Re: [PATCHSET 0/2] io_uring: handle short reads internally

[Jens Axboe <axboe@kernel.dk>]

On 8/13/20 2:49 PM, Jeff Moyer wrote:
> Jens Axboe <axboe@kernel.dk> writes:
> 
>> On 8/13/20 2:37 PM, Jens Axboe wrote:
>>> On 8/13/20 2:33 PM, Jens Axboe wrote:
>>>> On 8/13/20 2:25 PM, Jeff Moyer wrote:
>>>>> Jens Axboe <axboe@kernel.dk> writes:
>>>>>
>>>>>> Since we've had a few cases of applications not dealing with this
>>>>>> appopriately, I believe the safest course of action is to ensure that
>>>>>> we don't return short reads when we really don't have to.
>>>>>>
>>>>>> The first patch is just a prep patch that retains iov_iter state over
>>>>>> retries, while the second one actually enables just doing retries if
>>>>>> we get a short read back.
>>>>>
>>>>> Have you run this through the liburing regression tests?
>>>>>
>>>>> Tests  <eeed8b54e0df-test> <timeout-overflow> <read-write> failed
>>>>>
>>>>> I'll take a look at the failures, but wanted to bring it to your
>>>>> attention sooner rather than later.  I was using your io_uring-5.9
>>>>> branch.
>>>>
>>>> The eed8b54e0df-test failure is known with this one, pretty sure it
>>>> was always racy, but I'm looking into it.
>>>>
>>>> The timeout-overflow test needs fixing, it's just an ordering thing
>>>> with the batched completions done through submit. Not new with these
>>>> patches.
> 
> OK.
> 
>>>> The read-write one I'm interested in, what did you run it on? And
>>>> what was the failure?
>>>
>>> BTW, what git sha did you run?
>>
>> I do see a failure with dm on that, I'll take a look.
> 
> I ran it on a file system atop nvme with 8 poll queues.
> 
> liburing head: 9e1d69e078ee51f253a829ff421b17cfc996d158
> linux-block head: ff1353802d86a9d8e40ef1377efb12a1d3000a20

Fixed it, and actually enabled a further cleanup.

> The error I saw was:
> 
> Running test read-write:
> Non-vectored IO not supported, skipping
> cqe res -22, wanted 2048
> test_buf_select_short vec failed
> Test read-write failed with ret 1
> 
> But I don't think it was due to these two commits.

Yeah don't think so either.

-- 
Jens Axboe