- 21 Oct, 2020 7 commits
-
-
Ville Syrjälä authored
A bunch of the ips calculations require 64bit math. In particular 'corr' and 'corr2' look like they can overflow on 32bit systems. Switch to explicit u64 for those. Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20201021131443.25616-3-ville.syrjala@linux.intel.comReviewed-by:
Chris Wilson <chris@chris-wilson.co.uk>
-
Ville Syrjälä authored
There is no GEN6_RPSTAT1 on ILK. Instead of reading that let's try to get the same information from MEMSTAT_ILK. At least it seems to track MEMSWCTL frequency request perfectly on my ILK. It needs the same invert trick as the request value. We don't want to put the invert thing into intel_gpu_freq() and intel_freq_opcode() because that would incorrectly invert the min/max/etc frequencies also. One day someone might want to reverse engineer the formula for converting these numbers to Hz, but for now we'll just report them raw. Signed-off-by:
-
Ville Syrjälä authored
Restore RPS for ILK-M. We lost it when an extra HAS_RPS() check appeared in intel_rps_enable(). Unfortunaltey this just makes the performance worse on my ILK because intel_ips insists on limiting the GPU freq to the minimum. If we don't do the RPS init then intel_ips will not limit the frequency for whatever reason. Either it can't get at some required information and thus makes wrong decisions, or we mess up some weights/etc. and cause it to make the wrong decisions when RPS init has been done, or the entire thing is just wrong. Would require a bunch of reverse engineering to figure out what's going on. Cc: stable@vger.kernel.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Fixes: 9c878557 ("drm/i915/gt: Use the RPM config register to determine clk frequencies") Signed-off-by:
-
Pankaj Bharadiya authored
GEN >= 10 hardware supports the programmable scaler filter. Attach scaling filter property for CRTC and plane for GEN >= 10 hardwares and program scaler filter based on the selected filter type. changes since v3: * None changes since v2: * Use updated functions * Add ps_ctrl var to contain the full PS_CTRL register value (Ville) * Duplicate the scaling filter in crtc and plane hw state (Ville) changes since v1: * None Changes since RFC: * Enable properties for GEN >= 10 platforms (Ville) * Do not round off the crtc co-ordinate (Danial Stone, Ville) * Add new functions to handle scaling filter setup (Ville) * Remove coefficient set 0 hardcoding. Reviewed-by:
Uma Shankar <uma.shankar@intel.com> Signed-off-by:
Shashank Sharma <shashank.sharma@intel.com> Signed-off-by:
Ankit Nautiyal <ankit.k.nautiyal@intel.com> Signed-off-by:
Pankaj Bharadiya <pankaj.laxminarayan.bharadiya@intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20201020161427.6941-5-pankaj.laxminarayan.bharadiya@intel.com
-
Pankaj Bharadiya authored
Integer scaling (IS) is a nearest-neighbor upscaling technique that simply scales up the existing pixels by an integer (i.e., whole number) multiplier.Nearest-neighbor (NN) interpolation works by filling in the missing color values in the upscaled image with that of the coordinate-mapped nearest source pixel value. Both IS and NN preserve the clarity of the original image. Integer scaling is particularly useful for pixel art games that rely on sharp, blocky images to deliver their distinctive look. Introduce functions to configure the scaler filter coefficients to enable nearest-neighbor filtering. Bspec: 49247 changes since v6: * Trust compiler, remove pointless inline keyword from cnl_coef_tap() & cnl_nearest_filter_coef() functions (Ville) changes since v4: * Make cnl_coef_tap(), cnl_nearest_filter_coef() inline (Uma) changes since v3: * None changes since v2: * Move APIs from 5/5 into this patch. * Change filter programming related function names to cnl_*, move filter select bits related code into inline function (Ville) changes since v1: * Rearrange skl_scaler_setup_nearest_neighbor_filter() to iterate the registers directly instead of the phases and taps (Ville) changes since RFC: * Refine the skl_scaler_setup_nearest_neighbor_filter() logic (Ville) Reviewed-by:
