drm/i915: Don't flag both full modeset and fastset at the same time

Be consistent in whether we flag a full modeset or a
fastset for the pipe. intel_modeset_all_pipes() would
seem to be the only codepath not getting this right.

The other case is when we flag the fastset initially,
currently we just clear the mode_changed flag and set
the update_pipe flag. But we could still have
connectors_changed==true or active_changed==true forcing
a full modeset anyway. So check for that after clearing
the mode_changed flag.

And let's add a WARN to make sure we did get it right.

v2: Deal with {connectors,active}_changed

Reviewed-by: Jani Nikula <jani.nikula@intel.com> #v1
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021162442.27283-4-ville.syrjala@linux.intel.com

drm/i915: Remove some local 'mode_changed' bools

These 'mode_changed' booleans aren't very helpful. Just
replace them with direct intel_crtc_needs_modeset() calls
which is more descriptive.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021162442.27283-3-ville.syrjala@linux.intel.com

drm/i915: Introduce intel_crtc_needs_fastset()

Replace the somewhat obscure crtc_state.update_pipe checks
with a more descriptive thing. Also nicely matches the
intel_crtc_needs_modeset() counterpart for full modesets.

v2: Handle one more case in the fbc code

Reviewed-by: Jani Nikula <jani.nikula@intel.com> #v1
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221021162442.27283-2-ville.syrjala@linux.intel.com

drm/i915: Activate DRRS after state readout

On BDW+ we have just the one set of DP M/N registers. The
values we write into said registers depends on whether we
want DRRS to be in high or low gear. This causes issues
for the state checker which currently has to assume either
set of M/N (high or low refresh rate) values may appear there.
That sort of works for M/N itself, but all other values
derived from the M/N (dotclock, pixel rate) are not handled
correctly, leading to potential for state checker mismatches.

Let's avoid all those problems by simply keeping DRRS in
high gear until the state checker has done its hardware
state readout.

Note that hitting this issue presumable became very hard
after commit 1b333c679a0f ("drm/i915: Do DRRS disable/enable
during pre/post_plane_update()") since the state check would
have to laze about for one full second (delay used by
intel_drrs_schedule_work()) to see the low refresh rate.
But it is still theoretically possible.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020120706.25728-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Allow panel fixed modes to have differing sync polarities

Apparently some panels declare multiple modes with random
sync polarities. Seems a bit weird, but looks like Windows/GOP
doesn't care, so let follow suit and accept alternate fixed
modes regardless of their sync polarities.

v2: Don't pollute the DRM_ namespace with a define (Jani)

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6968
Acked-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020093938.27200-1-ville.syrjala@linux.intel.com

drm/i915: Remove one use macro

Remove one use macro for_each_connector_on_encoder which
is only being used at intel_encoder_find_connector.

Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Reviewed-by: Arun R Murthy <arun.r.murthy@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221020104635.874860-2-suraj.kandpal@intel.com

drm/i915/dp: Remove whitespace at the end of function.

Remove extraline left after intel_dp_configure_protocol_converter.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221011063447.904649-3-ankit.k.nautiyal@intel.com

drm/i915/dp: Reset frl trained flag before restarting FRL training

For cases where DP has HDMI2.1 sink and FRL Link issues are detected,
reset the flag to state FRL trained status before restarting FRL
training.

Fixes: 9488a030ac91 ("drm/i915: Add support for enabling link status and recovery")
Cc: Swati Sharma <swati2.sharma@intel.com>
Cc: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Cc: Uma Shankar <uma.shankar@intel.com> (v2)
Cc: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221011063447.904649-2-ankit.k.nautiyal@intel.com

drm/i915: Print return value on error

Print returned error code for better debuggability.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/7211
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221014154655.14075-1-nirmoy.das@intel.com

drm/i915/dgfx: Keep PCI autosuspend control 'on' by default on all dGPU

DGFX platforms has lmem and cpu can access the lmem objects
via mmap and i915 internal i915_gem_object_pin_map() for
i915 own usages. Both of these methods has pre-requisite
requirement to keep GFX PCI endpoint in D0 for a supported
iomem transaction over PCI link. (Refer PCIe specs 5.3.1.4.1)

Both DG1/DG2 have a known hardware bug that violates the PCIe specs
and support the iomem read write transaction over PCIe bus despite
endpoint is D3 state.
Due to above H/W bug, we had never observed any issue with i915 runtime
PM versus lmem access.
But this issue becomes visible when PCIe gfx endpoint's upstream
bridge enters to D3, at this point any lmem read/write access will be
returned as unsupported request. But again this issue is not observed
on every platform because it has been observed on few host machines
DG1/DG2 endpoint's upstream bridge does not bind with pcieport driver.
which really disables the PCIe  power savings and leaves the bridge
at D0 state.

We need a unique interface to read/write from lmem with runtime PM
wakeref protection something similar to intel_uncore_{read, write},
keep autosuspend control to 'on' on all discrete platforms,
until we have a unique interface to read/write from lmem.

This just change the default autosuspend setting of i915 on dGPU,
user can still change it to 'auto'.

v2:
- Modified the commit message and subject with more information.
- Changed the Fixes tag to LMEM support commit. [Joonas]
- Changed !HAS_LMEM() Cond to !IS_DGFX(). [Rodrigo]

Fixes: b908be543e44 ("drm/i915: support creating LMEM objects")
Suggested-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221014113258.1284226-1-anshuman.gupta@intel.com

drm/i915: fix clear mask in GEN7_MISCCPCTL update

GEN7_DOP_CLOCK_GATE_ENABLE bit should be cleared, not inverse.
The bug was introduced during conversion to intel_uncore_rmw helper.

