DPDK patches and discussions
 help / color / mirror / Atom feed
From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: "Jiang, Cheng1" <cheng1.jiang@intel.com>,
	"Xia, Chenbo" <chenbo.xia@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, "Hu, Jiayu" <jiayu.hu@intel.com>,
	"Yang, YvonneX" <yvonnex.yang@intel.com>,
	"Wang, Yinan" <yinan.wang@intel.com>
Subject: Re: [dpdk-dev] [PATCH v3] examples/vhost: remove async inflight packet counter
Date: Wed, 3 Feb 2021 15:21:20 +0100	[thread overview]
Message-ID: <956ba30b-7b99-1bc8-6c2d-20a6e65d576e@redhat.com> (raw)
In-Reply-To: <SJ0PR11MB50064B4C2CBB97473D9EBA89DCB49@SJ0PR11MB5006.namprd11.prod.outlook.com>



On 2/3/21 2:42 PM, Jiang, Cheng1 wrote:
> 
> 
>> -----Original Message-----
>> From: Maxime Coquelin <maxime.coquelin@redhat.com>
>> Sent: Wednesday, February 3, 2021 8:55 PM
>> To: Jiang, Cheng1 <cheng1.jiang@intel.com>; Xia, Chenbo
>> <chenbo.xia@intel.com>
>> Cc: dev@dpdk.org; Hu, Jiayu <jiayu.hu@intel.com>; Yang, YvonneX
>> <yvonnex.yang@intel.com>; Wang, Yinan <yinan.wang@intel.com>
>> Subject: Re: [PATCH v3] examples/vhost: remove async inflight packet
>> counter
>>
>>
>>
>> On 2/3/21 1:11 PM, Jiang, Cheng1 wrote:
>>> Hi,
>>>
>>>> -----Original Message-----
>>>> From: Maxime Coquelin <maxime.coquelin@redhat.com>
>>>> Sent: Wednesday, February 3, 2021 5:52 PM
>>>> To: Jiang, Cheng1 <cheng1.jiang@intel.com>; Xia, Chenbo
>>>> <chenbo.xia@intel.com>
>>>> Cc: dev@dpdk.org; Hu, Jiayu <jiayu.hu@intel.com>; Yang, YvonneX
>>>> <yvonnex.yang@intel.com>; Wang, Yinan <yinan.wang@intel.com>
>>>> Subject: Re: [PATCH v3] examples/vhost: remove async inflight packet
>>>> counter
>>>>
>>>>
>>>>
>>>> On 2/2/21 7:17 AM, Cheng Jiang wrote:
>>>>> Remove async inflight packet counter since there is no need to keep
>>>>> tracking it. Increase MAX_ENQUEUED_SIZE to prevent packet segment
>>>>> number tracking ring from being exhausted.
>>>>
>>>> Is that an optimization or a fix?
>>>>
>>>> If the former, let's move it to v21.05.
>>>>
>>>
>>> I think it's a fix since there is no need to keep the inflight packet counter,
>> sorry I forget adding the fixes, I can submit a v4 to fix it.
>> OK
>>>>>
>>>>> Fixes: a68ba8e0a6b6 ("examples/vhost: refactor vhost data path")
> 
> Oh, I got fixes here, sorry for the miss.
> 
>>>>>
>>>>> Signed-off-by: Cheng Jiang <Cheng1.jiang@intel.com>
>>>>> ---
>>>>> v3: fixed fixes commit id in git log
>>>>>
>>>>> v2: fixed a typo
>>>>>
>>>>>  examples/vhost/ioat.h |  2 +-
>>>>>  examples/vhost/main.c | 10 +---------  examples/vhost/main.h |  1 -
>>>>>  3 files changed, 2 insertions(+), 11 deletions(-)
>>>>>
>>>>> diff --git a/examples/vhost/ioat.h b/examples/vhost/ioat.h index
>>>>> 0a1dbb811..1aa28ed6a 100644
>>>>> --- a/examples/vhost/ioat.h
>>>>> +++ b/examples/vhost/ioat.h
>>>>> @@ -11,7 +11,7 @@
>>>>>
>>>>>  #define MAX_VHOST_DEVICE 1024
>>>>>  #define IOAT_RING_SIZE 4096
>>>>> -#define MAX_ENQUEUED_SIZE 512
>>>>> +#define It might be further simplified then 4096
>>>>
>>>> Basically, this the size of the ring size, correct?
>>>> It might be further simplified then.
>>>>
>>>
>>> Yes, it's a the size of packet tracking ring, and it should be no less then
>> IOAT_RING_SIZE for some corner cases.
>>> I'm not sure I understand what you mean by saying It might be further
>> simplified then.
>>
>> I meant maybe in this case just use IOAT_RING_SIZE?
> 
> I have thought about it before I use MAX_ENQUEUED_SIZE.
> But actually MAX_ENQUEUED_SIZE is used to for packet tracking ring, and IOAT_RING_SIZE is for ioat ring.
> Using IOAT_RING_SIZE for packet tracking ring, I think maybe it's not so logical and a little bit confusing. What do you think?

