Re: [PATCH v2 4/5] drm/i915: use display snapshot mechanism for display irq regs

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Wed, Apr 09, 2025 at 05:23:46PM +0300, Jani Nikula wrote:
> Move more display specific parts of GPU error logging behind the display
> snapshot interface.
> 
> With the display register capture reduced to just one register, DERRMR,
> there's quite a bit of boilerplate here. However, it's still a nice
> abstraction and removes a DISPLAY_VER() usage from core i915. With this
> approach, it's also easy to add to xe as needed.
> 
> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
> ---
>  .../gpu/drm/i915/display/intel_display_irq.c  | 28 +++++++++++++++++++
>  .../gpu/drm/i915/display/intel_display_irq.h  |  5 ++++
>  .../drm/i915/display/intel_display_snapshot.c |  5 ++++
>  drivers/gpu/drm/i915/i915_gpu_error.c         | 18 ------------
>  drivers/gpu/drm/i915/i915_gpu_error.h         |  1 -
>  5 files changed, 38 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.c b/drivers/gpu/drm/i915/display/intel_display_irq.c
> index d2a35e3630b1..f5dc050a6344 100644
> --- a/drivers/gpu/drm/i915/display/intel_display_irq.c
> +++ b/drivers/gpu/drm/i915/display/intel_display_irq.c
> @@ -2329,3 +2329,31 @@ void intel_display_irq_init(struct intel_display *display)
>  
>  	INIT_WORK(&display->irq.vblank_dc_work, intel_display_vblank_dc_work);
>  }
> +
> +struct intel_display_irq_snapshot {
> +	u32 derrmr;
> +};
> +
> +struct intel_display_irq_snapshot *
> +intel_display_irq_snapshot_capture(struct intel_display *display)
> +{
> +	struct intel_display_irq_snapshot *snapshot;
> +
> +	snapshot = kzalloc(sizeof(*snapshot), GFP_ATOMIC);
> +	if (!snapshot)
> +		return NULL;
> +
> +	if (DISPLAY_VER(display) >= 6 && DISPLAY_VER(display) < 20)

Technically we should exclude vlv/chv here (or just exclude all gmch).
IIRC vlv/chv have some kind of other register with a potentially similar
function, but we don't use that anywhere.