Suggested-by: Matt Roper <matthew.d.roper@intel.com>
Fixes: 8cee664d3eb6f8 ("drm/i915: use proper helper for register updates")
Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017085525.3898649-1-andrzej.hajda@intel.com

drm/i915: Use graphics ver, rel info for media on old platforms

Platforms prior to MTL do not have a separate media and graphics version.
On platforms where GMD id is not supported, reuse the graphics ip version,
release info for media.

The rest of the IP graphics, display versions would be copied during driver
creation.

While at it warn if GMD is not used for platforms greater than gen12.

v2:
 - Use simple assignment to copy contents of the structure(JaniN)

Fixes: c2c7075225ef ("drm/i915: Read graphics/media/display arch version from hw")
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221011153851.3781507-2-radhakrishna.sripada@intel.com

drm/i915: Add intel_ prefix to struct ip_version

Rename struct ip_version to intel_ip_version to comply with the
naming conventions for structures.

Suggested-by: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221011153851.3781507-1-radhakrishna.sripada@intel.com

drm/i915: Extend Wa_1607297627 to Alderlake-P

Workaround 1607297627 was missed for Alderlake-P, so here extending it
to it and adding the fixes tag so this WA is backported to all
stable kernels.

v2:
- fixed subject
- added Fixes tag

BSpec: 54369
Cc: <stable@vger.kernel.org> # v5.17+
Fixes: dfb924e33927 ("drm/i915/adlp: Remove require_force_probe protection")
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017132432.112850-1-jose.souza@intel.com

drm/i915: Fix simulated GPU reset wrt. encoder HW readout

The GPU reset involves a display suspend/resume sequence, but this is
done without suspending/resuming the encoders. The encoder HW readout
code during resume however assumes that the encoders were
suspended/resumed, at least on TypeC platforms where the TC PHYs must be
left in a disconnected state during encoder-suspend, and the PHY's TypeC
mode must be initialized already during encoder-resume.

Since the above issue occurs only in case the display recovery during
GPU reset is simulated in CI tests (on new platforms w/o the GPU reset
clobbering the display), this patch fixes the issue by simply restoring
the saved display state in this case w/o doing a display HW state
readout / sanitization first. This also fixes the WARN below introduced
by

commit a82796a2e332 ("drm/i915: Fix TypeC mode initialization during system resume")

<4> [319.983309] ------------[ cut here ]------------
<4> [319.983313] i915 0000:00:02.0: drm_WARN_ON(dig_port->tc_link_refcount != 1)
<4> [319.983341] WARNING: CPU: 10 PID: 268 at drivers/gpu/drm/i915/display/intel_tc.c:751
intel_tc_port_sanitize_mode+0x239/0x290 [i915]
<4> [319.983407] Modules linked in: fuse snd_hda_codec_hdmi i915 x86_pkg_temp_thermal mei_hdcp coretemp wmi_bmof
r8153_ecm cdc_ether kvm_intel usbnet r8152 mii kvm prime_numbers snd_hda_intel ttm snd_intel_dspcfg irqbypass
drm_buddy e1000e crct10dif_pclmul snd_hda_codec crc32_pclmul drm_display_helper ptp snd_hwdep ghash_clmulni_intel
snd_hda_core drm_kms_helper pps_core mei_me syscopyarea video i2c_i801 snd_pcm sysfillrect i2c_smbus sysimgblt mei
fb_sys_fops intel_lpss_pci wmi
<4> [319.983483] CPU: 10 PID: 268 Comm: kworker/10:1H Not tainted 6.0.0-rc7-CI_DRM_12200-g394e575b57e9+ #1
<4> [319.983486] Hardware name: Intel Corporation Alder Lake Client Platform/AlderLake-P LP5 RVP, BIOS
ADLPFWI1.R00.2313.A00.2107301001 07/30/2021
<4> [319.983488] Workqueue: events_highpri heartbeat [i915]
<4> [319.983536] RIP: 0010:intel_tc_port_sanitize_mode+0x239/0x290 [i915]
<4> [319.983600] Code: 85 d2 75 03 48 8b 17 48 89 14 24 e8 e1 dc 2d e1 48 8b 14 24 48 c7 c1 f8 db 5b a0 48 c7 c7 3e
3c 5e a0 48 89 c6 e8 45 d7 66 e1 <0f> 0b e9 20 fe ff ff 0f 0b 49 c7 c0 8b 3c 5e a0 e9 9e fe ff ff 48
<4> [319.983601] RSP: 0018:ffffc90001617a30 EFLAGS: 00010286
<4> [319.983604] RAX: 0000000000000000 RBX: ffff88811f9d2000 RCX: 0000000000000001
<4> [319.983606] RDX: 0000000080000001 RSI: ffffffff8231e8cd RDI: 00000000ffffffff
<4> [319.983607] RBP: ffff888121e98000 R08: 0000000000000000 R09: c0000000ffffc134
<4> [319.983608] R10: 00000000000d6078 R11: ffffc900016178c8 R12: ffff88811f9d3838
<4> [319.983609] R13: ffff88811f9d397d R14: ffff888121e98000 R15: 0000000000000000
<4> [319.983611] FS:  0000000000000000(0000) GS:ffff8882a7300000(0000) knlGS:0000000000000000
<4> [319.983612] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<4> [319.983613] CR2: 00007fe7397f1e18 CR3: 0000000006612003 CR4: 0000000000770ee0
<4> [319.983615] PKRU: 55555554
<4> [319.983616] Call Trace:
<4> [319.983617]  <TASK>
<4> [319.983621]  intel_ddi_sync_state+0x3f/0x90 [i915]
<4> [319.983698]  intel_modeset_setup_hw_state+0x3a3/0x1440 [i915]
<4> [319.983777]  ? intel_gt_reset_global+0xeb/0x160 [i915]
<4> [319.983839]  ? __intel_display_resume+0x15/0xe0 [i915]
<4> [319.983909]  __intel_display_resume+0x15/0xe0 [i915]
<4> [319.983979]  intel_display_finish_reset+0x58/0x130 [i915]
<4> [319.984048]  intel_gt_reset_global+0xf3/0x160 [i915]
<4> [319.984107]  ? intel_reset_guc.cold.62+0x5d/0x5d [i915]
<4> [319.984189]  ? 0xffffffff81000000
<4> [319.984192]  ? queue_work_node+0x90/0x90
<4> [319.984202]  intel_gt_handle_error+0x2c2/0x410 [i915]
<4> [319.984267]  ? _raw_spin_unlock_irqrestore+0x54/0x70
<4> [319.984271]  ? lockdep_hardirqs_on+0xbf/0x140
<4> [319.984276]  ? intel_guc_find_hung_context+0x19e/0x1d0 [i915]
<4> [319.984352]  reset_engine+0x99/0xd0 [i915]
<4> [319.984399]  ? __drm_printfn_seq_file+0x20/0x20
<4> [319.984406]  heartbeat+0x4cd/0x4f0 [i915]
<4> [319.984454]  process_one_work+0x272/0x5b0
<4> [319.984461]  worker_thread+0x37/0x370
<4> [319.984465]  ? process_one_work+0x5b0/0x5b0
<4> [319.984467]  kthread+0xed/0x120
<4> [319.984470]  ? kthread_complete_and_exit+0x20/0x20
<4> [319.984474]  ret_from_fork+0x1f/0x30
<4> [319.984484]  </TASK>
<4> [319.984485] irq event stamp: 36107
<4> [319.984487] hardirqs last  enabled at (36113): [<ffffffff811391d6>] __up_console_sem+0x66/0x70
<4> [319.984492] hardirqs last disabled at (36118): [<ffffffff811391bb>] __up_console_sem+0x4b/0x70
<4> [319.984494] softirqs last  enabled at (34316): [<ffffffff81e00323>] __do_softirq+0x323/0x48e
<4> [319.984497] softirqs last disabled at (34309): [<ffffffff810c16b8>] irq_exit_rcu+0xb8/0xe0
<4> [319.984499] ---[ end trace 0000000000000000 ]---

