Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <4cc86e76-46b7-09ce-65f9-cd27ffe4b26e@gmail.com>
Date:   Tue, 4 Apr 2023 14:21:41 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.7.2
Subject: Re: [PATCH 10/11] io_uring/rsrc: cache struct io_rsrc_node
To:     Gabriel Krisman Bertazi <krisman@suse.de>
Cc:     io-uring@vger.kernel.org, Jens Axboe <axboe@kernel.dk>,
        linux-kernel@vger.kernel.org
References: <cover.1680187408.git.asml.silence@gmail.com>
 <7f5eb1b89e8dcf93739607c79bbf7aec1784cbbe.1680187408.git.asml.silence@gmail.com>
 <87cz4p1083.fsf@suse.de> <6eaadad2-d6a6-dfbb-88aa-8ae68af2f89d@gmail.com>
 <87wn2wzcv3.fsf@suse.de>
Content-Language: en-US
From:   Pavel Begunkov <asml.silence@gmail.com>
In-Reply-To: <87wn2wzcv3.fsf@suse.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 4/1/23 01:04, Gabriel Krisman Bertazi wrote:
> Pavel Begunkov <asml.silence@gmail.com> writes:
> 
>> On 3/31/23 15:09, Gabriel Krisman Bertazi wrote:
>>> Pavel Begunkov <asml.silence@gmail.com> writes:
>>>
>>>> Add allocation cache for struct io_rsrc_node, it's always allocated and
>>>> put under ->uring_lock, so it doesn't need any extra synchronisation
>>>> around caches.
>>> Hi Pavel,
>>> I'm curious if you considered using kmem_cache instead of the custom
>>> cache for this case?  I'm wondering if this provokes visible difference in
>>> performance in your benchmark.
>>
>> I didn't try it, but kmem_cache vs kmalloc, IIRC, doesn't bring us
>> much, definitely doesn't spare from locking, and the overhead
>> definitely wasn't satisfactory for requests before.
> 
> There is no locks in the fast path of slub, as far as I know.  it has a
> per-cpu cache that is refilled once empty, quite similar to the fastpath
> of this cache.  I imagine the performance hit in slub comes from the
> barrier and atomic operations?

Yeah, I mean all kinds of synchronisation. And I don't think
that's the main offender here, the test is single threaded without
contention and the system was mostly idle.

> kmem_cache works fine for most hot paths of the kernel.  I think this

It doesn't for io_uring. There are caches for the net side and now
in the block layer as well. I wouldn't say it necessarily halves
performance but definitely takes a share of CPU.

> custom cache makes sense for the request cache, where objects are
> allocated at an incredibly high rate.  but is this level of update
> frequency a valid use case here?

I can think of some. For example it was of interest before to
install a file for just 2-3 IO operations and also fully bypassing
the normal file table. I rather don't see why we wouldn't use it.

> If it is indeed a significant performance improvement, I guess it is
> fine to have another user of the cache. But I'd be curious to know how
> much of the performance improvement you mentioned in the cover letter is
> due to this patch!

It was definitely sticking out in profiles, 5-10% of cycles, maybe more

-- 
Pavel Begunkov