drm/i915/gen12: Apply recommended L3 hashing mask

The TGL/RKL/DG1/ADL performance tuning guide suggests programming a
literal value of 0x2FC0100F for this register.  The register's hardware
default value is 0x2FC0108F, so this translates to just clearing one
bit.

Take this opportunity to also clean up the register definition and
re-write its existing bits/fields in the preferred notation.

Bspec: 31870
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221201222210.344152-1-matthew.d.roper@intel.com

drm/i915/mtl: Check full IP version when applying hw steering semaphore

When determining whether the platform has a hardware-level steering
semaphore (i.e., MTL and beyond), we need to use GRAPHICS_VER_FULL() to
compare the full version rather than just the major version number
returned by GRAPHICS_VER().

Reported-by: kernel test robot <lkp@intel.com>
Fixes: 3100240bf846 ("drm/i915/mtl: Add hardware-level lock for steering")
Cc: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221202223528.714491-1-matthew.d.roper@intel.com

drm/i915/mtl: Add hardware-level lock for steering

Starting with MTL, the driver needs to not only protect the steering
control register from simultaneous software accesses, but also protect
against races with hardware/firmware agents.  The hardware provides a
dedicated locking mechanism to support this via the MTL_STEER_SEMAPHORE
register.  Reading the register acts as a 'trylock' operation; the read
will return 0x1 if the lock is acquired or 0x0 if something else is
already holding the lock; once acquired, writing 0x1 to the register
will release the lock.

We'll continue to grab the software lock as well, just so lockdep can
track our locking; assuming the hardware lock is behaving properly,
there should never be any contention on the software lock in this case.

v2:
 - Extend hardware semaphore timeout and add a taint for CI if it ever
   happens (this would imply misbehaving hardware/firmware).  (Mika)
 - Add "MTL_" prefix to new steering semaphore register.  (Mika)

Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221128233014.4000136-5-matthew.d.roper@intel.com

drm/i915/mcr: Hold forcewake and MCR lock over PPAT setup

PPAT setup involves a series of multicast writes.  This can be optimized
slightly be acquiring forcewake and the steering lock just once for the
entire sequence.

v2:
 - We should use FW_REG_WRITE instead of FW_REG_READ.  (Bala)

Suggested-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221130155852.19601-1-matthew.d.roper@intel.com

drm/i915/pvc: Implement recommended caching policy

As per the performance tuning guide, set the HOSTCACHEEN bit to
implement the recommended caching policy on PVC.

Signed-off-by: Wayne Boyer <wayne.boyer@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221130170723.2460014-1-wayne.boyer@intel.com

drm/i915/guc: Use GuC submission API version number

The GuC firmware includes an extra version number to specify the
submission API level. So use that rather than the main firmware
version number for submission related checks.

Also, while it is guaranteed that GuC version number components are
only 8-bits in size, other firmwares do not have that restriction. So
stop making assumptions about them generically fitting in a u16
individually, or in a u32 as a combined 8.8.8.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221129232031.3401386-4-John.C.Harrison@Intel.com

drm/i915/uc: More refactoring of UC version numbers

As a precursor to a coming change (for adding a GuC submission API
version), abstract the UC version number into its own private
structure separate to the firmware filename.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221129232031.3401386-3-John.C.Harrison@Intel.com

drm/i915/uc: Rationalise delimiters in filename macros

The way delimiters (underscores and dots) were added to the UC
filenames was different for different types of delimiter. Rationalise
them to all be done the same way - implicitly in the concatenation
macro rather than explicitly in the file name prefix.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221129232031.3401386-2-John.C.Harrison@Intel.com

drm/i915/gt: Add dedicated MCR lock

We've been overloading uncore->lock to protect access to the MCR
steering register.  That's not really what uncore->lock is intended for,
and it would be better if we didn't need to hold such a high-traffic
spinlock for the whole sequence of (apply steering, access MCR register,
restore steering).  Let's create a dedicated MCR lock to protect the
steering control register over this critical section and stop relying on
the high-traffic uncore->lock.

For now the new lock is a software lock.  However some platforms (MTL
and beyond) have a hardware-provided locking mechanism that can be used
to serialize not only software accesses, but also hardware/firmware
accesses as well; support for that hardware level lock will be added in
a future patch.

v2:
 - Use irqsave/irqrestore spinlock calls; platforms using execlist
   submission rather than GuC submission can perform MCR accesses in
   interrupt context because reset -> errordump happens in a tasklet.

Cc: Chris Wilson <chris.p.wilson@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221128233014.4000136-4-matthew.d.roper@intel.com

drm/i915/gt: Pass gt rather than uncore to lowest-level reads/writes

Passing the GT rather than uncore to the lowest level MCR read and write
functions will make it easier to introduce dedicated MCR locking in a
following patch.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221128233014.4000136-3-matthew.d.roper@intel.com

drm/i915/gt: Correct kerneldoc for intel_gt_mcr_wait_for_reg()

The kerneldoc function name was not updated when this function was
converted to a non-fw form.

Fixes: 192bb40f030a ("drm/i915/gt: Manage uncore->lock while waiting on MCR register")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221128233014.4000136-2-matthew.d.roper@intel.com

drm/i915/huc: always init the delayed load fence

The fence is only tracking if the HuC load is in progress or not and
doesn't distinguish between already loaded, not supported or disabled,
so we can always initialize it to completed, no matter the actual
support. We already do that for most platforms, but we skip it on
GTs that lack VCS engines (e.g. MTL root GT), so fix that. Note that the
cleanup is already unconditional.

While at it, move the init/fini to helper functions.

Fixes: 02224691cb0f ("drm/i915/huc: fix leak of debug object in huc load fence on driver unload")
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221123235417.1475709-1-daniele.ceraolospurio@intel.com

drm/i915/mtl: Media GT and Render GT share common GGTT

On XE_LPM+ platforms the media engines are carved out into a separate
GT but have a common GGTMMADR address range which essentially makes
the GGTT address space to be shared between media and render GT. As a
result any updates in GGTT shall invalidate TLB of GTs sharing it and
similarly any operation on GGTT requiring an action on a GT will have to
involve all GTs sharing it. setup_private_pat was being done on a per
GGTT based as that doesn't touch any GGTT structures moved it to per GT
based.

BSPEC: 63834

v2:
1. Add details to commit msg
2. includes fix for failure to add item to ggtt->gt_list, as suggested
by Lucas
3. as ggtt_flush() is used only for ggtt drop i915_is_ggtt check within
it.
4. setup_private_pat moved out of intel_gt_tiles_init

v3:
1. Move out for_each_gt from i915_driver.c (Jani Nikula)

v4: drop using RCU primitives on ggtt->gt_list as it is not an RCU list
(Matt Roper)

Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221122070126.4813-1-aravind.iddamsetty@intel.com

drm/i915/dg2: Introduce Wa_18019271663

Wa_18019271663 applies to all DG2 steppings and skus.

Bspec: 66622

Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221123183648.407058-2-matthew.s.atwood@intel.com

drm/i915/dg2: Introduce Wa_18018764978

Wa_18018764978 applies to specific steppings of DG2 (G10 C0+,
G11 and G12 A0+). Clean up style in function at the same time.

Bspec: 66622

Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221123183648.407058-1-matthew.s.atwood@intel.com

drm/i915: Never return 0 if not all requests retired

Users of intel_gt_retire_requests_timeout() expect 0 return value on
success.  However, we have no protection from passing back 0 potentially
returned by a call to dma_fence_wait_timeout() when it succedes right
after its timeout has expired.

Replace 0 with -ETIME before potentially using the timeout value as return
code, so -ETIME is returned if there are still some requests not retired
after timeout, 0 otherwise.

v3: Use conditional expression, more compact but also better reflecting
    intention standing behind the change.

v2: Move the added lines down so flush_submission() is not affected.