v2:
- Instead of trying to fix the suspend/resume sequence, restore simply
  the state w/o the HW readout/sanitization step. (Ville)

References: https://lore.kernel.org/intel-gfx/20221005175251.3586272-1-imre.deak@intel.com/T/#mcfac180a67f6048096d09fa04347aa088291fafb
Closes: https://gitlab.freedesktop.org/drm/intel/issues/7021
Cc: Mika Kahola <mika.kahola@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221007133307.3805735-1-imre.deak@intel.com

drm/i915/display: Add DC5 counter and DMC debugfs entries for MTL

MTL and dgfx use the same DC5 counter.

While at it, this patch also adds the corresponding
debugfs entries. Some cleanup wrt dc3co register
which makes the code more readable.

Driver loads all firmware that it finds in the firmware
binary but platform doesn't *need* all of them. Cleaning the
previous debugs entries to reflect which firmware is needed
and if the needed firmware is loaded or not.

MTL needs both Pipe A and Pipe B DMC to be loaded
along with Main DMC.

BSpec: 49788
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Radhakrishna Sripada <radhakrishna.sripada@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/20221010202135.28388-1-anusha.srivatsa@intel.com

drm/i915: use proper helper for register updates

There is special helper for register read/modify/write.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221006163200.2803722-5-andrzej.hajda@intel.com

drm/i915: make intel_uncore_rmw() write unconditionally

Two small changes in intel_uncore_rmw will allow to use it more broadly:
- write register unconditionally, for use with latch registers,
- return old value of the register, IRQ cleanup and similar.

If we really want to keep write-only-if-changed feature maybe other
helper will be more suitable for it, intel_uncore_rmw name suggests
unconditional write.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221006163200.2803722-4-andrzej.hajda@intel.com

drm/i915/display: Use intel_uncore alias if defined

Alias is shorter and more readable.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221006163200.2803722-3-andrzej.hajda@intel.com

drm/i915/display: remove drm_device aliases

drm_device pointers are unwelcome.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221006204844.2831303-1-andrzej.hajda@intel.com

drm/i915: Enable atomic by default on ctg/elk

The watermark code for ctg/elk has been atomic ready for a long time
so let's just flip the switch now that some of the last CxSR issues
have been sorted out (which granted was a problem for vlv/chv as well
despite them already having atomic enabled by default).

v2: Rebase

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221007211108.3883-2-ville.syrjala@linux.intel.com
Acked-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Do the DRIVER_ATOMIC feature disable later

Currently we do the DRIVER_ATOMIC disable as almost the
first thing during pci probe. That involves the use of
DISPLAY_VER() which is perhaps a bit sketchy now that we
may need to read that out from the hardware itself.
Looks like we do populate a default value for it anyway
so the current does at least still work.

But let's make this safer anyway and move the code
into intel_device_info_runtime_init() where we also
handle the same thing for the !HAS_DISPLAY case.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221007211108.3883-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Write watermarks for disabled pipes on gmch platforms

We've excluded gmch platforms from writing the final watermarks
for any disabled pipe. IIRC the reason was perhaps some lingering
issue with the watermark merging across the pipes. But I can't
really see any reason for this anymore, so let's unify this behaviour.
The main benefit being more consistency in register dumps when
we don't have stale watermarks hanging around in the registers.
Functionally there should be no difference as the hardware just
ignore all of it when the pipe is disabled.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-9-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Fix pipe gamma enable/disable vs. CxSR on gmch platforms

Like most other plane control register bits, the pipe gamma
enable bit is also blocked by CxSR. So make sure we kick the
machine out of CxSR before trying to change that bit.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-8-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Fix g4x/vlv/chv CxSR vs. format/tiling/rotation changes

On g4x/vlv/chv the hardware seems incapable of changing the pixel
format, rotation, or YUV->RGB CSC matrix while in CxSR.

