From: Jens Axboe <axboe@kernel.dk>
To: Bart Van Assche <bvanassche@acm.org>,
Christoph Hellwig <hch@infradead.org>
Cc: Sagi Grimberg <sagig@mellanox.com>,
Sebastian Parschauer <sebastian.riemer@profitbricks.com>,
Robert Elliott <Elliott@hp.com>,
Ming Lei <ming.lei@canonical.com>,
"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>,
linux-rdma <linux-rdma@vger.kernel.org>
Subject: Re: [PATCH v2 01/12] blk-mq: Use all available hardware queues
Date: Tue, 07 Oct 2014 08:37:03 -0600 [thread overview]
Message-ID: <5433FA8F.3050100@kernel.dk> (raw)
In-Reply-To: <5433E473.4020102@acm.org>
On 10/07/2014 07:02 AM, Bart Van Assche wrote:
> Suppose that a system has two CPU sockets, three cores per socket,
> that it does not support hyperthreading and that four hardware
> queues are provided by a block driver. With the current algorithm
> this will lead to the following assignment of CPU cores to hardware
> queues:
>
> HWQ 0: 0 1
> HWQ 1: 2 3
> HWQ 2: 4 5
> HWQ 3: (none)
>
> This patch changes the queue assignment into:
>
> HWQ 0: 0 1
> HWQ 1: 2
> HWQ 2: 3 4
> HWQ 3: 5
>
> In other words, this patch has the following three effects:
> - All four hardware queues are used instead of only three.
> - CPU cores are spread more evenly over hardware queues. For the
> above example the range of the number of CPU cores associated
> with a single HWQ is reduced from [0..2] to [1..2].
> - If the number of HWQ's is a multiple of the number of CPU sockets
> it is now guaranteed that all CPU cores associated with a single
> HWQ reside on the same CPU socket.
>
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> Reviewed-by: Sagi Grimberg <sagig@mellanox.com>
> Cc: Jens Axboe <axboe@fb.com>
> Cc: Christoph Hellwig <hch@lst.de>
> Cc: Ming Lei <ming.lei@canonical.com>
> ---
> block/blk-mq-cpumap.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/block/blk-mq-cpumap.c b/block/blk-mq-cpumap.c
> index 1065d7c..8e56455 100644
> --- a/block/blk-mq-cpumap.c
> +++ b/block/blk-mq-cpumap.c
> @@ -17,7 +17,7 @@
> static int cpu_to_queue_index(unsigned int nr_cpus, unsigned int nr_queues,
> const int cpu)
> {
> - return cpu / ((nr_cpus + nr_queues - 1) / nr_queues);
> + return cpu * nr_queues / nr_cpus;
> }
>
> static int get_first_sibling(unsigned int cpu)
Lets do this separate, as explained last time, it needs to be evaluated
on its own and doesn't really belong in this series of patches.
