Posts by smp
-
-
wondering if we could disable this option at the kodi gui as it just creates problems ?
IMO this should be done upstream in Kodi, e.g. if platform supports VAAPI then DRM PRIME option should not be present in GUI.
-
What are the chances that the 11gen intel chips get fixed?
As you can see on this block diagram, NUC11PAxxx does not have a native HDMI port. It is using a DP->HDMI protocol converter (PCON) chip. It is more than likely the source of your audio passthrough issues. So this is not a gen11 issue but an issue specific to that NUC model. There is nothing to fix from LE/Kodi side.
-
No. DRM PRIME was never properly implemented for x86.
-
I simply downloaded the needed files and created the patch files with git diff.
e.g. for av201x .c:
git diff emptyfile av201x.c > 1.patch
then fixed the path/file name in .patch file.
There are probably better ways to do this (e.g. git cherry-pick) but it works for me.
-
Disable debug logging?
-
but sadly this is no longer available on dropbox
PostRE: [HOW-TO] TBS 5520SE Drivers for LE11, LE12, LE13 on RPi4
For those who only use the satellite tuner I modded the driver to disable the DVB-T/C tuner.
TBS 5520SE has 2 tuners attached to a single demod. DVB core apparently does not support 2 tuners on 1 demod so the driver used a hacky way to split the adapter in 2 virtual frontends. This cause all sorts of issues with tvheadend (e.g. "no free adapter available"). I modded the driver so si2157 DVB-T tuner is not attached and only 1 DVB-S frontend is created.
forum.libreelec.tv/core/attachment/10336/smpDecember 17, 2023 at 4:05 AM Looks like PCI-E cards are handled by tbsecp3 driver. You will also need to add av201x tuner driver and gx1133 demod driver.
-
-
Diff
Display Morediff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c index bc97591..3da321f 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.c +++ b/drivers/gpu/drm/i915/display/intel_hdmi.c @@ -1895,7 +1895,7 @@ static bool intel_hdmi_source_bpc_possible(struct drm_i915_private *i915, int bp { switch (bpc) { case 12: - return !HAS_GMCH(i915); + return !HAS_GMCH(i915) && DISPLAY_VER(i915) <= 10; case 10: return DISPLAY_VER(i915) >= 11; case 8:
^ This can be used to force 10-bit instead of proptest. This will disable 12-bit for newer Intel hw that can support 10-bit. 10-bit is only supported since Ice Lake.
-
What impact has it to only use 10 bit instead of 12?
None whatsoever.
Has anybody any pointers where i could open a bug for that?
Issues · drm / intel · GitLabIntel Graphics for Linux - drm/i915 kernel driver repository. For issues only for now.gitlab.freedesktop.org -
how do i force 8 Bit or 10 Bit
For 8-bit:
How do i edit the edid and change "SCDC Present"?
I used this. It's in CEA Extension 1 -> HDMI-Forum VSDB.
-
Try to force YCbCr 4:4:4 instead of RGB:
Diff
Display Morediff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c index 7816b2a..758f3dc 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.c +++ b/drivers/gpu/drm/i915/display/intel_hdmi.c @@ -2196,7 +2196,7 @@ intel_hdmi_output_format(struct intel_connector *connector, if (connector->base.ycbcr_420_allowed && ycbcr_420_output) return INTEL_OUTPUT_FORMAT_YCBCR420; else - return INTEL_OUTPUT_FORMAT_RGB; + return INTEL_OUTPUT_FORMAT_YCBCR444; } static int intel_hdmi_compute_output_format(struct intel_encoder *encoder, @@ -2215,8 +2215,8 @@ static int intel_hdmi_compute_output_format(struct intel_encoder *encoder, if (ycbcr_420_only && crtc_state->sink_format != INTEL_OUTPUT_FORMAT_YCBCR420) { drm_dbg_kms(&i915->drm, - "YCbCr 4:2:0 mode but YCbCr 4:2:0 output not possible. Falling back to RGB.\n"); - crtc_state->sink_format = INTEL_OUTPUT_FORMAT_RGB; + "YCbCr 4:2:0 mode but YCbCr 4:2:0 output not possible. Falling back to YCBCR444.\n"); + crtc_state->sink_format = INTEL_OUTPUT_FORMAT_YCBCR444; } crtc_state->output_format = intel_hdmi_output_format(crtc_state); @@ -2247,7 +2247,7 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder, adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) return -EINVAL; - pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB; + pipe_config->output_format = INTEL_OUTPUT_FORMAT_YCBCR444; if (pipe_config->has_hdmi_sink) pipe_config->has_infoframe = true;
Other things to try: force 10-bit and 8-bit 2160p @ 23,976 Hz mode instead of default 12-bit (with libdrm proptest tool).
Also worth trying a custom EDID with "SCDC Present" bit set to "disabled".
IMO those dropouts are data errors related to Intel HDMI 2.0 implementation.
-
Settings - System - Power Saving - Shutdown Function - Suspend
This is supposed to be used with Shutdown function timer. It works as intended.
-
-
-
-
-
smp do you know how to build a version with pipewire?
Edit distributions/LibreELEC/options
PIPEWIRE_SUPPORT="yes"
KODI_PIPEWIRE_SUPPORT="yes"
ALSA_SUPPORT="no"
KODI_ALSA_SUPPORT="no"
PULSEAUDIO_SUPPORT="no"
KODI_PULSEAUDIO_SUPPORT="no"