drm/i915/mtl: Define engine context layouts

The part of the media and blitter engine contexts that we care about for
setting up an initial state on MTL are nearly similar to DG2 (and PVC).
The difference being PRT_BB_STATE being replaced with NOP.

For render/compute engines, the part of the context images are nearly
the same, although the layout had a very slight change --- one POSH
register was removed and the placement of some LRI/noops adjusted
slightly to compensate.

v2:
 - Dg2, mtl xcs offsets slightly vary. Use a separate offsets array(Bala)
 - Add missing nop in xcs offsets(Bala)
v3:
 - Fix the spacing for nop in xcs offset(MattR)
v4:
 - Fix rcs register offset(MattR)
v4.1:
 - Fix commit message(Lucas)

Bspec: 46261, 46260, 45585
Cc: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Cc: Licas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220928155511.2379663-1-radhakrishna.sripada@intel.com

drm/i915/gt: Flush to global observation point before breadcrumb write

Add flag to pipecontrol instruction to ensure in-flight writes are
flushed to global observation point. Also split the pipecontrol
instruction like we have in gen8.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/5886
Signed-off-by: Prathap Kumar Valsan <prathap.kumar.valsan@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@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/20220927171313.6553-1-nirmoy.das@intel.com

drm/i915: remove excessive i915_gem_drain_freed_objects

i915_gem_drain_workqueue() call i915_gem_drain_freed_objects()
so no need to call that again.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220923073515.23093-2-nirmoy.das@intel.com

drm/i915: Fix a potential UAF at device unload

i915_gem_drain_freed_objects() might not be enough to
free all the objects and RCU delayed work might get
scheduled after the i915 device struct gets freed.

Call i915_gem_drain_workqueue() to catch all RCU delayed work.

Suggested-by: Chris Wilson <chris.p.wilson@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220923073515.23093-1-nirmoy.das@intel.com

drm/i915/guc/slpc: Add SLPC selftest live_slpc_power

A fundamental assumption is that at lower frequencies,
not only do we run slower, but we save power compared to
higher frequencies.
live_slpc_power checks if running at low frequency saves power

v2: re-use code to measure power
    fixed cosmetic review comments (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/20220923110043.789178-4-riana.tauro@intel.com

drm/i915/selftests: Add helper function measure_power

move the power measurement and the triangle filter
to a different function. No functional changes.

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/20220923110043.789178-3-riana.tauro@intel.com

drm/i915/guc/slpc: Run SLPC selftests on all tiles

Run slpc selftests on all tiles

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/20220923110043.789178-2-riana.tauro@intel.com

drm/i915: Remove unwanted pointer unpacking

In await_fence_array(), unpacking syncobj pointer is not needed.
Remove it.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220927181346.1187-1-niranjana.vishwanathapura@intel.com

drm/i915/guc: Enable compute scheduling on DG2

DG2 has issues. To work around one of these the GuC must schedule
apps in an exclusive manner across both RCS and CCS. That is, if a
context from app X is running on RCS then all CCS engines must sit
idle even if there are contexts from apps Y, Z, ... waiting to run. A
certain OS favours RCS to the total starvation of CCS. Linux does not.
Hence the GuC now has a scheduling policy setting to control this
abitration.

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/20220922201209.1446343-2-John.C.Harrison@Intel.com

drm/i915/gt: Bump the reset-failure timeout to 60s

If attempting to perform a GT reset takes long than 5 seconds (including
resetting the display for gen3/4), then we declare all hope lost and
discard all user work and wedge the device to prevent further
misbehaviour. 5 seconds is too short a time for such drastic action, as
we may be stuck on other timeouts and watchdogs. If we allow a little
bit longer before hitting the big red button, we should at the very
least capture other hung task indicators pointing towards the reason why
the reset was hanging; and allow more marginal cases the extra headroom
to complete the reset without further collateral damage.

Bug: https://gitlab.freedesktop.org/drm/intel/-/issues/6448
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
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/20220916204823.1897089-1-ashutosh.dixit@intel.com

drm/i915/gt: Move scratch page into system memory on all platforms

The scratch page should never be accessed, and is only assigned as a
filler page to redirection invalid userspace access. It is not of a
performance concern and so we prefer to have a single consistent
configuration across all platforms, reducing the pressure on device
memory and avoiding the direct device access that would be required to
initialise the scratch page.

Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Cc: Matthew Auld <matthew.auld@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/20220926155018.109678-1-matthew.auld@intel.com

drm/i915: Improve debug print in vm_fault_ttm

Print the error code returned by __i915_ttm_migrate()
for better debuggability.

v2: Fix kernel test robot warning.
v3: Fix dim checkpatch warning.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/6889
Acked-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220923143730.13498-1-nirmoy.das@intel.com

drm/i915/gt: Use i915_vm_put on ppgtt_create error paths

Now that the scratch page and page directories have a reference back to
the i915_address_space, we cannot do an immediate free of the ppgtt upon
error as those buffer objects will perform a later i915_vm_put in their
deferred frees.

