public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails
@ 2012-03-19  9:56 Jan Kiszka
  2012-03-19 15:52 ` Alex Williamson
  2012-03-21 13:13 ` Avi Kivity
  0 siblings, 2 replies; 3+ messages in thread
From: Jan Kiszka @ 2012-03-19  9:56 UTC (permalink / raw)
  To: Avi Kivity, Marcelo Tosatti
  Cc: kvm, Alex Williamson, Wen Congyang, Markus Armbruster

If the host or the device does not support INTx sharing, retry the IRQ
assignment with host-side MSI support enabled but warn about potential
consequences. This allows to preserve the previous behavior where we
defaulted to MSI and did not support INTx sharing at all.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
---

Detecting if the user actually specified prefer_msi=off as property of
pci-assign is non-trivial. So I decided to go for the retry approach,
ignoring potential user requests. The warning should attract the
attention.

 hw/device-assignment.c |   12 ++++++++++++
 1 files changed, 12 insertions(+), 0 deletions(-)

diff --git a/hw/device-assignment.c b/hw/device-assignment.c
index 89823f1..c953713 100644
--- a/hw/device-assignment.c
+++ b/hw/device-assignment.c
@@ -835,6 +835,7 @@ static int assign_irq(AssignedDevice *dev)
         dev->irq_requested_type = 0;
     }
 
+retry:
     assigned_irq_data.flags = KVM_DEV_IRQ_GUEST_INTX;
     if (dev->features & ASSIGNED_DEVICE_PREFER_MSI_MASK &&
         dev->cap.available & ASSIGNED_DEVICE_CAP_MSI)
@@ -844,6 +845,17 @@ static int assign_irq(AssignedDevice *dev)
 
     r = kvm_assign_irq(kvm_state, &assigned_irq_data);
     if (r < 0) {
+        if (r == -EIO && !(dev->features & ASSIGNED_DEVICE_PREFER_MSI_MASK) &&
+            dev->cap.available & ASSIGNED_DEVICE_CAP_MSI) {
+            /* Retry with host-side MSI. There might be an IRQ conflict and
+             * either the kernel or the device doesn't support sharing. */
+            fprintf(stderr,
+                    "Host-side INTx sharing not supported, "
+                    "using MSI instead.\n"
+                    "Some devices do not to work properly in this mode.\n");
+            dev->features |= ASSIGNED_DEVICE_PREFER_MSI_MASK;
+            goto retry;
+        }
         fprintf(stderr, "Failed to assign irq for \"%s\": %s\n",
                 dev->dev.qdev.id, strerror(-r));
         fprintf(stderr, "Perhaps you are assigning a device "
-- 
1.7.3.4

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails
  2012-03-19  9:56 [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails Jan Kiszka
@ 2012-03-19 15:52 ` Alex Williamson
  2012-03-21 13:13 ` Avi Kivity
  1 sibling, 0 replies; 3+ messages in thread
From: Alex Williamson @ 2012-03-19 15:52 UTC (permalink / raw)
  To: Jan Kiszka
  Cc: Avi Kivity, Marcelo Tosatti, kvm, Wen Congyang, Markus Armbruster

On Mon, 2012-03-19 at 10:56 +0100, Jan Kiszka wrote:
> If the host or the device does not support INTx sharing, retry the IRQ
> assignment with host-side MSI support enabled but warn about potential
> consequences. This allows to preserve the previous behavior where we
> defaulted to MSI and did not support INTx sharing at all.
> 
> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
> ---
> 
> Detecting if the user actually specified prefer_msi=off as property of
> pci-assign is non-trivial. So I decided to go for the retry approach,
> ignoring potential user requests. The warning should attract the
> attention.
> 
>  hw/device-assignment.c |   12 ++++++++++++
>  1 files changed, 12 insertions(+), 0 deletions(-)
> 
> diff --git a/hw/device-assignment.c b/hw/device-assignment.c
> index 89823f1..c953713 100644
> --- a/hw/device-assignment.c
> +++ b/hw/device-assignment.c
> @@ -835,6 +835,7 @@ static int assign_irq(AssignedDevice *dev)
>          dev->irq_requested_type = 0;
>      }
>  
> +retry:
>      assigned_irq_data.flags = KVM_DEV_IRQ_GUEST_INTX;
>      if (dev->features & ASSIGNED_DEVICE_PREFER_MSI_MASK &&
>          dev->cap.available & ASSIGNED_DEVICE_CAP_MSI)
> @@ -844,6 +845,17 @@ static int assign_irq(AssignedDevice *dev)
>  
>      r = kvm_assign_irq(kvm_state, &assigned_irq_data);
>      if (r < 0) {
> +        if (r == -EIO && !(dev->features & ASSIGNED_DEVICE_PREFER_MSI_MASK) &&
> +            dev->cap.available & ASSIGNED_DEVICE_CAP_MSI) {
> +            /* Retry with host-side MSI. There might be an IRQ conflict and
> +             * either the kernel or the device doesn't support sharing. */
> +            fprintf(stderr,
> +                    "Host-side INTx sharing not supported, "
> +                    "using MSI instead.\n"
> +                    "Some devices do not to work properly in this mode.\n");
> +            dev->features |= ASSIGNED_DEVICE_PREFER_MSI_MASK;
> +            goto retry;
> +        }
>          fprintf(stderr, "Failed to assign irq for \"%s\": %s\n",
>                  dev->dev.qdev.id, strerror(-r));
>          fprintf(stderr, "Perhaps you are assigning a device "

Acked-by: Alex Williamson <alex.williamson@redhat.com>


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails
  2012-03-19  9:56 [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails Jan Kiszka
  2012-03-19 15:52 ` Alex Williamson
@ 2012-03-21 13:13 ` Avi Kivity
  1 sibling, 0 replies; 3+ messages in thread
From: Avi Kivity @ 2012-03-21 13:13 UTC (permalink / raw)
  To: Jan Kiszka
  Cc: Marcelo Tosatti, kvm, Alex Williamson, Wen Congyang,
	Markus Armbruster

On 03/19/2012 11:56 AM, Jan Kiszka wrote:
> If the host or the device does not support INTx sharing, retry the IRQ
> assignment with host-side MSI support enabled but warn about potential
> consequences. This allows to preserve the previous behavior where we
> defaulted to MSI and did not support INTx sharing at all.
>
>

Applied, thanks.

-- 
error compiling committee.c: too many arguments to function


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2012-03-21 13:13 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-03-19  9:56 [PATCH] pci-assign: Fall back to host-side MSI if INTx sharing fails Jan Kiszka
2012-03-19 15:52 ` Alex Williamson
2012-03-21 13:13 ` Avi Kivity

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox