From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bart Van Assche Subject: Re: [PATCH 08/14] blk-mq: separate number of hardware queues from nr_cpu_ids Date: Mon, 29 Oct 2018 11:31:16 -0700 Message-ID: <1540837876.196084.75.camel@acm.org> References: <20181029163738.10172-1-axboe@kernel.dk> <20181029163738.10172-9-axboe@kernel.dk> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-7" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20181029163738.10172-9-axboe@kernel.dk> Sender: linux-kernel-owner@vger.kernel.org To: Jens Axboe , linux-block@vger.kernel.org, linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: linux-scsi@vger.kernel.org On Mon, 2018-10-29 at 10:37 -0600, Jens Axboe wrote: +AD4 With multiple maps, nr+AF8-cpu+AF8-ids is no longer the maximum number of +AD4 hardware queues we support on a given devices. The initializer of +AD4 the tag+AF8-set can have set -+AD4-nr+AF8-hw+AF8-queues larger than the available +AD4 number of CPUs, since we can exceed that with multiple queue maps. +AD4 +AD4 Reviewed-by: Hannes Reinecke +ADw-hare+AEA-suse.com+AD4 +AD4 Signed-off-by: Jens Axboe +ADw-axboe+AEA-kernel.dk+AD4 +AD4 --- +AD4 block/blk-mq.c +AHw 28 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-------- +AD4 1 file changed, 21 insertions(+-), 7 deletions(-) +AD4 +AD4 diff --git a/block/blk-mq.c b/block/blk-mq.c +AD4 index 0fab36372ace..60a951c4934c 100644 +AD4 --- a/block/blk-mq.c +AD4 +-+-+- b/block/blk-mq.c +AD4 +AEAAQA -2663,6 +-2663,19 +AEAAQA static void blk+AF8-mq+AF8-realloc+AF8-hw+AF8-ctxs(struct blk+AF8-mq+AF8-tag+AF8-set +ACo-set, +AD4 mutex+AF8-unlock(+ACY-q-+AD4-sysfs+AF8-lock)+ADs +AD4 +AH0 +AD4 +AD4 +-/+ACo +AD4 +- +ACo Maximum number of queues we support. For single sets, we'll never have +AF4 hardware? +AD4 +- +ACo more than the CPUs (software queues). For multiple sets, the tag+AF8-set +AD4 +- +ACo user may have set -+AD4-nr+AF8-hw+AF8-queues larger. +AD4 +- +ACo-/ Anyway: Reviewed-by: Bart Van Assche +ADw-bvanassche+AEA-acm.org+AD4