Pixel normalizer is enabled with normalization factor as 1.0 for
FP16 formats in order to support FBC for those formats in xe3p_lpd.
Previously pixel normalizer gets disabled during the plane disable
routine. But there could be plane format settings without explicitly
calling the plane disable in-between and we could endup keeping the
pixel normalizer enabled for formats which we don't require that.
This is causing crc mismatches in yuv formats and FIFO underruns in
planar formats like NV12. Fix this by updating the pixel normalizer
configuration based on the pixel formats explicitly during the plane
settings arm calls itself - enable it for FP16 and disable it for
other formats in HDR capable planes.
v2: avoid redundant pixel normalization setting updates
v3: moved the normalization factor definition to intel_fbc.c and some
updates to comments
v4: simplified the pixel normalizer setting handling
Fixes: 5298eea7ed ("drm/i915/xe3p_lpd: use pixel normalizer for fp16 formats for FBC")
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patch.msgid.link/20260130095919.107805-1-vinod.govindapillai@intel.com
(cherry picked from commit c0dc68f4e2)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Introduce INTEL_DISPLAY_DEVICE_*_OFFSET() macros to compute absolute
MMIO offsets for pipe, transcoder, and cursor registers.
Update _MMIO_PIPE2/_MMIO_TRANS2/_MMIO_CURSOR2 to use these macros
for cleaner abstraction and to prepare for external API usage (e.g. GVT).
Also move DISPLAY_MMIO_BASE() to intel_display_device.h so it can be
abstracted in GVT, allowing register macros to resolve via
exported helpers rather than peeking into struct intel_display.
v2: Wrap the macro argument usages in parenthesis. (Jani)
Suggested-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patch.msgid.link/20251219060302.2365123-2-ankit.k.nautiyal@intel.com
Whether AuxCCS can be properly supported depends on the support both from
the display side and non-display side of the driver.
Let us therefore allow for the non-display part to be queried via the
display parent interface.
The new interface replaces the HAS_AUX_CCS macro and we also remove the
FIXME from skl_universal_plane_create since now the xe will not advertise
the AuxCCS caps to start with so they do not need to be removed after
enumeration.
Also, by removing this build specific FIXME we come a step closer to fully
de-coupling display and non-display.
The existing HAS_AUX_CCS gets renamed to HAS_AUX_DIST since it is still
required for determining the need for PLANE_AUX_DIST programming.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
References: cf48bddd31 ("drm/i915/display: Disable AuxCCS framebuffers if built for Xe")
Cc: intel-gfx@lists.freedesktop.org
Cc: intel-xe@lists.freedesktop.org
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com> # v1
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251209120034.9143-1-tursulin@igalia.com
One of the FBC instances can utilize the reserved area of SoC
level cache for the fbc transactions to benefit reduced memory
system power especially in idle scenarios. Reserved area of the
system cache can be assigned to an fbc instance by configuring
the cacheability configuration register with offset of the
compressed frame buffer in stolen memoty of that fbc. There is
a limit to this reserved area which is programmable and for
xe3p_lpd the limit is defined as 2MB. The first FBC instance
enabled will utilize the system cache as of now.
v2: - better to track fbc sys cache usage from intel_display level,
sanitize the cacheability config register on probe (Matt)
- limit this for integrated graphics solutions, confirmed that
no default value set for cache range by hw (Gustavo)
v3: - changes related to the use of fbc substruct in intel_display
- use intel_de_write() instead of intel_rmw() by hardcoding the
default value fields
v4: - protect sys cache config accesses, sys cache usage status in
debugfs per fbc instance (Jani)
v5: - mutex_init and missing mutex_lock in sanitize call
v6: - changes to commit message and some obvious comments removed
Bspec: 68881, 74722
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patch.msgid.link/20251128113557.129192-1-vinod.govindapillai@intel.com
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Let's gradually start calling i915 and xe parent, or core, drivers from
display via function pointers passed at display probe.
Going forward, the struct intel_display_parent_interface is expected to
include const pointers to sub-structs by functionality, for example:
struct intel_display_rpm {
struct ref_tracker *(*get)(struct drm_device *drm);
/* ... */
};
struct intel_display_parent_interface {
/* ... */
const struct intel_display_rpm *rpm;
};
This is a baby step towards not building display as part of both i915
and xe drivers, but rather making it an independent driver interfacing
with the two.
v3: useless include additions dropped
v2: unrelated include removal dropped
Cc: Jouni Högander <jouni.hogander@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patch.msgid.link/20251030202836.1815680-2-jouni.hogander@intel.com
DRRS has been defeatured on LNL+. Adjust HAS_DOUBLE_BUFFERED_M_N()
to match.