The downside is that by replacing the onion unwind along the error
paths, the ppgtt cleanup must handle a partially constructed vm. This
includes ensuring that the vm->cleanup is set prior to the error path.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6900
Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Fixes: 4d8151ae5329 ("drm/i915: Don't free shared locks while shared")
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: <stable@vger.kernel.org> # v5.14+
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220926153333.102195-1-matthew.auld@intel.com

drm/i915/gt: Restrict forced preemption to the active context

When we submit a new pair of contexts to ELSP for execution, we start a
timer by which point we expect the HW to have switched execution to the
pending contexts. If the promotion to the new pair of contexts has not
occurred, we declare the executing context to have hung and force the
preemption to take place by resetting the engine and resubmitting the
new contexts.

This can lead to an unfair situation where almost all of the preemption
timeout is consumed by the first context which just switches into the
second context immediately prior to the timer firing and triggering the
preemption reset (assuming that the timer interrupts before we process
the CS events for the context switch). The second context hasn't yet had
a chance to yield to the incoming ELSP (and send the ACk for the
promotion) and so ends up being blamed for the reset.

If we see that a context switch has occurred since setting the
preemption timeout, but have not yet received the ACK for the ELSP
promotion, rearm the preemption timer and check again. This is
especially significant if the first context was not schedulable and so
we used the shortest timer possible, greatly increasing the chance of
accidentally blaming the second innocent context.

Fixes: 3a7a92aba8fb ("drm/i915/execlists: Force preemption")
Fixes: d12acee84ffb ("drm/i915/execlists: Cancel banned contexts on schedule-out")
Reported-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Tested-by: Andrzej Hajda <andrzej.hajda@intel.com>
Cc: <stable@vger.kernel.org> # v5.5+
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220921135258.1714873-1-andrzej.hajda@intel.com

drm/i915: move i915_coherent_map_type() to i915_gem_pages.c and un-inline

The inline function has no place in i915_drv.h. Move it away, un-inline,
and untangle some header dependencies while at it.

Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220914163514.1837467-1-jani.nikula@intel.com

drm/i915/dg2: introduce Wa_22015475538

Wa_22015475538 applies to all DG2 (and ATSM) skus. The workaround
implementation is identical to Wa_16011620976. LSC_CHICKEN_BIT_0_UDW is
a general render register instead of rcs so adding this move to the
proper wa init function.

bspec:54077

Signed-off-by: Matt Atwood <matthew.s.atwood@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/20220920204359.103370-1-matthew.s.atwood@intel.com

drm/i915/selftests: Remove flush_scheduled_work() from live_execlists

There are ongoing efforts to remove usages of flush_scheduled_work() from
drivers in order to avoid several cases of potentential problems when
flushing is done from certain contexts.

Remove the call from the live_execlists selftest. Its purpose was to be
thorough and sync with the execlists capture state handling, but that is
not strictly required for the test to function and can be removed.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220630125716.50835-1-tvrtko.ursulin@linux.intel.com

drm/i915: Noop lrc_init_wa_ctx() on recent/future platforms

Except for graphics version 8 and 9, nothing is done in
lrc_init_wa_ctx(). Assume this won't be needed on future platforms as
well and remove the warning.

Note that this function is not called for anything below version 8 since
those don't use either guc or execlist, i.e. HAS_EXECLISTS() is false.

Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220907230841.1703574-1-lucas.demarchi@intel.com

drm/i915/dgfx: Make failure to setup stolen non-fatal

There is no reason to consider the setup of Data Stolen Memory fatal on
dgfx and non-fatal on integrated. Move the debug and error propagation
around so both have the same behavior: non-fatal. Before this change,
loading i915 on a system with TGL + DG2 would result in just TGL
succeeding the initialization (without stolen).

Now loading i915 on the same system with an injected failure in
i915_gem_init_stolen():

$ dmesg | grep stolen
i915 0000:00:02.0: [drm] Injected failure, disabling use of stolen memory
i915 0000:00:02.0: [drm:init_stolen_smem [i915]] Skip stolen region: failed to setup
i915 0000:03:00.0: [drm] Injected failure, disabling use of stolen memory
i915 0000:03:00.0: [drm:init_stolen_lmem [i915]] Skip stolen region: failed to setup

Both GPUs are still available:

$ sudo build/tools/lsgpu
card1                    Intel Dg2 (Gen12)                 drm:/dev/dri/card1
└─renderD129                                               drm:/dev/dri/renderD129
card0                    Intel Tigerlake (Gen12)           drm:/dev/dri/card0
└─renderD128                                               drm:/dev/dri/renderD128

Reviewed-by: Wayne Boyer <wayne.boyer@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915-stolen-v2-3-20ff797de047@intel.com

drm/i915: Split i915_gem_init_stolen()

Add some helpers: adjust_stolen(), request_smem_stolen_() and
init_reserved_stolen() that are now called by i915_gem_init_stolen() to
initialize each part of the Data Stolen Memory region.

Main goal is to split the reserved part within the stolen, also known as
WOPCM, as its calculation changes often per platform and is a big source
of confusion when handling stolen memory.