Fixes: f33a8a51602c ("drm/i915: Merge wait_for_timelines with retire_request")
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: stable@vger.kernel.org # v5.5+
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221121145655.75141-3-janusz.krzysztofik@linux.intel.com

drm/i915: Fix negative value passed as remaining time

Commit b97060a99b01 ("drm/i915/guc: Update intel_gt_wait_for_idle to work
with GuC") extended the API of intel_gt_retire_requests_timeout() with an
extra argument 'remaining_timeout', intended for passing back unconsumed
portion of requested timeout when 0 (success) is returned.  However, when
request retirement happens to succeed despite an error returned by a call
to dma_fence_wait_timeout(), that error code (a negative value) is passed
back instead of remaining time.  If we then pass that negative value
forward as requested timeout to intel_uc_wait_for_idle(), an explicit BUG
will be triggered.

If request retirement succeeds but an error code is passed back via
remaininig_timeout, we may have no clue on how much of the initial timeout
might have been left for spending it on waiting for GuC to become idle.
OTOH, since all pending requests have been successfully retired, that
error code has been already ignored by intel_gt_retire_requests_timeout(),
then we shouldn't fail.

Assume no more time has been left on error and pass 0 timeout value to
intel_uc_wait_for_idle() to give it a chance to return success if GuC is
already idle.

v3: Don't fail on any error passed back via remaining_timeout.

v2: Fix the issue on the caller side, not the provider.

Fixes: b97060a99b01 ("drm/i915/guc: Update intel_gt_wait_for_idle to work with GuC")
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Cc: stable@vger.kernel.org # v5.15+
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221121145655.75141-2-janusz.krzysztofik@linux.intel.com

drm/i915/uc: Fix table order verification to check all FW types

It was noticed that the table order verification step was only being
run once rather than once per firmware type. Fix that.

Note that the long term plan is to convert this code to be a mock
selftest. It is already only compiled in when selftests are enabled.
And the work involved in the conversion was estimated to be
non-trivial. So that conversion is currently low on the priority list.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221122233328.854217-1-John.C.Harrison@Intel.com

drm/i915/huc: fix leak of debug object in huc load fence on driver unload

The fence is always initialized in huc_init_early, but the cleanup in
huc_fini is only being run if HuC is enabled. This causes a leaking of
the debug object when HuC is disabled/not supported, which can in turn
trigger a warning if we try to register a new debug offset at the same
address on driver reload.

To fix the issue, make sure to always run the cleanup code.

Reported-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reported-by: Brian Norris <briannorris@chromium.org>
Fixes: 27536e03271d ("drm/i915/huc: track delayed HuC load with a fence")
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Brian Norris <briannorris@chromium.org>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Tested-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221111005651.4160369-1-daniele.ceraolospurio@intel.com

drm/i915/guc: make default_lists const data

The default_lists array should be in rodata.

Fixes: dce2bd542337 ("drm/i915/guc: Add Gen9 registers for GuC error state capture.")
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221122141616.3469214-1-jani.nikula@intel.com

drm/i915/gsc: Only initialize GSC in tile 0

For multi-tile setups the GSC operational only on the tile 0.
Skip GSC auxiliary device creation for all other tiles
in GSC device init code.
Initialize basic GSC fields and use the same path
as HECI1 (HECI_PXP) device disable.

Cc: Tomas Winkler <tomas.winkler@intel.com>
Cc: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Reviewed-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221121092449.328674-1-alexander.usyskin@intel.com

drm/i915/selftest: Bump up sample period for busy stats selftest

Engine busyness samples around a 10ms period is failing with busyness
ranging approx. from 87% to 115% as shown below. The expected range is
+/- 5% of the sample period. Fail 10% of the time.

rcs0: reported 11716042ns [91%] busyness while spinning [for 12805719ns]

When determining busyness of active engine, the GuC based engine
busyness implementation relies on a 64 bit timestamp register read. The
latency incurred by this register read causes the failure.

On DG1, when the test fails, the observed latencies range from 900us -
1.5ms.

Optimizing the 2x32 read by acquiring the lock and forcewake prior to
all reg reads reduces the rate of failure to around 2%, but does not
eliminate it.

In order to make the selftest more robust and always account for such
latencies, increase the sample period to 100 ms. This eliminates the
issue as seen in a 1000 runs.

v2: (Ashutosh)
- Add error to commit msg
- Include gitlab bug
- Update commit for inclusion of 2x32 optimized read

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4418
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221110171913.670286-3-umesh.nerlige.ramappa@intel.com

i915/uncore: Acquire fw before loop in intel_uncore_read64_2x32

PMU reads the GT timestamp as a 2x32 mmio read and since upper and lower
32 bit registers are read in a loop, there is a latency involved between
getting the GT timestamp and the CPU timestamp. As part of the
resolution, refactor intel_uncore_read64_2x32 to acquire forcewake and
uncore lock prior to reading upper and lower regs.

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221110171913.670286-2-umesh.nerlige.ramappa@intel.com

drm/i915/mtl: Enable Idle Messaging for GSC CS

By defaut idle messaging is disabled for GSC CS so to unblock RC6
entry on media tile idle messaging need to be enabled.

v2:
 - Fix review comments (Vinay)
 - Set GSC idle hysteresis as per spec (Badal)
v3:
 - Fix review comments (Rodrigo)

Bspec: 71496

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221118183354.1047829-1-badal.nilawar@intel.com

drm/i915: Fix workarounds on Gen2-3

In 3653727560d0 ("drm/i915: Simplify internal helper function signature")
I broke the old platforms by not noticing engine workaround init does not
initialize the list on old platforms. Fix it by always initializing which
already does the right thing by mostly not doing anything if there aren't
any workarounds on the list.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 3653727560d0 ("drm/i915: Simplify internal helper function signature")
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221118115249.2683946-1-tvrtko.ursulin@linux.intel.com

drm/i915/gt: Manage uncore->lock while waiting on MCR register

The GT MCR code currently relies on uncore->lock to avoid race
conditions on the steering control register during MCR operations.  The
*_fw() versions of MCR operations expect the caller to already hold
uncore->lock, while the non-fw variants manage the lock internally.
However the sole callsite of intel_gt_mcr_wait_for_reg_fw() does not
currently obtain the forcewake lock, allowing a potential race condition
(and triggering an assertion on lockdep builds).  Furthermore, since
'wait for register value' requests may not return immediately, it is
undesirable to hold a fundamental lock like uncore->lock for the entire
wait and block all other MMIO for the duration; rather the lock is only
needed around the MCR read operations and can be released during the
delays.

Convert intel_gt_mcr_wait_for_reg_fw() to a non-fw variant that will
manage uncore->lock internally.  This does have the side effect of
causing an unnecessary lookup in the forcewake table on each read
operation, but since the caller is still holding the relevant forcewake
domain, this will ultimately just incremenent the reference count and
won't actually cause any additional MMIO traffic.

In the future we plan to switch to a dedicated MCR lock to protect the
steering critical section rather than using the overloaded and
high-traffic uncore->lock; on MTL and beyond the new lock can be
implemented on top of the hardware-provided synchonization mechanism for
steering.

