req, u64 extra1, u64 extra2)
>>>   inline void __io_req_complete(struct io_kiocb *req, unsigned issue_flags)
>>>   {
>>> -    if (issue_flags & IO_URING_F_COMPLETE_DEFER)
>>> -        io_req_complete_state(req);
>>> -    else
>>> -        io_req_complete_post(req);
>>> +    io_req_complete_post(req);
>>>   }
>>
>> io_read/write and provide_buffers/remove_buffers are still using
>> io_req_complete() in their own function. By removing the
>> IO_URING_F_COMPLETE_DEFER branch they will end in complete_post path
>> 100% which we shouldn't.
> 
> Old provided buffers are such a useful feature that Jens adds
> a new ring-based version of it, so I couldn't care less about
> those two.
> 
> I any case, let's leave it to follow ups. Those locking is a
> weird construct and shouldn't be done this ad-hook way, it's
> a potential bug nest

Ok, missed io_read/write part, that's a problem, agree

-- 
Pavel Begunkov