On Mon, Aug 6, 2018 at 3:39 PM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 3 August 2018 at 16:32, Marc-André Lureau > <marcandre.lur...@redhat.com> wrote: >> With vga=775 on the Linux command line a first boot of the VM running >> Linux works fine. After a warm reboot it crashes during Linux boot. >> >> Before that, valgrind points out bad memory write to console >> surface. The VGA code is not aware that virtio-gpu got a message >> surface scanout when the display is disabled. Let's reset VGA graphic >> mode when it is the case, so that a new display surface is created >> when doing further VGA operations. >> >> https://bugs.launchpad.net/qemu/+bug/1784900/ >> >> Reported-by: Stefan Berger <stef...@linux.vnet.ibm.com> >> Signed-off-by: Marc-André Lureau <marcandre.lur...@redhat.com> >> --- >> include/hw/virtio/virtio-gpu.h | 1 + >> hw/display/virtio-gpu.c | 5 +++++ >> hw/display/virtio-vga.c | 11 +++++++++++ >> 3 files changed, 17 insertions(+) >> >> diff --git a/include/hw/virtio/virtio-gpu.h b/include/hw/virtio/virtio-gpu.h >> index 9780f755ef..d0321672f4 100644 >> --- a/include/hw/virtio/virtio-gpu.h >> +++ b/include/hw/virtio/virtio-gpu.h >> @@ -125,6 +125,7 @@ typedef struct VirtIOGPU { >> uint32_t bytes_3d; >> } stats; >> >> + void (*disable_scanout)(struct VirtIOGPU *g, int scanout_id); >> Error *migration_blocker; >> } VirtIOGPU; > > I guess for a last-minute 3.0 fix this is ok, but it looks > a bit weird to have a function pointer in a device instance > struct, which the user of the device has reached in and > set in order to modify the behaviour of the object...
It's not much different than pointing to a static const structure with function pointers (I thought that was unnecessary for just one function) We could also have a VirtioGPUClass with virtual methods, and subclass TYPE_VIRTIO_GPU. Also a bit over-complicated, and conflicting with my vhost-user-gpu series. So we may eventually get there in 3.1 though. > > thanks > -- PMM
