mbox series

[v1,0/6] Support virtio-gpu DRM native context

Message ID 20241006232350.3198759-1-dmitry.osipenko@collabora.com (mailing list archive)
Headers show
Series Support virtio-gpu DRM native context | expand

Message

Dmitry Osipenko Oct. 6, 2024, 11:23 p.m. UTC
This patchset adds DRM native context support to VirtIO-GPU on Qemu.
It's based on the pending Venus v17 patches [1] that bring host blobs
support to virtio-gpu-gl device.

[1] https://lore.kernel.org/qemu-devel/20240822185110.1757429-1-dmitry.osipenko@collabora.com/

Contarary to Virgl and Venus contexts which mediate high level GFX APIs,
DRM native context [2] mediates lower level kernel driver UAPI, which
reflects in a less CPU overhead and less/simpler code needed to support it.
DRM context consists of a host and guest parts that have to be implemented
for each GPU driver. On a guest side, DRM context presents a virtual GPU as
a real/native host GPU device for GL/VK applications.

[2] https://www.youtube.com/watch?v=9sFP_yddLLQ

Today there are four known DRM native context drivers existing in a wild:

  - Freedreno (Qualcomm SoC GPUs), completely upstreamed
  - AMDGPU, mostly merged into upstreams
  - Intel (i915), merge requests are opened
  - Asahi (Apple SoC GPUs), WIP status


# How to try out DRM context:

1. Like Venus and Virgl context, DRM context requires applying WIP
KVM patches [3] to host kernel, otherwise mapping of GPU memory blobs
will likely fail.

[3] https://lore.kernel.org/all/20240726235234.228822-1-seanjc@google.com/

2. Use latest libvirglrenderer from upstream git/main for Freedreno
and AMDGPU native contexts. For Intel use patches [4].

[4] https://gitlab.freedesktop.org/virgl/virglrenderer/-/merge_requests/1384

3. On guest, use latest Mesa version for Freedreno. For AMDGPU use
Mesa patches [5], for Intel [6].

[5] https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21658
[6] https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29870

4. On guest, use latest Linux kernel v6.6+.

Example Qemu cmdline that enables DRM context:

  qemu-system-x86_64 -device virtio-vga-gl,hostmem=4G,blob=true,drm=true \
      -machine q35,accel=kvm,memory-backend=mem1 \
      -object memory-backend-memfd,id=mem1,size=8G -m 8G


# Note about known performance problem in Qemu:

DRM contexts are mapping host blobs extensively and these mapping
operations work slowly in Qemu. Exact reason is unkown. Mappings work
fast on Crosvm For DRM contexts this problem is more visible than for
Venus/Virgl.

Dmitry Osipenko (5):
  ui/sdl2: Restore original context after new context creation
  linux-headers: Update to Linux v6.12-rc1
  virtio-gpu: Handle virgl fence creation errors
  virtio-gpu: Support asynchronous fencing
  virtio-gpu: Support DRM native context

