[PATCH] virtio-balloon: correct used length

Jason Wang posted 1 patch 2 weeks ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20211124043255.6050-1-jasowang@redhat.com
Maintainers: David Hildenbrand <david@redhat.com>, "Michael S. Tsirkin" <mst@redhat.com>
hw/virtio/virtio-balloon.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)

[PATCH] virtio-balloon: correct used length

Posted by Jason Wang 2 weeks ago
Spec said:

"and len the total of bytes written into the buffer."

For inflateq, deflateq and statsq, we don't process in_sg so the used
length should be zero. For free_page_vq, though the free pages are
supplied via in_sgs, zero used length should still be fine since
anyway driver is expected to use the length in this case and it
simplifies the error handling path.

Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 hw/virtio/virtio-balloon.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
index c6962fcbfe..3e52daa793 100644
--- a/hw/virtio/virtio-balloon.c
+++ b/hw/virtio/virtio-balloon.c
@@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
         return;
     }
 
-    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
+    virtqueue_push(s->svq, s->stats_vq_elem, 0);
     virtio_notify(vdev, s->svq);
     g_free(s->stats_vq_elem);
     s->stats_vq_elem = NULL;
@@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
             memory_region_unref(section.mr);
         }
 
-        virtqueue_push(vq, elem, offset);
+        virtqueue_push(vq, elem, 0);
         virtio_notify(vdev, vq);
         g_free(elem);
         virtio_balloon_pbp_free(&pbp);
@@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
     }
 
 out:
-    virtqueue_push(vq, elem, 1);
+    virtqueue_push(vq, elem, 0);
     g_free(elem);
     return ret;
 }
-- 
2.25.1


Re: [PATCH] virtio-balloon: correct used length

Posted by Michael S. Tsirkin 2 weeks ago
On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> Spec said:
> 
> "and len the total of bytes written into the buffer."
> 
> For inflateq, deflateq and statsq, we don't process in_sg so the used
> length should be zero. For free_page_vq, though the free pages are
> supplied via in_sgs, zero used length should still be fine since
> anyway driver is expected to use the length in this case and it
> simplifies the error handling path.
> 
> Signed-off-by: Jason Wang <jasowang@redhat.com>

I think for free page vq the point is that the pages are
zeroed by hypervisor, so we must set used len accordingly. No?

> ---
>  hw/virtio/virtio-balloon.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> index c6962fcbfe..3e52daa793 100644
> --- a/hw/virtio/virtio-balloon.c
> +++ b/hw/virtio/virtio-balloon.c
> @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
>          return;
>      }
>  
> -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
>      virtio_notify(vdev, s->svq);
>      g_free(s->stats_vq_elem);
>      s->stats_vq_elem = NULL;
> @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
>              memory_region_unref(section.mr);
>          }
>  
> -        virtqueue_push(vq, elem, offset);
> +        virtqueue_push(vq, elem, 0);
>          virtio_notify(vdev, vq);
>          g_free(elem);
>          virtio_balloon_pbp_free(&pbp);
> @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
>      }
>  
>  out:
> -    virtqueue_push(vq, elem, 1);
> +    virtqueue_push(vq, elem, 0);
>      g_free(elem);
>      return ret;
>  }
> -- 
> 2.25.1


Re: [PATCH] virtio-balloon: correct used length

Posted by Jason Wang 2 weeks ago
On Wed, Nov 24, 2021 at 3:03 PM Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> > Spec said:
> >
> > "and len the total of bytes written into the buffer."
> >
> > For inflateq, deflateq and statsq, we don't process in_sg so the used
> > length should be zero. For free_page_vq, though the free pages are
> > supplied via in_sgs, zero used length should still be fine since
> > anyway driver is expected to use the length in this case and it
> > simplifies the error handling path.
> >
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
>
> I think for free page vq the point is that the pages are
> zeroed by hypervisor, so we must set used len accordingly. No?

I may miss something, I think it is used for clearing the dirty
bitmaps to avoid migrating unused pages, the zeroing is not needed.