Additionally on VLV/CHV the sprites seem incapable of tiling
changes while in CxSR. On g4x CxSR is not even possible with
the sprite enabled. Curiously the primary plane seems perfectly
happy when changing tiling during CxSR.

Pimp up the code to account for these when determining whether
CxSR needs to be disabled. Since it looks like most of the plane
control register bits are affected let's just compare that.
But in the name of efficiency we'll make an exception for the
primary plane tiling changes (avoids some extra vblank waits).

v2: Just use the pre-computed plane control register values

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-7-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Clean up some namespacing

Rename a few functions from intel_crtc_foo_init() to
intel_foo_crtc_init() so that the namespaec clearly
indicates what feature/file we're talking about.

I left out intel_crtc_crc_init() because the whole crc
stuff uses intel_crtc_ as its namespace currently.

Suggested-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221004150929.23910-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/mtl: Add MTP ddc pin configuration

Meteorlake PCH reuses Alderlake vbt, DE pin mapping. Extend
ADL-P pin mapping for Meteorlake.

Bspec: 20124 does not have the mapping for MTP. Based on Bspec:49306, 64051,
it is concluded that MTP and ADL-P PCH have the same vbt -> DE pin pair mapping.

Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221005002534.2966978-1-radhakrishna.sripada@intel.com

drm/i915/mtl: Extend PSR support

Meteorlake and display 14 platform don't have any PSR differences
when comparing to Alderlake-P display, so it was only necessary to
extend some checks to properly program hardware.

BSpec: 55229, 49196

Cc: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220907081543.92268-1-mika.kahola@intel.com

drm/i915: Setup final panel drrs_type already during init

Now that we track the VBT drrs type per-panel we can move the
has_drrs_modes() check to the panel init rather than doing it
for every intel_panel_drrs_type() call.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-7-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Tighten DRRS capability reporting

Only report DRRS capability for the connector if its fixed_modes
list contains at least two modes capable of seamless DRRS switch
between them.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-6-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Fix locking in DRRS debugfs

Grab the crtc mutex so that looking at the crtc state is
actually safe.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-5-ville.syrjala@linux.intel.com

drm/i915: Make DRRS debugfs per-crtc/connector

Since I already broke anything that relied on the old contents
of the DRRS debugfs files might as well finish the breakage
and convert the files to be per-crtc/connector so we don't need
to have annoying code in igt to parse these.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-4-ville.syrjala@linux.intel.com

drm/i915: Make the DRRS debugfs contents more consistent

The stuff in the DRRS debugs is currently a hodgepode mix
of camelcase, lowercase, spaces, undescores, you name it.
Convert over to a reasonably common style.

Also move the busy bits thing to be the last sine it's
generally the least interesting thing in there.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-3-ville.syrjala@linux.intel.com

drm/i915: Move DRRS debugfs next to the implementation

Move the DRRS debugfs stuff next to the actual implementation
so that it's easier to deal with the whole.

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003113249.16213-2-ville.syrjala@linux.intel.com

drm/i915: Reject excessive dotclocks early

Make sure modes with crazy big dotclocks are rejected early,
so as to not cause problems for subsequent code via integer
overflows and whatnot.

These would eventually be rejected in intel_crtc_compute_pipe_mode()
but that is now too late as we do the clock computations a bit
earlier than that. And we don't want to just reorder the two since
we still want to check the final computed dotclock against the
hardware limit to make sure we didn't end up above the limit due
to rounding/etc.

Fixes: 0ff0e219d9b8 ("drm/i915: Compute clocks earlier")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927182455.3422-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Simplify modifier lookup in watermark code

Replace the huge modifier lists in the watermark code with
a few calls to intel_fb.c.

Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-7-ville.syrjala@linux.intel.com

drm/i915: Fix watermark calculations for DG2 CCS+CC modifier

Take the DG2 CCS+CC modifier into account when calculating the
watermarks. Othwerwise we'll calculate the watermarks thinking this
tile-4 modifier is linear.

The rc_surface part is actually a nop since that is not used
for any glk+ platform.

Cc: stable@vger.kernel.org
Fixes: 680025dcc400 ("drm/i915/dg2: Add support for DG2 clear color compression")
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-6-ville.syrjala@linux.intel.com

drm/i915: Fix watermark calculations for DG2 CCS modifiers

Take the DG2 CCS modifiers into account when calculating the
watermarks. Othwerwise we'll calculate the watermarks thinking these
tile-4 modifiers are linear.

The rc_surface part is actually a nop since that is not used
for any glk+ platform.

Cc: stable@vger.kernel.org
Fixes: 4c3afa72138c ("drm/i915/dg2: Add support for DG2 render and media compression")
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-5-ville.syrjala@linux.intel.com

drm/i915: Fix watermark calculations for gen12+ CCS+CC modifier

Take the gen12+ CCS+CC modifier into account when calculating the
watermarks. Othwerwise we'll calculate the watermarks thinking this
Y-tiled modifier is linear.

The rc_surface part is actually a nop since that is not used
for any glk+ platform.

Cc: stable@vger.kernel.org
Fixes: d1e2775e9b96 ("drm/i915/tgl: Add Clear Color support for TGL Render Decompression")
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-4-ville.syrjala@linux.intel.com

drm/i915: Fix watermark calculations for gen12+ MC CCS modifier

Take the gen12+ MC CCS modifier into account when calculating the
watermarks. Othwerwise we'll calculate the watermarks thinking this
Y-tiled modifier is linear.

The rc_surface part is actually a nop since that is not used
for any glk+ platform.

v2: Split RC CCS vs. MC CCS to separate patches

Cc: stable@vger.kernel.org
Fixes: 2dfbf9d2873a ("drm/i915/tgl: Gen-12 display can decompress surfaces compressed by the media engine")
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-3-ville.syrjala@linux.intel.com

drm/i915: Fix watermark calculations for gen12+ RC CCS modifier

Take the gen12+ RC CCS modifier into account when calculating the
watermarks. Othwerwise we'll calculate the watermarks thinking this
Y-tiled modifier is linear.