-
Pankaj Bharadiya authored
Introduce scaler registers and bit fields needed to configure the scaling filter in prgrammed mode and configure scaling filter coefficients. changes since v3: * None changes since v2: * Change macro names to CNL_* and use +(set)*8 instead of adding another trip through _PICK_EVEN (Ville). changes since v1: * None changes since RFC: * Parametrize scaler coeffient macros by 'set' (Ville) Reviewed-by:
-
Pankaj Bharadiya authored
Introduce per-plane and per-CRTC scaling filter properties to allow userspace to select the driver's default scaling filter or Nearest-neighbor(NN) filter for upscaling operations on CRTC and plane. Drivers can set up this property for a plane by calling drm_plane_create_scaling_filter() and for a CRTC by calling drm_crtc_create_scaling_filter(). NN filter works by filling in the missing color values in the upscaled image with that of the coordinate-mapped nearest source pixel value. NN filter for integer multiple scaling can be particularly useful for for pixel art games that rely on sharp, blocky images to deliver their distinctive look. changes since: v6: * Move property doc to existing "Standard CRTC Properties" and "Plane Composition Properties" doc comments (Simon) changes since v3: * Refactor code, add new function for common code (Ville) changes since v2: * Create per-plane and per-CRTC scaling filter property (Ville) changes since v1: * None changes since RFC: * Add separate properties for plane and CRTC (Ville) Link: https://github.com/xbmc/xbmc/pull/18194 Link: https://github.com/xbmc/xbmc/pull/18567Reviewed-by:
Simon Ser <contact@emersion.fr> Acked-by:
Daniel Vetter <daniel@ffwll.ch> Signed-off-by:
-
- 20 Oct, 2020 23 commits
-
-
Manasi Navare authored
No functional changes in this patch. With Bigjoiner, there are 2 pipes driving 2 halfs of 1 transcoder. The transcoder_mode has the full timings, and is used for configuring the transcoder with the intended mode after joining the 2 halves. To clear the confusion, we rename intel_set_pipe_timings to intel_set_transcoder_timings v2: * Split the renaming into separate patch (Ville) Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Manasi Navare <manasi.d.navare@intel.com> Reviewed-by:
Animesh Manna <animesh.manna@intel.com> Reviewed-by:
-
Ville Syrjälä authored
Sort the ICL PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
Anusha Srivatsa <anusha.srivatsa@intel.com>
-
Ville Syrjälä authored
Sort the CNL PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Sort the CFL PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Sort the CML PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Sort the KBL PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Sort the SKL PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Sort the HSW PCI IDs numerically. Some order seems better than randomness. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Most of the HSW PCI IDs are upper case hex numbers, but a few are lower case. Make it consistent so these don't stick out like a sore thumb. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Ville Syrjälä authored
Bunch of the SKL SKUs currently documented as GT3/4 seem to actually be GT3e/4e. Fix up the comments. Cc: Alexei Podtelezhnikov <apodtele@gmail.com> Signed-off-by:
-
Alexei Podtelezhnikov authored
Add three new devices 0x1913, 0x1915, and 0x1917 also known as iSKLULTGT15, iSKLULXGT15, and iSKLDTGT15. Signed-off-by:
Alexei Podtelezhnikov <apodtele@gmail.com> [vsyrjala: Split separate changes into separate patchs, Sort the IDs] Signed-off-by:
-
Alexei Podtelezhnikov authored
Reclassify 0x1923, 0x1927 according to specifications. Of note, the second to last digit seems to correspond to GT#. Signed-off-by:
-
Alexei Podtelezhnikov authored
Reclassify 0x192A according to specifications. Of note, the second to last digit seems to correspond to GT#. Signed-off-by:
-
Alexei Podtelezhnikov authored
Reclassify 0x0426 as GT3 (GT2+) according to specifications and the second least significant digit. Signed-off-by:
