- 22 Jun, 2021 1 commit
-
-
Jani Nikula authored
Add a single point of truth for figuring out the primary/secondary crtc for bigjoiner instead of duplicating the magic pipe +/- 1 in multiple places. Also fix the pipe validity checks to properly take non-contiguous pipes into account. The current checks may theoretically overflow i915->pipe_to_crtc_mapping[pipe], albeit with a warning, due to fused off pipes, as INTEL_NUM_PIPES() returns the actual number of pipes on the platform, and the check is for INTEL_NUM_PIPES() == pipe + 1. Prefer primary/secondary terminology going forward. v2: - Improved abstractions for pipe validity etc. Fixes: 8a029c11 ("drm/i915/dp: Modify VDSC helpers to configure DSC for Bigjoiner slave") Fixes: d961eb20 ("drm/i915/bigjoiner: atomic commit changes for uncompressed joiner") Cc: Animesh Manna <animesh.manna@intel.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Vandita Kulkarni <vandita.kulkarni@intel.com> Reviewed-by:
Manasi Navare <manasi.dl.navare@intel.com> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210610090528.20511-1-jani.nikula@intel.com
-
- 21 Jun, 2021 5 commits
-
-
Anusha Srivatsa authored
Load DMC v2.10 on ADLP. The release notes mention that this version enables few power savings features. v2: Add DMC_PATH() for ADLP (Lucas) Cc: Lucas De Marchi <lucas.demarchi@intel.com> Cc: Clint Taylor <clinton.a.taylor@intel.com> Signed-off-by:
Anusha Srivatsa <anusha.srivatsa@intel.com> Reviewed-by:
Lucas De Marchi <lucas.demarchi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210621191415.29823-5-anusha.srivatsa@intel.com
-
Anusha Srivatsa authored
ADLP requires us to load both Pipe A and Pipe B. Plug Pipe B loading support. Cc: Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
-
Anusha Srivatsa authored
This patch adds Pipe A plumbing to the already existing parsing and loading functions which is taken care of in the prep patches. Adding MAX_DMC_FW to keep track for both Main and Pipe A DMC while loading the respective blobs. Also adding present field in dmc_info. s/find_dmc_fw_offset/csr_set_dmc_fw_offset. While at it add fw_info_matches_stepping() helper. CSR_PROGRAM() should now take the starting address of the particular blob (Main or Pipe) and not hardcode it. v2: Add dmc_offset and start_mmioaddr fields for dmc_info struct. v3: Add a missing corner cases of stepping-substepping combination in fw_info_matches_stepping() helper. v4: Add macro for start_mmioaddr for V1 package. Simplify code in dmc_set_fw_offset (Lucas) Cc: Souza, Jose <jose.souza@intel.com> Cc: Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
-
Anusha Srivatsa authored
This is a prep patch for Pipe DMC plugging. Add dmc_info struct in intel_dmc to have all common fields shared between all DMC's in the package. Add DMC_FW_MAIN(dmc_id 0) to refer to the blob. v2: Remove dmc_offset and start_mmioaddr from dmc_info struct (Jose) Cc: Souza, Jose <jose.souza@intel.com> Signed-off-by:
-
Kees Cook authored
intel_dp_vsc_sdp_unpack() was using a memset() size (36, struct dp_sdp) larger than the destination (24, struct drm_dp_vsc_sdp), clobbering fields in struct intel_crtc_state after infoframes.vsc. Use the actual target size for the memset(). Fixes: 1b404b7d ("drm/i915/dp: Read out DP SDPs") Cc: stable@vger.kernel.org Signed-off-by:
Kees Cook <keescook@chromium.org> Reviewed-by:
José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
- 16 Jun, 2021 2 commits
-
-
Imre Deak authored
Disconnect TypeC PHYs during system suspend and shutdown, even with the corresponding TypeC sink still plugged to its connector, since leaving the PHY connected causes havoc at least during system resume in the presence of an Nvidia card. Note that this will only make a difference in the TypeC DP alternate mode, since in Thunderbolt alternate mode the PHY is never owned by the display engine and there is no notion of PHY ownership in legacy mode (the display engine being the only possible owner in that mode and the TypeC subsystem not having anything to do with the port in that case). Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3500Reported-and-tested-by:
Chris Chiu <chris.chiu@canonical.com> Signed-off-by:
Imre Deak <imre.deak@intel.com> Reviewed-by:
Uma Shankar <uma.shankar@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210610174223.605904-1-imre.deak@intel.com
-
Tejas Upadhyay authored
When pipe A is disabled and MIPI DSI is enabled on pipe B, the AMT KVMR feature will incorrectly see pipe A as enabled. Set 0x42080 bit 23=1 before enabling DSI on pipe B and leave it set while DSI is enabled on pipe B. No impact to setting it all the time. Changes since V5: - Added reviewed-by - Removed redundant braces and debug message format - Imre Changes since V4: - Modified function comment Wa_<number>:icl,jsl,ehl - Lucas - Modified debug message in sync state - Imre Changes since V3: - More meaningful name to workaround - Imre - Remove boolean check clear flag - Add WA_verify hook in dsi sync_state Changes since V2: - Used REG_BIT, ignored pipe A and used sw state check - Jani - Made function wrapper - Jani Changes since V1: - ./dim checkpatch errors addressed Signed-off-by:Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com> Reviewed-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210615105613.851491-1-tejaskumarx.surendrakumar.upadhyay@intel.com
-
- 14 Jun, 2021 1 commit
-
-
Lucas De Marchi authored
It's becoming pretty cumbersome to track the features enabled going back to GEN7. Gather the XE_LPD display features together in XE_LPD_FEATURES macro so they are sufficient to describe the display features. In ADL-P's device_info we set has_psr_hw_tracking to 0 as it would otherwise be enabled since it is inheriting from GEN12_FEATURES. Signed-off-by:
Matt Roper <matthew.d.roper@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210612053531.1870920-1-lucas.demarchi@intel.com
-
- 11 Jun, 2021 1 commit
-
-
José Roberto de Souza authored
Looks this FIXME is still valid as we need a way to tell LSPCON to stop sending infoframes, so reverting it. This reverts commit 3f409e4c. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210610194527.84997-1-jose.souza@intel.com
-
- 10 Jun, 2021 1 commit
-
-
Emil Velikov authored
Currently as the workaround is applied the screen flickers. As a result we do not achieve seamless boot experience. Avoiding the issue in the common use-case might be hard, although we can resolve it for dual GPU setups - when the "other" GPU is primary and/or outputs are connected to it. With this I was able to get seamless experience on my Intel/Nvidia box, running systemd-boot and sddm/Xorg. Note that the i915 driver is within initrd while the Nvidia one is not. Without this patch, the splash presented by systemd-boot (UEFI BGRT) is torn down as the code-path kicks in, leaving the monitor blank until the login manager starts. Same issue were reported with plymouth/grub, although personally I wasn't able to get them to behave on my setup. Suggested-by:
Emil Velikov <emil.velikov@collabora.com> Signed-off-by:
-
- 09 Jun, 2021 3 commits
-
-
Stanislav Lisovskiy authored
We currently treat same slice mask as a same DBuf state and skip updating the Dbuf slices, if we detect that. This is wrong as if we have a multi to single pipe change or vice versa, that would be treated as a same Dbuf state and thus no changes required, so we don't get Mbus updated, causing issues. Solution: check also mbus_join, in addition to slices mask. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
Stanislav Lisovskiy <stanislav.lisovskiy@intel.com> Signed-off-by:
-
Stanislav Lisovskiy authored