Fixes: 3068bec83eea ("drm/i915/gt: Add intel_gt_mcr_wait_for_reg_fw()")
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221117173358.1980230-1-matthew.d.roper@intel.com

drm/i915/mtl: C6 residency and C state type for MTL SAMedia

Add support for C6 residency and C state type for MTL SAMedia. Also add
mtl_drpc.

v2: Fixed review comments (Ashutosh)
v3: Sort registers and fix whitespace errors in intel_gt_regs.h (Matt R)
    Remove MTL_CC_SHIFT (Ashutosh)
    Adapt to RC6 residency register code refactor (Jani N)
v4: Move MTL branch to top in drpc_show
v5: Use FORCEWAKE_MT identical to gen6_drpc (Ashutosh)
v6: Add MISSING_CASE for gt_core_status switch statement (Rodrigo)
    Change state name for MTL_CC0 to C0 (from "on") (Rodrigo)
v7: Change state name for MTL_CC0 to RC0 (Rodrigo)

Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221114123348.3474216-6-badal.nilawar@intel.com

drm/i915/gt: Use RC6 residency types as arguments to residency functions

Previously RC6 residency functions directly accepted RC6 residency register
MMIO offsets (there are four RC6 residency registers). This worked but
required an assumption on the residency register layout so was not future
proof.

Therefore change RC6 residency functions to accept RC6 residency types
instead of register MMIO offsets. The knowledge of register offsets as well
as ID to offset mapping is now maintained solely in intel_rc6 and can be
tailored for different platforms and different register layouts as need
arises.

v2: Address review comments by Jani N
    - Change residency functions to accept RC6 residency types instead of
      register ID's
    - s/intel_rc6_print_rc5_res/intel_rc6_print_residency/
    - Remove "const enum" in function arguments
    - Naming: intel_rc6_* for enum
    - Use INTEL_RC6_RES_MAX and other minor changes
v3: Don't include intel_rc6_types.h in intel_rc6.h (Jani)

Suggested-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Suggested-by: Jani Nikula <jani.nikula@linux.intel.com>
Reported-by: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221114123348.3474216-5-badal.nilawar@intel.com

drm/i915/mtl: Modify CAGF functions for MTL

Update CAGF functions for MTL to get actual resolved frequency of 3D and
SAMedia.

v2: Update MTL_MIRROR_TARGET_WP1 position/formatting (MattR)
    Move MTL branches in cagf functions to top (MattR)
    Fix commit message (Andi)
v3: Added comment about registers not needing forcewake for Gen12+ and
    returning 0 freq in RC6
v4: Use REG_FIELD_GET and uncore (Rodrigo)

Bspec: 66300

Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221114123348.3474216-4-badal.nilawar@intel.com

drm/i915: Use GEN12_RPSTAT register for GT freq

On GEN12+ use GEN12_RPSTAT register to get actual resolved GT
freq. GEN12_RPSTAT does not require a forcewake and will return 0 freq if
GT is in RC6.

v2:
  - Fixed review comments(Ashutosh)
  - Added function intel_rps_read_rpstat_fw to read RPSTAT without
    forcewake, required especially for GEN6_RPSTAT1 (Ashutosh, Tvrtko)
v3:
  - Updated commit title and message for more clarity (Ashutosh)
  - Replaced intel_rps_read_rpstat with direct read to GEN12_RPSTAT1 in
    read_cagf (Ashutosh)
v4: Remove GEN12_CAGF_SHIFT and use REG_FIELD_GET (Rodrigo)

Cc: Don Hiatt <donhiatt@gmail.com>
Cc: Andi Shyti <andi.shyti@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221114123348.3474216-3-badal.nilawar@intel.com

drm/i915/rps: Prefer REG_FIELD_GET in intel_rps_get_cagf

Instead of masks/shifts settle on REG_FIELD_GET as the standard way to
extract reg fields. This allows future patches touching this code to also
consistently use REG_FIELD_GET and friends.

Suggested-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221114123348.3474216-2-badal.nilawar@intel.com

drm/i915/guc: add the GSC CS to the GuC capture list

For the GSC engine we only want to capture the instance regs.

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221111001832.4144910-1-daniele.ceraolospurio@intel.com

drm/i915/pxp: Separate PXP FW interface structures for both v42 and 43

Previously, we only used PXP FW interface version-42 structures for
PXP arbitration session on ADL/TGL products and version-43 for HuC
authentication on DG2. That worked fine despite not differentiating such
versioning of the PXP firmware interaction structures. This was okay
back then because the only commands used via version 42 was not
used via version 43 and vice versa.

With MTL, we'll need both these versions side by side for the same
commands (PXP-session) with the older platform feature support. That
said, let's create separate files to define the structures and definitions
for both version-42 and 43 of PXP FW interfaces.

Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108045628.4187260-2-alan.previn.teres.alexis@intel.com

drm/i915/ttm: never purge busy objects

In i915_gem_madvise_ioctl() we immediately purge the object is not
currently used, like when the mm.pages are NULL.  With shmem the pages
might still be hanging around or are perhaps swapped out. Similarly with
ttm we might still have the pages hanging around on the ttm resource,
like with lmem or shmem, but here we need to be extra careful since
async unbinds are possible as well as in-progress kernel moves. In
i915_ttm_purge() we expect the pipeline-gutting to nuke the ttm resource
for us, however if it's busy the memory is only moved to a ghost object,
which then leads to broken behaviour when for example clearing the
i915_tt->filp, since the actual ttm_tt is still alive and populated,
even though it's been moved to the ghost object.  When we later destroy
the ghost object we hit the following, since the filp is now NULL:

[  +0.006982] #PF: supervisor read access in kernel mode
[  +0.005149] #PF: error_code(0x0000) - not-present page
[  +0.005147] PGD 11631d067 P4D 11631d067 PUD 115972067 PMD 0
[  +0.005676] Oops: 0000 [#1] PREEMPT SMP NOPTI
[  +0.012962] Workqueue: events ttm_device_delayed_workqueue [ttm]
[  +0.006022] RIP: 0010:i915_ttm_tt_unpopulate+0x3a/0x70 [i915]
[  +0.005879] Code: 89 fb 48 85 f6 74 11 8b 55 4c 48 8b 7d 30 45 31 c0 31 c9 e8 18 6a e5 e0 80 7d 60 00 74 20 48 8b 45 68
8b 55 08 4c 89 e7 5b 5d <48> 8b 40 20 83 e2 01 41 5c 89 d1 48 8b 70
 30 e9 42 b2 ff ff 4c 89
[  +0.018782] RSP: 0000:ffffc9000bf6fd70 EFLAGS: 00010202
[  +0.005244] RAX: 0000000000000000 RBX: ffff8883e12ae380 RCX: 0000000000000000
[  +0.007150] RDX: 000000008000000e RSI: ffffffff823559b4 RDI: ffff8883e12ae3c0
[  +0.007142] RBP: ffff888103b65d48 R08: 0000000000000001 R09: 0000000000000001
[  +0.007144] R10: 0000000000000001 R11: ffff88829c2c8040 R12: ffff8883e12ae3c0
[  +0.007148] R13: 0000000000000001 R14: ffff888115184140 R15: ffff888115184248
[  +0.007154] FS:  0000000000000000(0000) GS:ffff88844db00000(0000) knlGS:0000000000000000
[  +0.008108] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  +0.005763] CR2: 0000000000000020 CR3: 000000013fdb4004 CR4: 00000000003706e0
[  +0.007152] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  +0.007145] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  +0.007154] Call Trace:
[  +0.002459]  <TASK>
[  +0.002126]  ttm_tt_unpopulate.part.0+0x17/0x70 [ttm]
[  +0.005068]  ttm_bo_tt_destroy+0x1c/0x50 [ttm]
[  +0.004464]  ttm_bo_cleanup_memtype_use+0x25/0x40 [ttm]
[  +0.005244]  ttm_bo_cleanup_refs+0x90/0x2c0 [ttm]
[  +0.004721]  ttm_bo_delayed_delete+0x235/0x250 [ttm]
[  +0.004981]  ttm_device_delayed_workqueue+0x13/0x40 [ttm]
[  +0.005422]  process_one_work+0x248/0x560
[  +0.004028]  worker_thread+0x4b/0x390
[  +0.003682]  ? process_one_work+0x560/0x560
[  +0.004199]  kthread+0xeb/0x120
[  +0.003163]  ? kthread_complete_and_exit+0x20/0x20
[  +0.004815]  ret_from_fork+0x1f/0x30

v2:
 - Just use ttm_bo_wait() directly (Niranjana)
 - Add testcase reference

Testcase: igt@gem_madvise@dontneed-evict-race
Fixes: 213d50927763 ("drm/i915/ttm: Introduce a TTM i915 gem object backend")
Reported-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Cc: <stable@vger.kernel.org> # v5.15+
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Acked-by: Nirmoy Das <Nirmoy.Das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221115104620.120432-1-matthew.auld@intel.com

drm/i915: Remove unwanted ghost obj check

vm_fault_ttm() should not expect ttm ghost obj so remove that check.