-
Ville Syrjälä authored
Move the DSC stuff out from the middle of the ICP HPD register definitions. The location seems to have been selected by a dice roll. SHPD_FILTER_CNT addition also went astray due to the DSC mess, so we also fix that vs. ICP_TC_HPD_{SHORT,LONG}_DETECT(). Signed-off-by:Lucas De Marchi <lucas.demarchi@intel.com>
-
Ville Syrjälä authored
Add a small wrapper for .hpd_irq_setup() which does the "do we even have the hook?" and "are display interrupts enabled?" checks. Signed-off-by:
Imre Deak <imre.deak@intel.com>
-
Ville Syrjälä authored
Currently we call .hpd_irq_setup() directly just before display resume, and follow it with another call via intel_hpd_init() just afterwards. Assuming the hpd pins are marked as enabled during the open-coded call these two things do exactly the same thing (ie. enable HPD interrupts). Which even makes sense since we definitely need working HPD interrupts for MST sideband during the display resume. So let's nuke the open-coded call and move the intel_hpd_init() call earlier. However we need to leave the poll_init_work stuff behind after the display resume as that will trigger display detection while we're resuming. We don't want that trampling over the display resume process. To make this a bit more symmetric we turn this into a intel_hpd_poll_{enable,disable}() pair. So we end up with the following transformation: intel_hpd_poll_init() -> intel_hpd_poll_enable() lone intel_hpd_init() -> intel_hpd_init()+intel_hpd_poll_disable() .hpd_irq_setup()+resume+intel_hpd_init() -> intel_hpd_init()+resume+intel_hpd_poll_disable() If we really would like to prevent all *long* HPD processing during display resume we'd need some kind of software mechanism to simply ignore all long HPDs. Currently we appear to have that just for fbdev via ifbdev->hpd_suspended. Since we aren't exploding left and right all the time I guess that's mostly sufficient. For a bit of history on this, we first got a mechanism to block hotplug processing during suspend in commit 15239099 ("drm/i915: enable irqs earlier when resuming") on account of moving the irq enable earlier. This then got removed in commit 50c3dc97 ("drm/fb-helper: Fix hpd vs. initial config races") because the fdev initial config got pushed to a later point. The second ad-hoc hpd_irq_setup() for resume was added in commit 0e32b39c ("drm/i915: add DP 1.2 MST support (v0.7)") to be able to do MST sideband during the resume. And finally we got a partial resurrection of the hpd blocking mechanism in commit e8a8fedd ("drm/i915: Block fbdev HPD processing during suspend"), but this time it only prevent fbdev from handling hpd while resuming. v2: Leave the poll_init_work behind v3: Remove the extra intel_hpd_poll_disable() from display reset (Lyude) Add the missing intel_hpd_poll_disable() to display init (Imre) Cc: Lyude Paul <lyude@redhat.com> Cc: Imre Deak <imre.deak@intel.com> Signed-off-by:Lyude Paul <lyude@redhat.com>
-
Ville Syrjälä authored
Rename intel_dp_sink_dpms() to intel_dp_set_power() so one doesn't always have to convert from the DPMS enum values to the actual DP D-states. Also when dealing with a branch device this has nothing to do with any sink, so the old name was nonsense anyway. Also adjust the debug message accordingly, and pimp it with the standard encoder id+name thing. Trivial bits done with cocci: @@ expression DP; @@ ( - intel_dp_sink_dpms(DP, DRM_MODE_DPMS_OFF) + intel_dp_set_power(DP, DP_SET_POWER_D3) | - intel_dp_sink_dpms(DP, DRM_MODE_DPMS_ON) + intel_dp_set_power(DP, DP_SET_POWER_D0) ) Signed-off-by:
-
Ville Syrjälä authored
Rather that try to trick LSPCON back into PCON mode from the .reset() hook let's just do that as a regular part of the normal modeset sequence, which is going to take care of the system resume case. During a normal modeset this should normally be a nop as the mode should have already been switched by .detect(). Signed-off-by:
-
Chris Wilson authored