Pierre-Eric Pelloux-Prayer (1):
  ui/sdl2: Implement dpy dmabuf functions

 docs/system/devices/virtio-gpu.rst            |  11 +
 hw/display/virtio-gpu-gl.c                    |   5 +
 hw/display/virtio-gpu-virgl.c                 | 153 ++++++++++--
 hw/display/virtio-gpu.c                       |  15 ++
 include/hw/virtio/virtio-gpu.h                |  17 ++
 include/standard-headers/drm/drm_fourcc.h     |  43 ++++
 include/standard-headers/linux/const.h        |  17 ++
 include/standard-headers/linux/ethtool.h      | 226 ++++++++++++++++++
 include/standard-headers/linux/fuse.h         |  22 +-
 .../linux/input-event-codes.h                 |   2 +
 include/standard-headers/linux/pci_regs.h     |  41 +++-
 .../standard-headers/linux/virtio_balloon.h   |  16 +-
 include/standard-headers/linux/virtio_gpu.h   |   1 +
 include/ui/sdl2.h                             |   5 +
 linux-headers/asm-arm64/mman.h                |   9 +
 linux-headers/asm-arm64/unistd.h              |  25 +-
 linux-headers/asm-generic/unistd.h            |   6 +-
 linux-headers/asm-loongarch/kvm.h             |  24 ++
 linux-headers/asm-loongarch/unistd.h          |   4 +-
 linux-headers/asm-riscv/kvm.h                 |   7 +
 linux-headers/asm-riscv/unistd.h              |  41 +---
 linux-headers/asm-x86/kvm.h                   |   2 +
 linux-headers/asm-x86/unistd_64.h             |   1 +
 linux-headers/asm-x86/unistd_x32.h            |   1 +
 linux-headers/linux/bits.h                    |   3 +
 linux-headers/linux/const.h                   |  17 ++
 linux-headers/linux/iommufd.h                 | 143 +++++++++--
 linux-headers/linux/kvm.h                     |  23 +-
 linux-headers/linux/mman.h                    |   1 +
 linux-headers/linux/psp-sev.h                 |  28 +++
 ui/sdl2-gl.c                                  |  42 ++++
 ui/sdl2.c                                     |   8 +
 32 files changed, 851 insertions(+), 108 deletions(-)

Comments

Akihiko Odaki Oct. 7, 2024, 9:48 a.m. UTC | #1
On 2024/10/07 8:23, Dmitry Osipenko wrote:
> This patchset adds DRM native context support to VirtIO-GPU on Qemu.
> It's based on the pending Venus v17 patches [1] that bring host blobs
> support to virtio-gpu-gl device.

Hi Dmitry,

Thank you for submitting this series.

> 
> [1] https://lore.kernel.org/qemu-devel/20240822185110.1757429-1-dmitry.osipenko@collabora.com/