Suggested-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221024144558.27747-1-nirmoy.das@intel.com

drm/i915/selftests: add igt_vma_move_to_active_unlocked

All calls to i915_vma_move_to_active are surrounded by vma lock
and/or there are multiple local helpers for it in particular tests.
Let's replace it by common helper.
The patch should not introduce functional changes.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019215906.295296-3-andrzej.hajda@intel.com

drm/i915: call i915_request_await_object from _i915_vma_move_to_active

Since almost all calls to i915_vma_move_to_active are prepended with
i915_request_await_object, let's call the latter from
_i915_vma_move_to_active by default and add flag allowing bypassing it.
Adjust all callers accordingly.
The patch should not introduce functional changes.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221019215906.295296-2-andrzej.hajda@intel.com

drm/i915: Update workaround documentation

There were several updates in the driver on how the workarounds are
handled since its documentation was written. Update the documentation to
reflect the current reality.

v2:
  - Remove footnote that was wrongly referenced, adding back the
    reference in the correct paragraph.
  - Remove "Display workarounds" and just mention "display IP" under
    "Other" category since all of them are peppered around the driver.

Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com> # v1
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221115192611.179981-1-lucas.demarchi@intel.com

drm/i915: Fix vma allocator debug

Add a missing colon which I accidentally removed in the recent logging
changes.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: a10234fda466 ("drm/i915: Partial abandonment of legacy DRM logging macros")
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221115101730.394880-1-tvrtko.ursulin@linux.intel.com

Documentation/gpu: Fix section in the wrong scope

That section should still be inside "DRM client usage stats" rather than
as a sibling.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221107173209.2219571-2-lucas.demarchi@intel.com

drm/i915/guc: handle interrupts from media GuC

The render and media GuCs share the same interrupt enable register, so
we can no longer disable interrupts when we disable communication for
one of the GuCs as this would impact the other GuC. Instead, we keep the
interrupts always enabled in HW and use a variable in the GuC structure
to determine if we want to service the received interrupts or not.

v2: use MTL_ prefix for reg definition (Matt)

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-7-daniele.ceraolospurio@intel.com

drm/i915/guc: define media GT GuC send regs

The media GT shares the G-unit with the root GT, so a second set of
communication registers is required for the media GuC.

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-6-daniele.ceraolospurio@intel.com

drm/i915/mtl: Handle wopcm per-GT and limit calculations.

With MTL standalone media architecture the wopcm layout has changed,
with separate partitioning in WOPCM for the root GT GuC and the media
GT GuC. The size of WOPCM is 4MB with the lower 2MB reserved for the
media GT and the upper 2MB for the root GT.

Given that MTL has GuC deprivilege, the WOPCM registers are pre-locked
by the bios. Therefore, we can skip all the math for the partitioning
and just limit ourselves to sanity-checking the values.

v2: fix makefile file ordering (Jani)
v3: drop XELPM_SAMEDIA_WOPCM_SIZE, check huc instead of VDBOX (John)
v4: further clarify commit message, remove blank line (John)

Signed-off-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-5-daniele.ceraolospurio@intel.com

drm/i915/guc: Add GuC deprivilege feature to MTL

MTL supports GuC deprivilege. Add the feature flag to this platform.

Signed-off-by: Stuart Summers <stuart.summers@intel.com>
Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-4-daniele.ceraolospurio@intel.com

drm/i915/uc: use different ggtt pin offsets for uc loads

Our current FW loading process is the same for all FWs:

- Pin FW to GGTT at the start of the ggtt->uc_fw node
- Load the FW
- Unpin

This worked because we didn't have a case where 2 FWs would be loaded on
the same GGTT at the same time. On MTL, however, this can happen if both
GTs are reset at the same time, so we can't pin everything in the same
spot and we need to use separate offset. For simplicity, instead of
calculating the exact required size, we reserve a 2MB slot for each fw.

v2: fail fetch if FW is > 2MBs, improve comments (John)
v3: more comment improvements (John)

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-3-daniele.ceraolospurio@intel.com

drm/i915/uc: fetch uc firmwares for each GT

The FW binaries are independently loaded on each GT. On MTL, the memory
is shared so we could potentially re-use a single allocation, but on
discrete multi-gt platforms we are going to need independent copies,
so it is easier to do the same on MTL as well, given that the amount
of duplicated memory is relatively small (~500K).

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-2-daniele.ceraolospurio@intel.com

drm/i915/huc: only load HuC on GTs that have VCS engines

On MTL the primary GT doesn't have any media capabilities, so no video
engines and no HuC. We must therefore skip the HuC fetch and load on
that specific case. Given that other multi-GT platforms might have HuC
on the primary GT, we can't just check for that and it is easier to
instead check for the lack of VCS engines.

Based on code from Aravind Iddamsetty

v2: clarify which engine_mask is used for each GT and why (Tvrtko)

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Cc: John Harrison <john.c.harrison@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108020600.3575467-1-daniele.ceraolospurio@intel.com

drm/i915: Simplify internal helper function signature

Since we are now storing the GT backpointer in the wa list we can drop the
explicit struct intel_gt * argument to wa_list_apply.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221110124633.3135026-1-tvrtko.ursulin@linux.intel.com

drm/i915/mtl: Add Wa_14017073508 for SAMedia

This workaround is added for Media tile of MTL A step. It is to help
pcode workaround which handles the hardware issue seen during package C2/C3
transitions due to RC6 entry/exit transitions on Media tile. As a part of
workaround pcode expect kmd to send mailbox message "media busy" when
components of Media tile are in use and "media idle" otherwise.
As per workaround description gucrc need to be disabled so enabled
host based RC for Media tile.

v2:
 - Correct workaround id (Matt)
 - Fix review comments (Rodrigo)

Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Cc: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Badal Nilawar <badal.nilawar@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/20221103184559.2306481-1-badal.nilawar@intel.com

drm/i915/guc/slpc: Add selftest for slpc tile-tile interaction

Run a workload on tiles simultaneously by requesting for RP0 frequency.
Pcode can however limit the frequency being granted due to throttling
reasons. This test checks if there is any throttling but does not fail
if RP0 is not granted due to throttle reasons

v2: Fix build error
v3: Use IS_ERR_OR_NULL to check worker
    Addressed cosmetic review comments (Tvrtko)
v4: do not skip test on media engines if gt type is GT_MEDIA.
    Use correct PERF_LIMIT_REASONS register for MTL (Vinay)

Signed-off-by: Riana Tauro <riana.tauro@intel.com>
Reviewed-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221109112541.275021-2-riana.tauro@intel.com

drm/i915/perf: Fix kernel-doc warning

Fix kernel-doc issue from a previous commit.

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Fixes: 2db609c01495 ("drm/i915/perf: Replace gt->perf.lock with stream->lock for file ops")
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221107202410.1976895-1-umesh.nerlige.ramappa@intel.com

drm/i915: Partial abandonment of legacy DRM logging macros

Convert some usages of legacy DRM logging macros into versions which tell
us on which device have the events occurred.

v2:
 * Don't have struct drm_device as local. (Jani, Ville)

v3:
 * Store gt, not i915, in workaround list. (John)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221109104633.2579245-1-tvrtko.ursulin@linux.intel.com

drm/i915: use i915_sg_dma_sizes() for all backends

We rely on page_sizes.sg in setup_scratch_page() reporting the correct
value if the underlying sgl is not contiguous, however in
get_pages_internal() we are only looking at the layout of the created
pages when calculating the sg_page_sizes, and not the final sgl, which
could in theory be completely different. In such a situation we might
incorrectly think we have a 64K scratch page, when it is actually only
4K or similar split over multiple non-contiguous entries, which could
lead to broken behaviour when touching the scratch space within the
padding of a 64K GTT page-table. For most of the other backends we
already just call i915_sg_dma_sizes() on the final mapping, so rather
just move that into __i915_gem_object_set_pages() to avoid such issues
coming back to bite us later.

v2: Update missing conversion in gvt

Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Stuart Summers <stuart.summers@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221108103238.165447-1-matthew.auld@intel.com

drm/i915/ttm: add some sanity checks for lmem_userfault_list

Rather than getting some hard to debug uaf, add some warns to hopefully
catch issues with userfault_count being non-zero when destroying the
object. Also if we somehow add an object to lmem_userfault_list that
somehow doesn't map lmem.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/7469
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221107165414.56970-2-matthew.auld@intel.com

drm/i915/ttm: fix uaf with lmem_userfault_list handling

In the fault handler, make sure we check if the BO maps lmem after
we schedule the migration, since the current resource might change from
lmem to smem, if the pages are in the non-cpu visible portion of lmem.
This then leads to adding the object to the lmem_userfault_list even
though the current resource is no longer lmem. If we then destroy the
object, the list might still contain a link to the now free object, since
we only remove it if the object is still in lmem.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7469
Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221107165414.56970-1-matthew.auld@intel.com

drm/i915/pxp: use <> instead of "" for headers in include/

Headers in include/ should be included using the system header #include
syntax.

Fixes: 887a193b4fb1 ("drm/i915/pxp: add huc authentication and loading command")
Cc: Tomas Winkler <tomas.winkler@intel.com>
Cc: Vitaly Lubart <vitaly.lubart@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221107140454.2680954-1-jani.nikula@intel.com

drm/i915/guc: don't hardcode BCS0 in guc_hang selftest

On MTL there are no BCS engines on the media GT, so we can't always use
BCS0 in the test. There is no actual reason to use a BCS engine over an
engine of a different class, so switch to using any available engine.

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Acked-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102214310.2829310-1-daniele.ceraolospurio@intel.com

drm/i915/mtl: don't expose GSC command streamer to the user

There is no userspace user for this CS yet, we only need it for internal
kernel ops (e.g. HuC, PXP), so don't expose it.

v2: even if it's not exposed, rename the engine so it is easier to
identify in the debug logs (Matt)

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102171047.2787951-6-daniele.ceraolospurio@intel.com

drm/i915/mtl: add GSC CS reset support

The GSC CS has its own dedicated bit in the GDRST register.

Bspec: 52549
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102171047.2787951-5-daniele.ceraolospurio@intel.com

drm/i915/mtl: add GSC CS interrupt support

The GSC CS re-uses the same interrupt bits that the GSC used in older
platforms. This means that we can now have an engine interrupt coming
out of OTHER_CLASS, so we need to handle that appropriately.

v2: clean up the if statement for the engine irq (Tvrtko)

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102171047.2787951-4-daniele.ceraolospurio@intel.com

drm/i915/mtl: pass the GSC CS info to the GuC

We need to tell the GuC that the GSC CS is there.

Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102171047.2787951-3-daniele.ceraolospurio@intel.com

drm/i915/mtl: add initial definitions for GSC CS

Starting on MTL, the GSC is no longer managed with direct MMIO access,
but we instead have a dedicated command streamer for it. As a first step
for adding support for this CS, add the required definitions.
Note that, although it is now a CS, the GSC retains its old
class:instance value (OTHER_CLASS instance 6)

Bspec: 65308, 45605
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102171047.2787951-2-daniele.ceraolospurio@intel.com

drm/i915/guc: Don't deadlock busyness stats vs reset

The engine busyness stats has a worker function to do things like
64bit extend the 32bit hardware counters. The GuC's reset prepare
function flushes out this worker function to ensure no corruption
happens during the reset. Unforunately, the worker function has an
infinite wait for active resets to finish before doing its work. Thus
a deadlock would occur if the worker function had actually started
just as the reset starts.

The function being used to lock the reset-in-progress mutex is called
intel_gt_reset_trylock(). However, as noted it does not follow
standard 'trylock' conventions and exit if already locked. So rename
the current _trylock function to intel_gt_reset_lock_interruptible(),
which is the behaviour it actually provides. In addition, add a new
implementation of _trylock and call that from the busyness stats
worker instead.

v2: Rename existing trylock to interruptible rather than trying to
preserve the existing (confusing) naming scheme (review comments from
Tvrtko).

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102192109.2492625-3-John.C.Harrison@Intel.com

drm/i915/guc: Properly initialise kernel contexts

If a context has already been registered prior to first submission
then context init code was not being called. The noticeable effect of
that was the scheduling priority was left at zero (meaning super high
priority) instead of being set to normal. This would occur with
kernel contexts at start of day as they are manually pinned up front
rather than on first submission. So add a call to initialise those
when they are pinned.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102192109.2492625-2-John.C.Harrison@Intel.com

drm/i915/guc: Remove excessive line feeds in state dumps

Some of the GuC state dump messages were adding extra line feeds. When
printing via a DRM printer to dmesg, for example, that messes up the
log formatting as it loses any prefixing from the printer. Given that
the extra line feeds are just in the middle of random bits of GuC
state, there isn't any real need for them. So just remove them
completely.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031220007.4176835-1-John.C.Harrison@Intel.com

drm/i915/userptr: restore probe_range behaviour

The conversion looks harmless, however the addr value is updated inside
the loop with the previous vm_end, which then incorrectly leads to
for_each_vma_range() iterating over stuff outside the range we care
about. Fix this by storing the end value separately. Also fix the case
where the range doesn't intersect with any vma, or if the vma itself
doesn't extend the entire range, which must mean we have hole at the
end. Both should result in an error, as per the previous behaviour.

v2: Fix the cases where the range is empty, or if there's a hole at
the end of the range

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7247
Testcase: igt@gem_userptr_blits@probe
Fixes: f683b9d61319 ("i915: use the VMA iterator")
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Liam R. Howlett <Liam.Howlett@Oracle.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Yu Zhao <yuzhao@google.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028130635.465839-1-matthew.auld@intel.com

Merge drm/drm-next into drm-intel-gt-next

Needed to bring in v6.1-rc1 which contains commit f683b9d61319 ("i915: use the VMA iterator")
which is needed for series https://patchwork.freedesktop.org/series/110083/ .