--
Jens Axboe
next prev parent reply other threads:[~2014-10-07 14:37 UTC|newest]
Thread overview: 83+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-10-07 13:01 [PATCH v2 0/12] IB/srp: Add multichannel support Bart Van Assche
2014-10-07 13:03 ` [PATCH v2 02/12] blk-mq: Add blk_mq_unique_tag() Bart Van Assche
2014-10-11 11:08 ` Christoph Hellwig
2014-10-13 9:21 ` Bart Van Assche
[not found] ` <543B99B2.1010307-HInyCGIudOg@public.gmane.org>
2014-10-13 10:15 ` Christoph Hellwig
2014-10-19 16:14 ` Sagi Grimberg
[not found] ` <5433E493.9030304-HInyCGIudOg@public.gmane.org>
2014-10-28 1:55 ` Martin K. Petersen
2014-10-07 13:04 ` [PATCH v2 04/12] scsi_tcq.h: Add support for multiple hardware queues Bart Van Assche
2014-10-19 16:12 ` Sagi Grimberg
[not found] ` <5443E2DF.1040605-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-10-20 12:01 ` Bart Van Assche
[not found] ` <5444F995.5080407-HInyCGIudOg@public.gmane.org>
2014-10-21 8:49 ` Christoph Hellwig
2014-10-21 8:59 ` Sagi Grimberg
2014-10-28 2:06 ` Martin K. Petersen
2014-10-07 13:05 ` [PATCH v2 07/12] IB/srp: Avoid that I/O hangs due to a cable pull during LUN scanning Bart Van Assche
2014-10-19 16:27 ` Sagi Grimberg
[not found] ` <5443E66F.7050901-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-10-20 12:15 ` Bart Van Assche
2014-10-21 8:50 ` Christoph Hellwig
2014-10-07 13:05 ` [PATCH v2 08/12] IB/srp: Introduce two new srp_target_port member variables Bart Van Assche
2014-10-19 16:30 ` Sagi Grimberg
2014-10-07 13:06 ` [PATCH v2 10/12] IB/srp: Use block layer tags Bart Van Assche
[not found] ` <5433E557.3010505-HInyCGIudOg@public.gmane.org>
2014-10-17 10:58 ` Christoph Hellwig
[not found] ` <20141017105858.GA7819-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-20 11:44 ` Bart Van Assche
2014-10-22 22:03 ` Elliott, Robert (Server Storage)
[not found] ` <94D0CD8314A33A4D9D801C0FE68B4029593212E0-wwDBVnaDRpYSZAcGdq5asR6epYMZPwEe5NbjCUgZEJk@public.gmane.org>
2014-10-23 7:16 ` Bart Van Assche
2014-10-23 17:43 ` Webb Scales
[not found] ` <54493E5A.7050803-VXdhtT5mjnY@public.gmane.org>
2014-10-24 6:45 ` Bart Van Assche
[not found] ` <5449F571.7080308-HInyCGIudOg@public.gmane.org>
2014-10-24 15:40 ` Webb Scales
2014-10-23 8:47 ` Christoph Hellwig
2014-10-24 4:43 ` Elliott, Robert (Server Storage)
2014-10-24 6:45 ` Christoph Hellwig
[not found] ` <20141024064514.GA15654-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-31 17:34 ` Hannes Reinecke
2014-11-03 7:52 ` Kashyap Desai
2014-11-03 8:25 ` Christoph Hellwig
[not found] ` <5433E43D.3010107-HInyCGIudOg@public.gmane.org>
2014-10-07 13:02 ` [PATCH v2 01/12] blk-mq: Use all available hardware queues Bart Van Assche
2014-10-07 14:37 ` Jens Axboe [this message]
[not found] ` <5433FA8F.3050100-tSWWG44O7X1aa/9Udqfwiw@public.gmane.org>
2014-10-08 13:21 ` Bart Van Assche
[not found] ` <54353A74.7040406-HInyCGIudOg@public.gmane.org>
2014-10-11 11:11 ` Christoph Hellwig
[not found] ` <20141011111114.GB9593-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-13 9:45 ` Bart Van Assche
[not found] ` <543B9F47.2090204-HInyCGIudOg@public.gmane.org>
2014-10-17 13:20 ` Christoph Hellwig
[not found] ` <20141017132053.GF16538-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-17 14:11 ` Sagi Grimberg
2014-10-07 13:03 ` [PATCH v2 03/12] scsi-mq: Add support for multiple " Bart Van Assche