Reviewed-by: Wayne Boyer <wayne.boyer@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915-stolen-v2-2-20ff797de047@intel.com

drm/i915: Add missing mask when reading GEN12_DSMBASE

DSMBASE register is defined so BDSM bitfield contains the bits 63 to 20
of the base address of stolen. For the supported platforms bits 0-19 are
zero but that may not be true in future. Add the missing mask.

v2: Use REG_GENMASK64()

Acked-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Reviewed-by: Caz Yokoyama <caz@caztech.com>
Reviewed-by: Wayne Boyer <wayne.boyer@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915-stolen-v2-1-20ff797de047@intel.com

drm/i915: Split GAM and MSLICE steering

Although the bspec lists several MMIO ranges as "MSLICE," it turns out
that a subset of these are of a "GAM" subclass that has unique rules and
doesn't followed regular mslice steering behavior.

 * Xe_HP SDV:  GAM ranges must always be steered to 0,0.  These
   registers share the regular steering control register (0xFDC) with
   other steering types

 * DG2:  GAM ranges must always be steered to 1,0.  GAM registers have a
   dedicated steering control register (0xFE0) so we can set the value
   once at startup and rely on implicit steering.  Technically the
   hardware default should already be set to 1,0 properly, but it never
   hurts to ensure that in the driver.

Bspec: 66534
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Prathap Kumar Valsan <prathap.kumar.valsan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916014345.3317739-1-matthew.d.roper@intel.com

drm/i915: Do not cleanup obj with NULL bo->resource

For delayed BO release i915_ttm_delete_mem_notify()
gets called twice, once with proper bo->resource and
another time with NULL. We shouldn't do anything for
the 2nd time as we already cleaned up the obj once.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/6850
Fixes: ad74457a6b5a96 ("drm/i915/dgfx: Release mmap on rpm suspend")
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/20220920170628.3391-1-nirmoy.das@intel.com

drm/i915: Perf_limit_reasons are only available for Gen11+

Register GT0_PERF_LIMIT_REASONS (0x1381a8) is available only for
Gen11+. Therefore ensure perf_limit_reasons sysfs/debugfs files are created
only for Gen11+. Otherwise on Gen < 5 accessing these files results in the
following oops:

<1> [88.829420] BUG: unable to handle page fault for address: ffffc90000bb81a8
<1> [88.829438] #PF: supervisor read access in kernel mode
<1> [88.829447] #PF: error_code(0x0000) - not-present page

Bspec: 20008
Bug: https://gitlab.freedesktop.org/drm/intel/-/issues/6863
Fixes: fe5979665f64 ("drm/i915/debugfs: Add perf_limit_reasons in debugfs")
Fixes: fa68bff7cf27 ("drm/i915/gt: Add sysfs throttle frequency interfaces")
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@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/20220919162401.2077713-1-ashutosh.dixit@intel.com

drm/i915/gem: Really move i915_gem_context.link under ref protection

i915_perf assumes that it can use the i915_gem_context reference to
protect its i915->gem.contexts.list iteration. However, this requires
that we do not remove the context from the list until after we drop the
final reference and release the struct. If, as currently, we remove the
context from the list during context_close(), the link.next pointer may
be poisoned while we are holding the context reference and cause a GPF:

