Message-ID: <CAGxU2F4Eeb68J6PuapCneZtEqnyU2MVhCd4VZyVh-o-kxvQBkA@mail.gmail.com>
Date:   Tue, 15 Dec 2020 12:43:22 +0100
From:   Stefano Garzarella <sgarzare@...hat.com>
To:     Jason Wang <jasowang@...hat.com>,
        "Michael S. Tsirkin" <mst@...hat.com>
Cc:     Linux Virtualization <virtualization@...ts.linux-foundation.org>,
        Stefan Hajnoczi <stefanha@...hat.com>,
        Oren Duer <oren@...dia.com>,
        Laurent Vivier <lvivier@...hat.com>,
        kernel list <linux-kernel@...r.kernel.org>,
        Max Gurtovoy <mgurtovoy@...dia.com>,
        Shahaf Shuler <shahafs@...dia.com>, Eli Cohen <elic@...dia.com>
Subject: Re: [PATCH v3 13/19] vdpa_sim: add get_config callback in
 vdpasim_dev_attr


On Wed, Dec 9, 2020 at 12:07 PM Stefano Garzarella <sgarzare@...hat.com> wrote:
> On Mon, Dec 07, 2020 at 01:29:17PM +0800, Jason Wang wrote:
> >
> >On 2020/12/4 上午1:05, Stefano Garzarella wrote:
> >>The get_config callback can be used by the device to fill the
> >>config structure.
> >>The callback will be invoked in vdpasim_get_config() before copying
> >>bytes into caller buffer.
> >>
> >>Move vDPA-net config updates from vdpasim_set_features() in the
> >>new vdpasim_net_get_config() callback.
> >>
> >>Signed-off-by: Stefano Garzarella <sgarzare@...hat.com>
> >>---
> >>v3:
> >>- checked if get_config callback is set before call it
> >>---
> >>  drivers/vdpa/vdpa_sim/vdpa_sim.c | 35 +++++++++++++++++++-------------
> >>  1 file changed, 21 insertions(+), 14 deletions(-)
> >>
> >>diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >>index fe71ed7890e1..f935ade0806b 100644
> >>--- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >>+++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >>@@ -60,6 +60,8 @@ struct vdpasim_virtqueue {
> >>  #define VDPASIM_NET_FEATURES        (VDPASIM_FEATURES | \
> >>                               (1ULL << VIRTIO_NET_F_MAC))
> >>+struct vdpasim;
> >>+
> >>  struct vdpasim_dev_attr {
> >>      u64 supported_features;
> >>      size_t config_size;
> >>@@ -67,6 +69,7 @@ struct vdpasim_dev_attr {
> >>      u32 id;
> >>      work_func_t work_fn;
> >>+     void (*get_config)(struct vdpasim *vdpasim, void *config);
> >>  };
> >>  /* State of each vdpasim device */
> >>@@ -522,8 +525,6 @@ static u64 vdpasim_get_features(struct vdpa_device *vdpa)
> >>  static int vdpasim_set_features(struct vdpa_device *vdpa, u64 features)
> >>  {
> >>      struct vdpasim *vdpasim = vdpa_to_sim(vdpa);
> >>-     struct virtio_net_config *config =
> >>-             (struct virtio_net_config *)vdpasim->config;
> >>      /* DMA mapping must be done by driver */
> >>      if (!(features & (1ULL << VIRTIO_F_ACCESS_PLATFORM)))
> >>@@ -531,16 +532,6 @@ static int vdpasim_set_features(struct vdpa_device *vdpa, u64 features)
> >>      vdpasim->features = features & vdpasim->dev_attr.supported_features;
> >>-     /* We generally only know whether guest is using the legacy interface
> >>-      * here, so generally that's the earliest we can set config fields.
> >>-      * Note: We actually require VIRTIO_F_ACCESS_PLATFORM above which
> >>-      * implies VIRTIO_F_VERSION_1, but let's not try to be clever here.
> >>-      */
> >>-
> >>-     config->mtu = cpu_to_vdpasim16(vdpasim, 1500);
> >>-     config->status = cpu_to_vdpasim16(vdpasim, VIRTIO_NET_S_LINK_UP);
> >>-     memcpy(config->mac, macaddr_buf, ETH_ALEN);
> >
> >
> >Patch looks good to me.
> >
> >But we need Michael to confirm whether doing moving like this is safe.
> >I guess what has been done were trying to make sure get_config() fail
> >before set_features(), but it's not clear to me whether it's useful.
>
> IIUC, also looking the QEMU code, the set_features() should be called
> every time before get_config(), but to be sure, in get_config(), I can
> check for example if 'vdpasim->features' is not zero (we require
> VIRTIO_F_ACCESS_PLATFORM set).

Working on this I just realized that we already check in 
vdpa_get_config() that set_features() is called, so I think the moving 
is safe.

I'll put these considerations in the commit message.

Thanks,
Stefano

>
> @Michael any suggestion?
>
> Thanks,
> Stefano
>
> >
> >Thanks
> >
> >
> >>-
> >>      return 0;
> >>  }
> >>@@ -595,8 +586,13 @@ static void vdpasim_get_config(struct vdpa_device *vdpa, unsigned int offset,
> >>  {
> >>      struct vdpasim *vdpasim = vdpa_to_sim(vdpa);
> >>-     if (offset + len < vdpasim->dev_attr.config_size)
> >>-             memcpy(buf, vdpasim->config + offset, len);
> >>+     if (offset + len > vdpasim->dev_attr.config_size)
> >>+             return;
> >>+
> >>+     if (vdpasim->dev_attr.get_config)
> >>+             vdpasim->dev_attr.get_config(vdpasim, vdpasim->config);
> >>+
> >>+     memcpy(buf, vdpasim->config + offset, len);
> >>  }
> >>  static void vdpasim_set_config(struct vdpa_device *vdpa, unsigned int offset,
> >>@@ -739,6 +735,16 @@ static const struct vdpa_config_ops vdpasim_batch_config_ops = {
> >>      .free                   = vdpasim_free,
> >>  };
> >>+static void vdpasim_net_get_config(struct vdpasim *vdpasim, void *config)
> >>+{
> >>+     struct virtio_net_config *net_config =
> >>+             (struct virtio_net_config *)config;
> >>+
> >>+     net_config->mtu = cpu_to_vdpasim16(vdpasim, 1500);
> >>+     net_config->status = cpu_to_vdpasim16(vdpasim, VIRTIO_NET_S_LINK_UP);
> >>+     memcpy(net_config->mac, macaddr_buf, ETH_ALEN);
> >>+}
> >>+
> >>  static int __init vdpasim_dev_init(void)
> >>  {
> >>      struct vdpasim_dev_attr dev_attr = {};
> >>@@ -747,6 +753,7 @@ static int __init vdpasim_dev_init(void)
> >>      dev_attr.supported_features = VDPASIM_NET_FEATURES;
> >>      dev_attr.nvqs = VDPASIM_VQ_NUM;
> >>      dev_attr.config_size = sizeof(struct virtio_net_config);
> >>+     dev_attr.get_config = vdpasim_net_get_config;
> >>      dev_attr.work_fn = vdpasim_net_work;
> >>      vdpasim_dev = vdpasim_create(&dev_attr);
> >

Powered by blists - more mailing lists