[Qemu-devel] [PATCH] virtio-scsi: virtio_scsi_push_event() lacks VirtIOSCSIReq parsing

[Qemu-devel] [PATCH] virtio-scsi: virtio_scsi_push_event() lacks VirtIOSCSIReq parsing

[Greg Kurz]

Hotplug of a virtio scsi disk is currently broken: no disk appears in the
guest (verified with a fedora 20 host running a fedora 20 guest with KVM).
Bisect leeds to Paolo's patches to support any_layout, especially this
commit:

commit 36b15c79aa1bef5fe7543f9f2629b6413720bbfb
Author: Paolo Bonzini <pbonzini@redhat.com>
Date:   Tue Jun 10 16:21:18 2014 +0200

    virtio-scsi: start preparing for any_layout

It modifies virtio_scsi_pop_req() so that it is up to the callers to parse
the virtio scsi request. It seems that virtio_scsi_push_event() was not
modified accordingly...

This patch adds a call to virtio_scsi_parse_req(). It also drops the in size
sanity check since it is already done by virtio_scsi_parse_req(). This is
enough to have hotplug working again.

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
---

I guess the following check could also be handled by virtio_scsi_parse_req()
as it does not allow both in and out at the same time, but I am not sure:

    if (req->elem.out_num) {
        virtio_scsi_bad_req();
    }

Thoughts ?

--
Greg

 hw/scsi/virtio-scsi.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
index 04ecfa7..0b0a331 100644
--- a/hw/scsi/virtio-scsi.c
+++ b/hw/scsi/virtio-scsi.c
@@ -565,7 +565,6 @@ static void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
     VirtIOSCSIReq *req;
     VirtIOSCSIEvent *evt;
     VirtIODevice *vdev = VIRTIO_DEVICE(s);
-    int in_size;
 
     if (!(vdev->status & VIRTIO_CONFIG_S_DRIVER_OK)) {
         return;
@@ -586,8 +585,7 @@ static void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
         s->events_dropped = false;
     }
 
-    in_size = iov_size(req->elem.in_sg, req->elem.in_num);
-    if (in_size < sizeof(VirtIOSCSIEvent)) {
+    if (virtio_scsi_parse_req(req, 0, sizeof(VirtIOSCSIEvent))) {
         virtio_scsi_bad_req();
     }
 

Re: [Qemu-devel] [PATCH] virtio-scsi: virtio_scsi_push_event() lacks VirtIOSCSIReq parsing

[Paolo Bonzini]

Il 30/06/2014 17:09, Greg Kurz ha scritto:
> Hotplug of a virtio scsi disk is currently broken: no disk appears in the
> guest (verified with a fedora 20 host running a fedora 20 guest with KVM).
> Bisect leeds to Paolo's patches to support any_layout, especially this
> commit:
>
> commit 36b15c79aa1bef5fe7543f9f2629b6413720bbfb
> Author: Paolo Bonzini <pbonzini@redhat.com>
> Date:   Tue Jun 10 16:21:18 2014 +0200
>
>     virtio-scsi: start preparing for any_layout
>
> It modifies virtio_scsi_pop_req() so that it is up to the callers to parse
> the virtio scsi request. It seems that virtio_scsi_push_event() was not
> modified accordingly...
>
> This patch adds a call to virtio_scsi_parse_req(). It also drops the in size
> sanity check since it is already done by virtio_scsi_parse_req(). This is
> enough to have hotplug working again.
>
> Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
> ---
>
> I guess the following check could also be handled by virtio_scsi_parse_req()
> as it does not allow both in and out at the same time, but I am not sure:
>
>     if (req->elem.out_num) {
>         virtio_scsi_bad_req();
>     }
>
> Thoughts ?
>
> --
> Greg
>
>  hw/scsi/virtio-scsi.c |    4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
> index 04ecfa7..0b0a331 100644
> --- a/hw/scsi/virtio-scsi.c
> +++ b/hw/scsi/virtio-scsi.c
> @@ -565,7 +565,6 @@ static void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
>      VirtIOSCSIReq *req;
>      VirtIOSCSIEvent *evt;
>      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> -    int in_size;
>
>      if (!(vdev->status & VIRTIO_CONFIG_S_DRIVER_OK)) {
>          return;
> @@ -586,8 +585,7 @@ static void virtio_scsi_push_event(VirtIOSCSI *s, SCSIDevice *dev,
>          s->events_dropped = false;
>      }
>
> -    in_size = iov_size(req->elem.in_sg, req->elem.in_num);
> -    if (in_size < sizeof(VirtIOSCSIEvent)) {
> +    if (virtio_scsi_parse_req(req, 0, sizeof(VirtIOSCSIEvent))) {
>          virtio_scsi_bad_req();
>      }
>
>
>
>

Looks good.  I'll queue the fix.

Paolo