The "mmio" writes into vgpu registers are simple memory traps from the guest into the host. We do not need to assert in the guest that the device is awake for the io as we do not write to the device itself. However, over time we have refactored all the mmio accessors with the result that the vgpu reuses the gen2 accessors and so inherits the assert for runtime-pm of the native device. The assert though has actually been there since commit 3be0bf5a ("drm/i915: Create vGPU specific MMIO operations to reduce traps"). References: 3be0bf5a ("drm/i915: Create vGPU specific MMIO operations to reduce traps") Signed-off-by:
Zhenyu Wang <zhenyuw@linux.intel.com> Cc: stable@vger.kernel.org Link: https://patchwork.freedesktop.org/patch/msgid/20200811092532.13753-1-chris@chris-wilson.co.uk
-
Chris Wilson authored
If i915.ko is being used as a passthrough device, it does not know if the host is using intel_iommu. Mixing the iommu and gfx causes a few issues (such as scanout overfetch) which we need to workaround inside the driver, so if we detect we are running under a hypervisor, also assume the device access is being virtualised. Reported-by:
Stefan Fritsch <sf@sfritsch.de> Suggested-by:
-
José Roberto de Souza authored
Underruns happens when plane height + y offset is not a modulo of 4 when FBC is enabled. It happens when scanline is at vactive - 10 but that is not feasible to do from the software side so here completely disabling FBC when height + y offset matches to avoid visual glitches. Specification says that it only affects TGL display C stepping and newer but to simply the check and as TGL is already in final costumers hands, pre-production display stepping A and B was also included. BSpec: 52887 ICL BSpec: 52888 EHL/JSL BSpec: 52890/55378 TGL BSpec: 53508 DG1 BSpec: 53273 RKL 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/20201019175609.28715-1-jose.souza@intel.com
-
José Roberto de Souza authored
This sequence is not part of "Sequences to Initialize Display" but as noted in the MBus page the DBUF_CTL.Tracker_state_service needs to be set to 8. BSpec: 49213 Reviewed-by:
-
- 16 Oct, 2020 6 commits
-
-
Ville Syrjälä authored
intel_dp_ycbcr420_config() is rather pointless. Just inline it directly into intel_dp_compute_config(). This gets rid of the ugly double assignment of output_format. Not really sure what the best policy would be when the user supplies a mode classified by the display as "YCbCr 4:2:0 only", but we know that we can't do YCbCr 4:2:0 output. For now keep the current behaviour of just silently upgrade it to RGB 4:4:4. Signed-off-by:
-
Ville Syrjälä authored
Remove the lspcon special case from intel_dp_compute_config() and just treat it like any other DFP than can do 4:4:4->4:2:0 conversion. The only difference between the two codepaths was that the lspcon code tried to already halve port_clock. That was just total nonsense as we hadn't even computed the base port_clock at that time. All that stuff happens intel_dp_compute_link_config*() and it already takes care of the 4:2:0 clock reduction. Signed-off-by:
-
Ville Syrjälä authored
crtc_state->lspcon_downsampling isn't particularly useful at the moment since we can't even do proper readout for it. Let's get rid of it. Will help with unifying the LSPCON with the regular DFP YCbCr output support. Signed-off-by:
-
Ville Syrjälä authored
Currently we leave the cache_level of the initial fb obj set to NONE. This means on eLLC machines the first pin_to_display() will try to switch it to WT which requires a vma unbind+bind. If that happens during the fbdev initialization rcu does not seem operational which causes the unbind to get stuck. To most appearances this looks like a dead machine on boot. Avoid the unbind by already marking the object cache_level as WT when creating it. We still do an excplicit ggtt pin which will rewrite the PTEs anyway, so they will match whatever cache level we set. Cc: <stable@vger.kernel.org> # v5.7+ Suggested-by:
-
Ville Syrjälä authored