[ 4070.573157] i915 0000:00:02.0: [drm:i915_perf_open_ioctl [i915]] filtering on ctx_id=0x1fffff ctx_id_mask=0x1fffff
[ 4070.574881] general protection fault, probably for non-canonical address 0xdead000000000100: 0000 [#1] PREEMPT SMP
[ 4070.574897] CPU: 1 PID: 284392 Comm: amd_performance Tainted: G            E     5.17.9 #180
[ 4070.574903] Hardware name: Intel Corporation NUC7i5BNK/NUC7i5BNB, BIOS BNKBL357.86A.0052.2017.0918.1346 09/18/2017
[ 4070.574907] RIP: 0010:oa_configure_all_contexts.isra.0+0x222/0x350 [i915]
[ 4070.574982] Code: 08 e8 32 6e 10 e1 4d 8b 6d 50 b8 ff ff ff ff 49 83 ed 50 f0 41 0f c1 04 24 83 f8 01 0f 84 e3 00 00 00 85 c0 0f 8e fa 00 00 00 <49> 8b 45 50 48 8d 70 b0 49 8d 45 50 48 39 44 24 10 0f 85 34 fe ff
[ 4070.574990] RSP: 0018:ffffc90002077b78 EFLAGS: 00010202
[ 4070.574995] RAX: 0000000000000002 RBX: 0000000000000002 RCX: 0000000000000000
[ 4070.575000] RDX: 0000000000000001 RSI: ffffc90002077b20 RDI: ffff88810ddc7c68
[ 4070.575004] RBP: 0000000000000001 R08: ffff888103242648 R09: fffffffffffffffc
[ 4070.575008] R10: ffffffff82c50bc0 R11: 0000000000025c80 R12: ffff888101bf1860
[ 4070.575012] R13: dead0000000000b0 R14: ffffc90002077c04 R15: ffff88810be5cabc
[ 4070.575016] FS:  00007f1ed50c0780(0000) GS:ffff88885ec80000(0000) knlGS:0000000000000000
[ 4070.575021] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 4070.575025] CR2: 00007f1ed5590280 CR3: 000000010ef6f005 CR4: 00000000003706e0
[ 4070.575029] Call Trace:
[ 4070.575033]  <TASK>
[ 4070.575037]  lrc_configure_all_contexts+0x13e/0x150 [i915]
[ 4070.575103]  gen8_enable_metric_set+0x4d/0x90 [i915]
[ 4070.575164]  i915_perf_open_ioctl+0xbc0/0x1500 [i915]
[ 4070.575224]  ? asm_common_interrupt+0x1e/0x40
[ 4070.575232]  ? i915_oa_init_reg_state+0x110/0x110 [i915]
[ 4070.575290]  drm_ioctl_kernel+0x85/0x110
[ 4070.575296]  ? update_load_avg+0x5f/0x5e0
[ 4070.575302]  drm_ioctl+0x1d3/0x370
[ 4070.575307]  ? i915_oa_init_reg_state+0x110/0x110 [i915]
[ 4070.575382]  ? gen8_gt_irq_handler+0x46/0x130 [i915]
[ 4070.575445]  __x64_sys_ioctl+0x3c4/0x8d0
[ 4070.575451]  ? __do_softirq+0xaa/0x1d2
[ 4070.575456]  do_syscall_64+0x35/0x80
[ 4070.575461]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[ 4070.575467] RIP: 0033:0x7f1ed5c10397
[ 4070.575471] Code: 3c 1c e8 1c ff ff ff 85 c0 79 87 49 c7 c4 ff ff ff ff 5b 5d 4c 89 e0 41 5c c3 66 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d a9 da 0d 00 f7 d8 64 89 01 48
[ 4070.575478] RSP: 002b:00007ffd65c8d7a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[ 4070.575484] RAX: ffffffffffffffda RBX: 0000000000000006 RCX: 00007f1ed5c10397
[ 4070.575488] RDX: 00007ffd65c8d7c0 RSI: 0000000040106476 RDI: 0000000000000006
[ 4070.575492] RBP: 00005620972f9c60 R08: 000000000000000a R09: 0000000000000005
[ 4070.575496] R10: 000000000000000d R11: 0000000000000246 R12: 000000000000000a
[ 4070.575500] R13: 000000000000000d R14: 0000000000000000 R15: 00007ffd65c8d7c0
[ 4070.575505]  </TASK>
[ 4070.575507] Modules linked in: nls_ascii(E) nls_cp437(E) vfat(E) fat(E) i915(E) x86_pkg_temp_thermal(E) intel_powerclamp(E) crct10dif_pclmul(E) crc32_pclmul(E) crc32c_intel(E) aesni_intel(E) crypto_simd(E) intel_gtt(E) cryptd(E) ttm(E) rapl(E) intel_cstate(E) drm_kms_helper(E) cfbfillrect(E) syscopyarea(E) cfbimgblt(E) intel_uncore(E) sysfillrect(E) mei_me(E) sysimgblt(E) i2c_i801(E) fb_sys_fops(E) mei(E) intel_pch_thermal(E) i2c_smbus(E) cfbcopyarea(E) video(E) button(E) efivarfs(E) autofs4(E)
[ 4070.575549] ---[ end trace 0000000000000000 ]---

v3: fix incorrect syntax of spin_lock() replacing spin_lock_irqsave()

v2: irqsave not required in a worker, neither conversion to irq safe
    elsewhere (Tvrtko),
  - perf: it's safe to call gen8_configure_context() even if context has
    been closed, no need to check,
  - drop unrelated cleanup (Andi, Tvrtko)

Reported-by: Mark Janes <mark.janes@intel.com>
Closes: https://gitlab.freedesktop.org/drm/intel/issues/6222
References: a4e7ccdac38e ("drm/i915: Move context management under GEM")
Fixes: f8246cf4d9a9 ("drm/i915/gem: Drop free_work for GEM contexts")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: <stable@vger.kernel.org> # v5.12+
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916092403.201355-3-janusz.krzysztofik@linux.intel.com

drm/i915/gem: Flush contexts on driver release

Due to i915_perf assuming that it can use the i915_gem_context reference
to protect its i915->gem.contexts.list iteration, we need to defer removal
of the context from the list until last reference to the context is put.
However, there is a risk of triggering kernel warning on contexts list not
empty at driver release time if we deleagate that task to a worker for
i915_gem_context_release_work(), unless that work is flushed first.
Unfortunately, it is not flushed on driver release.  Fix it.

Instead of additionally calling flush_workqueue(), either directly or via
a new dedicated wrapper around it, replace last call to
i915_gem_drain_freed_objects() with existing i915_gem_drain_workqueue()
that performs both tasks.