CDCLK crawl feature allows to change CDCLK frequency without disabling the actual PLL and doesn't require a full modeset. v2: - Added has_cdclk_crawl as a feature flag to intel_device_info(Matt Roper) - s/gen13_cdclk_pll_crawl/adlp_cdclk_pll_crawl/ (Matt Roper) Cc: Mika Kahola <mika.kahola@intel.com> Reviewed-by:Mika Kahola <mika.kahola@intel.com> Signed-off-by:
Gwan-gyeong Mun <gwan-gyeong.mun@intel.com> Signed-off-by:
-
Vandita Kulkarni authored
This change takes care of resetting the dss_ctl registers in case of dsc_disable, bigjoiner disable and also uncompressed joiner disable. v2: Fix formatting v3: Fix the typo (Mansi) Suggested-by:
Vandita Kulkarni <vandita.kulkarni@intel.com> Reviewed-by:
Manasi Navare <manasi.d.navare@intel.com> Signed-off-by:
Anshuman Gupta <anshuman.gupta@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210609065914.4454-1-vandita.kulkarni@intel.com
-
- 08 Jun, 2021 3 commits
-
-
Ville Syrjälä authored
AUX logic is often clocked from cdclk. Disable PSR to make sure there are no hw initiated AUX transactions in flight while we change the cdclk frequency. Cc: Mika Kahola <mika.kahola@intel.com> Signed-off-by:
-
Gwan-gyeong Mun authored
This introduces the following function that can exit and activate a psr source when intel_psr is already enabled. - intel_psr_pause(): Pause current PSR. It deactivates current psr state. - intel_psr_resume(): Resume paused PSR. It activates paused psr state. v2: Address Jose's review comment. - Remove unneeded changes around the intel_psr_enable(). - Add intel_psr_post_exit() which processes waiting until PSR is idle and WA for SelectiveFetch. v3: Address Jose's review comment. - Rename intel_psr_post_exit() to intel_psr_wait_exit_locked(). - Move WA_1408330847 to intel_psr_disable_locked() - If the PSR is paused by an explicit intel_psr_paused() call, make the intel_psr_flush() not to activate PSR. v4: Address Jose's review comment. - In order to avoid the scenario of PSR is not active but there is a scheduled psr->work, it changes the check routine of intel_psr_pause() for PSR's enablement from "psr->active" to "psr->enable". Cc: José Roberto de Souza <jose.souza@intel.com> Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: -
Tejas Upadhyay authored
Let's reserve JSL stolen memory for graphics. JasperLake is a gen11 platform which is compatible with ICL/EHL changes. This was missed in commit 24ea098b ("drm/i915/jsl: Split EHL/JSL platform info and PCI ids") V2: - Added maintainer list in cc - Added patch ref in commit message V1: - Added Cc: x86@kernel.org Fixes: 24ea098b ("drm/i915/jsl: Split EHL/JSL platform info and PCI ids") Cc: <stable@vger.kernel.org> # v5.11+ Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@redhat.com> Cc: Borislav Petkov <bp@alien8.de> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: x86@kernel.org Cc: José Roberto de Souza <jose.souza@intel.com> Signed-off-by:
-
- 07 Jun, 2021 6 commits
-
-
José Roberto de Souza authored
intel_dp_set_infoframes() call in intel_ddi_post_disable_dp() will take care to disable all enabled infoframes. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Radhakrishna Sripada <radhakrishna.sripada@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210514232247.144542-4-jose.souza@intel.com
-
José Roberto de Souza authored
When PSR is enabled it handles DP_SDP_VSC, changing revision and all the other fields as necessary. It can also enabled and disable this SDP as needed without a full modeset. So here masking DP_SDP_VSC bit when previous and future state PSR enabled, it will still be checked when comparing the asked state to what was programmed to hardware. Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com> Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com> Reported-by:
-
José Roberto de Souza authored