Thanks

>
> > ---
> >  hw/virtio/virtio-balloon.c | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> > index c6962fcbfe..3e52daa793 100644
> > --- a/hw/virtio/virtio-balloon.c
> > +++ b/hw/virtio/virtio-balloon.c
> > @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
> >          return;
> >      }
> >
> > -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> > +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
> >      virtio_notify(vdev, s->svq);
> >      g_free(s->stats_vq_elem);
> >      s->stats_vq_elem = NULL;
> > @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
> >              memory_region_unref(section.mr);
> >          }
> >
> > -        virtqueue_push(vq, elem, offset);
> > +        virtqueue_push(vq, elem, 0);
> >          virtio_notify(vdev, vq);
> >          g_free(elem);
> >          virtio_balloon_pbp_free(&pbp);
> > @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
> >      }
> >
> >  out:
> > -    virtqueue_push(vq, elem, 1);
> > +    virtqueue_push(vq, elem, 0);
> >      g_free(elem);
> >      return ret;
> >  }
> > --
> > 2.25.1
>


Re: [PATCH] virtio-balloon: correct used length

Posted by Michael S. Tsirkin 2 weeks ago
On Wed, Nov 24, 2021 at 03:58:22PM +0800, Jason Wang wrote:
> On Wed, Nov 24, 2021 at 3:03 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> > > Spec said:
> > >
> > > "and len the total of bytes written into the buffer."
> > >
> > > For inflateq, deflateq and statsq, we don't process in_sg so the used
> > > length should be zero. For free_page_vq, though the free pages are
> > > supplied via in_sgs, zero used length should still be fine since
> > > anyway driver is expected to use the length in this case and it
> > > simplifies the error handling path.
> > >
> > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> >
> > I think for free page vq the point is that the pages are
> > zeroed by hypervisor, so we must set used len accordingly. No?
> 
> I may miss something, I think it is used for clearing the dirty
> bitmaps to avoid migrating unused pages, the zeroing is not needed.
> 
> Thanks

And, once the page is migrated, it's value on destination may differ
from one on source.

> >
> > > ---
> > >  hw/virtio/virtio-balloon.c | 6 +++---
> > >  1 file changed, 3 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> > > index c6962fcbfe..3e52daa793 100644
> > > --- a/hw/virtio/virtio-balloon.c
> > > +++ b/hw/virtio/virtio-balloon.c
> > > @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
> > >          return;
> > >      }
> > >
> > > -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> > > +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
> > >      virtio_notify(vdev, s->svq);
> > >      g_free(s->stats_vq_elem);
> > >      s->stats_vq_elem = NULL;
> > > @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
> > >              memory_region_unref(section.mr);
> > >          }
> > >
> > > -        virtqueue_push(vq, elem, offset);
> > > +        virtqueue_push(vq, elem, 0);
> > >          virtio_notify(vdev, vq);
> > >          g_free(elem);
> > >          virtio_balloon_pbp_free(&pbp);
> > > @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
> > >      }
> > >
> > >  out:
> > > -    virtqueue_push(vq, elem, 1);
> > > +    virtqueue_push(vq, elem, 0);
> > >      g_free(elem);
> > >      return ret;
> > >  }
> > > --
> > > 2.25.1
> >


Re: [PATCH] virtio-balloon: correct used length

Posted by Jason Wang 2 weeks ago
On Wed, Nov 24, 2021 at 4:25 PM Michael S. Tsirkin <mst@redhat.com> wrote:
>
> On Wed, Nov 24, 2021 at 03:58:22PM +0800, Jason Wang wrote:
> > On Wed, Nov 24, 2021 at 3:03 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> > >
> > > On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> > > > Spec said:
> > > >
> > > > "and len the total of bytes written into the buffer."
> > > >
> > > > For inflateq, deflateq and statsq, we don't process in_sg so the used
> > > > length should be zero. For free_page_vq, though the free pages are
> > > > supplied via in_sgs, zero used length should still be fine since
> > > > anyway driver is expected to use the length in this case and it
> > > > simplifies the error handling path.
> > > >
> > > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > >
> > > I think for free page vq the point is that the pages are
> > > zeroed by hypervisor, so we must set used len accordingly. No?
> >
> > I may miss something, I think it is used for clearing the dirty
> > bitmaps to avoid migrating unused pages, the zeroing is not needed.
> >
> > Thanks
>
> And, once the page is migrated, it's value on destination may differ
> from one on source.

Right, I will fix this in the next version.