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Merge tag 'drm-intel-gt-next-2022-11-03' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

Driver Changes:

- Fix for #7306: [Arc A380] white flickering when using arc as a
  secondary gpu (Matt A)
- Add Wa_18017747507 for DG2 (Wayne)
- Avoid spurious WARN on DG1 due to incorrect cache_dirty flag
  (Niranjana, Matt A)
- Corrections to CS timestamp support for Gen5 and earlier (Ville)

- Fix a build error used with clang compiler on hwmon (GG)
- Improvements to LMEM handling with RPM (Anshuman, Matt A)
- Cleanups in dmabuf code (Mike)

- Selftest improvements (Matt A)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Y2N11wu175p6qeEN@jlahtine-mobl.ger.corp.intel.com

Merge tag 'drm-misc-next-2022-11-03' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for 6.2:

UAPI Changes:

Cross-subsystem Changes:
- dma-buf: locking improvements
- firmware: New API in the RaspberryPi firmware driver used by vc4

Core Changes:
- client: Null pointer dereference fix in drm_client_buffer_delete()
- mm/buddy: Add back random seed log
- ttm: Convert ttm_resource to use size_t for its size, fix for an
  undefined behaviour

Driver Changes:
- bridge:
  - adv7511: use dev_err_probe
  - it6505: Fix return value check of pm_runtime_get_sync
- panel:
  - sitronix: Fixes and clean-ups
- lcdif: Increase DMA burst size
- rockchip: runtime_pm improvements
- vc4: Fix for a regression preventing the use of 4k @ 60Hz, and
  further HDMI rate constraints check.
- vmwgfx: Cursor improvements

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20221103083437.ksrh3hcdvxaof62l@houat

drm/i915/selftests: Reduce oversaturation of request smoketesting

The goal in launching the request smoketest is to have sufficient tasks
running across the system such that we are likely to detect concurrency
issues. We aim to have 2 tasks using the same engine, gt, device (each
level of locking around submission and signaling) running at the same
time. While tasks may not be running all the time as they synchronise
with the gpu, they will be running most of the time, in which case
having many more tasks than cores available is wasteful (and
dramatically increases the workload causing excess runtime). Aim to
limit the number of tasks such that there is at least 2 running per
engine, spreading surplus cores around the engines (rather than running
a task per core per engine.)

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Tested-by: Nirmoy Das <nirmoy.das@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221102155709.31717-1-nirmoy.das@intel.com

drm/i915/hwmon: Fix a build error used with clang compiler

Use REG_FIELD_PREP() and a constant value for hwm_field_scale_and_write()

If the first argument of FIELD_PREP() is not a compile-time constant value
or unsigned long long type, this routine of the __BF_FIELD_CHECK() macro
used internally by the FIELD_PREP() macro always returns false.

 BUILD_BUG_ON_MSG(__bf_cast_unsigned(_mask, _mask) >      \
                  __bf_cast_unsigned(_reg, ~0ull),        \
                  _pfx "type of reg too small for mask"); \

And it returns a build error by the option among the clang
compilation options. [-Werror,-Wtautological-constant-out-of-range-compare]

Reported build error while using clang compiler:

drivers/gpu/drm/i915/i915_hwmon.c:115:16: error: result of comparison of
constant 18446744073709551615 with expression of type 'typeof (_Generic((field_msk),
char: (unsigned char)0, unsigned char: (unsigned char)0, signed char: (unsigned char)0,
unsigned short: (unsigned short)0, short: (unsigned short)0, unsigned int:
(unsigned int)0, int: (unsigned int)0, unsigned long: (unsigned long)0, long:
(unsigned long)0, unsigned long long: (unsigned long long)0, long long:
(unsigned long long)0, default: (field_msk)))' (aka 'unsigned int') is always false
[-Werror,-Wtautological-constant-out-of-range-compare]
        bits_to_set = FIELD_PREP(field_msk, nval);
                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/bitfield.h:114:3: note: expanded from macro 'FIELD_PREP'
                __BF_FIELD_CHECK(_mask, 0ULL, _val, "FIELD_PREP: ");    \
                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/bitfield.h:71:53: note: expanded from macro '__BF_FIELD_CHECK'
                BUILD_BUG_ON_MSG(__bf_cast_unsigned(_mask, _mask) >     \
                ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~
./include/linux/build_bug.h:39:58: note: expanded from macro 'BUILD_BUG_ON_MSG'
                                    ~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~
./include/linux/compiler_types.h:357:22: note: expanded from macro 'compiletime_assert'
        _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
        ~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/compiler_types.h:345:23: note: expanded from macro '_compiletime_assert'
        __compiletime_assert(condition, msg, prefix, suffix)
        ~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
./include/linux/compiler_types.h:337:9: note: expanded from macro '__compiletime_assert'
                if (!(condition))                                       \

v2: Use REG_FIELD_PREP() macro instead of FIELD_PREP() (Jani)

Fixes: 99f55efb7911 ("drm/i915/hwmon: Power PL1 limit and TDP setting")
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Cc: Andi Shyti <andi.shyti@linux.intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Jani Nikula <jani.nikula@intel.com>
[Joonas: Wrapped commit message error line length to be more reasonable]
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221029044230.32128-1-gwan-gyeong.mun@intel.com

drm/i915: Do not set cache_dirty for DGFX

Currently on DG1, which does not have LLC, we hit the below
warning while rebinding an userptr invalidated object.

WARNING: CPU: 4 PID: 13008 at drivers/gpu/drm/i915/gem/i915_gem_pages.c:34 __i915_gem_object_set_pages+0x296/0x2d0 [i915]
...
RIP: 0010:__i915_gem_object_set_pages+0x296/0x2d0 [i915]
...
Call Trace:
 <TASK>
 i915_gem_userptr_get_pages+0x175/0x1a0 [i915]
 ____i915_gem_object_get_pages+0x32/0xb0 [i915]
 i915_gem_object_userptr_submit_init+0x286/0x470 [i915]
 eb_lookup_vmas+0x2ff/0xcf0 [i915]
 ? __intel_wakeref_get_first+0x55/0xb0 [i915]
 i915_gem_do_execbuffer+0x785/0x21d0 [i915]
 i915_gem_execbuffer2_ioctl+0xe7/0x3d0 [i915]

We shouldn't be setting the obj->cache_dirty for DGFX,
fix it.

Fixes: d70af57944a1 ("drm/i915/shmem: ensure flush during swap-in on non-LLC")
Suggested-by: Matthew Auld <matthew.auld@intel.com>
Reported-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Acked-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@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/20221102051416.27327-1-niranjana.vishwanathapura@intel.com

drm/tests: Add back seed value information

As reported by Michał, the drm_mm and drm_buddy unit tests lost the
printk with seed value after they were refactored into KUnit.

Add kunit_info with seed value information to assure reproducibility.

Reported-by: Michał Winiarski <michal.winiarski@intel.com>
Signed-off-by: Arthur Grillo <arthurgrillo@riseup.net>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028221755.340487-1-arthurgrillo@riseup.net

drm/client: Prevent NULL dereference in drm_client_buffer_delete()

The drm_gem_vunmap() will crash with a NULL dereference if the passed
object pointer is NULL. It wasn't a problem before we added the locking
support to drm_gem_vunmap function because the mapping argument was always
NULL together with the object. Make drm_client_buffer_delete() to check
whether GEM is NULL before trying to unmap the GEM, it will happen on
framebuffer creation error.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://lore.kernel.org/dri-devel/Y1kFEGxT8MVlf32V@kili/
Fixes: 79e2cf2e7a19 ("drm/gem: Take reservation lock for vmap/vunmap operations")
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221030154412.8320-3-dmitry.osipenko@collabora.com

dma-buf: Make locking consistent in dma_buf_detach()

The dma_buf_detach() locks attach->dmabuf->resv and then unlocks
dmabuf->resv, which could be a two different locks from a static
code checker perspective. In particular this triggers Smatch to
report the "double unlock" error. Make the locking pointers consistent.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://lore.kernel.org/dri-devel/Y1fLfsccW3AS%2Fo+%2F@kili/
Fixes: 809d9c72c2f8 ("dma-buf: Move dma_buf_attach() to dynamic locking specification")
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221030154412.8320-2-dmitry.osipenko@collabora.com

drm/ttm: fix undefined behavior in bit shift for TTM_TT_FLAG_PRIV_POPULATED

Shifting signed 32-bit value by 31 bits is undefined, so changing
significant bit to unsigned. The UBSAN warning calltrace like below:

