| Message ID | 1401972913-1771-1-git-send-email-shobhit.kumar@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Jun 05, 2014 at 06:25:13PM +0530, Shobhit Kumar wrote: > The DEVICE_TYPE_eDP has been changed to 0x1806 in case of BYT which > can causes wrong detection failures for eDP. Reduce the number of bits > of interest in DEVICE_TYPE_eDP_BITS to just check most relevant and > conclusive ones and ensure they are set in device_class from VBT. This > will ensure that eDP detection works across platforms The current bits will match correctly for 0x1806. I don't have a problem with the patch per-say but would be nice to know what the troublesome device type really was, if indeed there was one. I'm just slightly worried that we might start to get false positives on some platforms if we make it too relaxed. > > Signed-off-by: Shobhit Kumar <shobhit.kumar@intel.com> > --- > drivers/gpu/drm/i915/intel_bios.h | 10 +--------- > drivers/gpu/drm/i915/intel_dp.c | 4 ++-- > 2 files changed, 3 insertions(+), 11 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h > index b986677..30d02b7 100644 > --- a/drivers/gpu/drm/i915/intel_bios.h > +++ b/drivers/gpu/drm/i915/intel_bios.h > @@ -710,16 +710,8 @@ int intel_parse_bios(struct drm_device *dev); > */ > #define DEVICE_TYPE_eDP_BITS \ > (DEVICE_TYPE_INTERNAL_CONNECTOR | \ > - DEVICE_TYPE_NOT_HDMI_OUTPUT | \ > - DEVICE_TYPE_MIPI_OUTPUT | \ > - DEVICE_TYPE_COMPOSITE_OUTPUT | \ > - DEVICE_TYPE_DUAL_CHANNEL | \ > - DEVICE_TYPE_LVDS_SINGALING | \ > - DEVICE_TYPE_TMDS_DVI_SIGNALING | \ > - DEVICE_TYPE_VIDEO_SIGNALING | \ > DEVICE_TYPE_DISPLAYPORT_OUTPUT | \ > - DEVICE_TYPE_DIGITAL_OUTPUT | \ > - DEVICE_TYPE_ANALOG_OUTPUT) > + DEVICE_TYPE_DIGITAL_OUTPUT) > > /* define the DVO port for HDMI output type */ > #define DVO_B 1 > diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c > index dad3780..68f7380 100644 > --- a/drivers/gpu/drm/i915/intel_dp.c > +++ b/drivers/gpu/drm/i915/intel_dp.c > @@ -3842,8 +3842,8 @@ bool intel_dp_is_edp(struct drm_device *dev, enum port port) > p_child = dev_priv->vbt.child_dev + i; > > if (p_child->common.dvo_port == port_mapping[port] && > - (p_child->common.device_type & DEVICE_TYPE_eDP_BITS) == > - (DEVICE_TYPE_eDP & DEVICE_TYPE_eDP_BITS)) > + ((p_child->common.device_type & DEVICE_TYPE_eDP_BITS) == > + DEVICE_TYPE_eDP_BITS)) > return true; > } > return false; > -- > 1.9.1 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
On Thu, 5 Jun 2014 16:09:29 +0300 Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: > On Thu, Jun 05, 2014 at 06:25:13PM +0530, Shobhit Kumar wrote: > > The DEVICE_TYPE_eDP has been changed to 0x1806 in case of BYT which > > can causes wrong detection failures for eDP. Reduce the number of bits > > of interest in DEVICE_TYPE_eDP_BITS to just check most relevant and > > conclusive ones and ensure they are set in device_class from VBT. This > > will ensure that eDP detection works across platforms > > The current bits will match correctly for 0x1806. I don't have a problem > with the patch per-say but would be nice to know what the troublesome > device type really was, if indeed there was one. I'm just slightly > worried that we might start to get false positives on some platforms if > we make it too relaxed. Still waiting for results, but https://bugs.freedesktop.org/show_bug.cgi?id=78795 was what motivated this. Jesse
On 6/5/2014 9:34 PM, Jesse Barnes wrote: > On Thu, 5 Jun 2014 16:09:29 +0300 > Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: > >> On Thu, Jun 05, 2014 at 06:25:13PM +0530, Shobhit Kumar wrote: >>> The DEVICE_TYPE_eDP has been changed to 0x1806 in case of BYT which >>> can causes wrong detection failures for eDP. Reduce the number of bits >>> of interest in DEVICE_TYPE_eDP_BITS to just check most relevant and >>> conclusive ones and ensure they are set in device_class from VBT. This >>> will ensure that eDP detection works across platforms >> >> The current bits will match correctly for 0x1806. I don't have a problem >> with the patch per-say but would be nice to know what the troublesome >> device type really was, if indeed there was one. I'm just slightly >> worried that we might start to get false positives on some platforms if >> we make it too relaxed. > > Still