| Message ID | 20250319191508.2751216-5-animesh.manna@intel.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | LOBF enablement fix | expand |
On Thu, 2025-03-20 at 00:45 +0530, Animesh Manna wrote: > For every commit the dependent condition for LOBF is checked > and accordingly update has_lobf flag which will be used > to update the ALPM_CTL register during commit. > > v1: Initial version. > v2: Avoid reading h/w register without has_lobf check. [Jani] > v3: Update LOBF in post plane update instead of separate function. > [Jouni] > > Signed-off-by: Animesh Manna <animesh.manna@intel.com> > --- > drivers/gpu/drm/i915/display/intel_alpm.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c > b/drivers/gpu/drm/i915/display/intel_alpm.c > index c2862888466f..5df1253a6b6c 100644 > --- a/drivers/gpu/drm/i915/display/intel_alpm.c > +++ b/drivers/gpu/drm/i915/display/intel_alpm.c > @@ -378,9 +378,12 @@ void intel_alpm_post_plane_update(struct > intel_atomic_state *state, > struct intel_display *display = to_intel_display(state); > const struct intel_crtc_state *crtc_state = > intel_atomic_get_new_crtc_state(state, crtc); > + const struct intel_crtc_state *old_crtc_state = > + intel_atomic_get_old_crtc_state(state, crtc); > struct intel_encoder *encoder; > > - if (!crtc_state->has_lobf && !crtc_state->has_psr) > + if (!crtc_state->has_lobf && !crtc_state->has_psr && > + !old_crtc_state->has_lobf) I don't really understand this change? Where lobf is disabled when has_lobf is false? BR, Jouni Högander > return; > > for_each_intel_encoder_mask(display->drm, encoder,
> -----Original Message----- > From: Hogander, Jouni <jouni.hogander@intel.com> > Sent: Monday, March 24, 2025 1:36 PM > To: intel-xe@lists.freedesktop.org; Manna, Animesh > <animesh.manna@intel.com>; intel-gfx@lists.freedesktop.org > Cc: Nikula, Jani <jani.nikula@intel.com>; B, Jeevan <jeevan.b@intel.com> > Subject: Re: [PATCH v6 4/8] drm/i915/lobf: Update lobf if any change in > dependent parameters > > On Thu, 2025-03-20 at 00:45 +0530, Animesh Manna wrote: > > For every commit the dependent condition for LOBF is checked and > > accordingly update has_lobf flag which will be used to update the > > ALPM_CTL register during commit. > > > > v1: Initial version. > > v2: Avoid reading h/w register without has_lobf check. [Jani] > > v3: Update LOBF in post plane update instead of separate function. > > [Jouni] > > > > Signed-off-by: Animesh Manna <animesh.manna@intel.com> > > --- > > drivers/gpu/drm/i915/display/intel_alpm.c | 5 ++++- > > 1 file changed, 4 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c > > b/drivers/gpu/drm/i915/display/intel_alpm.c > > index c2862888466f..5df1253a6b6c 100644 > > --- a/drivers/gpu/drm/i915/display/intel_alpm.c > > +++ b/drivers/gpu/drm/i915/display/intel_alpm.c > > @@ -378,9 +378,12 @@ void intel_alpm_post_plane_update(struct > > intel_atomic_state *state, > > struct intel_display *display = to_intel_display(state); > > const struct intel_crtc_state *crtc_state = > > intel_atomic_get_new_crtc_state(state, crtc); > > + const struct intel_crtc_state *old_crtc_state = > > + intel_atomic_get_old_crtc_state(state, crtc); > > struct intel_encoder *encoder; > > > > - if (!crtc_state->has_lobf && !crtc_state->has_psr) > > + if (!crtc_state->has_lobf && !crtc_state->has_psr && > > + !old_crtc_state->has_lobf) > > I don't really understand this change? Where lobf is disabled when has_lobf > is false? When crtc_state->has_lobf = false and old_crtc_state->has_lobf = true, then (!crtc_state->has_lobf && !crtc_state->has_psr && !old_crtc_state->has_lobf) = false, so return will not happen. Next lnl_alpm_configure() will program other aux-wake/aux-less alpm related bits except lobf-enable. Do you see any issue here? Regards, Animesh > > BR, > > Jouni Högander > > return; > > > > for_each_intel_encoder_mask(display->drm, encoder,
