- Jun 27, 2022
-
-
Ville Syrjälä authored
Eliminate the PIPECONF RMWs from .comit_commit() so that we can finally declare the whole vblank evade part (and the noarm() part) of the pipe commit free of register reads. Or at least I hope that's the last read... Only the i9xx/ilk codepaths need this for now, but let's add the same thing for hsw+ just in case we want to start calling that during fastsets at some point (eg. to change dithering settings/etc.). Should open up the way to start experimenting with different DSB usage approaches for pipe commits. Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220413192607.27533-1-ville.syrjala@linux.intel.com Reviewed-by:
Uma Shankar <uma.shankar@intel.com>
-
Anshuman Gupta authored
Currently i915 disables d3cold for i915 pci dev. This blocks D3 for i915 gfx pci upstream bridge (VSP). Let's disable d3cold at gfx root port to make sure that i915 gfx VSP can transition to D3 to save some power. We don't need to disable/enable d3cold in rpm, s2idle suspend/resume handlers. Disabling/Enabling d3cold at gfx root port in probe/remove phase is sufficient. Fixes: 1a085e23 ("drm/i915: Disable D3Cold in s2idle and runtime pm") Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by:
Anshuman Gupta <anshuman.gupta@intel.com> Reviewed-by:
Badal Nilawar <badal.nilawar@intel.com> Acked-by:
: Rodrigo Vivi <rodrigo.vivi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220616122249.5007-1-anshuman.gupta@intel.com
-
- Jun 24, 2022
-
-
José Roberto de Souza authored
When the last reference of a gem object is removed it is added to the mm.free_list list and mm.free_work is queued to actually free the object. So gem objects that had their last reference removed by display during drm_atomic_helper_shutdown() are added to mm.free_list what could cause that mm.free_work is executed at the same time as intel_runtime_pm_driver_release() causing raw-wakerefs warning. So here only calling i915_gem_suspend() and by consequence i915_gem_drain_freed_objects() only after display is down making sure all display gem objecs are freed when intel_runtime_pm_driver_release() is executed. Signed-off-by:
José Roberto de Souza <jose.souza@intel.com> Reviewed-by:
Matt Roper <matthew.d.roper@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220617190629.355356-1-jose.souza@intel.com
-
- Jun 21, 2022
-
-
Animesh Manna authored
Each LFP may have different panel type which is stored in LFP data data block. Based on the child device index respective panel-type/ panel-type2 field will be used. v1: Initial rfc verion. v2: Based on review comments from Jani, - Used panel-type instead addition panel-index variable. - DEVICE_HANDLE_* name changed and placed before DEVICE_TYPE_* macro. v3: - passing intel_bios_encoder_data as argument of intel_bios_init_panel(). Passing NULL to indicate encoder is not initialized yet for dsi as current focus is to enable dual EDP. [Jani] v4: - encoder->devdata used which is initialized before from vbt structure. [Jani] Signed-off-by:
Animesh Manna <animesh.manna@intel.com> Reviewed-by:
Jani Nikula <jani.nikula@intel.com> Acked-by:
-
Jani Nikula authored
Pass struct drm_i915_private * instead of struct drm_device *, and rename dev_priv to i915. v2: Rebase Signed-off-by:
-
Jani Nikula authored
Split out the modeset hardware state readout and sanitize, or state setup, to a separate file. Do some drive-by checkpatch fixes while at it. v2: Rebase Signed-off-by:
-
Jason A. Donenfeld authored
This reverts commit 73867c87 ("drm/i915/display: Remove check for low voltage sku for max dp source rate"), which, on an i7-11850H iGPU with a Thinkpad X1 Extreme Gen 4, attached to a LG LP160UQ1-SPB1 embedded panel, causes wild flickering glitching technicolor pyrotechnics on resumption from suspend. The display shows strobing colors in an utter disaster explosion of pantone, as though bombs were dropped on the leprechauns at the base of the rainbow. Rebooting the machine fixes the issue, presumably because the display is initialized by firmware rather than by i915. Otherwise, the GPU appears to work fine. Bisection traced it back to this commit, which makes sense given the issues. Note: This re-opens, and puts back to the drawing board, https://gitlab.freedesktop.org/drm/intel/-/issues/5272 which was fixed by the regressing commit. Fixes: 73867c87 ("drm/i915/display: Remove check for low voltage sku for max dp source rate") Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6205 Cc: Ankit Nautiyal <ankit.k.nautiyal@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Uma Shankar <uma.shankar@intel.com> Cc: Animesh Manna <animesh.manna@intel.com> Cc: Jani Saarinen <jani.saarinen@intel.com> Signed-off-by:
Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by:
-
- Jun 17, 2022
-
-
Jani Nikula authored
Prefer struct drm_i915_private *i915 over struct drm_device or dev_priv. Signed-off-by:
-
Jani Nikula authored
Rename dev_priv to i915 in crtc state dumping code. Signed-off-by:
-
Jani Nikula authored
Add the brackets [] around crtc state dump context string in intel_crtc_state_dump() so the callers don't have to. Signed-off-by:
-
Jani Nikula authored
Declutter intel_display.c by splitting out crtc state dumping to a separate file. v2: intel_pipe_config_dump -> intel_crtc_state_dump Signed-off-by:
-
Jani Nikula authored
Add new file intel_modeset_verify.c for high level modeset verification code to declutter intel_display.h. The new file is supposed to be about crtc/encoder/connector verification; the state verification for very specific functionality such as plls or wm should be placed next to the code it verifies. Fix some minor checkpatch issues while at it. v2: Rebase Signed-off-by:
-
Jani Nikula authored
Keep the mpllb implementation details together in intel_snps_phy.c. Also declutter intel_display.c. v2: intel_mpllb_verify_state -> void intel_mpllb_state_verify (Ville) Signed-off-by:
-
Jani Nikula authored
The pipe_config_mismatch() function is primarily for logging comparison results. Signed-off-by:
-
Jani Nikula authored
Keep the shared dpll implementation details together by moving the dpll state verification to intel_dpll_mgr.c. Also declutter intel_display.c. v2: intel_shared_dpll_verify_state -> intel_shared_dpll_state_verify (Ville) Signed-off-by:
-
Jani Nikula authored
By moving wm state verification to intel_pm.c, we can make a bunch of functions static, hiding the wm details better. Also declutter intel_display.c. v2: intel_wm_state_verify -> intel_wm_verify_state (Ville) Signed-off-by:
-
Ville Syrjälä authored
adl-s needs the combo PLL DCO fraction w/a as well. Gets us slightly more accurate clock out of the PLL. Cc: stable@vger.kernel.org Signed-off-by:
-
- Jun 16, 2022
-
-
Ville Syrjälä authored
Abstract the bit extraction from the VBT per-panel bitfields slightly. Signed-off-by:
-
Ville Syrjälä authored
Older VBTs don't have all the stuff we've defined for the LVDS options block (40). In particular we're currently parsing the DPS panel type bits even though they may not exist, which could mean we end up flagging the machine as supporting static DRRS when the VBT declared no such thing. We don't actually have a clear idea which VBT versions have which bits so we rely on the block size instead. Here's a quick list from my VBT stash: mgm version 108 -> 4 bytes alv version 120 -> 4 bytes cst version 134 -> 14 bytes pnv version 144 -> 14 bytes cl version 142 -> 16 bytes ctg version 155 -> 24 bytes Signed-off-by:
-
Ville Syrjälä authored
Parsing the panel_type is a bit special and should be done before we parse anything else potentially panel-specific from the VBT. So move it out from parse_panel_options(). It doesn't neet to be there anyway since it'll do its own LVDS options block lookup. Signed-off-by:
-
Ville Syrjälä authored
The WARN shouldn't have been added yet. For the moment the clock that gets passed here is just what the user has requested (via the modeline) and may not be exactly what iCLKIP can generate. Later on the plan is to change things so that we already get passed the exact clock here, at which point the WARN should be reintroduced. Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6246 Fixes: 97708335 ("drm/i915: Introduce struct iclkip_params") Signed-off-by:
-
Ville Syrjälä authored
Make sure FIFO underrun reporting is flagged as disabled very early during the state readout so that we don't get any spurious FIFO underruns reports from intel_crtc_disable_noatomic(). Signed-off-by:
-
Ville Syrjälä authored
Pull the underrun status sanitation into its own helper. Signed-off-by:
-
Jouni Högander authored
Connector polling is waking up the polled device. Polling is unnecessary if our device is known to not have display. Fix this and save some power by disabling starting connector polling when we are having headless sku. Use information from opregion. v2: Move headless sku check into INTEL_DISPLAY_ENABLED macro Signed-off-by:
Jouni Högander <jouni.hogander@intel.com> Reviewed-by:
-
Jouni Högander authored
Currently we are starting connector polling if display is disabled using disable_display module parameter. Polling is just returning always "not connected" state. This can be optimized by not starting polling at all. Signed-off-by:
-
Jouni Högander authored