Fixes: 75eefd82581f ("drm/i915: Release i915_gem_context from a worker")
Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Cc: stable@kernel.org # v5.16+
Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220916092403.201355-2-janusz.krzysztofik@linux.intel.com

drm/i915/mtl: Add MTL forcewake support

MTL has separate forcewake tables for the primary/render GT and the
media GT; each GT's intel_uncore will use a separate forcewake table and
should only initialize the domains that are relevant to that GT.  The GT
ack register also moves to a new location of (GSI base + 0xDFC) on this
platform.

Note that although our uncore handlers take care of transparently
redirecting all register accesses in the media GT's GSI range to their
new offset at 0x380000, the forcewake ranges listed in the table should
use the final, post-translation offsets.

NOTE:  There are two ranges in the media IP that have multicast
registers where the two register instances reside in different power
wells (either VD0 or VD2).  We don't have an easy way to deal with this
today (and in fact we don't even access these register ranges in the
driver today), so for now we just mark those ranges as FORCEWAKE_ALL
which will cause all of the media power wells to be grabbed, ensuring
proper operation.  If we start reading/writing in those ranges in the
future, we can re-visit whether it's worth adding extra steering
complexity into our forcewake support.

Bspec: 67788, 67789, 52077
Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Harish Chegondi <harish.chegondi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220910001631.1986601-1-matthew.d.roper@intel.com

drm/i915/guc: Fix release build bug in 'remove log size module parameters'

A patch was merged to remove the GuC log size override module
parameters. That patch was broken and caused kernel error messages on
boot in non CONFIG_DEBUG_GUC|GEM builds:
[   12.085121] i915 0000:00:02.0: [drm] *ERROR* Zero GuC log crash dump size!
[   12.092035] i915 0000:00:02.0: [drm] *ERROR* Zero GuC log debug size!

So fit it.

Fixes: f54e515c9180 ("drm/i915/guc: Remove log size module parameters")
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Julia Lawall <Julia.Lawall@inria.fr>
Cc: Chris Wilson <chris.p.wilson@intel.com>
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/20220913010929.2734885-2-John.C.Harrison@Intel.com

drm/i915/rps: Freq caps for MTL

For MTL, when reading from HW, RP0, RP1 (actuall RPe) and RPn freq use an
entirely different set of registers with different fields, bitwidths and
units.

v2: Move MTL check into a separate function (Jani)

Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Badal Nilawar <badal.nilawar@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@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/20220910143844.1755324-4-ashutosh.dixit@intel.com

drm/i915/mtl: PERF_LIMIT_REASONS changes for MTL

PERF_LIMIT_REASONS register for MTL media gt is different now.

v2: Avoid static inline for intel_gt_perf_limit_reasons_reg() (Jani)

Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Badal Nilawar <badal.nilawar@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@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/20220910143844.1755324-3-ashutosh.dixit@intel.com

drm/i915/debugfs: Add perf_limit_reasons in debugfs

Add perf_limit_reasons in debugfs. The upper 16 perf_limit_reasons RW "log"
bits are identical to the lower 16 RO "status" bits except that the "log"
bits remain set until cleared, thereby ensuring the throttling occurrence
is not missed. The clear fop clears the upper 16 "log" bits, the get fop
gets all 32 "log" and "status" bits.

v2: Expand commit message and clarify "log" and "status" bits in
    comment (Rodrigo)

Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Tilak Tangudu <tilak.tangudu@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/20220910143844.1755324-2-ashutosh.dixit@intel.com

drm/i915: Handle all GTs on driver (un)load paths

This, along with the changes already landed in commit 1c66a12ab431
("drm/i915: Handle each GT on init/release and suspend/resume") makes
engines from all GTs actually known to the driver.

To accomplish this we need to sprinkle a lot of for_each_gt calls around
but is otherwise pretty un-eventuful.

v2:
 - Consolidate adjacent GT loops in a couple places.  (Daniele)

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915232654.3283095-5-matthew.d.roper@intel.com

drm/i915: Make GEM suspend all GTs

Walk all GTs when suspending.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915232654.3283095-4-matthew.d.roper@intel.com

drm/i915: Make GEM resume all engines

Walk all GTs from i915_gem_resume when resuming engines.

Cc: Andi Shyti <andi.shyti@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915232654.3283095-3-matthew.d.roper@intel.com

drm/i915/gt: Cleanup partial engine discovery failures

If we abort driver initialisation in the middle of gt/engine discovery,
some engines will be fully setup and some not. Those incompletely setup
engines only have 'engine->release == NULL' and so will leak any of the
common objects allocated.

v2:
 - Drop the destroy_pinned_context() helper for now.  It's not really
   worth it with just a single callsite at the moment.  (Janusz)

Signed-off-by: Chris Wilson <chris.p.wilson@intel.com>
Cc: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220915232654.3283095-2-matthew.d.roper@intel.com

drm/i915/uc: Update to latest GuC and use new-format GuC/HuC names

Going forwards, the intention is for GuC firmware files to be named
for their major version only and HuC firmware files to have no version
number in the name at all. This patch adds those entries for all
platforms that are officially GuC/HuC enabled.