waiting for results, but > https://bugs.freedesktop.org/show_bug.cgi?id=78795 was what motivated this. > Yeah and update from Jesse that forcefully returning true from the function make things work, I assumed this is the reason, without really checking that even the old values will result DEVIVE_TYPE_eDP & DEVICE_TYPE_eDP_BITS = 0x1806 :) Nevertheless I still feel that a simplified bit mask is good to have. We need to check further on this issue Jesse. Let me know once you have results. It would be still better to spit out the device_type being read from VBT as a log to confirm that indeed it is coming as 0x1806 or can just ask to get the opregion dump from debugfs and we can check ourselves. Regards Shobhit
On 6/6/2014 9:33 AM, Kumar, Shobhit wrote: > On 6/5/2014 9:34 PM, Jesse Barnes wrote: >> On Thu, 5 Jun 2014 16:09:29 +0300 >> Ville Syrjälä <ville.syrjala@linux.intel.com> wrote: >> >>> On Thu, Jun 05, 2014 at 06:25:13PM +0530, Shobhit Kumar wrote: >>>> The DEVICE_TYPE_eDP has been changed to 0x1806 in case of BYT which >>>> can causes wrong detection failures for eDP. Reduce the number of bits >>>> of interest in DEVICE_TYPE_eDP_BITS to just check most relevant and >>>> conclusive ones and ensure they are set in device_class from VBT. This >>>> will ensure that eDP detection works across platforms >>> >>> The current bits will match correctly for 0x1806. I don't have a problem >>> with the patch per-say but would be nice to know what the troublesome >>> device type really was, if indeed there was one. I'm just slightly >>> worried that we might start to get false positives on some platforms if >>> we make it too relaxed. >> >> Still waiting for results, but >> https://bugs.freedesktop.org/show_bug.cgi?id=78795 was what motivated >> this. >> > > Yeah and update from Jesse that forcefully returning true from the > function make things work, I assumed this is the reason, without really > checking that even the old values will result DEVIVE_TYPE_eDP & > DEVICE_TYPE_eDP_BITS = 0x1806 :) > > Nevertheless I still feel that a simplified bit mask is good to have. We > need to check further on this issue Jesse. Let me know once you have > results. It would be still better to spit out the device_type being read > from VBT as a log to confirm that indeed it is coming as 0x1806 or can > just ask to get the opregion dump from debugfs and we can check ourselves. I see above are already been asked in BZ. Will track there for updates. Regards Shobhit
diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h index b986677..30d02b7 100644 --- a/drivers/gpu/drm/i915/intel_bios.h +++ b/drivers/gpu/drm/i915/intel_bios.h @@ -710,16 +710,8 @@ int intel_parse_bios(struct drm_device *dev); */ #define DEVICE_TYPE_eDP_BITS \ (DEVICE_TYPE_INTERNAL_CONNECTOR | \ - DEVICE_TYPE_NOT_HDMI_OUTPUT | \ - DEVICE_TYPE_MIPI_OUTPUT | \ - DEVICE_TYPE_COMPOSITE_OUTPUT | \ - DEVICE_TYPE_DUAL_CHANNEL | \ - DEVICE_TYPE_LVDS_SINGALING | \ - DEVICE_TYPE_TMDS_DVI_SIGNALING | \ - DEVICE_TYPE_VIDEO_SIGNALING | \ DEVICE_TYPE_DISPLAYPORT_OUTPUT | \ - DEVICE_TYPE_DIGITAL_OUTPUT | \ - DEVICE_TYPE_ANALOG_OUTPUT) + DEVICE_TYPE_DIGITAL_OUTPUT) /* define the DVO port for HDMI output type */ #define DVO_B 1 diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index dad3780..68f7380 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -3842,8 +3842,8 @@ bool intel_dp_is_edp(struct drm_device *dev, enum port port) p_child = dev_priv->vbt.child_dev + i; if (p_child->common.dvo_port == port_mapping[port] && - (p_child->common.device_type & DEVICE_TYPE_eDP_BITS) == - (DEVICE_TYPE_eDP & DEVICE_TYPE_eDP_BITS)) + ((p_child->common.device_type & DEVICE_TYPE_eDP_BITS) == + DEVICE_TYPE_eDP_BITS)) return true; } return false;
The DEVICE_TYPE_eDP has been changed to 0x1806 in case of BYT which can causes wrong detection failures for eDP. Reduce the number of bits of interest in DEVICE_TYPE_eDP_BITS to just check most relevant and conclusive ones and ensure they are set in device_class from VBT. This will ensure that eDP detection works across platforms Signed-off-by: Shobhit Kumar <shobhit.kumar@intel.com> --- drivers/gpu/drm/i915/intel_bios.h | 10 +--------- drivers/gpu/drm/i915/intel_dp.c | 4 ++-- 2 files changed, 3 insertions(+), 11 deletions(-)