Export headless sku bit (bit 13) from opregion->header->pcon as an interface to check if our device is headless configuration. This is mainly targeted for hybrid gfx systems. E.g. when display is not supposed to be connected discrete graphics card it's opregion can inform this is headless graphics card. v3: Dummy version is now static inline function v2: Check also opregion version Bspec: 53441 Reviewed-by:
-
- Jun 15, 2022
-
-
Jani Nikula authored
Split ddi port parsing and debug printing to clarify the functional parts of parse_ddi_port(), which are quite small nowadays. Signed-off-by:
-
Jani Nikula authored
i915 is available via devdata, grab it there instead of passing. Signed-off-by:
-
Jani Nikula authored
Improve clarity by using the helpers we have. Signed-off-by:
-
- Jun 14, 2022
-
-
Ville Syrjälä authored
The VBIOS/GOP may not program the FDI M/n vs. dotclock entirely consistently. Eg. on a SNB Thinkpad X220 LVDS I see dotclock of 69.286 MHz (the best the DPLL can do) vs. FDI M/N 69.3 MHz (matches what the EDID actually declares). Signed-off-by:
-
Ville Syrjälä authored
Pull the various iCLKIP parameters into a struct. Later on we'll reuse this during the state computation to determine the exact dotclock the hardware will be generating for us. v2: Don't lose the phaseinc calculation v3: Drop the misplaced '#include <intel_pch_refclk.h>' from intel_crt.c (Jani) Signed-off-by:
-
Ville Syrjälä authored
Extract intel_crtc_dotclock() from ddi_dotclock_get(). We'll reuse this during state computation in order to determine the actual final dotclcok after the DPLL computation has been done (which may not give us the exact same port_clock that we fed in). v2: Add the prototype Signed-off-by:
-
- Jun 10, 2022
-
-
José Roberto de Souza authored
Commit 3cf05076 ("drm/i915/bios: Split VBT data into per-panel vs. global parts") cause PSR to be disabled when enable_psr has the default value and there is at least one DP port that do not supports PSR. That was happening because intel_psr_init() is called for every DP port and then enable_psr is globaly set to 0 based on the PSR support of the DP port. Here dropping the enable_psr overwritten and using the VBT PSR value when enable_psr is set as default. Fixes: 3cf05076 ("drm/i915/bios: Split VBT data into per-panel vs. global parts") Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Jouni Högander <jouni.hogander@intel.com> Cc: Mika Kahola <mika.kahola@intel.com> Signed-off-by:
-
- Jun 08, 2022
-
-
Ville Syrjälä authored
The eDP BDB block has gained yet another max link rate field. Let's parse it and consult it during the source rate filtering. v2: *20 instead of *2 to get the correct units (Jani) Reviewed-by:
-
Ville Syrjälä authored
We're not parsing the 5.4 Gbps value for the old eDP fast link training link rate, nor are we parsing the new fast link training link rate field. Remedy both. Also we'll now use the actual link rate instead of the DPCD BW register value. Note that we're not even using this information for anything currently, so should perhaps just nuke it all unless someone is planning on implementing fast link training finally... v2: Stop using the DPCD BW values (Jani) *20 instead of *2 to get the rate in correct units (Jani) Cc: Jani Nikula <jani.nikula@intel.com> Signed-off-by: -
Ville Syrjälä authored
We'll need to know the VBT panel_type before we can determine the maximum link rate for eDP. To that end move intel_dp_set_source_rates() & co. to be called after the per-panel VBT parsing has been done. intel_dp_mst_encoder_init() depends on the source rates so we'll have to do it a bit later as well. v2: Fix the intel_dp_mst_encoder_init() oops Reviewed-by: Jani Nikula <jani.nikula@intel.com> #v1 Signed-off-by:
-
Jani Nikula authored
There's an early return for !engine->kernel_context. Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220516081015.1058987-1-jani.nikula@intel.com
-
Jani Nikula authored
Fix: drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.c:61:6: warning: symbol 'intel_pxp_debugfs_register' was not declared. Should it be static? Sort and remove the redundant pxp prefixes from the includes while at it. Signed-off-by:
-
Luca Coelho authored
The intel_dp_dsc_get_output_bpp() function outputs two lines of unconditional logs, which was okay when it was called only once. But now, we also call this function from intel_dp_mode_valid(), which is in turn called for every mode we need to validate. This causes a lot of useless noise. Remove the unconditional prints to avoid spamming the logs. Also remove one more print that is not unconditional, but is related. Signed-off-by:
Luca Coelho <luciano.coelho@intel.com> Signed-off-by:
-
Jani Nikula authored
Remove some of the unused helpers from i915_utils.h. Signed-off-by:
-
