[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7b28925a-cbee-620f-fde7-d16f256836cc@linux.alibaba.com>
Date: Mon, 19 Sep 2022 00:10:15 +0800
From: Liu Song <liusong@...ux.alibaba.com>
To: Jens Axboe <axboe@...nel.dk>, kbusch@...nel.org, axboe@...com,
hch@....de, sagi@...mberg.me
Cc: linux-nvme@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH] nvme: request remote is usually not involved for nvme
devices
On 2022/9/18 00:50, Jens Axboe wrote:
> On 9/17/22 10:40 AM, Liu Song wrote:
>> From: Liu Song <liusong@...ux.alibaba.com>
>>
>> NVMe devices usually have a 1:1 mapping between "ctx" and "hctx",
>> so when "nr_ctx" is equal to 1, there is no possibility of remote
>> request, so the corresponding process can be simplified.
> If the worry is the call overhead of blk_mq_complete_request_remote(),
> why don't we just make that available as an inline instead? That seems
> vastly superior to providing a random shortcut in a driver to avoid
> calling it.
Hi
This is what I think about it. If it is an SSD with only one hw queue,
remote requests will
appear occasionally. As a real multi-queue device, nvme usually does not
have remote
requests, so we don't need to care about it. So even if
"blk_mq_complete_request_remote"
is called, there is a high probability that it will return false, and
the cost of changing the call
to "blk_mq_complete_request_remote" to determine whether
"req->mq_hctx->nr_ctx" is 1 is
not only a function call, but more judgments in
"blk_mq_complete_request_remote".
If "blk_mq_complete_request_remote" is decorated as inline, it also
depends on the compiler,
there is uncertainty.
Thanks
Powered by blists - more mailing lists