| Message ID | 20161201153957.13390-1-hdegoede@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Dec 01, 2016 at 04:39:57PM +0100, Hans de Goede wrote: > On my Cherrytrail CUBE iwork8 Air tablet PIPE-A would get stuck on loading > i915 at boot 1 out of every 3 boots, resulting in a non functional LCD. > Once the i915 driver has successfully loaded, the panel can be disabled / > enabled without hitting this issue. > > The getting stuck is caused by vlv_init_display_clock_gating() clearing > the DPOUNIT_CLOCK_GATE_DISABLE bit in DSPCLK_GATE_D when called from > chv_pipe_power_well_ops.enable() on driver load, while PIPE-A is enabled > driving the DSI LCD by the BIOS. > > Clearing this bit while DSI is in use is a known issue and > intel_dsi_pre_enable() / intel_dsi_post_disable() already set / clear it > as appropriate. > > This commit modifies vlv_init_display_clock_gating() to leave the > DPOUNIT_CLOCK_GATE_DISABLE bit alone fixing PIPE-A getting stuck. > > BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=97330 > Cc: stable@vger.kernel.org > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > --- > drivers/gpu/drm/i915/intel_runtime_pm.c | 13 ++++++++++++- > 1 file changed, 12 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c > index 356c662..b5ce7cb 100644 > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c > @@ -1039,7 +1039,18 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, > > static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv) > { > - I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE); > + u32 val; > + > + /* > + * When on driver load, PIPE A may be active and driving a DSI display. > + * Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid PIPE A getting stuck > + * (and never recovering) in this case. intel_dsi_post_disable() will > + * clear it when we turn off the display. Why are you talking only about pipe A here? > + */ > + val = I915_READ(DSPCLK_GATE_D); > + val &= DPOUNIT_CLOCK_GATE_DISABLE; > + val |= VRHUNIT_CLOCK_GATE_DISABLE; > + I915_WRITE(DSPCLK_GATE_D, val); > > /* > * Disable trickle feed and enable pnd deadline calculation > -- > 2.9.3 > > _______________________________________________ > dri-devel mailing list > dri-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/dri-devel
Hi, On 02-12-16 13:40, Ville Syrjälä wrote: > On Thu, Dec 01, 2016 at 04:39:57PM +0100, Hans de Goede wrote: >> On my Cherrytrail CUBE iwork8 Air tablet PIPE-A would get stuck on loading >> i915 at boot 1 out of every 3 boots, resulting in a non functional LCD. >> Once the i915 driver has successfully loaded, the panel can be disabled / >> enabled without hitting this issue. >> >> The getting stuck is caused by vlv_init_display_clock_gating() clearing >> the DPOUNIT_CLOCK_GATE_DISABLE bit in DSPCLK_GATE_D when called from >> chv_pipe_power_well_ops.enable() on driver load, while PIPE-A is enabled >> driving the DSI LCD by the BIOS. >> >> Clearing this bit while DSI is in use is a known issue and >> intel_dsi_pre_enable() / intel_dsi_post_disable() already set / clear it >> as appropriate. >> >> This commit modifies vlv_init_display_clock_gating() to leave the >> DPOUNIT_CLOCK_GATE_DISABLE bit alone fixing PIPE-A getting stuck. >> >> BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=97330 >> Cc: stable@vger.kernel.org >> Signed-off-by: Hans de Goede <hdegoede@redhat.com> >> --- >> drivers/gpu/drm/i915/intel_runtime_pm.c | 13 ++++++++++++- >> 1 file changed, 12 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c >> index 356c662..b5ce7cb 100644 >> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c >> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c >> @@ -1039,7 +1039,18 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, >> >> static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv) >> { >> - I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE); >> + u32 val; >> + >> + /* >> + * When on driver load, PIPE A may be active and driving a DSI display. >> + * Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid PIPE A getting stuck >> + * (and never recovering) in this case. intel_dsi_post_disable() will >> + * clear it when we turn off the display. > > Why are you talking only about pipe A here? Because that is the pipe which was getting stuck on my tablet. I see that the comment for the same workaround in intel_dsi.c just says: " Disable DPOunit clock gating, can stall pipe" so I guess I should update the comment s/PIPE A/pipe(s)/. Any other remarks before I send a v2 ? Regards, Hans > >> + */ >> + val = I915_READ(DSPCLK_GATE_D); >> + val &= DPOUNIT_CLOCK_GATE_DISABLE; >> + val |= VRHUNIT_CLOCK_GATE_DISABLE; >> + I915_WRITE(DSPCLK_GATE_D, val); >> >> /* >> * Disable trickle feed and enable pnd deadline calculation >> -- >> 2.9.3 >> >> _______________________________________________ >> dri-devel mailing list >> dri-devel@lists.freedesktop.org >> https://lists.freedesktop.org/mailman/listinfo/dri-devel >
