Re: [PATCH 4/4] drm/i915/reset: Add "display_reset_count" debugfs file

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

On Fri, Apr 10, 2026 at 10:42:23AM +0300, Jani Nikula wrote:
> On Fri, 10 Apr 2026, Ville Syrjala <ville.syrjala@linux.intel.com> wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >
> > Expose the number of display resets performed in a new
> > "display_reset_count" debugfs file. kms_busy can use this to
> > confirm that the kernel actually took the full display reset path.
> >
> > Cc: Jani Nikula <jani.nikula@intel.com>
> > Cc: Jouni Högander <jouni.hogander@intel.com>
> > Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_display_core.h    |  4 ++++
> >  drivers/gpu/drm/i915/display/intel_display_debugfs.c |  2 ++
> >  drivers/gpu/drm/i915/display/intel_display_reset.c   | 10 ++++++++++
> >  drivers/gpu/drm/i915/display/intel_display_reset.h   |  2 ++
> >  drivers/gpu/drm/xe/Makefile                          |  1 +
> >  5 files changed, 19 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_core.h b/drivers/gpu/drm/i915/display/intel_display_core.h
> > index 9e77003addd0..38535d1056d1 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_core.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_core.h
> > @@ -556,6 +556,10 @@ struct intel_display {
> >  		unsigned long mask;
> >  	} quirks;
> >  
> > +	struct {
> > +		u32 count;
> > +	} reset;
> > +
> >  	struct {
> >  		/* restore state for suspend/resume and display reset */
> >  		struct drm_atomic_state *modeset_state;
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.c b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
> > index f244a2b5d139..81bef000a4e3 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_debugfs.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
> > @@ -27,6 +27,7 @@
> >  #include "intel_display_power.h"
> >  #include "intel_display_power_well.h"
> >  #include "intel_display_regs.h"
> > +#include "intel_display_reset.h"
> >  #include "intel_display_rpm.h"
> >  #include "intel_display_types.h"
> >  #include "intel_dmc.h"
> > @@ -838,6 +839,7 @@ void intel_display_debugfs_register(struct intel_display *display)
> >  
> >  	intel_bios_debugfs_register(display);
> >  	intel_cdclk_debugfs_register(display);
> > +	intel_display_reset_debugfs_register(display);
> >  	intel_dmc_debugfs_register(display);
> >  	intel_dp_test_debugfs_register(display);
> >  	intel_fbc_debugfs_register(display);
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_reset.c b/drivers/gpu/drm/i915/display/intel_display_reset.c
> > index ca15dc18ef0f..79c2e77ca137 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_reset.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display_reset.c
> > @@ -3,6 +3,8 @@
> >   * Copyright © 2023 Intel Corporation
> >   */
> >  
> > +#include <linux/debugfs.h>
> > +
> >  #include <drm/drm_atomic_helper.h>
> >  #include <drm/drm_print.h>
> >  
> > @@ -66,6 +68,7 @@ void intel_display_reset_prepare(struct intel_display *display)
> >  		return;
> >  	}
> >  
> > +	display->reset.count++;
> >  	display->restore.modeset_state = state;
> >  	state->acquire_ctx = ctx;
> >  }
> > @@ -114,3 +117,10 @@ void intel_display_reset_finish(struct intel_display *display, bool test_only)
> >  	drm_modeset_acquire_fini(ctx);
> >  	mutex_unlock(&display->drm->mode_config.mutex);
> >  }
> > +
> > +void intel_display_reset_debugfs_register(struct intel_display *display)
> > +{
> > +	debugfs_create_u32("display_reset_count", 0400,
> > +			   display->drm->debugfs_root,
> > +			   &display->reset.count);
> 
> I'm wondering about the names of the debugfs files. We've used the i915_
> prefix so far, but it's obviously misleading nowadays. I've started
> using intel_ in some places.
> 
> I primarily worry about the potential clash with drm core debugfs files,
> which leads to failures to create the file, and clash with other
> drivers, where the files are created all right, but the contents differ
> driver to driver.

Fair point. I suppose I'll just stick an "intel_" prefix on it.

Another option could be use a subdirectory to separate the driver
specific stuff from the core stuff, but dunno if we want to start
down that path. And I guess we'd then need similar subdirectories
inside the crtc/connector/etc. subdirectories.

> 
> *shrug*
> 
> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
> 
> 
> > +}
> > diff --git a/drivers/gpu/drm/i915/display/intel_display_reset.h b/drivers/gpu/drm/i915/display/intel_display_reset.h
> > index a8aa7729d33f..b88c330a3441 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display_reset.h
> > +++ b/drivers/gpu/drm/i915/display/intel_display_reset.h
> > @@ -15,4 +15,6 @@ bool intel_display_reset_test(struct intel_display *display);
> >  void intel_display_reset_prepare(struct intel_display *display);
> >  void intel_display_reset_finish(struct intel_display *display, bool test_only);
> >  
> > +void intel_display_reset_debugfs_register(struct intel_display *display);
> > +
> >  #endif /* __INTEL_RESET_H__ */
> > diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
> > index 110fef511fe2..1a85dfe457f0 100644
> > --- a/drivers/gpu/drm/xe/Makefile
> > +++ b/drivers/gpu/drm/xe/Makefile
> > @@ -262,6 +262,7 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
> >  	i915-display/intel_display_power.o \
> >  	i915-display/intel_display_power_map.o \
> >  	i915-display/intel_display_power_well.o \
> > +	i915-display/intel_display_reset.o \
> >  	i915-display/intel_display_rpm.o \
> >  	i915-display/intel_display_rps.o \
> >  	i915-display/intel_display_trace.o \
> 
> -- 
> Jani Nikula, Intel

-- 
Ville Syrjälä
Intel