linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jason Wang <jasowang@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: virtualization@lists.linux-foundation.org,
	linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	netdev@vger.kernel.org, xieyongji@bytedance.com,
	stefanha@redhat.com
Subject: Re: [PATCH 2/2] vdpa: vp_vdpa: don't use hard-coded maximum virtqueue size
Date: Tue, 6 Jul 2021 10:30:59 +0800	[thread overview]
Message-ID: <e0744583-fcdc-07d0-a414-31d660089e2b@redhat.com> (raw)
In-Reply-To: <20210705065534-mutt-send-email-mst@kernel.org>


在 2021/7/6 上午1:59, Michael S. Tsirkin 写道:
> On Mon, Jul 05, 2021 at 03:29:47PM +0800, Jason Wang wrote:
>> 在 2021/7/5 下午3:26, Michael S. Tsirkin 写道:
>>> On Mon, Jul 05, 2021 at 03:19:10PM +0800, Jason Wang wrote:
>>>> This patch switch to read virtqueue size from the capability instead
>>>> of depending on the hardcoded value. This allows the per virtqueue
>>>> size could be advertised.
>>>>
>>>> Signed-off-by: Jason Wang <jasowang@redhat.com>
>>> So let's add an ioctl for this? It's really a bug we don't..
>>
>> As explained in patch 1. Qemu doesn't use VHOST_VDPA_GET_VRING_NUM actually.
>> Instead it checks the result VHOST_VDPA_SET_VRING_NUM.
>>
>> So I change VHOST_VDPA_GET_VRING_NUM to return the minimal size of all the
>> virtqueues.
>>
>> If you wish we can add a VHOST_VDPA_GET_VRING_NUM2, but I'm not sure it will
>> have a user or not.
>>
>> Thanks
> Question is how do we know returning the minimal and not e.g. the max
> size is the right thing to do?


For the new ioctl, it will return the max queue size per vq.

It's probably too late to fix the old one, so it's only safe to return 
the minimal one.

Actually, most of the vDPA parents should be fine except for the 
vp_vdpa. When running in a nested environment, Qemu only advertise cvq 
with 64 entries.

Thanks


>
>
>>>> ---
>>>>    drivers/vdpa/virtio_pci/vp_vdpa.c | 6 ++++--
>>>>    1 file changed, 4 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/vdpa/virtio_pci/vp_vdpa.c b/drivers/vdpa/virtio_pci/vp_vdpa.c
>>>> index 2926641fb586..198f7076e4d9 100644
>>>> --- a/drivers/vdpa/virtio_pci/vp_vdpa.c
>>>> +++ b/drivers/vdpa/virtio_pci/vp_vdpa.c
>>>> @@ -18,7 +18,6 @@
>>>>    #include <linux/virtio_pci.h>
>>>>    #include <linux/virtio_pci_modern.h>
>>>> -#define VP_VDPA_QUEUE_MAX 256
>>>>    #define VP_VDPA_DRIVER_NAME "vp_vdpa"
>>>>    #define VP_VDPA_NAME_SIZE 256
>>>> @@ -197,7 +196,10 @@ static void vp_vdpa_set_status(struct vdpa_device *vdpa, u8 status)
>>>>    static u16 vp_vdpa_get_vq_num_max(struct vdpa_device *vdpa, u16 qid)
>>>>    {
>>>> -	return VP_VDPA_QUEUE_MAX;
>>>> +	struct vp_vdpa *vp_vdpa = vdpa_to_vp(vdpa);
>>>> +	struct virtio_pci_modern_device *mdev = &vp_vdpa->mdev;
>>>> +
>>>> +	return vp_modern_get_queue_size(mdev, qid);
>>>>    }
>>>>    static int vp_vdpa_get_vq_state(struct vdpa_device *vdpa, u16 qid,
>>>> -- 
>>>> 2.25.1


  reply	other threads:[~2021-07-06  2:36 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-05  7:19 [PATCH 1/2] vdpa: support per virtqueue max queue size Jason Wang
2021-07-05  7:19 ` [PATCH 2/2] vdpa: vp_vdpa: don't use hard-coded maximum virtqueue size Jason Wang
2021-07-05  7:26   ` Michael S. Tsirkin
2021-07-05  7:29     ` Jason Wang
2021-07-05 17:59       ` Michael S. Tsirkin
2021-07-06  2:30         ` Jason Wang [this message]
2021-07-06 13:14 ` [PATCH 1/2] vdpa: support per virtqueue max queue size Stefan Hajnoczi
2021-07-07  4:04 ` Yongji Xie
2021-07-07  5:38   ` Jason Wang

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=e0744583-fcdc-07d0-a414-31d660089e2b@redhat.com \
    --to=jasowang@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mst@redhat.com \
    --cc=netdev@vger.kernel.org \
    --cc=stefanha@redhat.com \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=xieyongji@bytedance.com \
    --subject='Re: [PATCH 2/2] vdpa: vp_vdpa: don'\''t use hard-coded maximum virtqueue size' \
    /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

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).