On Fri, Dec 02, 2016 at 02:27:43PM +0100, Hans de Goede wrote: > Hi, > > On 02-12-16 13:40, Ville Syrjälä wrote: > > On Thu, Dec 01, 2016 at 04:39:57PM +0100, Hans de Goede wrote: > >> On my Cherrytrail CUBE iwork8 Air tablet PIPE-A would get stuck on loading > >> i915 at boot 1 out of every 3 boots, resulting in a non functional LCD. > >> Once the i915 driver has successfully loaded, the panel can be disabled / > >> enabled without hitting this issue. > >> > >> The getting stuck is caused by vlv_init_display_clock_gating() clearing > >> the DPOUNIT_CLOCK_GATE_DISABLE bit in DSPCLK_GATE_D when called from > >> chv_pipe_power_well_ops.enable() on driver load, while PIPE-A is enabled > >> driving the DSI LCD by the BIOS. > >> > >> Clearing this bit while DSI is in use is a known issue and > >> intel_dsi_pre_enable() / intel_dsi_post_disable() already set / clear it > >> as appropriate. > >> > >> This commit modifies vlv_init_display_clock_gating() to leave the > >> DPOUNIT_CLOCK_GATE_DISABLE bit alone fixing PIPE-A getting stuck. > >> > >> BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=97330 > >> Cc: stable@vger.kernel.org > >> Signed-off-by: Hans de Goede <hdegoede@redhat.com> > >> --- > >> drivers/gpu/drm/i915/intel_runtime_pm.c | 13 ++++++++++++- > >> 1 file changed, 12 insertions(+), 1 deletion(-) > >> > >> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c > >> index 356c662..b5ce7cb 100644 > >> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c > >> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c > >> @@ -1039,7 +1039,18 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, > >> > >> static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv) > >> { > >> - I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE); > >> + u32 val; > >> + > >> + /* > >> + * When on driver load, PIPE A may be active and driving a DSI display. > >> + * Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid PIPE A getting stuck > >> + * (and never recovering) in this case. intel_dsi_post_disable() will > >> + * clear it when we turn off the display. > > > > Why are you talking only about pipe A here? > > Because that is the pipe which was getting stuck on my tablet. I see that > the comment for the same workaround in intel_dsi.c just says: > " Disable DPOunit clock gating, can stall pipe" so I guess I should > update the comment s/PIPE A/pipe(s)/. Any other remarks before I send a v2 ? Nope. With that this is Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > Regards, > > Hans > > > > > > >> + */ > >> + val = I915_READ(DSPCLK_GATE_D); > >> + val &= DPOUNIT_CLOCK_GATE_DISABLE; > >> + val |= VRHUNIT_CLOCK_GATE_DISABLE; > >> + I915_WRITE(DSPCLK_GATE_D, val); > >> > >> /* > >> * Disable trickle feed and enable pnd deadline calculation > >> -- > >> 2.9.3 > >> > >> _______________________________________________ > >> dri-devel mailing list > >> dri-devel@lists.freedesktop.org > >> https://lists.freedesktop.org/mailman/listinfo/dri-devel > >
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 356c662..b5ce7cb 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -1039,7 +1039,18 @@ static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv, static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv) { - I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE); + u32 val; + + /* + * When on driver load, PIPE A may be active and driving a DSI display. + * Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid PIPE A getting stuck + * (and never recovering) in this case. intel_dsi_post_disable() will + * clear it when we turn off the display. + */ + val = I915_READ(DSPCLK_GATE_D); + val &= DPOUNIT_CLOCK_GATE_DISABLE; + val |= VRHUNIT_CLOCK_GATE_DISABLE; + I915_WRITE(DSPCLK_GATE_D, val); /* * Disable trickle feed and enable pnd deadline calculation
On my Cherrytrail CUBE iwork8 Air tablet PIPE-A would get stuck on loading i915 at boot 1 out of every 3 boots, resulting in a non functional LCD. Once the i915 driver has successfully loaded, the panel can be disabled / enabled without hitting this issue. The getting stuck is caused by vlv_init_display_clock_gating() clearing the DPOUNIT_CLOCK_GATE_DISABLE bit in DSPCLK_GATE_D when called from chv_pipe_power_well_ops.enable() on driver load, while PIPE-A is enabled driving the DSI LCD by the BIOS. Clearing this bit while DSI is in use is a known issue and intel_dsi_pre_enable() / intel_dsi_post_disable() already set / clear it as appropriate. This commit modifies vlv_init_display_clock_gating() to leave the DPOUNIT_CLOCK_GATE_DISABLE bit alone fixing PIPE-A getting stuck. BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=97330 Cc: stable@vger.kernel.org Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- drivers/gpu/drm/i915/intel_runtime_pm.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)