Re: [Intel-gfx] [PATCH v4 06/10] drm/i915: Separate wakeref tracking

[Andrzej Hajda <andrzej.hajda@intel.com>]

On 20.03.2023 00:57, Andi Shyti wrote:
> Hi Andrzej,
> 
> On Mon, Mar 06, 2023 at 05:32:02PM +0100, Andrzej Hajda wrote:
>> From: Chris Wilson <chris@chris-wilson.co.uk>
>>
>> Extract the callstack tracking of intel_runtime_pm.c into its own
>> utility so that that we can reuse it for other online debugging of
>> scoped wakerefs.
>>
>> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
>> ---
>>   drivers/gpu/drm/i915/Kconfig.debug           |   9 ++
>>   drivers/gpu/drm/i915/Makefile                |   4 +
>>   drivers/gpu/drm/i915/intel_runtime_pm.c      | 222 +++----------------------
>>   drivers/gpu/drm/i915/intel_wakeref.h         |   2 +-
>>   drivers/gpu/drm/i915/intel_wakeref_tracker.c | 234 +++++++++++++++++++++++++++
>>   drivers/gpu/drm/i915/intel_wakeref_tracker.h |  52 ++++++
>>   6 files changed, 319 insertions(+), 204 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
>> index 93dfb7ed970547..5fde52107e3b44 100644
>> --- a/drivers/gpu/drm/i915/Kconfig.debug
>> +++ b/drivers/gpu/drm/i915/Kconfig.debug
>> @@ -25,6 +25,7 @@ config DRM_I915_DEBUG
>>   	select PREEMPT_COUNT
>>   	select I2C_CHARDEV
>>   	select STACKDEPOT
>> +	select STACKTRACE
>>   	select DRM_DP_AUX_CHARDEV
>>   	select X86_MSR # used by igt/pm_rpm
>>   	select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks)
>> @@ -37,6 +38,7 @@ config DRM_I915_DEBUG
>>   	select DRM_I915_DEBUG_GEM
>>   	select DRM_I915_DEBUG_GEM_ONCE
>>   	select DRM_I915_DEBUG_MMIO
>> +	select DRM_I915_TRACK_WAKEREF
>>   	select DRM_I915_DEBUG_RUNTIME_PM
>>   	select DRM_I915_SW_FENCE_DEBUG_OBJECTS
>>   	select DRM_I915_SELFTEST
>> @@ -227,11 +229,18 @@ config DRM_I915_DEBUG_VBLANK_EVADE
>>   
>>   	  If in doubt, say "N".
>>   
>> +config DRM_I915_TRACK_WAKEREF
>> +	depends on STACKDEPOT
>> +	depends on STACKTRACE
>> +	bool
>> +
>>   config DRM_I915_DEBUG_RUNTIME_PM
>>   	bool "Enable extra state checking for runtime PM"
>>   	depends on DRM_I915
>>   	default n
>>   	select STACKDEPOT
>> +	select STACKTRACE
>> +	select DRM_I915_TRACK_WAKEREF
>>   	help
>>   	  Choose this option to turn on extra state checking for the
>>   	  runtime PM functionality. This may introduce overhead during
>> diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
>> index b2f91a1f826858..42daff6d575a82 100644
>> --- a/drivers/gpu/drm/i915/Makefile
>> +++ b/drivers/gpu/drm/i915/Makefile
>> @@ -81,6 +81,10 @@ i915-$(CONFIG_DEBUG_FS) += \
>>   	i915_debugfs_params.o \
>>   	display/intel_display_debugfs.o \
>>   	display/intel_pipe_crc.o
>> +
>> +i915-$(CONFIG_DRM_I915_TRACK_WAKEREF) += \
>> +	intel_wakeref_tracker.o
>> +
> 
> This patch, along with the previous one and two following it, is
> a bit confusing. We add this file only to remove it later and
> the code hops from file to file. There seem to be some extra
> steps that could be avoided.
> 
> Is there room for simplification?

The reason behind this was that i915 had it's own tracker integrated 
with i915_runtime_pm, then it was abstracted out (05,06) to allow track 
gt->wakerefs (07) and then I proposed replacement of internal tracker 
with ref_tracker (09). I wanted to keep original history of development.
I can squash all/some of this work, but I am afraid it will generate 
less readable patches - now we have separated abstract-out and replace 
steps.

Probably sending patches 05-08 1st, then proposing conversion to 
ref_tracker in another patchset would make it more clear.

Regards
Andrzej


> 
> Thanks,
> Andi