On Thu, 2025-03-27 at 10:43 +0000, Manna, Animesh wrote: > > > > -----Original Message----- > > From: Hogander, Jouni <jouni.hogander@intel.com> > > Sent: Monday, March 24, 2025 1:36 PM > > To: intel-xe@lists.freedesktop.org; Manna, Animesh > > <animesh.manna@intel.com>; intel-gfx@lists.freedesktop.org > > Cc: Nikula, Jani <jani.nikula@intel.com>; B, Jeevan > > <jeevan.b@intel.com> > > Subject: Re: [PATCH v6 4/8] drm/i915/lobf: Update lobf if any > > change in > > dependent parameters > > > > On Thu, 2025-03-20 at 00:45 +0530, Animesh Manna wrote: > > > For every commit the dependent condition for LOBF is checked and > > > accordingly update has_lobf flag which will be used to update the > > > ALPM_CTL register during commit. > > > > > > v1: Initial version. > > > v2: Avoid reading h/w register without has_lobf check. [Jani] > > > v3: Update LOBF in post plane update instead of separate > > > function. > > > [Jouni] > > > > > > Signed-off-by: Animesh Manna <animesh.manna@intel.com> > > > --- > > > drivers/gpu/drm/i915/display/intel_alpm.c | 5 ++++- > > > 1 file changed, 4 insertions(+), 1 deletion(-) > > > > > > diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c > > > b/drivers/gpu/drm/i915/display/intel_alpm.c > > > index c2862888466f..5df1253a6b6c 100644 > > > --- a/drivers/gpu/drm/i915/display/intel_alpm.c > > > +++ b/drivers/gpu/drm/i915/display/intel_alpm.c > > > @@ -378,9 +378,12 @@ void intel_alpm_post_plane_update(struct > > > intel_atomic_state *state, > > > struct intel_display *display = to_intel_display(state); > > > const struct intel_crtc_state *crtc_state = > > > intel_atomic_get_new_crtc_state(state, crtc); > > > + const struct intel_crtc_state *old_crtc_state = > > > + intel_atomic_get_old_crtc_state(state, crtc); > > > struct intel_encoder *encoder; > > > > > > - if (!crtc_state->has_lobf && !crtc_state->has_psr) > > > + if (!crtc_state->has_lobf && !crtc_state->has_psr && > > > + !old_crtc_state->has_lobf) > > > > I don't really understand this change? Where lobf is disabled when > > has_lobf > > is false? > > When crtc_state->has_lobf = false and old_crtc_state->has_lobf = > true, > then (!crtc_state->has_lobf && !crtc_state->has_psr && > !old_crtc_state->has_lobf) = false, so return will not happen. Ok, I see. How about crtc_state->has_lobf = true and old_crtc_state- >has_lobf = true. Do you still want to configure it? How about crtc_state->has_lobf != old_crtc_state->has_lobf? > Next lnl_alpm_configure() will program other aux-wake/aux-less alpm > related bits except lobf-enable. Do you see any issue here? I think it should be. Thank you for the clarification. BR, Jouni Högander > > Regards, > Animesh > > > > BR, > > > > Jouni Högander > > > return; > > > > > > for_each_intel_encoder_mask(display->drm, encoder, >
diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c b/drivers/gpu/drm/i915/display/intel_alpm.c index c2862888466f..5df1253a6b6c 100644 --- a/drivers/gpu/drm/i915/display/intel_alpm.c +++ b/drivers/gpu/drm/i915/display/intel_alpm.c @@ -378,9 +378,12 @@ void intel_alpm_post_plane_update(struct intel_atomic_state *state, struct intel_display *display = to_intel_display(state); const struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); + const struct intel_crtc_state *old_crtc_state = + intel_atomic_get_old_crtc_state(state, crtc); struct intel_encoder *encoder; - if (!crtc_state->has_lobf && !crtc_state->has_psr) + if (!crtc_state->has_lobf && !crtc_state->has_psr && + !old_crtc_state->has_lobf) return; for_each_intel_encoder_mask(display->drm, encoder,
For every commit the dependent condition for LOBF is checked and accordingly update has_lobf flag which will be used to update the ALPM_CTL register during commit. v1: Initial version. v2: Avoid reading h/w register without has_lobf check. [Jani] v3: Update LOBF in post plane update instead of separate function. [Jouni] Signed-off-by: Animesh Manna <animesh.manna@intel.com> --- drivers/gpu/drm/i915/display/intel_alpm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)