Commit 78b772e1 ("drm/i915/display: Fill PSR state during hardware configuration read out") is not allowing fastsets to happen when PSR states changes but PSR is a feature that can be enabled and disabled during fastsets. So here moving the PSR pipe conf checks to a block that is only executed when checking if HW state matches with requested state, not during the phase where it checks if fastset is possible or not. There still a state mismatch not allowing fastsets between states turning off or on PSR because of crtc_state->infoframes.enable BIT(DP_SDP_VSC) but at least for now it will allow a fastset between PSR1 <-> PSR2, that is a case heavilly used by CI due to pipe CRC not work with PSR2, but the remaning issue will be fixed in a future patch. Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com> Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com> Reported-by:
-
Wan Jiabing authored
Fix the following coccicheck warning: ./drivers/gpu/drm/i915/display/intel_display_power.c:3081:1-28: duplicated argument to & or | This commit fixes duplicate argument. It might be a typo. But what I can do is to remove it now. Signed-off-by:
Wan Jiabing <wanjiabing@vivo.com> Signed-off-by:
-
Lucas De Marchi authored
For some reason coccinelle misses a few cases in header files with calls to INTEL_GEN()/IS_GEN(). Do a manual conversion for those. Signed-off-by:
-
Lucas De Marchi authored
This was done by the following semantic patch: @@ expression i915; @@ - INTEL_GEN(i915) + GRAPHICS_VER(i915) @@ expression i915; expression E; @@ - INTEL_GEN(i915) >= E + GRAPHICS_VER(i915) >= E @@ expression dev_priv; expression E; @@ - !IS_GEN(dev_priv, E) + GRAPHICS_VER(dev_priv) != E @@ expression dev_priv; expression E; @@ - IS_GEN(dev_priv, E) + GRAPHICS_VER(dev_priv) == E @@ expression dev_priv; expression from, until; @@ - IS_GEN_RANGE(dev_priv, from, until) + IS_GRAPHICS_VER(dev_priv, from, until) @def@ expression E; identifier id =~ "^gen$"; @@ - id = GRAPHICS_VER(E) + ver = GRAPHICS_VER(E) @@ identifier def.id; @@ - id + ver It also takes care of renaming the variable we assign to GRAPHICS_VER() so to use "ver" rather than "gen". Signed-off-by:
-
- 04 Jun, 2021 3 commits
-
-
Lucas De Marchi authored
Since we are replacing IS_GEN() with GRAPHICS_VER(), make sure we take care of the comments as well. Signed-off-by:
-
Lucas De Marchi authored
This was done by the following semantic patch: @@ expression i915; @@ - INTEL_GEN(i915) + GRAPHICS_VER(i915) @@ expression i915; expression E; @@ - INTEL_GEN(i915) >= E + GRAPHICS_VER(i915) >= E @@ expression dev_priv; expression E; @@ - !IS_GEN(dev_priv, E) + GRAPHICS_VER(dev_priv) != E @@ expression dev_priv; expression E; @@ - IS_GEN(dev_priv, E) + GRAPHICS_VER(dev_priv) == E @@ expression dev_priv; expression from, until; @@ - IS_GEN_RANGE(dev_priv, from, until) + IS_GRAPHICS_VER(dev_priv, from, until) @def@ expression E; identifier id =~ "^gen$"; @@ - id = GRAPHICS_VER(E) + ver = GRAPHICS_VER(E) @@ identifier def.id; @@ - id + ver It also takes care of renaming the variable we assign to GRAPHICS_VER() so to use "ver" rather than "gen". Cc: intel-gvt-dev@lists.freedesktop.org Cc: Zhenyu Wang <zhenyuw@linux.intel.com> Signed-off-by:
-
Manasi Navare authored
Static analysis identified an issue in skl_crtc_allocate_ddb where mbus_offset may be used uninitialized. This patch fixes it. Fixes: 835c176c ("drm/i915: Introduce MBUS relative dbuf offsets") Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
- 03 Jun, 2021 6 commits
-
-
Imre Deak authored
On ADL_P the power well->PHY mapping doesn't follow the mapping on previous platforms, fix this up. While at it remove the redundant dev_priv param from icl_tc_phy_aux_ch(). Signed-off-by:
-
Imre Deak authored