> +		snapshot->derrmr = intel_de_read(display, DERRMR);
> +
> +	return snapshot;
> +}
> +
> +void intel_display_irq_snapshot_print(const struct intel_display_irq_snapshot *snapshot,
> +				      struct drm_printer *p)
> +{
> +	if (!snapshot)
> +		return;
> +
> +	drm_printf(p, "DERRMR: 0x%08x\n", snapshot->derrmr);
> +}
> diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.h b/drivers/gpu/drm/i915/display/intel_display_irq.h
> index f72727768351..4f0e27a37854 100644
> --- a/drivers/gpu/drm/i915/display/intel_display_irq.h
> +++ b/drivers/gpu/drm/i915/display/intel_display_irq.h
> @@ -12,7 +12,9 @@
>  
>  enum pipe;
>  struct drm_crtc;
> +struct drm_printer;
>  struct intel_display;
> +struct intel_display_irq_snapshot;
>  
>  void valleyview_enable_display_irqs(struct intel_display *display);
>  void valleyview_disable_display_irqs(struct intel_display *display);
> @@ -82,4 +84,7 @@ void intel_display_irq_init(struct intel_display *display);
>  
>  void i915gm_irq_cstate_wa(struct intel_display *display, bool enable);
>  
> +struct intel_display_irq_snapshot *intel_display_irq_snapshot_capture(struct intel_display *display);
> +void intel_display_irq_snapshot_print(const struct intel_display_irq_snapshot *snapshot, struct drm_printer *p);
> +
>  #endif /* __INTEL_DISPLAY_IRQ_H__ */
> diff --git a/drivers/gpu/drm/i915/display/intel_display_snapshot.c b/drivers/gpu/drm/i915/display/intel_display_snapshot.c
> index 25ba043cbb65..66087302fdbc 100644
> --- a/drivers/gpu/drm/i915/display/intel_display_snapshot.c
> +++ b/drivers/gpu/drm/i915/display/intel_display_snapshot.c
> @@ -7,6 +7,7 @@
>  
>  #include "intel_display_core.h"
>  #include "intel_display_device.h"
> +#include "intel_display_irq.h"
>  #include "intel_display_params.h"
>  #include "intel_display_snapshot.h"
>  #include "intel_dmc.h"
> @@ -20,6 +21,7 @@ struct intel_display_snapshot {
>  	struct intel_display_params params;
>  	struct intel_overlay_snapshot *overlay;
>  	struct intel_dmc_snapshot *dmc;
> +	struct intel_display_irq_snapshot *irq;
>  };
>  
>  struct intel_display_snapshot *intel_display_snapshot_capture(struct intel_display *display)
> @@ -38,6 +40,7 @@ struct intel_display_snapshot *intel_display_snapshot_capture(struct intel_displ
>  
>  	intel_display_params_copy(&snapshot->params);
>  
> +	snapshot->irq = intel_display_irq_snapshot_capture(display);
>  	snapshot->overlay = intel_overlay_snapshot_capture(display);
>  	snapshot->dmc = intel_dmc_snapshot_capture(display);
>  
> @@ -57,6 +60,7 @@ void intel_display_snapshot_print(const struct intel_display_snapshot *snapshot,
>  	intel_display_device_info_print(&snapshot->info, &snapshot->runtime_info, p);
>  	intel_display_params_dump(&snapshot->params, display->drm->driver->name, p);
>  
> +	intel_display_irq_snapshot_print(snapshot->irq, p);
>  	intel_overlay_snapshot_print(snapshot->overlay, p);
>  	intel_dmc_snapshot_print(snapshot->dmc, p);
>  }
> @@ -68,6 +72,7 @@ void intel_display_snapshot_free(struct intel_display_snapshot *snapshot)
>  
>  	intel_display_params_free(&snapshot->params);
>  
> +	kfree(snapshot->irq);
>  	kfree(snapshot->overlay);
>  	kfree(snapshot->dmc);
>  	kfree(snapshot);
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 50cd0c433e6d..62e4a293ed9b 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -726,12 +726,6 @@ static void err_print_gt_info(struct drm_i915_error_state_buf *m,
>  	intel_sseu_print_topology(gt->_gt->i915, &gt->info.sseu, &p);
>  }
>  
> -static void err_print_gt_display(struct drm_i915_error_state_buf *m,
> -				 struct intel_gt_coredump *gt)
> -{
> -	err_printf(m, "DERRMR: 0x%08x\n", gt->derrmr);
> -}
> -
>  static void err_print_gt_global_nonguc(struct drm_i915_error_state_buf *m,
>  				       struct intel_gt_coredump *gt)
>  {
> @@ -877,7 +871,6 @@ static void __err_print_to_sgl(struct drm_i915_error_state_buf *m,
>  		if (error->gt->uc && error->gt->uc->guc.is_guc_capture)
>  			print_guc_capture = true;
>  
> -		err_print_gt_display(m, error->gt);
>  		err_print_gt_global_nonguc(m, error->gt);
>  		err_print_gt_fences(m, error->gt);
>  
> @@ -1766,16 +1759,6 @@ gt_record_uc(struct intel_gt_coredump *gt,
>  	return error_uc;
>  }
>  
> -/* Capture display registers. */
> -static void gt_record_display_regs(struct intel_gt_coredump *gt)
> -{
> -	struct intel_uncore *uncore = gt->_gt->uncore;
> -	struct drm_i915_private *i915 = uncore->i915;
> -
> -	if (DISPLAY_VER(i915) >= 6 && DISPLAY_VER(i915) < 20)
> -		gt->derrmr = intel_uncore_read(uncore, DERRMR);
> -}
> -
>  /* Capture all other registers that GuC doesn't capture. */
>  static void gt_record_global_nonguc_regs(struct intel_gt_coredump *gt)
>  {
> @@ -2034,7 +2017,6 @@ intel_gt_coredump_alloc(struct intel_gt *gt, gfp_t gfp, u32 dump_flags)
>  	gc->_gt = gt;
>  	gc->awake = intel_gt_pm_is_awake(gt);
>  
> -	gt_record_display_regs(gc);
>  	gt_record_global_nonguc_regs(gc);
>  
>  	/*
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h
> index ac55603a81d7..004e7ad52a4a 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.h
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.h
> @@ -164,7 +164,6 @@ struct intel_gt_coredump {
>  	u32 clock_period_ns;
>  
>  	/* Display related */

That comment should be nuked as well.

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

> -	u32 derrmr;
>  	u32 sfc_done[I915_MAX_SFC]; /* gen12 */
>  
>  	u32 nfence;
> -- 
> 2.39.5

-- 
Ville Syrjälä
Intel