From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: dev@dpdk.org, Ilya Maximets <i.maximets@samsung.com>
Subject: Re: [dpdk-dev] [PATCH 4/8] net/virtio: allocate queue at init stage
Date: Fri, 4 Nov 2016 09:08:23 +0100 [thread overview]
Message-ID: <16f9caf1-423d-ceee-2b26-811f196201a7@redhat.com> (raw)
In-Reply-To: <20161104015048.GS16751@yliu-dev.sh.intel.com>
On 11/04/2016 02:50 AM, Yuanhan Liu wrote:
> On Thu, Nov 03, 2016 at 10:11:43PM +0100, Maxime Coquelin wrote:
>>
>>
>> On 11/03/2016 05:09 PM, Yuanhan Liu wrote:
>>> Queue allocation should be done once, since the queue related info (such
>>> as vring addreess) will only be informed to the vhost-user backend once
>>> without virtio device reset.
>>>
>>> That means, if you allocate queues again after the vhost-user negotiation,
>>> the vhost-user backend will not be informed any more. Leading to a state
>>> that the vring info mismatches between virtio PMD driver and vhost-backend:
>>> the driver switches to the new address has just been allocated, while the
>>> vhost-backend still sticks to the old address has been assigned in the init
>>> stage.
>>>
>>> Unfortunately, that is exactly how the virtio driver is coded so far: queue
>>> allocation is done at queue_setup stage (when rte_eth_tx/rx_queue_setup is
>>> invoked). This is wrong, because queue_setup can be invoked several times.
>>> For example,
>>>
>>> $ start_testpmd.sh ... --txq=1 --rxq=1 ...
>>> > port stop 0
>>> > port config all txq 1 # just trigger the queue_setup callback again
>>> > port config all rxq 1
>>> > port start 0
>>>
>>> The right way to do is allocate the queues in the init stage, so that the
>>> vring info could be persistent with the vhost-user backend.
>>>
>>> Besides that, we should allocate max_queue pairs the device supports, but
>>> not nr queue pairs firstly configured, to make following case work.
>> I understand, but how much memory overhead does that represent?
>
> We are allocating max queue pairs the device supports, but not the
> virtio-net spec supports, which, as you stated, would be too much.
Oh, ok. In this case this is good to me.
Thanks,
Maxime
next prev parent reply other threads:[~2016-11-04 8:08 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-03 16:09 [dpdk-dev] [PATCH 0/8] net/virtio: fix queue reconfigure issue Yuanhan Liu
2016-11-03 16:09 ` [dpdk-dev] [PATCH 1/8] net/virtio: revert "virtio: fix restart" Yuanhan Liu
2016-11-03 20:36 ` Maxime Coquelin
2016-11-04 2:00 ` Yuanhan Liu
2016-11-04 8:09 ` Maxime Coquelin
2016-11-04 14:28 ` Yuanhan Liu
2016-11-04 8:10 ` Maxime Coquelin
2016-11-03 16:09 ` [dpdk-dev] [PATCH 2/8] net/virtio: simplify queue memzone name Yuanhan Liu
2016-11-03 20:41 ` Maxime Coquelin
2016-11-03 16:09 ` [dpdk-dev] [PATCH 3/8] net/virtio: simplify queue allocation Yuanhan Liu
2016-11-03 20:48 ` Maxime Coquelin
2016-11-04 1:51 ` Yuanhan Liu
2016-11-03 16:09 ` [dpdk-dev] [PATCH 4/8] net/virtio: allocate queue at init stage Yuanhan Liu
2016-11-03 21:11 ` Maxime Coquelin
2016-11-04 1:50 ` Yuanhan Liu
2016-11-04 8:08 ` Maxime Coquelin [this message]
2016-11-04 8:25 ` Maxime Coquelin
2016-11-04 15:21 ` Kevin Traynor
2016-11-04 20:30 ` Kevin Traynor
2016-11-05 6:15 ` Yuanhan Liu
2016-11-03 16:09 ` [dpdk-dev] [PATCH 5/8] net/virtio: initiate vring " Yuanhan Liu
2016-11-04 8:34 ` Maxime Coquelin
2016-11-03 16:09 ` [dpdk-dev] [PATCH 6/8] net/virtio: move queue configure code to proper place Yuanhan Liu
2016-11-04 8:39 ` Maxime Coquelin
2016-11-03 16:09 ` [dpdk-dev] [PATCH 7/8] net/virtio: complete init stage at the right place Yuanhan Liu
2016-11-04 8:44 ` Maxime Coquelin
2016-11-03 16:10 ` [dpdk-dev] [PATCH 8/8] net/virtio: remove started field Yuanhan Liu
2016-11-04 8:46 ` Maxime Coquelin
2016-11-05 9:40 ` [dpdk-dev] [PATCH v2 00/10] net/virtio: fix queue reconfigure issue Yuanhan Liu
2016-11-05 9:40 ` [dpdk-dev] [PATCH v2 01/10] net/virtio: revert fix restart Yuanhan Liu
2016-11-05 9:40 ` [dpdk-dev] [PATCH v2 02/10] net/virtio: simplify queue memzone name Yuanhan Liu
2016-11-05 9:40 ` [dpdk-dev] [PATCH v2 03/10] net/virtio: simplify queue allocation Yuanhan Liu
2016-11-05 9:40 ` [dpdk-dev] [PATCH v2 04/10] net/virtio: allocate queue at init stage Yuanhan Liu
2016-11-07 14:23 ` Thomas Monjalon
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 05/10] net/virtio: initiate vring " Yuanhan Liu
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 06/10] net/virtio: move queue configure code to proper place Yuanhan Liu
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 07/10] net/virtio: complete init stage at the right place Yuanhan Liu
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 08/10] net/virtio: remove started field Yuanhan Liu
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 09/10] net/virtio: fix less queues being enabled issue Yuanhan Liu
2016-11-05 9:41 ` [dpdk-dev] [PATCH v2 10/10] net/virtio: fix multiple queue enabling Yuanhan Liu
2016-11-07 9:25 ` Yuanhan Liu
2016-11-07 14:44 ` [dpdk-dev] [PATCH v2 00/10] net/virtio: fix queue reconfigure issue Thomas Monjalon
2016-11-07 15:05 ` Yao, Lei A
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=16f9caf1-423d-ceee-2b26-811f196201a7@redhat.com \
--to=maxime.coquelin@redhat.com \
--cc=dev@dpdk.org \
--cc=i.maximets@samsung.com \
--cc=yuanhan.liu@linux.intel.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).