Thanks


>
> > >
> > > > ---
> > > >  hw/virtio/virtio-balloon.c | 6 +++---
> > > >  1 file changed, 3 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> > > > index c6962fcbfe..3e52daa793 100644
> > > > --- a/hw/virtio/virtio-balloon.c
> > > > +++ b/hw/virtio/virtio-balloon.c
> > > > @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
> > > >          return;
> > > >      }
> > > >
> > > > -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> > > > +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
> > > >      virtio_notify(vdev, s->svq);
> > > >      g_free(s->stats_vq_elem);
> > > >      s->stats_vq_elem = NULL;
> > > > @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
> > > >              memory_region_unref(section.mr);
> > > >          }
> > > >
> > > > -        virtqueue_push(vq, elem, offset);
> > > > +        virtqueue_push(vq, elem, 0);
> > > >          virtio_notify(vdev, vq);
> > > >          g_free(elem);
> > > >          virtio_balloon_pbp_free(&pbp);
> > > > @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
> > > >      }
> > > >
> > > >  out:
> > > > -    virtqueue_push(vq, elem, 1);
> > > > +    virtqueue_push(vq, elem, 0);
> > > >      g_free(elem);
> > > >      return ret;
> > > >  }
> > > > --
> > > > 2.25.1
> > >
>


Re: [PATCH] virtio-balloon: correct used length

Posted by Jason Wang 2 weeks ago
On Wed, Nov 24, 2021 at 4:28 PM Jason Wang <jasowang@redhat.com> wrote:
>
> On Wed, Nov 24, 2021 at 4:25 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > On Wed, Nov 24, 2021 at 03:58:22PM +0800, Jason Wang wrote:
> > > On Wed, Nov 24, 2021 at 3:03 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> > > >
> > > > On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> > > > > Spec said:
> > > > >
> > > > > "and len the total of bytes written into the buffer."
> > > > >
> > > > > For inflateq, deflateq and statsq, we don't process in_sg so the used
> > > > > length should be zero. For free_page_vq, though the free pages are
> > > > > supplied via in_sgs, zero used length should still be fine since
> > > > > anyway driver is expected to use the length in this case and it
> > > > > simplifies the error handling path.
> > > > >
> > > > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > > >
> > > > I think for free page vq the point is that the pages are
> > > > zeroed by hypervisor, so we must set used len accordingly. No?
> > >
> > > I may miss something, I think it is used for clearing the dirty
> > > bitmaps to avoid migrating unused pages, the zeroing is not needed.
> > >
> > > Thanks
> >
> > And, once the page is migrated, it's value on destination may differ
> > from one on source.
>
> Right, I will fix this in the next version.
>
> Thanks

Btw in the get_free_page_hints() I see this:

    if (elem->in_num && dev->free_page_hint_status == FREE_PAGE_HINT_S_START) {
qemu_guest_free_page_hint(elem->in_sg[0].iov_base,
                                  elem->in_sg[0].iov_len);
    }

It looks like only the first in sg is used, it looks like a bug?

Thanks

>
>
> >
> > > >
> > > > > ---
> > > > >  hw/virtio/virtio-balloon.c | 6 +++---
> > > > >  1 file changed, 3 insertions(+), 3 deletions(-)
> > > > >
> > > > > diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> > > > > index c6962fcbfe..3e52daa793 100644
> > > > > --- a/hw/virtio/virtio-balloon.c
> > > > > +++ b/hw/virtio/virtio-balloon.c
> > > > > @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
> > > > >          return;
> > > > >      }
> > > > >
> > > > > -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> > > > > +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
> > > > >      virtio_notify(vdev, s->svq);
> > > > >      g_free(s->stats_vq_elem);
> > > > >      s->stats_vq_elem = NULL;
> > > > > @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
> > > > >              memory_region_unref(section.mr);
> > > > >          }
> > > > >
> > > > > -        virtqueue_push(vq, elem, offset);
> > > > > +        virtqueue_push(vq, elem, 0);
> > > > >          virtio_notify(vdev, vq);
> > > > >          g_free(elem);
> > > > >          virtio_balloon_pbp_free(&pbp);
> > > > > @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
> > > > >      }
> > > > >
> > > > >  out:
> > > > > -    virtqueue_push(vq, elem, 1);
> > > > > +    virtqueue_push(vq, elem, 0);
> > > > >      g_free(elem);
> > > > >      return ret;
> > > > >  }
> > > > > --
> > > > > 2.25.1
> > > >
> >