Note that the M/N registers still appear to be double buffered under
the hood but the double buffer update point is now documented to be
just the last register write to the M/N registers, so it no longer
happens synchronously with the vblank/MSA transmission. We should
perhaps rename HAS_DOUBLE_BUFFERED_M_N() to more accurately reflect
reality, but couldn't come up with a decent name right now...
Bspec: 68917
HSD: 14016007525
Cc: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250912135926.18910-1-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
With addition of double buffered GAMMA registers in PTL, we can now
program them in the active region. Use GOSUB instruction of DSB to
program them.
It is done in the following steps:
1. intel_color_prepare_commit()
- If the platform supports, prepare a dsb instance (dsb_color)
hooked to DSB0.
- Add all the register write instructions to dsb_color through
the load_lut() hook
- Do not add the vrr_send_push() logic to the buffer as it
should be taken care by dsb_commit instance of DSB0
- Finish preparation of the buffer by aligning it to 64 bit
2. intel_atomic_dsb_finish()
- Add the gosub instruction into the dsb_commit instance of DSB0
using intel_dsb_gosub()
- If needed, add the vrr_send_push() logic to dsb_commit after it
v2: Refactor code to simplify commit completion flow.
Add some helpers along the way (Ville)
v3: s/doubled/double and add display to commit message prefix (Uma)
Signed-off-by: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Animesh Manna <animesh.manna@intel.com>
Link: https://lore.kernel.org/r/20250523062041.166468-9-chaitanya.kumar.borah@intel.com
Hook up PIPEDMC interrupts. We'll need these for:
- flip queue signalling
- GTT/ATS faults on LNL+
- unclaimed register access errors (supposedly that is what
the error interrupt indicated according to Windows code).
On LNL+ we get a new level of interrupts registers PIPEDMC_INTERRUPT*.
On earlier platforms we only have the INT_VECTOR field in the
PIPEDMC_STATUS registers, whose values are defined by the firmware.
For now we'll enable the interrupts on LNL+ only. For earlier platforms
it's not clear that there is any use for these interrupts, and some
ADL machines have exhibited spurious DE_PIPE interrupts with the
PIPEDMC interrupts unmasked/enabled. We can revisit enabling these
for earlier platforms in the future.
For some unknown reason LNL pipe B triggers the error interrupt
during the first DC state transition (subsequent transitions are
maybe OK?). No clear idea what's going on here yet, so keep the
error interrupt disabled for now.
Similar to DSB interrupt registers, the unused bits in
PIPEDMC_INTERRUPT* seem to act like randomg r/w bits (instead
of being hardwired to 0 like one would expect), and so we'll try
to avoid setting them so that we don't mistake them for real
interrupts.
v2: Only enable/unmask for LNL+
Keep the flip queue interrupt masked off for now since
we don't have a use for it yet
v3: Also keep the error interrupt masked off for now due to
LNL pipe B triggering it
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250514174257.8708-1-ville.syrjala@linux.intel.com
Add support for forcing the link bpp on a connector via a connector
debugfs entry. During reducing link bpps due to a link BW limit, keep
bpps close to their forced value.
Add the debugfs entry to all relevant connectors: all DP connectors and
on an FDI link CRT/SDVO/LVDS/HDMI connectors.
v2:
- Move adding the debugfs entries to this patch.
- Rename i915_force_link_bpp to intel_force_link_bpp. (Jani)
- Select the relevant connectors via platform checks. (Jani)
- Use for_each_new_intel_connector_in_state(). (Jani)
- Fix 64 bit division vs. 32 bit build when converting str to q4. (lkp)
- Avoid division and addition overflow when converting str to q4.
v3:
- Add TODO: to make the non-DSC min bpp value connector specific. (Ankit)
Cc: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://lore.kernel.org/r/20250509180340.554867-12-imre.deak@intel.com
The CMTG is a timing generator that runs in parallel with transcoders
timing generators and can be used as a reference for synchronization.
We have observed that we are inheriting from GOP a display configuration
with the CMTG enabled. Because our driver doesn't currently implement
any CMTG sequences, the CMTG ends up still enabled after our driver
takes over.
We need to make sure that the CMTG is not enabled if we are not going to
use it. For that, let's add a partial implementation in our driver that
only cares about disabling the CMTG if it was found enabled during
initial hardware readout. In the future, we can also implement sequences
for using the CMTG if that becomes a needed feature.
