[v6,0/5] drm/v3d: Fix GPU reset issues on the Raspberry Pi 5

Message

Maíra Canal March 18, 2025, 1:01 a.m. UTC

This series addresses GPU reset issues reported in [1], where running a
long compute job would trigger repeated GPU resets, leading to a UI
freeze.

The patches that prevent the same faulty job from being resubmitted in a
loop were merged in drm-misc/drm-misc-fixes in v4.

However, those patches don't solve the issue entirely. Even with a single
GPU reset, the UI still freezes on the Raspberry Pi 5, indicating a GPU
hang. Patches #1, #3, and #5 address this by properly configuring the
V3D_SMS registers, which are required for power management and resets
in V3D 7.1.

Patches #2 and #4 are associated changes related to the robustness
of the DT bindings. Patch #3 added a new register bank to the DT binding
and during the reviewing process, we identified points that could to be
improved in the DT binding.

Patch #2 started by adding per-compatible register restrictions to ensure
that the DTB accurately reflects the hardware design and #5 updates the DT
maintainership, replacing Emma with the current v3d driver maintainer.

[1] https://github.com/raspberrypi/linux/issues/6660

Best Regards,
- Maíra

---
v1 -> v2:
- [1/6, 2/6, 5/6] Add Iago's R-b (Iago Toral)
- [3/6] Use V3D_GEN_* macros consistently throughout the driver (Phil Elwell)
- [3/6] Don't add Iago's R-b in 3/6 due to changes in the patch
- [4/6] Add per-compatible restrictions to enforce per‐SoC register rules (Conor Dooley)
- [6/6] Add Emma's A-b, collected through IRC (Emma Anholt)
- [6/6] Add Rob's A-b (Rob Herring)
- Link to v1: https://lore.kernel.org/r/[e-mail]

v2 -> v3:
- [3/7] Add Iago's R-b (Iago Toral)
- [4/7, 5/7] Separate the patches to ease the reviewing process -> Now,
  PATCH 4/7 only adds the per-compatible rules and PATCH 5/7 adds the
  SMS registers
- [4/7] `allOf` goes above `additionalProperties` (Krzysztof Kozlowski)
- [4/7, 5/7] Sync `reg` and `reg-names` items (Krzysztof Kozlowski)
- Link to v2: https://lore.kernel.org/r/[e-mail]

v3 -> v4:
- [4/7] BCM2712 has an external reset controller, therefore the "bridge"
	register is not needed (Krzysztof Kozlowski)
- [4/7] Remove the word "required" from the reg descriptions (Rob Herring)
- [5/7] Improve commit message (Rob Herring)
- Link to v3: https://lore.kernel.org/r/[e-mail]

v4 -> v5:
- [1/7, 2/7] Applied to drm-misc/drm-misc-fixes
- Changed the commit order to aggregate all DT binding related patches
- [2/6] Add the widest constraints to `reg` and `reg-names` (Krzysztof Kozlowski)
- [2/6] Only declare the registers that exists in a SoC, don't declare all
  the possible registers for a V3D generation (Krzysztof Kozlowski)
- [2/6] Enforce register order by ascending address/offset (Florian Fainelli)
- [2/6] Fix the example to match the new stable ABI
- [3/6] Add Krzysztof's A-b (Krzysztof Kozlowski)
- [4/6] New PATCH: Add more examples to the DT binding
- [5/6] Add "v3d: " to commit title
- Link to v4: https://lore.kernel.org/r/[e-mail]

v5 -> v6:
- Drop "[PATCH v5 4/6] dt-bindings: gpu: v3d: Add additional examples
  to improve binding checks" (Krzysztof Kozlowski)
- [1/5] Improve commit message to mention cases where unexisting V3D
  generations were replaced by supported generations (Stefan Wahren)
- [2/5] Don't break the ABI, enforce the order established by reg (Krzysztof Kozlowski)
- [3/5] Add SMS's functionality in the description (Stefan Wahren)
- Link to v5: https://lore.kernel.org/r/[e-mail]

---
Maíra Canal (5):
      drm/v3d: Associate a V3D tech revision to all supported devices
      dt-bindings: gpu: v3d: Add per-compatible register restrictions
      dt-bindings: gpu: v3d: Add SMS register to BCM2712 compatible
      dt-bindings: gpu: v3d: Add V3D driver maintainer as DT maintainer
      drm/v3d: Use V3D_SMS registers for power on/off and reset on V3D 7.x

