| Message ID | 1453447210-13259-1-git-send-email-famz@redhat.com |
|---|---|
| State | New |
| Headers | show |
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c index 94667e6..f0ab812 100644 --- a/hw/virtio/virtio-pci.c +++ b/hw/virtio/virtio-pci.c @@ -1310,6 +1310,8 @@ static uint64_t virtio_pci_common_read(void *opaque, hwaddr addr, return val; } +static void virtio_pci_reset(DeviceState *qdev); + static void virtio_pci_common_write(void *opaque, hwaddr addr, uint64_t val, unsigned size) { @@ -1351,7 +1353,7 @@ static void virtio_pci_common_write(void *opaque, hwaddr addr, } if (vdev->status == 0) { - virtio_reset(vdev); + virtio_pci_reset(DEVICE(proxy)); msix_unuse_all_vectors(&proxy->pci_dev); }
Virtio 1.0 requires queue_enable status in virtio_pci_common_cfg to be cleared upon device reset (virtio-v1.0.pdf, section 4.1.4.3.1), but we don't do it. This causes trouble on Linux's virtio-blk-pci probe since seabios 1.9.0 update (commit ad30c0b0d), if modern enabled. Kernel spits: ... [ 5.624492] virtio_blk: probe of virtio0 failed with error -2 ... To fix it, call the virtio-pci's reset handler instead of the device attached on the bus. Signed-off-by: Fam Zheng <famz@redhat.com> --- hw/virtio/virtio-pci.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)