OK, let's keep it as is then.

Thanks,
Maxime

> Thanks,
> Cheng
> 
>>
>>> Thanks,
>>> Cheng
>>>
>>>>>
>>>>>  struct dma_info {
>>>>>  	struct rte_pci_addr addr;
>>>>> diff --git a/examples/vhost/main.c b/examples/vhost/main.c index
>>>>> e74fc8750..ca73e7086 100644
>>>>> --- a/examples/vhost/main.c
>>>>> +++ b/examples/vhost/main.c
>>>>> @@ -831,11 +831,8 @@ complete_async_pkts(struct vhost_dev *vdev)
>>>>>
>>>>>  	complete_count = rte_vhost_poll_enqueue_completed(vdev->vid,
>>>>>  					VIRTIO_RXQ, p_cpl,
>>>> MAX_PKT_BURST);
>>>>> -	if (complete_count) {
>>>>> -		__atomic_sub_fetch(&vdev->nr_async_pkts,
>>>> complete_count,
>>>>> -			__ATOMIC_SEQ_CST);
>>>>> +	if (complete_count)
>>>>>  		free_pkts(p_cpl, complete_count);
>>>>> -	}
>>>>>  }
>>>>>
>>>>>  static __rte_always_inline void
>>>>> @@ -878,8 +875,6 @@ drain_vhost(struct vhost_dev *vdev)
>>>>>  		complete_async_pkts(vdev);
>>>>>  		ret = rte_vhost_submit_enqueue_burst(vdev->vid,
>>>> VIRTIO_RXQ,
>>>>>  					m, nr_xmit, m_cpu_cpl,
>>>> &cpu_cpl_nr);
>>>>> -		__atomic_add_fetch(&vdev->nr_async_pkts, ret -
>>>> cpu_cpl_nr,
>>>>> -				__ATOMIC_SEQ_CST);
>>>>>
>>>>>  		if (cpu_cpl_nr)
>>>>>  			free_pkts(m_cpu_cpl, cpu_cpl_nr); @@ -1210,9
>> +1205,6 @@
>>>>> drain_eth_rx(struct vhost_dev *vdev)
>>>>>  		enqueue_count = rte_vhost_submit_enqueue_burst(vdev-
>>>>> vid,
>>>>>  					VIRTIO_RXQ, pkts, rx_count,
>>>>>  					m_cpu_cpl, &cpu_cpl_nr);
>>>>> -		__atomic_add_fetch(&vdev->nr_async_pkts,
>>>>> -					enqueue_count - cpu_cpl_nr,
>>>>> -					__ATOMIC_SEQ_CST);
>>>>>  		if (cpu_cpl_nr)
>>>>>  			free_pkts(m_cpu_cpl, cpu_cpl_nr);
>>>>>
>>>>> diff --git a/examples/vhost/main.h b/examples/vhost/main.h index
>>>>> 2d6c05fd7..0ccdce4b4 100644
>>>>> --- a/examples/vhost/main.h
>>>>> +++ b/examples/vhost/main.h
>>>>> @@ -51,7 +51,6 @@ struct vhost_dev {
>>>>>  	uint64_t features;
>>>>>  	size_t hdr_len;
>>>>>  	uint16_t nr_vrings;
>>>>> -	uint64_t nr_async_pkts;
>>>>>  	struct rte_vhost_memory *mem;
>>>>>  	struct device_statistics stats;
>>>>>  	TAILQ_ENTRY(vhost_dev) global_vdev_entry;
>>>>> --
>>>>> 2.29.2
>>>>>
>>>
> 


  reply	other threads:[~2021-02-03 14:21 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-31 12:56 [dpdk-dev] [PATCH] " Cheng Jiang
2021-01-31 13:06 ` [dpdk-dev] [PATCH v2] " Cheng Jiang
2021-02-02  6:17 ` [dpdk-dev] [PATCH v3] " Cheng Jiang
2021-02-03  9:51   ` Maxime Coquelin
2021-02-03 12:11     ` Jiang, Cheng1
2021-02-03 12:55       ` Maxime Coquelin
2021-02-03 13:42         ` Jiang, Cheng1
2021-02-03 14:21           ` Maxime Coquelin [this message]
2021-02-03 14:33             ` Jiang, Cheng1
2021-02-03 16:27   ` Maxime Coquelin
2021-02-03 17:21     ` Maxime Coquelin

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=956ba30b-7b99-1bc8-6c2d-20a6e65d576e@redhat.com \
    --to=maxime.coquelin@redhat.com \
    --cc=chenbo.xia@intel.com \
    --cc=cheng1.jiang@intel.com \
    --cc=dev@dpdk.org \
    --cc=jiayu.hu@intel.com \
    --cc=yinan.wang@intel.com \
    --cc=yvonnex.yang@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).