Re: [Qemu-devel] [PATCH] libqos: Solve bug in interrupt checking when using MSIX in virtio-pci.c

[John Snow <jsnow@redhat.com>]

On 02/24/2015 11:34 AM, Marc Marí wrote:
> The MSIX interrupt was always acked without checking its value, which caused a
> race condition. If the ISR was raised between the read and the acking, the ISR
> was never detected and it timed out.
>
> Signed-off-by: Marc Marí <marc.mari.barcelo@gmail.com>
> ---
>   tests/libqos/virtio-pci.c |   16 ++++++++++++----
>   1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/tests/libqos/virtio-pci.c b/tests/libqos/virtio-pci.c
> index 788ebaf..c74a669 100644
> --- a/tests/libqos/virtio-pci.c
> +++ b/tests/libqos/virtio-pci.c
> @@ -140,8 +140,12 @@ static bool qvirtio_pci_get_queue_isr_status(QVirtioDevice *d, QVirtQueue *vq)
>               return qpci_msix_pending(dev->pdev, vqpci->msix_entry);
>           } else {
>               data = readl(vqpci->msix_addr);
> -            writel(vqpci->msix_addr, 0);
> -            return data == vqpci->msix_data;
> +            if (data == vqpci->msix_data) {
> +                writel(vqpci->msix_addr, 0);
> +                return true;
> +            } else {
> +                return false;
> +            }
>           }
>       } else {
>           return qpci_io_readb(dev->pdev, dev->addr + QVIRTIO_ISR_STATUS) & 1;
> @@ -160,8 +164,12 @@ static bool qvirtio_pci_get_config_isr_status(QVirtioDevice *d)
>               return qpci_msix_pending(dev->pdev, dev->config_msix_entry);
>           } else {
>               data = readl(dev->config_msix_addr);
> -            writel(dev->config_msix_addr, 0);
> -            return data == dev->config_msix_data;
> +            if (data == dev->config_msix_data) {
> +                writel(dev->config_msix_addr, 0);
> +                return true;
> +            } else {
> +                return false;
> +            }
>           }
>       } else {
>           return qpci_io_readb(dev->pdev, dev->addr + QVIRTIO_ISR_STATUS) & 2;
>

1,600+ runs and no hang, thanks :)

Tested-by: John Snow <jsnow@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>