public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Dragos Tatulea <dtatulea@nvidia.com>
To: "eperezma@redhat.com" <eperezma@redhat.com>
Cc: "virtualization@lists.linux-foundation.org" 
	<virtualization@lists.linux-foundation.org>,
	Gal Pressman <gal@nvidia.com>, "mst@redhat.com" <mst@redhat.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"jasowang@redhat.com" <jasowang@redhat.com>,
	"xuanzhuo@linux.alibaba.com" <xuanzhuo@linux.alibaba.com>
Subject: Re: [PATCH vhost 16/16] vdpa/mlx5: Update cvq iotlb mapping on ASID change
Date: Thu, 5 Oct 2023 13:12:11 +0000	[thread overview]
Message-ID: <b6f90921bc239fc0e1fc04dc62e4376fcd622376.camel@nvidia.com> (raw)
In-Reply-To: <CAJaqyWe9iYnh_DwZbU4986bSAZ0C1=Y9Th6-vv_ZRTTYwQqhYA@mail.gmail.com>

On Thu, 2023-10-05 at 12:41 +0200, Eugenio Perez Martin wrote:
> On Thu, Sep 28, 2023 at 6:50 PM Dragos Tatulea <dtatulea@nvidia.com> wrote:
> > 
> > For the following sequence:
> > - cvq group is in ASID 0
> > - .set_map(1, cvq_iotlb)
> > - .set_group_asid(cvq_group, 1)
> > 
> > ... the cvq mapping from ASID 0 will be used. This is not always correct
> > behaviour.
> > 
> > This patch adds support for the above mentioned flow by saving the iotlb
> > on each .set_map and updating the cvq iotlb with it on a cvq group change.
> > 
> > Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
> > ---
> >  drivers/vdpa/mlx5/core/mlx5_vdpa.h |  2 ++
> >  drivers/vdpa/mlx5/core/mr.c        | 26 ++++++++++++++++++++++++++
> >  drivers/vdpa/mlx5/net/mlx5_vnet.c  |  9 ++++++++-
> >  3 files changed, 36 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/vdpa/mlx5/core/mlx5_vdpa.h
> > b/drivers/vdpa/mlx5/core/mlx5_vdpa.h
> > index ae09296f4270..db988ced5a5d 100644
> > --- a/drivers/vdpa/mlx5/core/mlx5_vdpa.h
> > +++ b/drivers/vdpa/mlx5/core/mlx5_vdpa.h
> > @@ -32,6 +32,8 @@ struct mlx5_vdpa_mr {
> >         unsigned long num_directs;
> >         unsigned long num_klms;
> > 
> > +       struct vhost_iotlb *iotlb;
> > +
> >         bool user_mr;
> >  };
> > 
> > diff --git a/drivers/vdpa/mlx5/core/mr.c b/drivers/vdpa/mlx5/core/mr.c
> > index a4135c16b5bf..403c08271489 100644
> > --- a/drivers/vdpa/mlx5/core/mr.c
> > +++ b/drivers/vdpa/mlx5/core/mr.c
> > @@ -499,6 +499,8 @@ static void _mlx5_vdpa_destroy_mr(struct mlx5_vdpa_dev
> > *mvdev, struct mlx5_vdpa_
> >                 destroy_user_mr(mvdev, mr);
> >         else
> >                 destroy_dma_mr(mvdev, mr);
> > +
> > +       vhost_iotlb_free(mr->iotlb);
> >  }
> > 
> >  void mlx5_vdpa_destroy_mr(struct mlx5_vdpa_dev *mvdev,
> > @@ -558,6 +560,30 @@ static int _mlx5_vdpa_create_mr(struct mlx5_vdpa_dev
> > *mvdev,
> >         else
> >                 err = create_dma_mr(mvdev, mr);
> > 
> > +       if (err)
> > +               return err;
> > +
> > +       mr->iotlb = vhost_iotlb_alloc(0, 0);
> > +       if (!mr->iotlb) {
> > +               err = -ENOMEM;
> > +               goto err_mr;
> > +       }
> > +
> > +       err = dup_iotlb(mr->iotlb, iotlb);
> > +       if (err)
> > +               goto err_iotlb;
> > +
> > +       return 0;
> > +
> > +err_iotlb:
> > +       vhost_iotlb_free(mr->iotlb);
> > +
> > +err_mr:
> > +       if (iotlb)
> > +               destroy_user_mr(mvdev, mr);
> > +       else
> > +               destroy_dma_mr(mvdev, mr);
> > +
> >         return err;
> >  }
> > 
> > diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > index 46441e41892c..fc5d6b989a5a 100644
> > --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> > @@ -3154,12 +3154,19 @@ static int mlx5_set_group_asid(struct vdpa_device
> > *vdev, u32 group,
> >                                unsigned int asid)
> >  {
> >         struct mlx5_vdpa_dev *mvdev = to_mvdev(vdev);
> > +       int err = 0;
> > 
> >         if (group >= MLX5_VDPA_NUMVQ_GROUPS || asid >= MLX5_VDPA_NUM_AS)
> >                 return -EINVAL;
> > 
> >         mvdev->group2asid[group] = asid;
> > -       return 0;
> > +
> > +       mutex_lock(&mvdev->mr_mtx);
> > +       if (group == MLX5_VDPA_CVQ_GROUP && mvdev->mr[asid])
> > +               err = mlx5_vdpa_update_cvq_iotlb(mvdev, mvdev->mr[asid]-
> > >iotlb, asid);
> 
> Do we need to protect here in case userspace sets the same ASID twice?
> mlx5_vdpa_update_cvq_iotlb shouldn't call dup_iotlb with the same src
> and dst.

