From: yangerkun <yangerkun@huawei.com>
To: <yukuai@fygo.io>, <josef@toxicpanda.com>, <axboe@kernel.dk>,
<hch@lst.de>
Cc: <yi.zhang@huawei.com>, <chengzhihao1@huawei.com>,
<echo.chenlin@huawei.com>, <leo.lilong@huaweicloud.com>,
<wangkefeng.wang@huawei.com>, <huawei.libin@huawei.com>,
<leijitang@huawei.com>, <linux-block@vger.kernel.org>,
<nbd@other.debian.org>
Subject: Re: [PATCH v8 8/8] nbd: add pre_defined_connections module parameter for pre-created devices
Date: Thu, 6 Aug 2026 19:23:06 +0800 [thread overview]
Message-ID: <ac98320c-fa54-42d9-b792-ed0c8841920f@huawei.com> (raw)
In-Reply-To: <22a4c13f-517a-42c3-9079-15dd20940424@fygo.io>
在 2026/8/6 17:26, yu kuai 写道:
> Hi,
>
> 在 2026/8/5 20:29, Yang Erkun 写道:
>> blk_mq_update_nr_hw_queues() in nbd_start_device() may cause a queue
>> freeze. The previous commit addressed this for newly created nbd
>> devices by setting the expected nr_hw_queues in nbd_dev_add(). However,
>> when reusing an old inactive nbd device, the queue freeze can still
>> occur if the old nbd->tag_set->nr_hw_queues does not match the new
>> socket connection count. Inactive nbd devices can originate from two
>> sources: loading the nbd module with nbds_max, which sets the default
>> nr_hw_queues to 1, and the netlink method, which sets nr_hw_queues
>> according to the expected number of socket connections. For the first
>> case, add a module parameter so the default nr_hw_queues can be
>> changed. Users who know their expected number of connections can then
>> prevent queue freezes on pre-created devices via nbds_max.
>>
>> Before this patchset:
>> real 0m2.195s
>> user 0m0.005s
>> sys 0m0.022s
>>
>> After this patchset:
>> real 0m0.090s
>> user 0m0.004s
>> sys 0m0.018s
>>
>> Signed-off-by: Yang Erkun <yangerkun@huawei.com>
>> ---
>> drivers/block/nbd.c | 16 +++++++++++++++-
>> 1 file changed, 15 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
>> index 6c4aa19ea11a..ffce519bf008 100644
>> --- a/drivers/block/nbd.c
>> +++ b/drivers/block/nbd.c
>> @@ -166,6 +166,7 @@ static struct dentry *nbd_dbg_dir;
>>
>> static unsigned int nbds_max = 16;
>> static int max_part = 16;
>> +static int pre_defined_connections = 1;
>> static int part_shift;
>>
>> static int nbd_dev_dbg_init(struct nbd_device *nbd);
>> @@ -2712,6 +2713,12 @@ static int __init nbd_init(void)
>> if (nbds_max > 1UL << (MINORBITS - part_shift))
>> return -EINVAL;
>>
>> + /* An excessively large value will be adjusted in blk_mq_alloc_tag_set */
>> + if (pre_defined_connections < 1) {
>> + pr_err("pre_defined_connections must be >= 1\n");
>
> It's also better to add NR_CPUS checking for upper bound, because the nr_hw_queues is passed
> directly to __blk_mq_realloc_hw_ctxs() for memory allocation.
Hi,
Like the comment above said, nbd_dev_add->blk_mq_alloc_tag_set
4876 if (set->nr_maps == 1 && set->nr_hw_queues > nr_cpu_ids)
4877 set->nr_hw_queues = nr_cpu_ids;
this will help do the same thing. So I does not the check here.
Thanks,
Erkun.
>
>> + return -EINVAL;
>> + }
>> +
>> if (register_blkdev(NBD_MAJOR, "nbd"))
>> return -EIO;
>>
>> @@ -2728,8 +2735,12 @@ static int __init nbd_init(void)
>> }
>> nbd_dbg_init();
>>
>> + /*
>> + * Set to the intended connection count so nbd_start_device() can skip
>> + * the queue-freezing blk_mq_update_nr_hw_queues() call.
>> + */
>> for (i = 0; i < nbds_max; i++)
>> - nbd_dev_add(i, 1, 1);
>> + nbd_dev_add(i, 1, pre_defined_connections);
>> return 0;
>> }
>>
>> @@ -2790,3 +2801,6 @@ module_param(nbds_max, int, 0444);
>> MODULE_PARM_DESC(nbds_max, "number of network block devices to initialize (default: 16)");
>> module_param(max_part, int, 0444);
>> MODULE_PARM_DESC(max_part, "number of partitions per device (default: 16)");
>> +module_param(pre_defined_connections, int, 0444);
>> +MODULE_PARM_DESC(pre_defined_connections,
>> +"number of connections for devices pre-created at module load (default: 1)");
>
next prev parent reply other threads:[~2026-08-06 11:23 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-05 12:29 [PATCH v8 0/8] nbd: eliminate queue freeze/unfreeze overhead in connection setup Yang Erkun
2026-08-05 12:29 ` [PATCH v8 1/8] nbd: simplify find_fallback() by removing redundant logic Yang Erkun
2026-08-05 12:29 ` [PATCH v8 2/8] nbd: disallow NBD_SET_SOCK on an active device Yang Erkun
2026-08-05 12:29 ` [PATCH v8 3/8] nbd: clear queue limits on disconnect Yang Erkun
2026-08-05 12:29 ` [PATCH v8 4/8] nbd: remove queue freeze in nbd_add_socket Yang Erkun
2026-08-05 12:29 ` [PATCH v8 5/8] nbd: skip queue freeze when setting size at device startup Yang Erkun
2026-08-05 12:29 ` [PATCH v8 6/8] nbd: factor out a nbd_genl_foreach_sock Yang Erkun
2026-08-05 12:29 ` [PATCH v8 7/8] nbd: remove queue freeze for newly created nbd from netlink path Yang Erkun
2026-08-05 12:29 ` [PATCH v8 8/8] nbd: add pre_defined_connections module parameter for pre-created devices Yang Erkun
2026-08-06 9:26 ` yu kuai
2026-08-06 11:23 ` yangerkun [this message]
2026-08-07 8:12 ` yu kuai
2026-08-07 9:26 ` yangerkun
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=ac98320c-fa54-42d9-b792-ed0c8841920f@huawei.com \
--to=yangerkun@huawei.com \
--cc=axboe@kernel.dk \
--cc=chengzhihao1@huawei.com \
--cc=echo.chenlin@huawei.com \
--cc=hch@lst.de \
--cc=huawei.libin@huawei.com \
--cc=josef@toxicpanda.com \
--cc=leijitang@huawei.com \
--cc=leo.lilong@huaweicloud.com \
--cc=linux-block@vger.kernel.org \
--cc=nbd@other.debian.org \
--cc=wangkefeng.wang@huawei.com \
--cc=yi.zhang@huawei.com \
--cc=yukuai@fygo.io \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox