On PTL, no combo PHY is connected to PORT B. However, PORT B can
still be used for Type-C and will utilize the C20 PHY for eDP
over Type-C. In such configurations, VBTs also enumerate PORT B.
This leads to issues where PORT B is incorrectly identified as using the
C10 PHY, due to the assumption that returning true for PORT B in
intel_encoder_is_c10phy() would not cause problems.
From PTL's perspective, only PORT A/PHY A uses the C10 PHY.
Update the helper intel_encoder_is_c10phy() to return true only for
PORT A/PHY on PTL.
v2: Change the condition code style for ptl/wcl
Bspec: 72571,73944
Fixes: 9d10de78a3 ("drm/i915/wcl: C10 phy connected to port A and B")
Signed-off-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Link: https://lore.kernel.org/r/20250922150317.2334680-4-dnyaneshwar.bhadane@intel.com
We will need to differentiate between WCL and PTL in
intel_encoder_is_c10phy(). Since WCL and PTL use the same display
architecture, let's define WCL as a subplatform of PTL to allow the
differentiation.
v2: Update commit message and reorder wcl define (Gustavo)
Signed-off-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Link: https://lore.kernel.org/r/20250922150317.2334680-3-dnyaneshwar.bhadane@intel.com
To form the WCL platform as a subplatform of PTL in definition,
WCL pci ids are splited into saparate group from PTL.
So update the pciidlist struct to cover all the pci ids.
v2:
- Squash wcl description in single patch for display and xe.(jani,gustavo)
Signed-off-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Link: https://lore.kernel.org/r/20250922150317.2334680-2-dnyaneshwar.bhadane@intel.com
Some systems (eg. LNL Lenovo Thinkapd X1 Carbon) declare
semi-bogus non-monotonic WM latency values:
WM0 latency not provided
WM1 latency 100 usec
WM2 latency 100 usec
WM3 latency 100 usec
WM4 latency 93 usec
WM5 latency 100 usec
Apparently Windows just papers over the issue by bumping the
latencies for the higher watermark levels to make them monotonic
again. Do the same.
Cc: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919193000.17665-12-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Move the inner loop out from the outer loop in
sanitize_wm_latency() to flatten things a bit.
Easier to read flat code.
v2: Move the inner loop out completely (Luca)
Cc: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919193000.17665-11-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Bump the latency for all watermark levels in the
16Gb+ DIMM w/a. The spec does ask us to do it only for level
0, but it seems more sane to bump all the levels. If the actual
memory access is slower then the wakeup (WM1+) should also
potentially happen earlier.
This also avoids the theoretical case that WM0 would get bumped
higher than WM1+. Not that it is likely to happen because the WM0
latency is always significantly lower than the WM1 latency.
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/20250919193000.17665-9-ville.syrjala@linux.intel.com
{mtl,skl}_read_wm_latency() are doing way too many things for
my liking. Move the adjustment stuff out into the caller.
This also gives us one place where we specify the 'read_latency'
for all the platforms, instead of two places.
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/20250919193000.17665-6-ville.syrjala@linux.intel.com
If WM0 latency is zero we need to bump it (and the WM1+ latencies)
but a fixed amount. But any WM1+ level with zero latency must
not be touched since that indicates that corresponding WM level
isn't supported.
Currently the loop doing that adjustment does work, but only because
the previous loop modified the num_levels used as the loop boundary.
This all seems a bit too fragile. Remove the num_levels adjustment
and instead adjust the read latency loop to abort when it encounters
a zero latency value.
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/20250919193000.17665-4-ville.syrjala@linux.intel.com
Currently the code assumes that every platform except dg2 need the
16Gb DIMM w/a, while in reality it's only needed by skl and icl (and
derivatives). Switch to a more specific platform check.
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/20250919193000.17665-3-ville.syrjala@linux.intel.com
While the spec only asks us to do the WM0 latency bump for 16Gb
DRAM devices I believe we should apply it for larger DRAM chips.
At the time the w/a was added there were no larger chips on
the market, but I think I've seen at least 32Gb DDR4 chips
being available these days.
Whether it's possible to actually find suitable DIMMs for the
affected systems with largers chips I don't know. Also it's
not known whether the 1 usec latency bump would be sufficient
for larger chips. Someone would need to find such DIMMs and
test this. Fortunately we do have a bit of extra latency already
with the 1 usec bump, as the actual requirement was .4 usec for
for 16Gb chips.
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/20250919193000.17665-2-ville.syrjala@linux.intel.com
I don't think there should be any need for us to call any of
pci_enable_device(), pci_disable_device() or pci_set_master()
from the suspend/resume paths. The config space save/restore should
take care of all of this.