The rc_surface part is actually a nop since that is not used
for any glk+ platform.

v2: Split RC CCS vs. MC CCS to separate patches

Cc: stable@vger.kernel.org
Fixes: b3e57bccd68a ("drm/i915/tgl: Gen-12 render decompression")
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003111544.8007-2-ville.syrjala@linux.intel.com

drm/i915/psr: Fix PSR_IMR/IIR field handling

Current PSR code is supposed to use TRANSCODER_EDP to force 0 shift for
bits in PSR_IMR/IIR registers:

/*
 * gen12+ has registers relative to transcoder and one per transcoder
 * using the same bit definition: handle it as TRANSCODER_EDP to force
 * 0 shift in bit definition
 */

At the time of writing the code assumption "TRANSCODER_EDP == 0" was made.
This is not the case and all fields in PSR_IMR and PSR_IIR are shifted
incorrectly if DISPLAY_VER >= 12.

Fix this by adding separate register field defines for >=12 and add bit
getter functions to keep code readability.

v4:
 - Remove EDP from TGL definitions (José)
 - Use REG_BIT and REG_GENMASK (José)
v3:
 - Add separate register field defines (José)
 - Add bit getter functions (José)
v2:
 - Improve commit message (José)

Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Mika Kahola <mika.kahola@intel.com>
Fixes: 8241cfbe67f4 ("drm/i915/tgl: Access the right register when handling PSR interruptions")
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221003072011.72408-1-jouni.hogander@intel.com

drm/i915: Add some debug prints for intel_modeset_all_pipes()

Print out on which pipes, and for what reason, we are forcing a
full modeset.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220928060813.23264-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Change glk_load_degamma_lut() calling convention

Make glk_load_degamma_lut() more like most everyone else and
pass in the LUT explicitly.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220929071521.26612-6-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Clean up intel_color_init_hooks()

Remove a bunch of pointless curly brackets and do
the s/dev_priv/i915/ while at it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220929071521.26612-5-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Simplify the intel_color_init_hooks() if ladder

Get rid of the funny hsw vs. ivb extra indentation level in
intel_color_init_hooks().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220929071521.26612-4-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Split up intel_color_init()

intel_color_init() does both device level and crtc level stuff.
Split it up accordingly.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220929071521.26612-3-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Remove PLL asserts from .load_luts()

.load_luts() potentially runs from the vblank worker, and is
under a deadline to complete within the vblank. Thus we can't
do expesive stuff like talk to the Punit, etc.

To that end get rid of the assert_dsi_pll_enabled() call for
vlv/chv. We'll just have to trust that the PLL is already enabled
here.

And I don't think the normal assert_pll_enabled() really buys us
anything useful on gmch platforms either, so nuke that one too.
We don't have corresponding asserts in the ilk+ codepaths anyway
despite the hardware (IIRC) still requiring the clock to be
enabled when we access the LUT.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220929071521.26612-2-ville.syrjala@linux.intel.com
Acked-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Allow alternate fixed modes always for LVDS

As with eDP let's stop considering the VBTs DRRS knobs and
just always accept all otherwise suitable EDID modes. This
appears to be how Windows does it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927180615.25476-4-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Allow alternate fixed modes always for eDP

Stop considering VBT's static DRRS support when deciding whether
to use alternate fixed modes or not. It looks like Windows more
or less just uses that to decide whether to automagically switch
refresh rates on AC<->battery changes, or perhaps whether to
even expose a control for that in some UI thing. Either way it
seems happy to always use all EDID modes, and I guess the
DRRS/VRR stuff more or less adjusts how said modes get
actually used.

Let's do the same and just accept all the suitable looking
modes from EDID, whether we have DRRS or VRR.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6323
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6484
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927180615.25476-3-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Simplify intel_panel_add_edid_alt_fixed_modes()

Since commit a5810f551d0a ("drm/i915: Allow more varied alternate
fixed modes for panels") intel_panel_add_edid_alt_fixed_modes()
no longer considers vrr vs. drrs separately. So no reason to
pass them as separate parameters either.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927180615.25476-2-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/pxp: Add firmware status when ARB session fails

Add firmware status using a drm_warn when ARB session fails
or else a drm_dbg when the ARB session register slot bit did
get set.

Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Juston Li <justonli@chromium.org>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220923064542.415252-2-alan.previn.teres.alexis@intel.com

drm/i915: Round to closest in g4x+ HDMI clock readout

On pre-ddi platforms we have slightly different code being
used for HDMI TMDS clock to dotclock conversion between the
state computation and state readout. Both of these need to
round the same way in order to not get a mismatch between
the computed and read out states. Fix up the rounding
direction in the readout path to match what is used during
state computation.

Another option would to just use intel_crtc_dotclock()
in the readout path as well, but I don't really want to
do that as the current code more accurately represents
how the hardware really works; The HDMI port register
defines whether we're actually outputting 8bpc or 12bpc
over HDMI, and the PIPECONF bpc setting just defines what
goes over FDI between the CPU and PCH. The fact that we
try to cram all that into a single pipe_bpp during state
computation is perhaps not entirely great...

Fixes: f2c9df101095 ("drm/i915: Round TMDS clock to nearest")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220926193021.23287-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Fix TypeC mode initialization during system resume

During system resume DP MST requires AUX to be working already before
the HW state readout of the given encoder. Since AUX requires the
encoder/PHY TypeC mode to be initialized, which atm only happens during
HW state readout, these AUX transfers can change the TypeC mode
incorrectly (disconnecting the PHY for an enabled encoder) and trigger
the state check WARNs in intel_tc_port_sanitize().