UBSAN: shift-out-of-bounds in ./include/drm/ttm/ttm_tt.h:122:26
left shift of 1 by 31 places cannot be represented in type 'int'
Call Trace:
 <TASK>
 dump_stack_lvl+0x7d/0xa5
 dump_stack+0x15/0x1b
 ubsan_epilogue+0xe/0x4e
 __ubsan_handle_shift_out_of_bounds+0x1e7/0x20c
 ttm_bo_move_memcpy+0x3b4/0x460 [ttm]
 bo_driver_move+0x32/0x40 [drm_vram_helper]
 ttm_bo_handle_move_mem+0x118/0x200 [ttm]
 ttm_bo_validate+0xfa/0x220 [ttm]
 drm_gem_vram_pin_locked+0x70/0x1b0 [drm_vram_helper]
 drm_gem_vram_pin+0x48/0xb0 [drm_vram_helper]
 drm_gem_vram_plane_helper_prepare_fb+0x53/0xe0 [drm_vram_helper]
 drm_gem_vram_simple_display_pipe_prepare_fb+0x26/0x30 [drm_vram_helper]
 drm_simple_kms_plane_prepare_fb+0x4d/0xe0 [drm_kms_helper]
 drm_atomic_helper_prepare_planes+0xda/0x210 [drm_kms_helper]
 drm_atomic_helper_commit+0xc3/0x1e0 [drm_kms_helper]
 drm_atomic_commit+0x9c/0x160 [drm]
 drm_client_modeset_commit_atomic+0x33a/0x380 [drm]
 drm_client_modeset_commit_locked+0x77/0x220 [drm]
 drm_client_modeset_commit+0x31/0x60 [drm]
 __drm_fb_helper_restore_fbdev_mode_unlocked+0xa7/0x170 [drm_kms_helper]
 drm_fb_helper_set_par+0x51/0x90 [drm_kms_helper]
 fbcon_init+0x316/0x790
 visual_init+0x113/0x1d0
 do_bind_con_driver+0x2a3/0x5c0
 do_take_over_console+0xa9/0x270
 do_fbcon_takeover+0xa1/0x170
 do_fb_registered+0x2a8/0x340
 fbcon_fb_registered+0x47/0xe0
 register_framebuffer+0x294/0x4a0
 __drm_fb_helper_initial_config_and_unlock+0x43c/0x880 [drm_kms_helper]
 drm_fb_helper_initial_config+0x52/0x80 [drm_kms_helper]
 drm_fbdev_client_hotplug+0x156/0x1b0 [drm_kms_helper]
 drm_fbdev_generic_setup+0xfc/0x290 [drm_kms_helper]
 bochs_pci_probe+0x6ca/0x772 [bochs]
 local_pci_probe+0x4d/0xb0
 pci_device_probe+0x119/0x320
 really_probe+0x181/0x550
 __driver_probe_device+0xc6/0x220
 driver_probe_device+0x32/0x100
 __driver_attach+0x195/0x200
 bus_for_each_dev+0xbb/0x120
 driver_attach+0x27/0x30
 bus_add_driver+0x22e/0x2f0
 driver_register+0xa9/0x190
 __pci_register_driver+0x90/0xa0
 bochs_pci_driver_init+0x52/0x1000 [bochs]
 do_one_initcall+0x76/0x430
 do_init_module+0x61/0x28a
 load_module+0x1f82/0x2e50
 __do_sys_finit_module+0xf8/0x190
 __x64_sys_finit_module+0x23/0x30
 do_syscall_64+0x58/0x80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
 </TASK>

Fixes: 3312be8f6fc8 ("drm/ttm: move populated state into page flags")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031113350.4180975-1-cuigaosheng1@huawei.com
Signed-off-by: Christian König <christian.koenig@amd.com>

drm/i915/selftests: Run the perf MI_BB tests on gen4/5

Now that we know the ring timestamp frequency on gen4/5 we
can run the perf tests that depend on sampling the timestamp.

On g4x/ilk we must read the udw of the 64bit timestamp
register. Details in {g4x,gen5)_read_clock_frequency().

When executing the read via the CS i965 doesn't seem to need
the double read trick that CPU mmio reads need.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-7-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/i915/selftests: Test RING_TIMESTAMP on gen4/5

Now that we actually know the cs timestamp frequency on gen4/5
let's run the corresponding test.

On g4x/ilk we must read the udw of the 64bit timestamp
register. Details in {g4x,gen5)_read_clock_frequency().

The one extra caveat is that on i965 (or at least CL, don't
recall if I ever tested on BW) we must read the register
twice to get an up to date value. For some unknown reason
the first read tends to return a stale value.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-6-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/i915/selftests: Run MI_BB perf selftests on SNB

SNB does have the RING_TIMESTAMP register on the RCS engine.
Run the MI_BB perf tests on it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-5-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/i915: Fix cs timestamp frequency for cl/bw

Despite what the spec says the TIMESTAMP register seems to
tick once every hrawclk (confirmed on i965gm and g35).

v2: Rebase

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-4-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/i915: Stop claiming cs timestamp frquency on gen2/3

Gen2/3 have no TIMESTAMP registers to sample so no point in thinking
we have any frequency for it either.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-3-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/i915: Fix cs timestamp frequency for ctg/elk/ilk

On ilk the UDW of TIMESTAMP increments every 1000 ns,
LDW is mbz. In order to represent that we'd need 52 bits,
but we only have 32 bits. Even worse most things want to
only deal with 32 bits of timestamp. So let's just set
up the timestamp frequency as if we only had the UDW.

On ctg/elk 63:20 of TIMESTAMP increments every 1/4 ns, 19:0
are mbz. To make life simpler let's ignore the LDW and set up
timestamp frequency based on the UDW only (increments every
1024 ns).

v2: Rebase

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031135703.14670-2-ville.syrjala@linux.intel.com
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

drm/panel/panel-sitronix-st7701: Remove panel on DSI attach failure

In case mipi_dsi_attach() fails, call drm_panel_remove() to
avoid memory leak.

Fixes: 849b2e3ff969 ("drm/panel: Add Sitronix ST7701 panel driver")
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221014231106.468063-1-marex@denx.de

drm/panel/panel-sitronix-st7701: Clean up CMDnBKx selection

There are two command register files, CMD1 and CMD2, where only the CMD2
contains additional register sub-files BK0..3 . Pull the register file
selection call into separate function instead of duplicating it all over
the driver. The CMD2BK2 file is undocumented in datasheet, and is used
for BIST. No functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221014231042.468033-1-marex@denx.de

drm/i915/dg2: Introduce Wa_18017747507

WA 18017747507 applies to all DG2 skus.

BSpec: 56035, 46121, 68173

Signed-off-by: Wayne Boyer <wayne.boyer@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221031131509.3411195-1-wayne.boyer@intel.com

drm/panel/panel-sitronix-st7701: Fix RTNI calculation

The RTNI field is multiplied by 16 and incremented by 512 before being
used as the minimum number of pixel clock per horizontal line, hence
it is necessary to subtract those 512 bytes from htotal and then divide
the result by 16 before writing the value into the RTNI field. Fix the
calculation.

Fixes: de2b4917843c ("drm/panel/panel-sitronix-st7701: Infer horizontal pixel count from TFT mode")
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20221012221159.88397-1-marex@denx.de

drm: lcdif: change burst size to 256B

If a axi bus master with a higher priority do a lot of memory access
FIFO underruns can be inspected. Increase the burst size to 256B to
avoid such underruns and to improve the memory access efficiency.