Why would src and dst be identical? Doesn't dup_iotlb create a copy of src?

That being said, there should be a check in dup_iotlb for dst == src. Not sure
if this is an error or success though...

Thanks,
Dragos
- 
> Apart from that:
> 
> Acked-by: Eugenio Pérez <eperezma@redhat.com>
> 
> > +       mutex_unlock(&mvdev->mr_mtx);
> > +
> > +       return err;
> >  }
> > 
> >  static const struct vdpa_config_ops mlx5_vdpa_ops = {
> > --
> > 2.41.0
> > 
> 


  reply	other threads:[~2023-10-05 14:09 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-28 16:45 [PATCH vhost v2 00/16] vdpa: Add support for vq descriptor mappings Dragos Tatulea
2023-09-28 16:45 ` [PATCH mlx5-next 01/16] vdpa/mlx5: Expose descriptor group mkey hw capability Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 02/16] vdpa: introduce dedicated descriptor group for virtqueue Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 03/16] vhost-vdpa: introduce descriptor group backend feature Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 04/16] vhost-vdpa: uAPI to get dedicated descriptor group id Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 05/16] vdpa/mlx5: Create helper function for dma mappings Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 06/16] vdpa/mlx5: Decouple cvq iotlb handling from hw mapping code Dragos Tatulea
2023-10-04 18:39   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 07/16] vdpa/mlx5: Take cvq iotlb lock during refresh Dragos Tatulea
2023-10-04 18:40   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 08/16] vdpa/mlx5: Collapse "dvq" mr add/delete functions Dragos Tatulea
2023-10-04 18:41   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 09/16] vdpa/mlx5: Rename mr destroy functions Dragos Tatulea
2023-10-04 18:41   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 10/16] vdpa/mlx5: Allow creation/deletion of any given mr struct Dragos Tatulea
2023-10-03  8:11   ` Dragos Tatulea
2023-10-04 18:45   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 11/16] vdpa/mlx5: Move mr mutex out of " Dragos Tatulea
2023-10-04 18:46   ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 12/16] vdpa/mlx5: Improve mr update flow Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 13/16] vdpa/mlx5: Introduce mr for vq descriptor Dragos Tatulea
2023-10-04 18:53   ` Eugenio Perez Martin
2023-10-05 12:09     ` Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 14/16] vdpa/mlx5: Enable hw support for vq descriptor mapping Dragos Tatulea
2023-10-05  9:42   ` Eugenio Perez Martin
2023-10-05 12:16     ` Dragos Tatulea
2023-10-05 14:48       ` Eugenio Perez Martin
2023-09-28 16:45 ` [PATCH vhost 15/16] vdpa/mlx5: Make iotlb helper functions more generic Dragos Tatulea
2023-10-05 10:14   ` Eugenio Perez Martin
2023-10-05 12:16     ` Dragos Tatulea
2023-09-28 16:45 ` [PATCH vhost 16/16] vdpa/mlx5: Update cvq iotlb mapping on ASID change Dragos Tatulea
2023-10-05 10:41   ` Eugenio Perez Martin
2023-10-05 13:12     ` Dragos Tatulea [this message]
2023-10-05 14:18       ` Eugenio Perez Martin
2023-10-02 10:16 ` [PATCH vhost v2 00/16] vdpa: Add support for vq descriptor mappings Leon Romanovsky
2023-10-02 10:16 ` (subset) " Leon Romanovsky
2023-10-05 17:31 ` Michael S. Tsirkin
2023-10-05 17:44   ` Dragos Tatulea
2023-10-05 19:18     ` Michael S. Tsirkin

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=b6f90921bc239fc0e1fc04dc62e4376fcd622376.camel@nvidia.com \
    --to=dtatulea@nvidia.com \
    --cc=eperezma@redhat.com \
    --cc=gal@nvidia.com \
    --cc=jasowang@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mst@redhat.com \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=xuanzhuo@linux.alibaba.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