Cc: Karthik Poosa <karthik.poosa@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-7-ville.syrjala@linux.intel.com
Stop doing the pci_save_state(), except when we need to prevent
D3 due to BIOS bugs, so that the code in drivers/pci is allowed
to manage the state of the PCI device. Less chance something
getting left by the wayside by i915 if/when the things change in
drivers/pci.
Cc: Badal Nilawar <badal.nilawar@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-6-ville.syrjala@linux.intel.com
Since this switcheroo garbage bypasses all the core pm we
have to manually manage the pci state. To that end add the
missing pci_restore_state() to the switcheroo resume hook.
We already have the pci_save_state() counterpart on the
suspend side.
Arguably none of this code should exist in the driver
in the first place, and instead the entire switcheroo
mechanism should be rewritten and properly integrated into
core pm code...
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-5-ville.syrjala@linux.intel.com
If the driver doesn't call pci_save_state() drivers/pci will
normally save+power manage the device from the _noirq() pm hooks.
We can't let that happen as some old BIOSes fail to hibernate
when the device is in D3. However, we can get very close to
the standard behaviour by doing our explicit pci_save_state()
and pci_set_power_state() stuff from driver provided _noirq()
hooks.
This results in a change of behaviour where we no longer go
into D3 at the end of freeze_late, so when it comes time
to thaw() we'll already be in D0, and thus we can drop the
explicit pci_set_power_state(D0) call.
Presumably switcheroo suspend will want to go into D3 so
call the _noirq() stuff from the switcheroo suspend hook,
and since we dropped the pci_set_power_state(D0) from
resume_early() we'll need to add one back into the
switcheroo resume hook.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-4-ville.syrjala@linux.intel.com
drivers/pci does the pci_save_state()+pci_set_power_state() from
the _noirq() pm hooks. Move our manual calls (needed for the
hibernate vs. D3 workaround with buggy BIOSes) towards that same
point. We currently have no _noirq() hooks, so end of _late()
hooks is the best we can do right now.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-3-ville.syrjala@linux.intel.com
Stop spelling out each variant of the hook ("" vs. "_late" vs.
"_early") and just say eg. "@thaw*" to indicate all of them.
Avoids having to update the docs whenever we start/stop using
one of the variants.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250919185015.14561-2-ville.syrjala@linux.intel.com
Currently we check if the encoder is INVALID or -1 and throw a
WARN_ON but we still end up writing the temp value which will
overflow and corrupt the whole programmed value.
--v2
-Assign a bogus transcoder to master in case we get a INVALID
TRANSCODER [Jani]
Fixes: 6671c367a9 ("drm/i915/tgl: Select master transcoder for MST stream")
Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Link: https://lore.kernel.org/r/20250908042208.1011144-1-suraj.kandpal@intel.com
ICL/TGL VRR hardware won't allow us to program flipline==vmin. If we do
that the actual effect will be the same as if we had programmed
flipline=vmin+1, which would make the minimum vtotal one scanline taller
than expected.
To compensate for this we reduce vmin by one, and then program
flipline=vmin+1. So we end up with a flipline value that matches
the expected minimum vtotal. Currently this adjustment happens
in intel_vrr_compute_config() which means that crtc_state->vrr.vmin
will no longer be directly usable for the remainder of the high
level VRR code. That is annoying at best, fragile at worst.
Hide the adjustment in low level code instead. This will allow most
of the higher level VRR code to remain blissfully ignorant about this
fact. Afterwards crtc_state->vrr.{vmin,flipline} will be equal
and match the minimum vtotal, exactly how things already work
on ADL+.
The only slight downside is that the actual register value will no
longer match crtc_state->vrr.vmin on ICL/TGL, but that may already
be the case on TGL because the register value will also have been
adjusted by the SCL.
Note that we must change the guardband calculation to account
for intel_vrr_extra_vblank_delay() explicitly. Previously that
was accidentally handled by the earlier vmin reduction by
intel_vrr_flipline_offset().
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250918232226.25295-2-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
num_entries comes from package_header, which is read from an external
firmware blob and thus untrusted. In parse_dmc_fw_package() we assign
package_header->num_entries to a local variable, but the range check
still uses the struct field directly.
Switch the check to use the local copy instead. This makes the
sanitization explicit and avoids a redundant dereference.
Reviewed-by: Mitul Golani <mitulkumar.ajitkumar.golani@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/20250909083042.1292672-1-luciano.coelho@intel.com
From display 14 onward do not enable the cursor
size reduction bit as it has been defeatured.