Please use Based-on: tag.
For details, see: docs/devel/submitting-a-patch.rst

Regards,
Akihiko Odaki

> 
> Contarary to Virgl and Venus contexts which mediate high level GFX APIs,
> DRM native context [2] mediates lower level kernel driver UAPI, which
> reflects in a less CPU overhead and less/simpler code needed to support it.
> DRM context consists of a host and guest parts that have to be implemented
> for each GPU driver. On a guest side, DRM context presents a virtual GPU as
> a real/native host GPU device for GL/VK applications.
> 
> [2] https://www.youtube.com/watch?v=9sFP_yddLLQ
> 
> Today there are four known DRM native context drivers existing in a wild:
> 
>    - Freedreno (Qualcomm SoC GPUs), completely upstreamed
>    - AMDGPU, mostly merged into upstreams
>    - Intel (i915), merge requests are opened
>    - Asahi (Apple SoC GPUs), WIP status
> 
> 
> # How to try out DRM context:
> 
> 1. Like Venus and Virgl context, DRM context requires applying WIP
> KVM patches [3] to host kernel, otherwise mapping of GPU memory blobs
> will likely fail.
> 
> [3] https://lore.kernel.org/all/20240726235234.228822-1-seanjc@google.com/
> 
> 2. Use latest libvirglrenderer from upstream git/main for Freedreno
> and AMDGPU native contexts. For Intel use patches [4].
> 
> [4] https://gitlab.freedesktop.org/virgl/virglrenderer/-/merge_requests/1384
> 
> 3. On guest, use latest Mesa version for Freedreno. For AMDGPU use
> Mesa patches [5], for Intel [6].
> 
> [5] https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21658
> [6] https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29870
> 
> 4. On guest, use latest Linux kernel v6.6+.
> 
> Example Qemu cmdline that enables DRM context:
> 
>    qemu-system-x86_64 -device virtio-vga-gl,hostmem=4G,blob=true,drm=true \
>        -machine q35,accel=kvm,memory-backend=mem1 \
>        -object memory-backend-memfd,id=mem1,size=8G -m 8G
> 
> 
> # Note about known performance problem in Qemu:
> 
> DRM contexts are mapping host blobs extensively and these mapping
> operations work slowly in Qemu. Exact reason is unkown. Mappings work
> fast on Crosvm For DRM contexts this problem is more visible than for
> Venus/Virgl.
> 
> Dmitry Osipenko (5):
>    ui/sdl2: Restore original context after new context creation
>    linux-headers: Update to Linux v6.12-rc1
>    virtio-gpu: Handle virgl fence creation errors
>    virtio-gpu: Support asynchronous fencing
>    virtio-gpu: Support DRM native context
> 
> Pierre-Eric Pelloux-Prayer (1):
>    ui/sdl2: Implement dpy dmabuf functions
> 
>   docs/system/devices/virtio-gpu.rst            |  11 +
>   hw/display/virtio-gpu-gl.c                    |   5 +
>   hw/display/virtio-gpu-virgl.c                 | 153 ++++++++++--
>   hw/display/virtio-gpu.c                       |  15 ++
>   include/hw/virtio/virtio-gpu.h                |  17 ++
>   include/standard-headers/drm/drm_fourcc.h     |  43 ++++
>   include/standard-headers/linux/const.h        |  17 ++
>   include/standard-headers/linux/ethtool.h      | 226 ++++++++++++++++++
>   include/standard-headers/linux/fuse.h         |  22 +-
>   .../linux/input-event-codes.h                 |   2 +
>   include/standard-headers/linux/pci_regs.h     |  41 +++-
>   .../standard-headers/linux/virtio_balloon.h   |  16 +-
>   include/standard-headers/linux/virtio_gpu.h   |   1 +
>   include/ui/sdl2.h                             |   5 +
>   linux-headers/asm-arm64/mman.h                |   9 +
>   linux-headers/asm-arm64/unistd.h              |  25 +-
>   linux-headers/asm-generic/unistd.h            |   6 +-
>   linux-headers/asm-loongarch/kvm.h             |  24 ++
>   linux-headers/asm-loongarch/unistd.h          |   4 +-
>   linux-headers/asm-riscv/kvm.h                 |   7 +
>   linux-headers/asm-riscv/unistd.h              |  41 +---
>   linux-headers/asm-x86/kvm.h                   |   2 +
>   linux-headers/asm-x86/unistd_64.h             |   1 +
>   linux-headers/asm-x86/unistd_x32.h            |   1 +
>   linux-headers/linux/bits.h                    |   3 +
>   linux-headers/linux/const.h                   |  17 ++
>   linux-headers/linux/iommufd.h                 | 143 +++++++++--
>   linux-headers/linux/kvm.h                     |  23 +-
>   linux-headers/linux/mman.h                    |   1 +
>   linux-headers/linux/psp-sev.h                 |  28 +++
>   ui/sdl2-gl.c                                  |  42 ++++
>   ui/sdl2.c                                     |   8 +
>   32 files changed, 851 insertions(+), 108 deletions(-)
>
Dmitry Osipenko Oct. 7, 2024, 12:02 p.m. UTC | #2
On 10/7/24 12:48, Akihiko Odaki wrote:
> On 2024/10/07 8:23, Dmitry Osipenko wrote:
>> This patchset adds DRM native context support to VirtIO-GPU on Qemu.
>> It's based on the pending Venus v17 patches [1] that bring host blobs
>> support to virtio-gpu-gl device.
> 
> Hi Dmitry,
> 
> Thank you for submitting this series.
> 
>>
>> [1]
>> https://lore.kernel.org/qemu-devel/20240822185110.1757429-1-dmitry.osipenko@collabora.com/
> 
> Please use Based-on: tag.
> For details, see: docs/devel/submitting-a-patch.rst

Thanks for a very quick review! All the comments are good to me, will
address them in v2.