[PATCH v3] doc: fix vhost multi-queue reconnection

Xia, Chenbo chenbo.xia at intel.com
Fri May 6 10:40:55 CEST 2022


> -----Original Message-----
> From: Ding, Xuan <xuan.ding at intel.com>
> Sent: Friday, April 15, 2022 10:08 AM
> To: maxime.coquelin at redhat.com; Xia, Chenbo <chenbo.xia at intel.com>
> Cc: dev at dpdk.org; Hu, Jiayu <jiayu.hu at intel.com>; Ding, Xuan
> <xuan.ding at intel.com>; stable at dpdk.org
> Subject: [PATCH v3] doc: fix vhost multi-queue reconnection
> 
> From: Xuan Ding <xuan.ding at intel.com>
> 
> Since QEMU 5.2.0 fixes the vhost split ring multi-queue reconnection
> issue in commit f66337bdbfda ("vhost-user: save features of multiqueues
> if chardev is closed"), this patch updates known issue to indicate
> the range of affeacted QEMU versions.
> 
> Fixes: b37e95507e1b ("doc: add vhost multi-queue reconnection issue")
> Cc: stable at dpdk.org
> 
> Signed-off-by: Xuan Ding <xuan.ding at intel.com>
> ---
> v3:
> * Updated affected QEMU versions.
> 
> v2:
> * Add fixes in commit title and log.
> ---
>  doc/guides/rel_notes/known_issues.rst | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/doc/guides/rel_notes/known_issues.rst
> b/doc/guides/rel_notes/known_issues.rst
> index 187d9c942e..bd89d58d24 100644
> --- a/doc/guides/rel_notes/known_issues.rst
> +++ b/doc/guides/rel_notes/known_issues.rst
> @@ -885,14 +885,15 @@ Unsuitable IOVA mode may be picked as the default
>  **Driver/Module**:
>     ALL.
> 
> -Vhost multi-queue reconnection failed with QEMU version >= 4.2.0
> -----------------------------------------------------------------
> +Vhost multi-queue reconnection failed with QEMU version 4.2.0 to 5.1.0-
> rc3

So 5.1.0 fixes the problem or not? As 5.1.0 is later than 5.1.0-rc3 but you said
from 5.2.0 it's fixed. And I think we should just mention formal release, not any
rc, qemu users should not use any rc for formal usage.

Thanks,
Chenbo

> +-------------------------------------------------------------------------
> -
> 
>  **Description**
>     It's a QEMU regression bug (bad commit: c6beefd674ff). QEMU only saves
>     acked features for one vhost-net when vhost quits. When vhost
> reconnects
>     to virtio-net/virtio-pmd in multi-queue situations, the features been
> -   set multiple times are not consistent.
> +   set multiple times are not consistent. QEMU-5.2.0 fixes this issue in
> commit
> +   f66337bdbfda ("vhost-user: save features of multiqueues if chardev is
> closed").
> 
>  **Implication**
>     Vhost cannot reconnect back to virtio-net/virtio-pmd normally.
> --
> 2.17.1



More information about the stable mailing list