mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] vfio: Fix PCI 2.3 shared interrupt
@ 2010-11-03 20:17 Alex Williamson
  2010-11-03 21:03 ` Tom Lyon
  0 siblings, 1 reply; 2+ messages in thread
From: Alex Williamson @ 2010-11-03 20:17 UTC (permalink / raw)
  To: pugs; +Cc: linux-kernel, linux-pci, kvm, alex.williamson

Trying to be too clever with setting the irq_disabled flag.  PCI 2.3
disabled devices can still share IRQs, which can lead to clearing
the irq_disabled flag, preventing the EOI from registering, and leaving
the device without interrupts.  Interrupt handler should only ever
set irq_disabled and we can exit earlier to avoid the config space
access if we know we're disabled.

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

 drivers/vfio/vfio_intrs.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/drivers/vfio/vfio_intrs.c b/drivers/vfio/vfio_intrs.c
index 604082c..73e3deb 100644
--- a/drivers/vfio/vfio_intrs.c
+++ b/drivers/vfio/vfio_intrs.c
@@ -57,6 +57,12 @@ irqreturn_t vfio_interrupt(int irq, void *dev_id)
 
 	spin_lock_irq(&vdev->irqlock);
 
+	/* INTX disabled interrupts can still be shared */
+	if (vdev->irq_disabled) {
+		spin_unlock_irq(&vdev->irqlock);
+		return ret;
+	}
+
 	if (vdev->pci_2_3) {
 		pci_block_user_cfg_access(pdev);
 
@@ -87,7 +93,8 @@ done:
 		ret = IRQ_HANDLED;
 	}
 
-	vdev->irq_disabled = (ret == IRQ_HANDLED);
+	if (ret == IRQ_HANDLED)
+		vdev->irq_disabled = true;
 
 	spin_unlock_irq(&vdev->irqlock);
 


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] vfio: Fix PCI 2.3 shared interrupt
  2010-11-03 20:17 [PATCH] vfio: Fix PCI 2.3 shared interrupt Alex Williamson
@ 2010-11-03 21:03 ` Tom Lyon
  0 siblings, 0 replies; 2+ messages in thread
From: Tom Lyon @ 2010-11-03 21:03 UTC (permalink / raw)
  To: Alex Williamson; +Cc: linux-kernel, linux-pci, kvm

Applied.

On Wednesday, November 03, 2010 01:17:33 pm Alex Williamson wrote:
> Trying to be too clever with setting the irq_disabled flag.  PCI 2.3
> disabled devices can still share IRQs, which can lead to clearing
> the irq_disabled flag, preventing the EOI from registering, and leaving
> the device without interrupts.  Interrupt handler should only ever
> set irq_disabled and we can exit earlier to avoid the config space
> access if we know we're disabled.
> 
> Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> ---
> 
>  drivers/vfio/vfio_intrs.c |    9 ++++++++-
>  1 files changed, 8 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/vfio/vfio_intrs.c b/drivers/vfio/vfio_intrs.c
> index 604082c..73e3deb 100644
> --- a/drivers/vfio/vfio_intrs.c
> +++ b/drivers/vfio/vfio_intrs.c
> @@ -57,6 +57,12 @@ irqreturn_t vfio_interrupt(int irq, void *dev_id)
> 
>  	spin_lock_irq(&vdev->irqlock);
> 
> +	/* INTX disabled interrupts can still be shared */
> +	if (vdev->irq_disabled) {
> +		spin_unlock_irq(&vdev->irqlock);
> +		return ret;
> +	}
> +
>  	if (vdev->pci_2_3) {
>  		pci_block_user_cfg_access(pdev);
> 
> @@ -87,7 +93,8 @@ done:
>  		ret = IRQ_HANDLED;
>  	}
> 
> -	vdev->irq_disabled = (ret == IRQ_HANDLED);
> +	if (ret == IRQ_HANDLED)
> +		vdev->irq_disabled = true;
> 
>  	spin_unlock_irq(&vdev->irqlock);

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2010-11-03 21:03 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-11-03 20:17 [PATCH] vfio: Fix PCI 2.3 shared interrupt Alex Williamson
2010-11-03 21:03 ` Tom Lyon

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

all inboxes | Powered by JetHome®