[not found] ` <5433E4AB.8030306-HInyCGIudOg@public.gmane.org>
2014-10-19 15:54 ` Sagi Grimberg
2014-10-28 2:01 ` Martin K. Petersen
2014-10-29 12:22 ` Bart Van Assche
2014-10-29 12:27 ` Bart Van Assche
[not found] ` <5450DD49.6090108-HInyCGIudOg@public.gmane.org>
2014-10-30 0:53 ` Martin K. Petersen
2014-10-07 13:04 ` [PATCH v2 05/12] IB/srp: Move ib_destroy_cm_id() call into srp_free_ch_ib() Bart Van Assche
2014-10-07 13:04 ` [PATCH v2 06/12] IB/srp: Remove stale connection retry mechanism Bart Van Assche
2014-10-07 13:05 ` [PATCH v2 09/12] IB/srp: Separate target and channel variables Bart Van Assche
2014-10-19 16:48 ` Sagi Grimberg
2014-10-07 13:06 ` [PATCH v2 11/12] IB/srp: Eliminate free_reqs list Bart Van Assche
[not found] ` <5433E56E.6010600-HInyCGIudOg@public.gmane.org>
2014-10-17 10:59 ` Christoph Hellwig
[not found] ` <20141017105939.GB7819-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-19 16:59 ` Sagi Grimberg
2014-10-20 11:47 ` Bart Van Assche
2014-10-21 8:49 ` Christoph Hellwig
2014-10-07 13:07 ` [PATCH v2 12/12] IB/srp: Add multichannel support Bart Van Assche
2014-10-17 11:01 ` EH action after scsi_remove_host, was: " Christoph Hellwig
2014-10-20 13:53 ` Bart Van Assche
2014-10-21 8:51 ` Christoph Hellwig
2014-10-17 11:06 ` Christoph Hellwig
[not found] ` <20141017110627.GD7819-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2014-10-20 11:57 ` Bart Van Assche
2014-10-21 8:49 ` Christoph Hellwig
[not found] ` <5433E585.607-HInyCGIudOg@public.gmane.org>
2014-10-19 17:36 ` Sagi Grimberg
2014-10-20 12:56 ` Bart Van Assche
[not found] ` <54450690.709-HInyCGIudOg@public.gmane.org>
2014-10-21 9:10 ` Sagi Grimberg
[not found] ` <544622FE.5040906-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-10-28 18:32 ` Sagi Grimberg
[not found] ` <544FE13A.60807-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-10-29 10:52 ` Bart Van Assche
2014-10-30 14:19 ` Sagi Grimberg
2014-10-30 14:36 ` Bart Van Assche
[not found] ` <54524D08.4040203-HInyCGIudOg@public.gmane.org>
2014-10-30 15:06 ` Sagi Grimberg
[not found] ` <545253E3.7000009-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-10-30 15:19 ` Bart Van Assche
[not found] ` <545256E5.9010501-HInyCGIudOg@public.gmane.org>
2014-10-30 17:33 ` Sagi Grimberg
2014-10-31 9:19 ` Bart Van Assche
[not found] ` <5453541D.7040206-HInyCGIudOg@public.gmane.org>
2014-11-02 13:03 ` Sagi Grimberg
2014-11-03 1:46 ` Elliott, Robert (Server Storage)
2014-11-04 11:46 ` Bart Van Assche
[not found] ` <5458BC8B.40202-HInyCGIudOg@public.gmane.org>
2014-11-04 12:15 ` Sagi Grimberg
[not found] ` <5458C344.2040109-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2014-11-05 4:57 ` Elliott, Robert (Server Storage)
[not found] ` <94D0CD8314A33A4D9D801C0FE68B40295937104F-2m9nI20wMFwSZAcGdq5asR6epYMZPwEe5NbjCUgZEJk@public.gmane.org>
2014-11-05 11:22 ` Sagi Grimberg
2014-10-21 9:14 ` Sagi Grimberg
2014-10-29 12:36 ` Bart Van Assche
2014-10-30 14:22 ` Sagi Grimberg
2014-10-08 13:16 ` [PATCH] blk-mq: Use all available hardware queues Bart Van Assche
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5433FA8F.3050100@kernel.dk \
--to=axboe@kernel.dk \
--cc=Elliott@hp.com \
--cc=bvanassche@acm.org \
--cc=hch@infradead.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=ming.lei@canonical.com \
--cc=sagig@mellanox.com \
--cc=sebastian.riemer@profitbricks.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.