[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ebb71378-7224-45f6-b47e-3f89eb446fc2@kernel.dk>
Date: Mon, 4 Nov 2024 09:21:17 -0700
From: Jens Axboe <axboe@...nel.dk>
To: hexue <xue01.he@...sung.com>
Cc: asml.silence@...il.com, io-uring@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v9 1/1] io_uring: releasing CPU resources when polling
On 11/4/24 12:29 AM, hexue wrote:
> On 11/1/2024 08:06, Jens Axboe wrote:
>> On 11/1/24 3:19 AM, hexue wrote:
>>> A new hybrid poll is implemented on the io_uring layer. Once IO issued,
>>> it will not polling immediately, but block first and re-run before IO
>>> complete, then poll to reap IO. This poll function could be a suboptimal
>>> solution when running on a single thread, it offers the performance lower
>>> than regular polling but higher than IRQ, and CPU utilization is also lower
>>> than polling.
>>
>> This looks much better now.
>>
>> Do you have a patch for liburing to enable testing of hybrid polling
>> as well? Don't care about perf numbers for that, but it should get
>> exercised.
>
> Sure, I'll add some liburing test cases and submit patch soon.
Thanks! Bonus for also documenting the setup flag in
man/io_uring_setup.2 as well in the liburing repo. If things don't get
documented, then people don't know they exist...
--
Jens Axboe
Powered by blists - more mailing lists