Re: [PATCH] virtio-balloon: correct used length

Posted by Michael S. Tsirkin 2 weeks ago
On Wed, Nov 24, 2021 at 04:32:06PM +0800, Jason Wang wrote:
> On Wed, Nov 24, 2021 at 4:28 PM Jason Wang <jasowang@redhat.com> wrote:
> >
> > On Wed, Nov 24, 2021 at 4:25 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> > >
> > > On Wed, Nov 24, 2021 at 03:58:22PM +0800, Jason Wang wrote:
> > > > On Wed, Nov 24, 2021 at 3:03 PM Michael S. Tsirkin <mst@redhat.com> wrote:
> > > > >
> > > > > On Wed, Nov 24, 2021 at 12:32:55PM +0800, Jason Wang wrote:
> > > > > > Spec said:
> > > > > >
> > > > > > "and len the total of bytes written into the buffer."
> > > > > >
> > > > > > For inflateq, deflateq and statsq, we don't process in_sg so the used
> > > > > > length should be zero. For free_page_vq, though the free pages are
> > > > > > supplied via in_sgs, zero used length should still be fine since
> > > > > > anyway driver is expected to use the length in this case and it
> > > > > > simplifies the error handling path.
> > > > > >
> > > > > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > > > >
> > > > > I think for free page vq the point is that the pages are
> > > > > zeroed by hypervisor, so we must set used len accordingly. No?
> > > >
> > > > I may miss something, I think it is used for clearing the dirty
> > > > bitmaps to avoid migrating unused pages, the zeroing is not needed.
> > > >
> > > > Thanks
> > >
> > > And, once the page is migrated, it's value on destination may differ
> > > from one on source.
> >
> > Right, I will fix this in the next version.
> >
> > Thanks
> 
> Btw in the get_free_page_hints() I see this:
> 
>     if (elem->in_num && dev->free_page_hint_status == FREE_PAGE_HINT_S_START) {
> qemu_guest_free_page_hint(elem->in_sg[0].iov_base,
>                                   elem->in_sg[0].iov_len);
>     }
> 
> It looks like only the first in sg is used, it looks like a bug?
> 
> Thanks

It sure does.


> >
> >
> > >
> > > > >
> > > > > > ---
> > > > > >  hw/virtio/virtio-balloon.c | 6 +++---
> > > > > >  1 file changed, 3 insertions(+), 3 deletions(-)
> > > > > >
> > > > > > diff --git a/hw/virtio/virtio-balloon.c b/hw/virtio/virtio-balloon.c
> > > > > > index c6962fcbfe..3e52daa793 100644
> > > > > > --- a/hw/virtio/virtio-balloon.c
> > > > > > +++ b/hw/virtio/virtio-balloon.c
> > > > > > @@ -231,7 +231,7 @@ static void balloon_stats_poll_cb(void *opaque)
> > > > > >          return;
> > > > > >      }
> > > > > >
> > > > > > -    virtqueue_push(s->svq, s->stats_vq_elem, s->stats_vq_offset);
> > > > > > +    virtqueue_push(s->svq, s->stats_vq_elem, 0);
> > > > > >      virtio_notify(vdev, s->svq);
> > > > > >      g_free(s->stats_vq_elem);
> > > > > >      s->stats_vq_elem = NULL;
> > > > > > @@ -438,7 +438,7 @@ static void virtio_balloon_handle_output(VirtIODevice *vdev, VirtQueue *vq)
> > > > > >              memory_region_unref(section.mr);
> > > > > >          }
> > > > > >
> > > > > > -        virtqueue_push(vq, elem, offset);
> > > > > > +        virtqueue_push(vq, elem, 0);
> > > > > >          virtio_notify(vdev, vq);
> > > > > >          g_free(elem);
> > > > > >          virtio_balloon_pbp_free(&pbp);
> > > > > > @@ -549,7 +549,7 @@ static bool get_free_page_hints(VirtIOBalloon *dev)
> > > > > >      }
> > > > > >
> > > > > >  out:
> > > > > > -    virtqueue_push(vq, elem, 1);
> > > > > > +    virtqueue_push(vq, elem, 0);
> > > > > >      g_free(elem);
> > > > > >      return ret;
> > > > > >  }
> > > > > > --
> > > > > > 2.25.1
> > > > >
> > >