Bspec: 50372
Signed-off-by: Nemesa Garg <nemesa.garg@intel.com>
Reviewed-by: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Signed-off-by: Animesh Manna <animesh.manna@intel.com>
Link: https://lore.kernel.org/r/20250813052017.3591331-1-nemesa.garg@intel.com
Abstract ilk_display_irq_reset(), moving display related reset
there. This results in a slightly different order between GT and PCH
reset, hopefully with no impact.
v3: Reset display first (Ville)
v2: Also move GEN7_ERR_INT (Ville)
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250918133835.2412980-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
There are three groups of platforms using i915->irq_mask independently:
gen 2-4, VLV/CHV, and gen 5-7.
The gen 5-7 usage is primarily limited to display. Move its irq_mask
usage to struct intel_display as ilk_de_imr_mask for gen 5-7.
ilk_de_imr_mask could be put inside a union with with vlv_imr_mask and
de_irq_mask[], but keep them separate to avoid accidental aliasing of
the values.
With this, we can also drop the irq_mask member from struct xe_device.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/adf60e74b890d52dd20ab4673111ae2063d33b49.1758198300.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
There are three groups of platforms using i915->irq_mask independently:
gen 2-4, VLV/CHV, and gen 5-7.
The VLV/CHV usage is purely limited to display. Move its irq_mask usage
to struct intel_display as vlv_imr_mask for VLV/CHV.
vlv_imr_mask could be put inside a union with de_irq_mask[], but keep
them separate to avoid accidental aliasing of the values.
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/cef6dee8d0b02ff76180c5879f3056e102947a57.1758198300.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
i915 and xe do different things on the failure path; i915 calls
drm_gem_object_put() while xe calls xe_bo_unpin_map_no_vm(). Add a
helper to enable further refactoring.
v2: Call drm_gem_object_put() on intel_fbdev_fb_bo_destroy()
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/22bc3c3158f5a22ab258ada8684766fdf75fefec.1758184771.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
For reasons unknown, xe uses XE_PAGE_SIZE alignment for
stride. Presumably it's just a confusion between stride alignment and bo
allocation size alignment. Switch to 64 byte alignment to, uh, align
with i915.
This will also be helpful in deduplicating and unifying the xe and i915
framebuffer allocation.
Link: https://lore.kernel.org/r/aLqsC87Ol_zCXOkN@intel.com
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/7f4972104de8b179d5724ae83892ee294d3f3fd3.1758184771.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Currently our crtc_state->vrr.{vmin.vmax,flipline} are mangled on
TGL to account for the SCL delay (the hardware requires this mangling
or the actual vtotals will become incorrect). Unfortunately this
means that one can't simply use these values directly in many places,
and instead we always have to go through functions that undo the
damage first. This is all rather fragile.
Simplify our lives a bit by hiding this mangling deeper inside
the low level VRR code, leaving the number stored in the crtc
state actually something that humans can use.
This does introduce a dependdency as intel_vrr_get_config()
will now need to know the SCL value, which is read out in
intel_get_transcoder_timings(). I suppose we could simply
duplicate the SCL readout in both places should this become
a real hinderance. For now just leave a note around the
intel_get_transcoder_timings() call to remind us.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250917203446.14374-6-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
intel_vrr_fixed_rr_*() return values that have had the TGL
SCL adjustment applied to them. So we should indicate that these
values are only really useful when fed to the hardware. Add
a "_hw_" indicator to the function names to reflect that fact.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250917203446.14374-5-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
While ICL/TGL VRR hardware doesn't have a register for the guardband
value, our lives will be simpler if we pretend that it does. Start
by computing the guardband the same as on ADL+ and storing it in
the state, and only then we convert it into the corresponding
pipeline_full value that the hardware can consume. During readout we
do the opposite.
I was debating whether to completely remove pipeline_full from the
crtc state, but decided to keep it for now. Mainly because we check
it in vrr_params_changed() and simply checking the guardband instead
isn't 100% equivalent; Theoretically, framestart_delay may have
changed in the opposite direction to pipeline_full, keeping the
derived guardband value unchaged. One solution would be to also check
framestart_delay, but that feels a bit leaky abstraction wise.
Also note that we don't currently handle the maximum limit of 255
scanlines for the pipeline_full in a very nice way. The actual
position of the delayed vblank will move because of that clamping,
and so some of our code may get confused. But fixing this shall
wait a for now.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250917203446.14374-4-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
In order to pretend that ICL/TGL VRR hardware has a similar guardband
as on ADL+ we'll need access to framestart_delay already during
intel_vrr_get_config(). Hoist the framestart_delay to an earlier point
to make that possible.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250917203446.14374-3-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
I'd like to move towards a world where we can't more or less
pretend that the ICl/TGL VRR hardware works the same way as
ADL+. To that end extract some helpers to convert between
the guardband and pipeline_full representations.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250917203446.14374-2-ville.syrjala@linux.intel.com
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@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>