It's possible that an already dequeued put_async_work() will release the reference (*) that was put asynchronously after the dequeue happened. This leaves an async-put work pending, without any reference to release. A subsequent async-put may trigger the drm_WARN_ON(!queue_delayed_work(&power_domains->async_put_work)); warn due to async_put_work() still pending. To avoid the warn, cancel the pending async_put_work() when releasing the reference at (*) above. Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3421 Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2289Signed-off-by:
-
Imre Deak authored
An async-put on an encoder specific power domain (for instance the AUX PW domain) may be pending when removing the encoder. Make sure any such async-puts are complete while the corresponding encoder is still in place since at least AUX power wells require this to do a power well->PHY lookup. Signed-off-by:
-
Anusha Srivatsa authored
Move struct intel_dmc from i915_drv.h to intel_dmc.h. v2: Add includes along with moving the struct. Reviewed-by:
-
Anusha Srivatsa authored
We check for dmc_payload being there at various points in the driver. Replace it with the helper. v2: rebased. v3: Move intel_dmc to intel_dmc.h in another patch (Lucas) v4: Remove headers not needed from intel_dmc.h Cc: Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by:
-
Anusha Srivatsa authored
Use new format of debug messages across intel_csr. While at it, change some function definitions which now need dev_priv for drm_err and drm_info etc. v2: use container_of() (Jani) v3: Indentation fixes. (Jani) Cc: Jani Nikula <jani.nikula@linux.intel.com> Suggested-by:
-
- 27 May, 2021 4 commits
-
-
Ville Syrjälä authored
Implement Wa_22012358565 to avoid underrun with 32bpp cursor in some high bandwidth scenarios. The implementation calls for overriding the arbitration slots for the planes. v2: Fix adlp_plane_ctl_arb_slots() return type Signed-off-by:
-
Ville Syrjälä authored
The FIFO underrun recovery mechanism has a boatload of cases where it can't be used. The description is also a bit ambiguous as it doesn't specify whether plane downscaling needs to be considered or just pipe downscaling. We may not even have sufficient state tracking to decide this on demand, so for now just disable the whole thing. Signed-off-by:
-
Imre Deak authored
On ADL_P the TBT AUX power wells depend on the PW#2 power well, add the corresponding power domain dependencies. Signed-off-by:
-
Chris Wilson authored
The kernel prefers enabling fbc over the initial fb, since this leads to actual runtime power savings, so if the initial fb is deemed too big using some heuristic, then we simply skip allocating stolen for it. However if the kernel is not configured with fbcon then it should be possible to relax this, since unlike with fbcon the display server shouldn't preserve it when later replacing it, and so we should be able to re-use the stolen memory for fbc and friends. This patch is reported to fix some flicker seen during boot splash on some devices. v2: s/FRAMEBUFFER_CONSOLE/CONFIG_FRAMEBUFFER_CONSOLE Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Cc: Lee Shawn C <shawn.c.lee@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Reviewed-by:
-
- 26 May, 2021 4 commits
-
-
Ville Syrjälä authored
intel_hdmi_bpc_possible() will check has_hdmi_sink for us, so no need to check it in intel_hdmi_mode_clock_valid() anymore. Cc: Werner Sembach <wse@tuxedocomputers.com> Signed-off-by:
-
Ville Syrjälä authored
Now that we have to tell intel_hdmi_mode_clock_valid() whether we're asking about 4:4:4 or 4:2:0 output it can take care of the dotclock->TMDS clock conversion. Cc: Werner Sembach <wse@tuxedocomputers.com> Signed-off-by:
-
Ville Syrjälä authored
Currently HDMI .mode_valid() only checks whether the source can do deep color. Let's check whether the sink can do it as well. Cc: Werner Sembach <wse@tuxedocomputers.com> Signed-off-by:
-
Ville Syrjälä authored
Let's put the platform checks into intel_hdmi_bpc_possible() so that it'll confirm both the source and sink capabilities. Cc: Werner Sembach <wse@tuxedocomputers.com> Signed-off-by:
-