Fix this by initializing the TypeC mode earlier both during driver
loading and system resume and making sure that the mode can't change
until the encoder's state is read out. While at it add the missing
DocBook comments and rename
intel_tc_port_sanitize()->intel_tc_port_sanitize_mode() for consistency.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Mika Kahola <mika.kahola@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922172148.2913088-1-imre.deak@intel.com

drm/i915: Nuke intel_get_shared_dpll_id()

Each PLL knows its own ID so intel_get_shared_dpll_id() is
pointless. Get rid of it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220921122343.13061-3-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Always initialize dpll.lock

Initialize the dll.lock mutex whether or not we manage to
initialize the rest of the dpll mgr.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220921122343.13061-2-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: WARN if PLL ref/unref got messed up

Spew a WARN if we try to ref/unref the same DPLL multiple
times for the same pipe.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220921211525.10675-5-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Pimp DPLL ref/unref debugs

We currently have a debug message in intel_reference_shared_dpll()
but no counterpart in intel_unreference_shared_dpll(). Add one.

Switch to the [CRTC:...] notation for the pipe name while at it.

v2: Use drm_dbg_kms() instead of drm_dbg() (Jani)

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922191350.4303-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Don't bail early from intel_dp_initial_fastset_check()

Do all the checks in intel_dp_initial_fastset_check() instead
of bailing out on the first condition that triggers.

This makes for better debug logs since we see all the reasons
why the full modeset computation is forced.

Also avoid the risk of someone accidentally adding a check
later in the function that would require connectors_changed=true
(ie. no fastset at all), but an earlier check may have already
bailed out with just mode_changed=true (ie. fastset is still
possible).

Pimp the debugs with the encoder id+name while at it.

v2: Call the return variable 'fastset' to convey its meaning

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922191314.4252-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Force DPLL calculation for TC ports after readout

We always allocate two DPLLs (TC and TBT) for TC ports. This
is because we can't know ahead of time wherher we need to put
the PHY into DP-Alt or TBT mode.

However during readout we can obviously only read out the state
of the DPLL that the port is actually using. Thus the state after
readout will not have both DPLLs populated.

We run into problems if during readout the TC port is in DP-Alt
mode, but we then perform a modeset on the port without going
through the full .compute_config() machinery, and during said
modeset the port cannot be switched back into DP-Alt mode and
we need to take the TBT fallback path. Such a modeset can
happen eg. due to cdclk reprogramming.

This wasn't a problem earlier because we did all the DPLL
calculations much later in the modeset. So even if flagged
a modeset very late we'd still have gone through the DPLL
calculations. But now all the DPLL calculations happen much
earlier and so we need to deal with it, or else we'll attempt
a modeset without a DPLL.

To guarantee that we always have both DPLLs fully cal/ulated
for TC ports force a full modeset computation during the
initial commit.

v2: Avoid bitwise operation on bool (Jani)
    Call the return variable 'fastset' to convey its meaning

Reported-by: Lee Shawn C <shawn.c.lee@intel.com>
Fixes: b000abd3b3d2 ("drm/i915: Do .crtc_compute_clock() earlier")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922191236.4194-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: fix device info for devices without display

Commit 00c6cbfd4e8a ("drm/i915: move pipe_mask and cpu_transcoder_mask
to runtime info") moved the pipe_mask member from struct
intel_device_info to intel_runtime_info, but overlooked some of our
platforms initializing device info .display = {}. This is significant,
as pipe_mask is the single point of truth for a device having a display
or not; the platforms in question left pipe_mask to whatever was set for
the platforms they "inherit" from in the complex macro scheme we have.

Add new NO_DISPLAY macro initializing .__runtime.pipe_mask = 0, which
will cause the device info .display sub-struct to be zeroed in
intel_device_info_runtime_init(). A better solution (or simply audit of
proper use of HAS_DISPLAY() checks) is required before moving forward
with [1].

Also clear all the display related members in runtime info if there's no
display. The latter is a bit tedious, but it's for completeness at this
time, to ensure similar functionality as before.

[1] https://lore.kernel.org/r/dfda1bf67f02ceb07c280b7a13216405fd1f7a34.1660137416.git.jani.nikula@intel.com

Fixes: 00c6cbfd4e8a ("drm/i915: move pipe_mask and cpu_transcoder_mask to runtime info")
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Maarten Lankhort <maarten.lankhorst@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Acked-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916082642.3451961-1-jani.nikula@intel.com

drm/i915: Add missing invalidate to g4x wm readout

Let's not forget to mark the unused watermark levels as invalid
after the readout. The vlv/chv codepath has this but the g4x
didn't for some reason.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-6-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Simplify up vlv watermark sanitation

We can simplify the vlv watermark sanitation by reusing the
second half of vlv_compute_pipe_wm() to convert the sanitized
raw watermarks into the proper form to be used as the
optimal/intermediate watermarks.

Also to be consistent with normal watermark computation the sanitized
watermarks should be all 0 for any disabled plane. Previously we
zeroed out the watermarks only up to the level (ie. PM2/5/DVDFS)
that was enabled.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-5-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Simplify up g4x watermark sanitation

We can simplify the g4x watermark sanitation by reusing the
second half of g4x_compute_pipe_wm() to convert the sanitized
raw watermarks into the proper form to be used as the
optimal/intermediate watermarks.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-4-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>

drm/i915: Split vlv_compute_pipe_wm() into two

Split vlv_compute_pipe_wm() into two halves. The first half computes
the new raw watermarks, and the second half munges those up into real
watermarks for the particular pipe.

We can reuse the second half for watermark sanitation as well.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-3-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanslav.lisovskiy@intel.com>

drm/i915: Split g4x_compute_pipe_wm() into two

Split g4x_compute_pipe_wm() into two halves. The first half computes
the new raw watermarks, and the second half munges those up into real
watermarks for the particular pipe.

We can reuse the second half for watermark sanitation as well.

Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220622155452.32587-2-ville.syrjala@linux.intel.com

drm/i915: Don't init eDP if we can't find a fixed mode

In the unlikely case of not finding a fixed mode don't register
the eDP connector. I think there are some places where we'd oops
if we didn't have a fixed mode for eDP so presumable this doesn't
typically happen. But better safe than sorry.

Also pimp the debugs with the encoder id+name. I think dumping
the encoder rather than the connector provides more information
here (eg. to match against the port information in the VBT).

We can also drop the extra check from intel_edp_add_properties().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-14-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Clean up connector->*_allowed setup

All the connectors are zero initialized so no need to clear
the *_allowed flags we don't support. Only leave the ones we want
to set. And while at it switch to booleans instead of ints.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-13-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Extract intel_attach_scaling_mode_property()

Consolidate the scaling_mode property setup into a single
place.

The one slight complicateion here is that GMCH platforms can't
do the CENTER mode except on the LVDS port. But we can deal with
that by just checking the connector type.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-12-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Move eDP scaling_mode prop setup to the proper place

We have an eDP specific intel_edp_add_properties() so move
the eDP scaling_mode property setup there from
intel_dp_add_properties().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-11-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Extract intel_lvds_add_properties()

Move the LVDS connector property setup to a dedicated
function to depollute intel_lvds_init() a bit.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-10-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Extract intel_dp_mst_add_properties()

Move the DP MST connector property setup into its own function
so it's not spread all over intel_dp_add_mst_connector().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-9-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Extract intel_tv_add_properties()

Pull all the TV connector property setup into its own neat function.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-8-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Extract intel_edp_backlight_setup()

Pull the eDP backlight setup into its own function. No
reason to pollute intel_edp_init_connector() with all
the mundane details.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-7-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Pass intel_encoder to to_lvds_encoder()

Most of our encoder type cast stuff already operates on
intel_encoder rather than drm_encoder. Switch to_lvds_encoder()
over as well.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-6-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Use BIT() when dealing with output types

Most places that deal with output types already use BIT()
but a few places still use manual shifts. Convert the
stragglers over to BIT().

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-5-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Drop pointless 'budget' variable

Just return the thing directly from the switch statement.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-4-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Clean up transcoder_to_stream_enc_status()

Drop the pointless middle man variable and just return the
correct thing directly.  And while at it change the return
type to u32 since this is a register value we're returning.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-3-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Drop pointless middle man variable

No need for the 'procmon' variable here. Just return the correct
thing from the switch statement directly.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912111814.17466-2-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915: Move hotplug inversion logic into separate helper

Use *_hpd_invert() helpers whenever possible to isolate logic specific
to hotplug inversion from common HPD setup logic to improve readability
and maintainability of the source code.

While we only define dg1_hpd_invert() here, future platforms are likely
to have different hotplug inversion needs, thus it makes sense grouping
different implementations under a common suffix.

v2: Fix coding style and prefer to use small *_hdp_invert() helpers
    instead of a generic one.

CC: Jani Nikula <jani.nikula@linux.intel.com>
CC: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922164949.163985-1-gustavo.sousa@intel.com

drm/i915: Remove unused function parameter

The function parameter 'exclude' in funciton
i915_sw_fence_await_reservation() is not used.
Remove it.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922213916.12112-1-niranjana.vishwanathapura@intel.com

drm/i915: Parse and set stepping for platforms with GMD

Expand the current stepping convention to accommodate the GMD
stepping info. Typically GMD step maps to letter stepping
by "A + step %4" and number to "A + step /4" i.e, GMD step
0 maps to STEP_A0, 1 to _A1, 2 to _A2, 3 to _A3, 4 to STEP_B0...

Future platforms might break this formulae and may require a table
mapping to decode GMD step compatible with the convention.

v2:
 - Pass the updated ip version structure
v3:
 - Skip using GMD to step table(MattR)

Cc: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916014648.1310346-3-radhakrishna.sripada@intel.com

drm/i915: Read graphics/media/display arch version from hw

Going forward, the hardware teams no longer consider new platforms to
have a "generation" in the way we've defined it for past platforms.
Instead, each IP block (graphics, media, display) will have their own
architecture major.minor versions and stepping ID's which should be read
directly from a register in the MMIO space.

Bspec: 63361, 64111

v2:
  - Move the IP version readout to intel_device_info.c
  - Convert the macro into a function

v3:
  - Move subplatform init to runtime early init
  - Cache runtime ver, release info to compare with hardware values.
  - Use IP_VER for snaity check(MattR)

v4:
  - Minor doccumentation changes.
  - Normalize HAS_GMD_ID macro value.(JaniN)

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916014648.1310346-2-radhakrishna.sripada@intel.com

drm/i915/fbc: Remove stale FIXME

Remove the old tales about 90/270 degree rotation
effectively preventing FBC. That hasn't been true since
we stopped demanding the fence is present in
commit 691f7ba58d52 ("drm/i915/display/fbc: Make fences
a nice-to-have for GEN9+")

Also fix up the multiline comment formatting while at it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916165206.1499-2-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coeho <luciano.coelho@intel.com>

drm/i915: Nuke stale plane cdclk ratio FIXMEs

The plane ratio stuff got implemented in
commit bb6ae9e653dc ("drm/i915: Allow planes to
declare their minimum acceptable cdclk") so these
FIXMEs have no business being here.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916165206.1499-1-ville.syrjala@linux.intel.com
Reviewed-by: Luca Coelho <luciano.coelho@intel.com>

drm/i915/fbc: Use lockdep_assert_held()

Replace the mutex_is_locked() stuff with lockdep_assert_held() since
that's what it's there for.

Also sprinkle these around so that we have more or less
mirrored coverage for the enable vs. disable instead of the
current situation where the asserts seem to be more or less
randomly thrown around.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220620182917.10765-2-ville.syrjala@linux.intel.com
Reviewed-by: Mika Kahola <mika.kahola@intel.com>

drm/i915/fbc: Move flip_pending assignmnt

Move the flip_pending assignment into __intel_fbc_post_update()
from intel_fbc_post_update(). Now mirrors the pre_update() side.