Fixes: 9db35bb349a0 ("drm: lcdif: Add support for i.MX8MP LCDIF variant")
Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
Reviewed-by: Marek Vasut <marex@denx.de>
Signed-off-by: Marek Vasut <marex@denx.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221101164615.778299-1-m.felsch@pengutronix.de

Merge tag 'drm-intel-next-2022-10-28' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

- Hotplug code clean-up and organization (Jani, Gustavo)
- More VBT specific code clean-up, doc, organization,
  and improvements (Ville)
- More MTL enabling work (Matt, RK, Anusha, Jose)
- FBC related clean-ups and improvements (Ville)
- Removing unused sw_fence_await_reservation (Niranjana)
- Big chunch of display house clean-up (Ville)
- Many Watermark fixes and clean-ups (Ville)
- Fix device info for devices without display (Jani)
- Fix TC port PLLs after readout (Ville)
- DPLL ID clean-ups (Ville)
- Prep work for finishing (de)gamma readout (Ville)
- PSR fixes and improvements (Jouni, Jose)
- Reject excessive dotclocks early (Ville)
- DRRS related improvements (Ville)
- Simplify uncore register updates (Andrzej)
- Fix simulated GPU reset wrt. encoder HW readout (Imre)
- Add a ADL-P workaround (Jose)
- Fix clear mask in GEN7_MISCCPCTL update (Andrzej)
- Temporarily disable runtime_pm for discrete (Anshuman)
- Improve fbdev debugs (Nirmoy)
- Fix DP FRL link training status (Ankit)
- Other small display fixes (Ankit, Suraj)
- Allow panel fixed modes to have differing sync
  polarities (Ville)
- Clean up crtc state flag checks (Ville)
- Fix race conditions during DKL PHY accesses (Imre)
- Prep-work for cdclock squash and crawl modes (Anusha)
- ELD precompute and readout (Ville)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Y1wd6ZJ8LdJpCfZL@intel.com

drm/vmwgfx: Cleanup the cursor snooping code

Cursor snooping depended on implicit size and format which made debugging
quite difficult. Make the code easier to following by making everything
explicit and instead of using magic numbers predefine all the
parameters the code depends on.

Also fixes incorrectly computed pitches for non-aligned cursor snoops.
Fix which has no practical effect because non-aligned cursor snoops
are not used by the X11 driver and Wayland cursors will go through
mob cursors, instead of surface dma's.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Michael Banack <banackm@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221026031936.1004280-2-zack@kde.org

drm/vmwgfx: Validate the box size for the snooped cursor

Invalid userspace dma surface copies could potentially overflow
the memcpy from the surface to the snooped image leading to crashes.
To fix it the dimensions of the copybox have to be validated
against the expected size of the snooped cursor.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Fixes: 2ac863719e51 ("vmwgfx: Snoop DMA transfers with non-covering sizes")
Cc: <stable@vger.kernel.org> # v3.2+
Reviewed-by: Michael Banack <banackm@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221026031936.1004280-1-zack@kde.org

drm/i915/dmabuf: Use scatterlist for_each_sg API

Update open coded for loop to use the standard scatterlist
for_each_sg API.

Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028155029.494736-4-matthew.auld@intel.com

drm/i915/dmabuf: dmabuf cleanup

Some minor cleanup of some variables for consistency.

Normalize struct sg_table to sgt.
Normalize struct dma_buf_attachment to attach.
checkpatch issues sizeof(), !NULL updates.

Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028155029.494736-3-matthew.auld@intel.com

drm/i915/selftests: exercise GPU access from the importer

Using PAGE_SIZE here potentially hides issues so bump that to something
larger. This should also make it possible for iommu to coalesce entries
for us. With that in place verify we can write from the GPU using the
importers sg_table, followed by checking that our writes match when read
from the CPU side.

v2: Switch over to igt_gpu_fill_dw(), which looks to be more widely
supported than the migrate stuff (at least OOTB).

References: https://gitlab.freedesktop.org/drm/intel/-/issues/7306
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Michael J. Ruhl <michael.j.ruhl@intel.com>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028155029.494736-2-matthew.auld@intel.com

drm/i915/dmabuf: fix sg_table handling in map_dma_buf

We need to iterate over the original entries here for the sg_table,
pulling out the struct page for each one, to be remapped. However
currently this incorrectly iterates over the final dma mapped entries,
which is likely just one gigantic sg entry if the iommu is enabled,
leading to us only mapping the first struct page (and any physically
contiguous pages following it), even if there is potentially lots more
data to follow.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7306
Fixes: 1286ff739773 ("i915: add dmabuf/prime buffer sharing support.")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Michael J. Ruhl <michael.j.ruhl@intel.com>
Cc: <stable@vger.kernel.org> # v3.5+
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221028155029.494736-1-matthew.auld@intel.com

drm/i915/dgfx: Grab wakeref at i915_ttm_unmap_virtual

We had already grabbed the rpm wakeref at obj destruction path,
but it also required to grab the wakeref when object moves.
When i915_gem_object_release_mmap_offset() gets called by
i915_ttm_move_notify(), it will release the mmap offset without
grabbing the wakeref. We want to avoid that therefore,
grab the wakeref at i915_ttm_unmap_virtual() accordingly.

While doing that also changed the lmem_userfault_lock from
mutex to spinlock, as spinlock widely used for list.

Also changed if (obj->userfault_count) to
GEM_BUG_ON(!obj->userfault_count).

v2:
- Removed lmem_userfault_{list,lock} from intel_gt. [Matt Auld]

Fixes: ad74457a6b5a ("drm/i915/dgfx: Release mmap on rpm suspend")
Suggested-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221027092242.1476080-3-anshuman.gupta@intel.com

drm/i915: Encapsulate lmem rpm stuff in intel_runtime_pm

Runtime pm is not really per GT, therefore it make sense to
move lmem_userfault_list, lmem_userfault_lock and
userfault_wakeref from intel_gt to intel_runtime_pm structure,
which is embedded to i915.

No functional change.

v2:
- Fixes the code comment nit. [Matt Auld]

Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221027092242.1476080-2-anshuman.gupta@intel.com

drm/rockchip: lvds: fix PM usage counter unbalance in poweron

pm_runtime_get_sync will increment pm usage counter even it failed.
Forgetting to putting operation will result in reference leak here.
We fix it by replacing it with the newest pm_runtime_resume_and_get
to keep usage counter balanced.

Fixes: 34cc0aa25456 ("drm/rockchip: Add support for Rockchip Soc LVDS")
Fixes: cca1705c3d89 ("drm/rockchip: lvds: Add PX30 support")
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220922132107.105419-3-zhangqilong3@huawei.com

drm/rockchip: use pm_runtime_resume_and_get() instead of pm_runtime_get_sync()

Replace pm_runtime_get_sync() with pm_runtime_resume_and_get() to avoid
device usage counter leak.

Signed-off-by: Yuan Can <yuancan@huawei.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220615062644.96837-1-yuancan@huawei.com

drm/rockchip: dsi: Remove the unused function dsi_update_bits()

The function dsi_update_bits() is defined in the dw-mipi-dsi-rockchip.c
file, but not called elsewhere, so delete this unused function.

drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c:367:20: warning: unused function 'dsi_update_bits'.

https://bugzilla.openanolis.cn/show_bug.cgi?id=2414

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221017084330.94117-1-jiapeng.chong@linux.alibaba.com

drm/rockchip: vop2: Register Esmart0-win0 as primary plane

Esmart0-win0 could serve as primary plane, so mark it as such. On
RK3568 this window will never be used as primary plane, because the
three windows at the beginning of the rk3568_vop_win_data[] array
will be used. On RK3566 however, two of the windows at the beginning
of the rk3568_vop_win_data[] array cannot not be used due to hardware
limitations, so without this patch we end up with CRTCs without primary
planes when multiple VPs are active.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Chris Morgan <macromorgan@hotmail.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20220926081643.304759-1-s.hauer@pengutronix.de