On Thu, 10 Oct 2024, "Hogander, Jouni" <jouni.hogander@xxxxxxxxx> wrote:
> On Wed, 2024-10-09 at 17:15 +0300, Jani Nikula wrote:
>> On Wed, 09 Oct 2024, Jouni Högander <jouni.hogander@xxxxxxxxx> wrote:
>> > We are about to change meaning of vblank_enabled to fix Panel
>> > Replay vblank
>> > workaround. For sake of clarity we need to rename it.
>> > Vblank_enabled is
>> > used for i915gm/i945gm vblank irq workaround as well -> instead of
>> > rename
>> > add new counter named as vblank_wa_pipes.
>> >
>> > v2:
>> > - s/vblank_wa_pipes/vblank_wa_num_pipes/
>> > - use int as a type for the counter
>> >
>> > Signed-off-by: Jouni Högander <jouni.hogander@xxxxxxxxx>
>> > ---
>> > drivers/gpu/drm/i915/display/intel_display_core.h | 2 ++
>> > drivers/gpu/drm/i915/display/intel_display_irq.c | 8 ++++----
>> > 2 files changed, 6 insertions(+), 4 deletions(-)
>> >
>> > diff --git a/drivers/gpu/drm/i915/display/intel_display_core.h
>> > b/drivers/gpu/drm/i915/display/intel_display_core.h
>> > index 982dd9469195..45697af25fa9 100644
>> > --- a/drivers/gpu/drm/i915/display/intel_display_core.h
>> > +++ b/drivers/gpu/drm/i915/display/intel_display_core.h
>> > @@ -455,6 +455,8 @@ struct intel_display {
>> > /* For i915gm/i945gm vblank irq workaround */
>> > u8 vblank_enabled;
>>
>> Maybe we want to rename this one too?
>
> I tried to explain it in commit message. Seems I didn't succeed in it.
>
> vblank_enabled is currently used by two workarounds:
>
> 1. i915gm/i945gm vblank irq workaround
> 2. LNL Panel Replay vblank workaround
>
> 1. and 2. are currently using vblank_enabled in a way they are
> respecting it's meaning: is vblank enabled/disabled.
Crucially, they're both using the same member, but never on the same
platform. The member is "overloaded" for two different things. (See the
otherwise unrelated series [1], it's the same thing with sb_lock.)
> It was found out that 2. as it is implemented currently doesn't work
> properly. Instead of information on vblank enabled/disabled we need to
> know if there are pipes that need the workaround.
>
> Renaming vblank_enabled wouldn't reflect how it's used in 1. This is
> why I decided to add own variable for 2. and this is what this patch is
> about: Keep vblank_enabled as it is for 1. and add own variable for 2.
My point is more generic, and it was not intended as a comment on *this*
patch. Having a "vblank_enabled" member is misleading, when it's really
just used for a very narrow case on a few platforms.
Does that make more sense?
BR,
Jani.
[1] https://lore.kernel.org/r/cover.1727890136.git.jani.nikula@xxxxxxxxx
>
> BR,
>
> Jouni Högander
>
>>
>> BR,
>> Jani.
>>
>> >
>> > + int vblank_wa_num_pipes;
>> > +
>> > struct work_struct vblank_dc_work;
>> >
>> > u32 de_irq_mask[I915_MAX_PIPES];
>> > diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.c
>> > b/drivers/gpu/drm/i915/display/intel_display_irq.c
>> > index a4367ddc7a44..8226ea218d77 100644
>> > --- a/drivers/gpu/drm/i915/display/intel_display_irq.c
>> > +++ b/drivers/gpu/drm/i915/display/intel_display_irq.c
>> > @@ -1424,7 +1424,7 @@ static void
>> > intel_display_vblank_dc_work(struct work_struct *work)
>> > struct intel_display *display =
>> > container_of(work, typeof(*display),
>> > irq.vblank_dc_work);
>> > struct drm_i915_private *i915 = to_i915(display->drm);
>> > - u8 vblank_enabled = READ_ONCE(display->irq.vblank_enabled);
>> > + int vblank_wa_num_pipes = READ_ONCE(display-
>> > >irq.vblank_wa_num_pipes);
>> >
>> > /*
>> > * NOTE: intel_display_power_set_target_dc_state is used
>> > only by PSR
>> > @@ -1432,7 +1432,7 @@ static void
>> > intel_display_vblank_dc_work(struct work_struct *work)
>> > * PSR code. If DC3CO is taken into use we need take that
>> > into account
>> > * here as well.
>> > */
>> > - intel_display_power_set_target_dc_state(i915,
>> > vblank_enabled ? DC_STATE_DISABLE :
>> > + intel_display_power_set_target_dc_state(i915,
>> > vblank_wa_num_pipes ? DC_STATE_DISABLE :
>> > DC_STATE_EN_UPTO_DC
>> > 6);
>> > }
>> >
>> > @@ -1447,7 +1447,7 @@ int bdw_enable_vblank(struct drm_crtc *_crtc)
>> > if (gen11_dsi_configure_te(crtc, true))
>> > return 0;
>> >
>> > - if (display->irq.vblank_enabled++ == 0 && crtc-
>> > >block_dc_for_vblank)
>> > + if (display->irq.vblank_wa_num_pipes++ == 0 && crtc-
>> > >block_dc_for_vblank)
>> > schedule_work(&display->irq.vblank_dc_work);
>> >
>> > spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
>> > @@ -1478,7 +1478,7 @@ void bdw_disable_vblank(struct drm_crtc
>> > *_crtc)
>> > bdw_disable_pipe_irq(dev_priv, pipe, GEN8_PIPE_VBLANK);
>> > spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
>> >
>> > - if (--display->irq.vblank_enabled == 0 && crtc-
>> > >block_dc_for_vblank)
>> > + if (--display->irq.vblank_wa_num_pipes == 0 && crtc-
>> > >block_dc_for_vblank)
>> > schedule_work(&display->irq.vblank_dc_work);
>> > }
>>
>
--
Jani Nikula, Intel