The only reason the assignment was in the higher level function
is that we used to call __intel_fbc_post_update() from elsewhere
as well. That got cleaned up in commit b39d2c620242 ("drm/i915/fbc:
Call intel_fbc_activate() directly from frontbuffer flush")

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220620182917.10765-1-ville.syrjala@linux.intel.com
Reviewed-by: Mika Kahola <mika.kahola@intel.com>

drm/i915: WARN if a port should use VBT provided vswing tables

We don't parse the VBT vswing/preemphassis tables at all currently.
Let's WARN if a port wants to use them so we get a heads up that
whether we really need to implement this stuff or not. My
current stash contains no VBTs with this bit set.

v2: Move to print_ddi_port() (Jani)

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916204132.10469-1-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/display: remove ipc_enabled from struct drm_i915_private

The ipc_enabled member was supposed to be moved under the display wm
sub-struct, but due to a rebase fail only the new one was added and the
old one was left behind. Finish the job.

Fixes: 70296670f672 ("drm/i915/display: move IPC under display wm sub-struct")
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916113850.3712354-1-jani.nikula@intel.com

drm/i915/hotplug: refactor hotplug init slightly

Rename intel_hpd_init_work() to the more generic intel_hpd_init_early(),
and move the hotplug storm initialization there. This lets us move the
HPD_STORM_DEFAULT_THRESHOLD macro to intel_hotplug.c too.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916130634.3781122-2-jani.nikula@intel.com

drm/i915/hotplug: move hotplug storm debugfs to intel_hotplug.c

The debugfs should be where the implementation details are.

v2: Rebase

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916130634.3781122-1-jani.nikula@intel.com

drm/i915: Fix return type of mode_valid function hook

All of the functions used for intel_dvo_dev_ops.mode_valid have a return
type of enum drm_mode_status, but the mode_valid field in the struct
definition has a return type of int.

The mismatched return type breaks forward edge kCFI since the underlying
function definitions do not match the function hook definition.

The return type of the mode_valid field should be changed from int to
enum drm_mode_status.

Reported-by: Dan Carpenter <error27@gmail.com>
Link: https://github.com/ClangBuiltLinux/linux/issues/1703
Cc: llvm@lists.linux.dev
Signed-off-by: Nathan Huckleberry <nhuck@google.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913205531.155046-1-nhuck@google.com

drm/i915/mtl: Update CHICKEN_TRANS* register addresses

In Display version 14, Transcoder Chicken Registers have updated address.
This patch performs checks to use the right register when required.

v2: Omit display version check in i915_reg.h(Jani)
v3:
 - Remove extra whitespace introduced
 - Fix reg definitions for MTL_CHICKEN_TRANS(MattR)

Bspec: 34387, 50054
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Madhumitha Tolakanahalli Pradeep <madhumitha.tolakanahalli.pradeep@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913183341.908028-6-radhakrishna.sripada@intel.com

drm/i915/mtl: Update MBUS_DBOX credits

Display version 14 platforms have different credits values
compared to ADL-P. Update the credits based on pipe usage.

v2: Simplify DBOX BW Credit definition(MattR)
v3:
 - Simplify only pipe per dbuf bank check(MattR)
 - Skip modeset check to handle the case when a new pipe within
   dbuf bank gets added/removed.(MattR)

Bspec: 49213

Cc: Jose Roberto de Souza <jose.souza@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Original Author: Caz Yokoyama
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913183341.908028-5-radhakrishna.sripada@intel.com

drm/i915/dsc: convert dsc debugfs entry from output_bpp to input_bpc

Convert dsc debugfs entry from output_bpp to input_bpc. The rationale
is to validate different input bpc across various platforms.

v2: -improved commit message (Jani N)
    -styling fixes (Jani N)

Signed-off-by: Swati Sharma <swati2.sharma@intel.com>
Reviewed-by: Manasi Navare <manasi.d.navare@intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Uma Shankar <uma.shankar@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220902190658.9459-2-swati2.sharma@intel.com

drm/i915/ipc: use intel_uncore_rmw() to enable/disable

Don't duplicate the rmw function.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/113a17cd18401b0e4c83396575b67aa6efb07346.1662983005.git.jani.nikula@intel.com

drm/i915/display: move IPC under display wm sub-struct

Move display IPC related member under drm_i915_private display
sub-struct.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/04ccaaceee9293e5a6c75761ba9d36792c36f095.1662983005.git.jani.nikula@intel.com

drm/i915/ipc: register debugfs only if IPC available

It looks like trying to enable IPC via debugfs on platforms that don't
have IPC resulted in dmesg info message about IPC being enabled, which
is clearly not possible and would not happen.

Seems sensible to register IPC debugfs only on platforms that have IPC.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/b18edb4f96c9d2ec728ef04e6f99d161fe5641d1.1662983005.git.jani.nikula@intel.com

drm/i915/ipc: move IPC debugfs to skl_watermark.c

Follow the new direction for debugfs files, moving the details where the
implementation is. It seems quite natural skl_watermark.c is the place
that controls IPC details, even for debugfs, not
intel_display_debugfs.c.

Rename the functions and convert dev_priv->i915 while at it.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/9d59b18f5dc06e86a48c1ce0f40d625f51e3e85a.1662983005.git.jani.nikula@intel.com

drm/i915/ipc: refactor and rename IPC functions

Rename the IPC functions to have skl_watermark_ipc_ prefix, rename
enable to update to reflect what the function actually does, and add
enabled function to abstract direct ->ipc_enabled access for state
query.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/536237d5bc919e8c97a96796f235f5bb264ceff2.1662983005.git.jani.nikula@intel.com