 .../devicetree/bindings/gpu/brcm,bcm-v3d.yaml      |  90 ++++++++++++---
 drivers/gpu/drm/v3d/v3d_debugfs.c                  | 126 ++++++++++-----------
 drivers/gpu/drm/v3d/v3d_drv.c                      |  62 +++++++++-
 drivers/gpu/drm/v3d/v3d_drv.h                      |  22 +++-
 drivers/gpu/drm/v3d/v3d_gem.c                      |  27 ++++-
 drivers/gpu/drm/v3d/v3d_irq.c                      |   6 +-
 drivers/gpu/drm/v3d/v3d_perfmon.c                  |   4 +-
 drivers/gpu/drm/v3d/v3d_regs.h                     |  26 +++++
 drivers/gpu/drm/v3d/v3d_sched.c                    |   6 +-
 9 files changed, 271 insertions(+), 98 deletions(-)
---
base-commit: 83a0237859bc5a9e0a716e1db8e7fd3cafd63259
change-id: 20250224-v3d-gpu-reset-fixes-2d21fc70711d

Comments

Maíra Canal March 23, 2025, 2:45 p.m. UTC | #1

On 17/03/25 22:01, Maíra Canal wrote:
> This series addresses GPU reset issues reported in [1], where running a
> long compute job would trigger repeated GPU resets, leading to a UI
> freeze.
> 
> The patches that prevent the same faulty job from being resubmitted in a
> loop were merged in drm-misc/drm-misc-fixes in v4.
> 
> However, those patches don't solve the issue entirely. Even with a single
> GPU reset, the UI still freezes on the Raspberry Pi 5, indicating a GPU
> hang. Patches #1, #3, and #5 address this by properly configuring the
> V3D_SMS registers, which are required for power management and resets
> in V3D 7.1.
> 
> Patches #2 and #4 are associated changes related to the robustness
> of the DT bindings. Patch #3 added a new register bank to the DT binding
> and during the reviewing process, we identified points that could to be
> improved in the DT binding.
> 
> Patch #2 started by adding per-compatible register restrictions to ensure
> that the DTB accurately reflects the hardware design and #5 updates the DT
> maintainership, replacing Emma with the current v3d driver maintainer.
> 
> [1] https://github.com/raspberrypi/linux/issues/6660
> 
> Best Regards,
> - Maíra
> 

[...]

Applied to misc/kernel.git (drm-misc-next). Before applying PATCH 1/5, I
fixed the W=1 warning reported by the kernel test robot with the
following diff:

-       gen = (enum v3d_gen)of_device_get_match_data(dev);
+       gen = (uintptr_t)of_device_get_match_data(dev);

Thanks for all involved in the reviewing process!

Best Regards,
- Maíra

> ---
> Maíra Canal (5):
>        drm/v3d: Associate a V3D tech revision to all supported devices
>        dt-bindings: gpu: v3d: Add per-compatible register restrictions
>        dt-bindings: gpu: v3d: Add SMS register to BCM2712 compatible
>        dt-bindings: gpu: v3d: Add V3D driver maintainer as DT maintainer
>        drm/v3d: Use V3D_SMS registers for power on/off and reset on V3D 7.x
> 
>   .../devicetree/bindings/gpu/brcm,bcm-v3d.yaml      |  90 ++++++++++++---
>   drivers/gpu/drm/v3d/v3d_debugfs.c                  | 126 ++++++++++-----------
>   drivers/gpu/drm/v3d/v3d_drv.c                      |  62 +++++++++-
>   drivers/gpu/drm/v3d/v3d_drv.h                      |  22 +++-
>   drivers/gpu/drm/v3d/v3d_gem.c                      |  27 ++++-
>   drivers/gpu/drm/v3d/v3d_irq.c                      |   6 +-
>   drivers/gpu/drm/v3d/v3d_perfmon.c                  |   4 +-
>   drivers/gpu/drm/v3d/v3d_regs.h                     |  26 +++++
>   drivers/gpu/drm/v3d/v3d_sched.c                    |   6 +-
>   9 files changed, 271 insertions(+), 98 deletions(-)
> ---
> base-commit: 83a0237859bc5a9e0a716e1db8e7fd3cafd63259
> change-id: 20250224-v3d-gpu-reset-fixes-2d21fc70711d
>