From: Jason Gunthorpe <jgg@ziepe.ca>
To: "Liuyixian (Eason)" <liuyixian@huawei.com>
Cc: dledford@redhat.com, leon@kernel.org, linux-rdma@vger.kernel.org,
linuxarm@huawei.com
Subject: Re: [PATCH for-next 1/2] RDMA/hns: Add the workqueue framework for flush cqe handler
Date: Thu, 7 Nov 2019 14:28:50 -0400 [thread overview]
Message-ID: <20191107182850.GB6730@ziepe.ca> (raw)
In-Reply-To: <e2e0f478-a057-c297-7e1e-d9b09eee2986@huawei.com>
On Thu, Nov 07, 2019 at 08:48:25PM +0800, Liuyixian (Eason) wrote:
>
>
> On 2019/11/7 4:40, Jason Gunthorpe wrote:
> > On Mon, Oct 28, 2019 at 05:45:44PM +0800, Yixian Liu wrote:
> >> @@ -1998,6 +2000,17 @@ static int hns_roce_v2_init(struct hns_roce_dev *hr_dev)
> >> }
> >> }
> >>
> >> + snprintf(workq_name, HNS_ROCE_WORKQ_NAME_LEN - 1,
> >> + "hns_roce_%d_flush_wq", device_id);
> >> + device_id++;
> >> +
> >> + hr_dev->flush_workq = alloc_workqueue(workq_name, WQ_HIGHPRI, 0);
> >> + if (!hr_dev->flush_workq) {
> >
> > Why is this so time critical?
>
> Hi Jason,
>
> I am not quite sure whether you concerned with the flag "WQ_HIGHPRI" or
> why WQ is created in hns_roce_v2_init.
Yes, why do you need a dedicated HIGHPRI work queue.
> If it is WQ_HIGHPRI, yes, it is much better to implement flush operation
> ASAP to help generating flushed cqe as ULP may poll cqe urgently. If you
> concerned allocation stage, as flush operation is support for hip08 only,
> there is no other place proper than here I think.
Why? This is only something that happens in error cases.
> > Why don't you do this from hns_roce_irq_work_handle() ?
>
> As described in the cover letter, here we used CMWQ (concurrency management workqueue)
> to make sure that flush operation can be implemented ASAP. Current irq workqueue is
> a singlethread workqueue, which may delay the flush too long when the system is heavy.
>
> Do you mean we can change irq workqueue to CMWQ to put flush work into it?
As far as I could tell the only thing the triggered the work to run
was some variable which was only set in another work queue 'hns_roce_irq_work_handle()'
> >> +}
> >> +
> >> +void init_flush_work(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
> >> +{
> >> + struct hns_roce_flush_work *flush_work;
> >> +
> >> + flush_work = kzalloc(sizeof(struct hns_roce_flush_work), GFP_ATOMIC);
> >> + if (!flush_work)
> >> + return;
> >
> > Don't do things that can fail here
>
> Do you mean that as "GFP_ATOMIC" is used, the if branch can be deleted?
No, don't do allocations at all if you can't allow them to fail.
Jason
next prev parent reply other threads:[~2019-11-07 18:28 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-28 9:45 [PATCH for-next 0/2] Fix crash due to sleepy mutex while holding lock in post_{send|recv|poll} Yixian Liu
2019-10-28 9:45 ` [PATCH for-next 1/2] RDMA/hns: Add the workqueue framework for flush cqe handler Yixian Liu
2019-11-06 20:40 ` Jason Gunthorpe
2019-11-07 12:48 ` Liuyixian (Eason)
2019-11-07 18:28 ` Jason Gunthorpe [this message]
2019-11-09 10:30 ` Liuyixian (Eason)
2019-10-28 9:45 ` [PATCH for-next 2/2] RDMA/hns: Delayed flush cqe process with workqueue Yixian Liu
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=20191107182850.GB6730@ziepe.ca \
--to=jgg@ziepe.ca \
--cc=dledford@redhat.com \
--cc=leon@kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linuxarm@huawei.com \
--cc=liuyixian@huawei.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).