WAC6entrylatency is trying to fix excessive rc6 entry latency caused by the extra delay from FBC_LLC_READ_CTRL, which is there for some extra sync with uncore for frame buffer caching in LLC. Reading through the hsd the recommendation was to set the FBC_LLC_FULLY_OPEN bit to disable this extra delay entirely. This can be done whenever fb LLC caching is not used. The alternative suggestion was to reduce the delay to eg. 0x5 via updated BIOS programming instructions. But all the kbl/cfl machines I've seen still have the default 0xff programmed. As we never use fb LLC caching let's just apply the w/a to all skl derivatives to get consistent rc6 latencies. I was able to measure the effect of FBC_LLC_READ_CTRL to rc6 latency via forcewake. Here's a graph of some of the results: sleep;fw_req=1;wait fw_ack==1;sleep;fw_req=0;wait fw_ack==0 fw_ack==1 duration 160us +----------------------------------------------------------------+ | + + $$+ + + | | $$ $ $ ******$$ ** $ $**$* #########$$######| 140us |-$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$*$$$$$$$$$$$$$$$$ $$$$$$| | $ * # | | $ * # | 120us |$+ * # +-| |$ * # | |$ * # # | 100us |$+ ************######################## +-| |$ * *# | |$ ***** ######### | 80us |$+ * # #### ## +-| |$ **** ### # # | | ** #### FBC_LLC_READ_CTRL: 0x8000 ******* | 60us |-###### FBC_LLC_READ_CTRL: 0xffff #######-| |## + + FBC_LLC_READ_CTRL: 0x400000ff $$$$$$$ | +----------------------------------------------------------------+ 0ms 10ms 20ms 30ms 40ms 50ms 60ms sleep duration The default FBC_LLC_READ_CTRL value of 0xff is documented to give us a 170usec delay. That tracks well with the knees at 0xffff->~44msec and 0x8000->~22msec we see in the graph. We can see that if we sleep longer than the FBC_LLC_READ_CTRL delay we always observe the full (~145usec) rc6 wakeup latency. But if we sleep for less than the FBC_LLC_READ_CTRL delay we see a quicker fw wakeup, presumably due the hardware not having yet entered rc6 fully. The other plateaus in the graph I suspect correspond to some shallower internal rc states. v2: s/usec/msec/ typo in commit msg Signed-off-by: -
Matt Roper authored
A recent bspec update has provided a new cdclk table for RKL. All of the cdclk values are the same as those we've been using on ICL, TGL, etc., but we obtain them by doubling both the PLL ratio and CD2X divider numbers. Bspec: 49202 Cc: Anusha Srivatsa <anusha.srivatsa@intel.com> Signed-off-by:
-
- 15 Oct, 2020 4 commits
-
-
Michel Thierry authored
While we do lack the faster shared LLC, we should still have support for snooping over PCIe. Signed-off-by:
Michel Thierry <michel.thierry@intel.com> Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Reviewed-by:
-
Anshuman Gupta authored
Update the DMC_DEBUG_DC5 register to its new location and do not try reading the DC6 counter since DG1 doesn't support DC6. v2: Use IS_DGFX() instead of IS_DG1(). Even if not having DC6 is not directly related to DGFX, the register move to a new location is. So in future, if there is one supporting DC6, it would just need to add the other register rather than fixing the case of a wrong register being read (Matt) Cc: Uma Shankar <uma.shankar@intel.com> Signed-off-by:
Anshuman Gupta <anshuman.gupta@intel.com> Signed-off-by:
-
Anshuman Gupta authored
DC6 is not supported on DG1, so change the allowed DC mask for DG1. This is not yet on bspec, but it has been confirmed by HW engineers. Cc: Uma Shankar <uma.shankar@intel.com> Signed-off-by:
-
Stuart Summers authored
DG1 shares some workarounds with TGL and RKL and also has some additional workarounds of its own. v2: Correct location of Wa_1408615072 (JohnH). v3: Apply WAs 1606700617, 18011464164 and 22010931296 to DG1 (José) v4 (Anusha) - Add Wa_22010271021 - s/Wa_14010096844/Wa_1409836686 v5: - Extend Wa_14010919138 to all revs (Matt Atwood) - Power gate media is global gen12 design. (Rodrigo) - Rebase (Lucas) v6: use REG_BIT() to fix checkpatch warning (Lucas) BSpec: 53508 Cc: Matt Atwood <matthew.s.atwood@intel.com> Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com> Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
Stuart Summers <stuart.summers@intel.com> Signed-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by:
-
