[SLOF] [PATCH 0/3] Fix virtio-serial
aik at ozlabs.ru
Tue Mar 10 14:43:15 AEDT 2020
On 10/03/2020 09:42, Greg Kurz wrote:
> This fixes the virtio-serial failures we're observing since the VQs
> were moved to the device.
> Patch 1 addresses the issue of the missing current instance, ie. my-self,
> when calling "dma-map-in". This fixes IOMMU support.
> Patch 2 reworks the shutdown sequence of the device. Note that the guest
> seems to be booting okay with or without IOMMU at this point.
> Patch 3 is Alexey's "virtio-serial: Close device completely" patch
> rebased on top of the other two.
> Lightly tested with a fedora31 guest. Doesn't break spapr-vty.
> Please comment.
Lovely, thanks a lot for fixing this up! I'll push this out tonight,
> Alexey Kardashevskiy (1):
> virtio-serial: Close device completely
> Greg Kurz (2):
> virtio-serial: Don't override some words
> virtio-serial: Rework shutdown sequence
> board-qemu/slof/virtio-serial.fs | 18 ++++++------------
> lib/libvirtio/virtio-serial.c | 7 +++++++
> slof/fs/client.fs | 5 +++++
> 3 files changed, 18 insertions(+), 12 deletions(-)
More information about the SLOF