Received-SPF: pass (google.com: domain of linux-kernel+bounces-109685-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Precedence: bulk
MIME-Version: 1.0
References: <20240320101912.28210-1-w_angrong@163.com>
In-Reply-To: <20240320101912.28210-1-w_angrong@163.com>
From: Jason Wang <jasowang@redhat.com>
Date: Thu, 21 Mar 2024 14:11:13 +0800
Message-ID: <CACGkMEst2ixZrtBUEWArQT+CkDqzSr9E3V7qMyVU6xX+FnBChA@mail.gmail.com>
Subject: Re: [PATCH v3] vhost/vdpa: Add MSI translation tables to iommu for
 software-managed MSI
To: Wang Rong <w_angrong@163.com>
Cc: mst@redhat.com, kvm@vger.kernel.org, virtualization@lists.linux.dev, 
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Mar 20, 2024 at 6:20=E2=80=AFPM Wang Rong <w_angrong@163.com> wrote=
:
>
> From: Rong Wang <w_angrong@163.com>
>
> Once enable iommu domain for one device, the MSI
> translation tables have to be there for software-managed MSI.
> Otherwise, platform with software-managed MSI without an
> irq bypass function, can not get a correct memory write event
> from pcie, will not get irqs.
> The solution is to obtain the MSI phy base address from
> iommu reserved region, and set it to iommu MSI cookie,
> then translation tables will be created while request irq.
>
> Change log
> ----------
>
> v1->v2:
> - add resv iotlb to avoid overlap mapping.
> v2->v3:
> - there is no need to export the iommu symbol anymore.
>
> Signed-off-by: Rong Wang <w_angrong@163.com>
> ---
>  drivers/vhost/vdpa.c | 59 +++++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 56 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c
> index ba52d128aeb7..28b56b10372b 100644
> --- a/drivers/vhost/vdpa.c
> +++ b/drivers/vhost/vdpa.c
> @@ -49,6 +49,7 @@ struct vhost_vdpa {
>         struct completion completion;
>         struct vdpa_device *vdpa;
>         struct hlist_head as[VHOST_VDPA_IOTLB_BUCKETS];
> +       struct vhost_iotlb resv_iotlb;

Is it better to introduce a reserved flag like VHOST_MAP_RESERVED,
which means it can't be modified by the userspace but the kernel.

So we don't need to have two IOTLB. But I guess the reason you have
this is because we may have multiple address spaces where the MSI
routing should work for all of them?

Another note, vhost-vDPA support virtual address mapping, so this
should only work for physicall address mapping. E.g in the case of
SVA, MSI iova is a valid IOVA for the driver/usrespace.

>         struct device dev;
>         struct cdev cdev;
>         atomic_t opened;
> @@ -247,6 +248,7 @@ static int _compat_vdpa_reset(struct vhost_vdpa *v)
>  static int vhost_vdpa_reset(struct vhost_vdpa *v)
>  {
>         v->in_batch =3D 0;
> +       vhost_iotlb_reset(&v->resv_iotlb);

We try hard to avoid this for performance, see this commit:

commit 4398776f7a6d532c466f9e41f601c9a291fac5ef
Author: Si-Wei Liu <si-wei.liu@oracle.com>
Date:   Sat Oct 21 02:25:15 2023 -0700

    vhost-vdpa: introduce IOTLB_PERSIST backend feature bit

Any reason you need to do this?

>         return _compat_vdpa_reset(v);
>  }
>
> @@ -1219,10 +1221,15 @@ static int vhost_vdpa_process_iotlb_update(struct=
 vhost_vdpa *v,
>             msg->iova + msg->size - 1 > v->range.last)
>                 return -EINVAL;
>
> +       if (vhost_iotlb_itree_first(&v->resv_iotlb, msg->iova,
> +                                       msg->iova + msg->size - 1))
> +               return -EINVAL;
> +
>         if (vhost_iotlb_itree_first(iotlb, msg->iova,
>                                     msg->iova + msg->size - 1))
>                 return -EEXIST;
>
> +
>         if (vdpa->use_va)
>                 return vhost_vdpa_va_map(v, iotlb, msg->iova, msg->size,
>                                          msg->uaddr, msg->perm);
> @@ -1307,6 +1314,45 @@ static ssize_t vhost_vdpa_chr_write_iter(struct ki=
ocb *iocb,
>         return vhost_chr_write_iter(dev, from);
>  }
>
> +static int vhost_vdpa_resv_iommu_region(struct iommu_domain *domain, str=
uct device *dma_dev,
> +       struct vhost_iotlb *resv_iotlb)
> +{
> +       struct list_head dev_resv_regions;
> +       phys_addr_t resv_msi_base =3D 0;
> +       struct iommu_resv_region *region;
> +       int ret =3D 0;
> +       bool with_sw_msi =3D false;
> +       bool with_hw_msi =3D false;
> +
> +       INIT_LIST_HEAD(&dev_resv_regions);
> +       iommu_get_resv_regions(dma_dev, &dev_resv_regions);
> +
> +       list_for_each_entry(region, &dev_resv_regions, list) {
> +               ret =3D vhost_iotlb_add_range_ctx(resv_iotlb, region->sta=
rt,
> +                               region->start + region->length - 1,
> +                               0, 0, NULL);

I think MSI should be write-only?

> +               if (ret) {
> +                       vhost_iotlb_reset(resv_iotlb);

Need to report an error here.

Thanks