[PATCH 0/2] virtio_gpu driver and relevant fix

[Jiaxun Yang]

在2024年5月17日五月 上午12:03，Jiaxun Yang写道：
> Hi all,
>
> This series added support for virtio-gpu device, it has been
> tested on QEMU riscv64 and aarch64.
>
> I also fixed a problem in EFI GOP so it works with EFI applications
> as well.
>
> There is a problem remaining. virtio_init() is being called too
> late to allow virtio-gpu to be selected as a vidconsole. My current
> appraoch is to register virtio_init in init_sequence_r, just after
> pci_init. Is this a viable approach? Should I introduce a config
> like CONFIG_VIRTIO_INIT_R to allow it to be selected by boards?

Just realized I made a mistake on preparing patches to be sent.

Two tiny modifications are missing from this series, dual endian
handling in patch 1 and removing an unused variable in patch 2.

I'll update a reversion, meanwhile fell free to review the current
version.

Apologize for the noise.

Thanks
- Jiaxun

>
> Thanks
>
> Signed-off-by: Jiaxun Yang <jiaxun.yang at flygoat.com>
> ---
> Jiaxun Yang (2):
>       virtio: New virtio_gpu driver
>       efi: gop: Mark pixel_format as BLTONLY if we have sync hook
>
>  drivers/virtio/Kconfig         |  29 +++
>  drivers/virtio/Makefile        |   1 +
>  drivers/virtio/virtio-uclass.c |   1 +
>  drivers/virtio/virtio_gpu.c    | 298 ++++++++++++++++++++++++++++
>  drivers/virtio/virtio_gpu.h    | 428 +++++++++++++++++++++++++++++++++++++++++
>  include/efi_api.h              |   1 +
>  include/virtio.h               |   4 +-
>  lib/efi_loader/efi_gop.c       |   9 +-
>  8 files changed, 769 insertions(+), 2 deletions(-)
> ---
> base-commit: 21aa37ec1cd4a12e01e0e863ec9e99d703ce4d52
> change-id: 20240513-virtio_gpu-abb4de685808
>
> Best regards,
> -- 
> Jiaxun Yang <jiaxun.yang at flygoat.com>

-- 
- Jiaxun