For now, we only deal with cases when it is possible to disable the CMTG
without requiring a modeset. For earlier display versions, we simply
skip if we find the CMTG enabled and we can't disable it without a
proper modeset. In the future, we need to properly handle that case.
v2:
- DG2 does not have the CMTG. Update HAS_CMTG() accordingly.
- Update logic to force disabling of CMTG only for initial commit.
v3:
- Add missing changes for v2 that were staged but not committed.
v4:
- Avoid if/else duplication in intel_cmtg_dump_state() by using "n/a"
for CMTG B enabled/disabled string for platforms without it. (Jani)
- Prefer intel_cmtg_readout_hw_state() over intel_cmtg_readout_state().
(Jani)
- Use display struct instead of i915 as first parameter for
TRANS_DDI_FUNC_CTL2(). (Jani)
- Fewer continuation lines in variable declaration/initialization for
better readability. (Jani)
- Coding style improvements. (Jani)
- Use drm_dbg_kms() instead of drm_info() for logging the disabling
of the CMTG.
- Make struct intel_cmtg_state entirely private to intel_cmtg.c.
v5:
- Do the disable sequence as part of the sanitization step after
hardware readout instead of initial modeset commit. (Jani)
- Adapt to commit 1513358246 ("drm/i915/display: convert global state
to struct intel_display") by using a display struct instead of i915
as argument for intel_atomic_global_obj_init().
v6:
- Do not track CMTG state as a global state. (Ville)
- Simplify the driver logic by only disabling the CMTG only on cases
when a modeset is not required. (Ville)
v7:
- Remove the call to drm_WARN_ON() when checking
intel_cmtg_disable_requires_modeset() and use a FIXME in the comment
instead.
- Remove the !HAS_CMTG() guard from intel_cmtg_get_config(), which is
static and its caller is already protected by that same condition.
- Also take the opportunity to put some Bspec references in the commit
trailers section.
v8:
- Use HAS_TRANSCODER() instead of intel_crtc_for_pipe(). (Ville)
- Ensure transcoder power well is enabled before reading
TRANS_DDI_FUNC_CTL2. (Ville)
Bspec: 68915, 49262
Cc: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250124173956.46534-1-gustavo.sousa@intel.com
In the C20 algorithm for HDMI TMDS, certain fields have been updated
in the BSpec to set values for SRAM_GENERIC_<A/B>_TX_CNTX_CFG_1,
such as tx_misc and dac_ctrl_range for Xe2LPD, Xe2HPD and MTL/ARL.
This patch covers fields that need to be set based on the platform type.
Some ARLs SoCs cannot be directly distinguished by their GMD version Id,
Specifically to set value of tx_misc, so PCI Host Bridge IDs are used
for differentiation.
v2:
- Relocate defines and Restructure the code(Jani)
v3:
- Replace conditions with display.platform.<platform> (jani)
- Move host bridge check to new function (Jani)
v4:
- Identify/Replace arrowlake_u as meteorlake_u(Jani)
Bspec:74165,74491
Signed-off-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Reviewed-by: Suraj Kandpal <suraj.kandpal@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241217201301.3593054-3-dnyaneshwar.bhadane@intel.com
From LNL onwards there is a new hardware feature, which
allows to detect if the driver wrongly allocated DBuf
entries and they happen to overlap. If enabled this will
cause a specific interrupt to occur.
We now handle it in the driver, by writing correspondent
error message to kernel log.
v2: Initialize dbuf overlap flag in runtime_defaults (Jani Nikula)
v3: Unmask the overlap detection interrupt (Uma)
v4: use display over i915 (Jani Nikula)
v5: Use display instead of dev_priv (Jani Nikula)
v6: rebased to resolve merge conflicts
Bspec: 69450, 69464
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Vinod Govindapillai <vinod.govindapillai@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241030103319.207235-1-vinod.govindapillai@intel.com
Use x100, or ver * 100 + rel, versions for full IP version checks,
similar to what xe driver does:
- Replace IP_VER(14, 1) inline with 1401, etc.
- Convert DISPLAY_VER_FULL() to DISPLAY_VERx100()
- Convert IS_DISPLAY_VER_FULL() to IS_DISPLAY_VERx100()
- Convert IS_DISPLAY_VER_STEP() to IS_DISPLAY_VERx100_STEP()
This makes ver.rel versions easier to use, follows the xe driver
pattern, and drops the dependency on the IP_VER() macro.
v2: Rebase, drop IP_VER() from xe compat headers
v3: Rebase
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Suraj Kandpal <suraj.kandpal@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241029155536.753413-1-jani.nikula@intel.com