Also, update the expected GuC version numbers to the latest firmware
release for those platforms.

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/20220914234605.622342-2-John.C.Harrison@Intel.com

drm/i915: Invert if/else ladder for stolen init

Continue converting the driver to the convention of last version first,
extending it to the future platforms. Now, any GRAPHICS_VER >= 11 will
be handled by the first branch.

Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220908-if-ladder-v2-3-7a7b15545c93@intel.com
Link: https://patchwork.freedesktop.org/patch/msgid/20220912-copy-engine-v1-1-ef92fd81758d@intel.com

drm/i915/gt: Extract per-platform function for frequency read

Instead of calling read_clock_frequency() to walk the if/else ladder
per platform, move the ladder to intel_gt_init_clock_frequency() and
use one function per branch.

With the new logic, it's now clear the call to
gen9_get_crystal_clock_freq() was just dead code, as gen9 is handled by
another function and there is no version 10. Remove that function and
the caller.

v2: Correctly handle intel_gt_check_clock_frequency() that also calls
the function to read clock frequency (Gustavo)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220908-if-ladder-v2-2-7a7b15545c93@intel.com

drm/i915: Invert if/else ladder for frequency read

Continue converting the driver to the convention of last version first,
extending it to the future platforms. Now, any GRAPHICS_VER >= 11 will
be handled by the first branch.

With the new ranges it's easier to see what platform a branch started to
be taken. Besides the >= 11 change, the branch taken for GRAPHICS_VER == 10
is also different, but currently there is no such platform in i915.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220908-if-ladder-v2-1-7a7b15545c93@intel.com

drm/i915/uc: Fix issues with overriding firmware files

The earlier update to support reduced versioning of firmware files
introduced an issue with the firmware override module parameter. A
self test would specify an invalid file name (invalid meaning not in
the table) both with and without setting the override flag. The
*non-override* case would cause an infinite loop. I.e. a situation
that is impossible to hit outside of the selftest because either the
file name has come from the table in first place or it came from an
override. However, the override case was also broken in that it would
bypass some of the later processing.

The first fix is to update the scanning loop code so that if an
invalid file is passed in, it will exit rather than loop forever. So
if the impossible situation did somehow occur in the future, it
wouldn't be such a big problem.

The second flips the logic on the override early exit to be negative
rather than positive. That way if an explicit override has been set,
then it won't try to scan for backup options (because there is no
point anyway - the user wanted X and if X is not available, that's
their problem). It also means that it won't skip code that still needs
to be run once a valid firmware file has been selected.

v2: Also remove ANSI colour codes that accidentally got left in an
error message in the original patch.

Fixes: 665ae9c9ca79 ("drm/i915/uc: Support for version reduced and multiple firmware files")
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Cc: Venkata Sandeep Dhanalakota <venkata.s.dhanalakota@intel.com>
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/20220914005821.3702446-2-John.C.Harrison@Intel.com

drm/i915/dgfx: Release mmap on rpm suspend

Release all mmap mapping for all lmem objects which are associated
with userfault such that, while pcie function in D3hot, any access
to memory mappings will raise a userfault.

Runtime resume the dgpu(when gem object lies in lmem).
This will transition the dgpu graphics function to D0
state if it was in D3 in order to access the mmap memory
mappings.

v2:
- Squashes the patches. [Matt Auld]
- Add adequate locking for lmem_userfault_list addition. [Matt Auld]
- Reused obj->userfault_count to avoid double addition. [Matt Auld]
- Added i915_gem_object_lock to check
  i915_gem_object_is_lmem. [Matt Auld]

v3:
- Use i915_ttm_cpu_maps_iomem. [Matt Auld]
- Fix 'ret == 0 to ret == VM_FAULT_NOPAGE'. [Matt Auld]
- Reuse obj->userfault_count as a bool 0 or 1. [Matt Auld]
- Delete the mmaped obj from lmem_userfault_list in obj
  destruction path. [Matt Auld]
- Get a wakeref for object destruction patch. [Matt Auld]
- Use intel_wakeref_auto to delay runtime PM. [Matt Auld]

v4:
- Avoid using mmo offset to get the vma_node. [Matt Auld]
- Added comment to use the lmem_userfault_lock. [Matt Auld]
- Get lmem_userfault_lock in i915_gem_object_release_mmap_offset.
  [Matt Auld]
- Fixed kernel test robot generated warning.

v5:
- Addressed the cosmetics comments. [Andi]
- Changed i915_gem_runtime_pm_object_release_mmap_offset() name to
  i915_gem_object_runtime_pm_release_mmap_offset() to be rhythmic.

PCIe Specs 5.3.1.4.1

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6331
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913152714.16541-3-anshuman.gupta@intel.com

drm/i915: Refactor userfault_wakeref to re-use

Refactor userfault_wakeref to re-use for discrete lmem mmap mapping
as well, as on discrete GTT mmap are not supported. Moving
userfault_wakeref from ggtt to gt structure.

Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913152714.16541-2-anshuman.gupta@intel.com

drm/i915/selftest: Clear the output buffers before GPU writes

When testing whether we can get the GPU to leak information about
non-privileged state, we first need to ensure that the output buffer is
set to a known value as the HW may opt to skip the write into memory for
a non-privileged read of a sensitive register. We chose POISON_INUSE (0x5a)
so that is both non-zero and distinct from the poison values used during
the test.

v2:
  Use i915_gem_object_pin_map_unlocked

Reported-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: CQ Tang <cq.tang@intel.com>
cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Thomas Hellstrom <thomas.hellstrom@linux.intel.com>
Signed-off-by: Karolina Drobnik <karolina.drobnik@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/5cebab02d182c171cf40cb5b73d6c3eeb7619360.1663081418.git.karolina.drobnik@intel.com

drm/i915/selftest: Always cancel semaphore on error

Ensure that we always signal the semaphore when timing out, so that if it
happens to be stuck waiting for the semaphore we will quickly recover
without having to wait for a reset.

Reported-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: CQ Tang <cq.tang@intel.com>
cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Thomas Hellstrom <thomas.hellstrom@linux.intel.com>
Signed-off-by: Karolina Drobnik <karolina.drobnik@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/8b7781f7dbaf2791156491b76d5faa7852e5cbbb.1663081418.git.karolina.drobnik@intel.com

drm/i915/selftests: Check for incomplete LRI from the context image

In order to keep the context image parser simple, we assume that all
commands follow a similar format. A few, especially not MI commands on
the render engines, have fixed lengths not encoded in a length field.
This caused us to incorrectly skip over 3D state commands, and start
interpreting context data as instructions. Eventually, as Daniele
discovered, this would lead us to find addition LRI as part of the data
and mistakenly add invalid LRI commands to the context probes.

Stop parsing after we see the first !MI command, as we know we will have
seen all the context registers by that point. (Mostly true for all gen
so far, though the render context does have LRI after the first page
that we have been ignoring so far. It would be useful to extract those
as well so that we have the full list of user accessible registers.)

Similarly, emit a warning if we do try to emit an invalid zero-length
LRI.

Testcase: igt@i915_selftest@live@gt_lrc
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6580
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6670

Reported-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Acked-by: Thomas Hellstrom <thomas.hellstrom@linux.intel.com>
Signed-off-by: Karolina Drobnik <karolina.drobnik@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@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/7377cb3b371a983dce02be69f6611fcf85c822bb.1663081418.git.karolina.drobnik@intel.com

drm/i915/gt: Explicitly clear BB_OFFSET for new contexts

Even though the initial protocontext we load onto HW has the register
cleared, by the time we save it into the default image, BB_OFFSET has
had the enable bit set. Reclear BB_OFFSET for each new context.

Testcase: igt/i915_selftests/gt_lrc

v2:
  Extend it for gen8.
v3:
  BB_OFFSET is recorded per engine from Gen9 onwards

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Reviewed-by: Thomas Hellstrom <thomas.hellstrom@linux.intel.com>
Signed-off-by: Karolina Drobnik <karolina.drobnik@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/37c67abb3303852f06a570a4360addf52bf941c1.1663081418.git.karolina.drobnik@intel.com

drm/i915: Skip applying copy engine fuses

Support for reading the fuses to check what are the Link Copy engines
was added in commit ad5f74f34201 ("drm/i915/pvc: read fuses for link
copy engines"). However they were added unconditionally because the
FUSE3 register is present since graphics version 10.

However the bitfield with meml3 fuses only exists since graphics version
12. Moreover, Link Copy engines are currently only available in PVC.
Tying additional copy engines to the meml3 fuses is not correct for
other platforms.

Make sure there is a check for  `12.60 <= ver < 12.70`. Later platforms
may extend this function later if it's needed to fuse off copy engines.

Currently it's harmless as the Link Copy engines are still not exported:
info->engine_mask only has BCS0 set and the register is only read for
platforms that do have it.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220912-copy-engine-v1-1-ef92fd81758d@intel.com

drm/i915: remove unused i915_gem_lmem_obj_ops declaration

i915_gem_lmem_obj_ops has been removed since
commit 213d50927763 ("drm/i915/ttm: Introduce a TTM i915
gem object backend"), so remove it.

Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220913024847.552254-7-cuigaosheng1@huawei.com

drm/i915/gt: Extract function to apply media fuses

Just like is done for compute and copy engines, extract a function to
handle media engines. While at it, be consistent on using or not the
uncore/gt/info variable aliases.

Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220909-media-v2-2-6f20f322b4ef@intel.com

drm/i915/gt: Use MEDIA_VER() when handling media fuses

Check for media IP version instead of graphics since this is figuring
out the media engines' configuration. Currently the only platform with
non-matching graphics/media version is Meteor Lake: update the check in
gen11_vdbox_has_sfc() so it considers not only version 12, but also any
later version which then includes that platform.

Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220909-media-v2-1-6f20f322b4ef@intel.com

drm/i915/mtl: Hook up interrupts for standalone media

Top-level handling of standalone media interrupts will be processed as
part of the primary GT's interrupt handler (since primary and media GTs
share an MMIO space, unlike remote tile setups).  When we get down to
the point of handling engine interrupts, we need to take care to lookup
VCS and VECS engines in the media GT rather than the primary.

There are also a couple of additional "other" instance bits that
correspond to the media GT's GuC and media GT's power management
interrupts; we need to direct those to the media GT instance as well.

Bspec: 45605
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-15-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/mtl: Use primary GT's irq lock for media GT

When we hook up interrupts (in the next patch), interrupts for the media
GT are still processed as part of the primary GT's interrupt flow.  As
such, we should share the same IRQ lock with the primary GT.  Let's
convert gt->irq_lock into a pointer and just point the media GT's
instance at the same lock the primary GT is using.

v2:
 - Point media's gt->irq_lock at the primary GT lock properly.  (Daniele)
 - Fix jump target for intel_root_gt_init_early errors.  (Daniele)

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-14-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/xelpmp: Expose media as another GT

Xe_LPM+ platforms have "standalone media."  I.e., the media unit is
designed as an additional GT with its own engine list, GuC, forcewake,
etc.  Let's allow platforms to include media GTs in their device info.

v2:
 - Simplify GSI register handling and split it out to a separate patch
   for ease of review.  (Daniele)

Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Acked-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-13-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/mtl: Add gsi_offset when emitting aux table invalidation

The aux table invalidation registers are a bit unique --- they're
engine-centric registers that reside in the GSI register space rather
than within the engines' regular MMIO ranges.  That means that when
issuing invalidation on engines in the standalone media GT, the GSI
offset must be added to the regular MMIO offset for the invalidation
registers.

Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-12-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/uncore: Add GSI offset to uncore

GT non-engine registers (referred to as "GSI" registers by the spec)
have the same relative offsets on standalone media as they do on the
primary GT, just with an additional "GSI offset" added to their MMIO
address.  If we store this GSI offset in the standalone media's
intel_uncore structure, it can be automatically applied to all GSI reg
reads/writes that happen on that GT, allowing us to re-use our existing
GT code with minimal changes.

Forcewake and shadowed register tables for the media GT (which will be
added in a future patch) are listed as final addresses that already
include the GSI offset, so we also need to add the GSI offset before
doing lookups of registers in one of those tables.

v2:
 - Add comment on raw_reg_*() macros explaining why we don't bother with
   GSI offsets in them.  (Daniele)

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220908224550.821257-1-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Handle each GT on init/release and suspend/resume

In preparation for enabling a second GT, there are a number of GT/uncore
operations that happen during initialization or suspend flows that need
to be performed on each GT, not just the primary,

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-10-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Initialize MMIO access for each GT

In a multi-GT system we need to initialize MMIO access for each GT, not
just the primary GT.

Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-9-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Use a DRM-managed action to release the PCI bridge device

As we start supporting multiple uncore structures in future patches, the
MMIO cleanup (which may also get called mid-init if there's a failure)
will become more complicated.  Moving to DRM-managed actions will help
keep things simple.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-8-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Rename and expose common GT early init routine

The common early GT init is needed for initialization of all GT types
(root/primary, remote tile, standalone media).  Since standalone media
(coming in a future patch) will be implemented in a separate file,
rename and expose the function for use.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-7-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Prepare more multi-GT initialization

We're going to introduce an additional intel_gt for MTL's media unit
soon.  Let's provide a bit more multi-GT initialization framework in
preparation for that.  The initialization will pull the list of GTs for
a platform from the device info structure.  Although necessary for the
immediate MTL media enabling, this same framework will also be used
farther down the road when we enable remote tiles on xehpsdv and pvc.

v2:
 - Re-add missing test for !HAS_EXTRA_GT_LIST in intel_gt_probe_all().

v3:
 - Move intel_gt_definition struct to intel_gt_types.h.  (Jani)
 - Drop gtdef->setup().  For now we'll just use a switch() based on GT
   type since we don't have too many different handlers for the
   foreseeable future.  (Jani)

Cc: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Aravind Iddamsetty <aravind.iddamsetty@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-6-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Drop intel_gt_tile_cleanup()

Unmapping of the MMIO range can be done as a DRM-managed action, which
will take care of the unmapping on device teardown and error paths.
This will also ensure proper ordering with respect to other DRM-managed
actions that we'll be using to clean up non-primary GTs in upcoming
patches.

We have not yet enabled any non-root GTs in the driver yet, so the
kfree() of the GT structure is effectively dead code.  When we do start
enabling non-root GTs in upcoming patches, those are going to be using
DRM-managed allocations tied to the device lifetime, so we don't need to
explicitly free them (and kfree would be incorrect anyway).

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/20220906234934.3655440-5-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Use managed allocations for extra uncore objects

We're slowly transitioning the init-time kzalloc's of the driver over to
DRM-managed allocations; let's make sure the uncore objects allocated
for non-root GTs are thus allocated.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-4-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Only hook up uncore->debug for primary uncore

The original intent of intel_uncore_mmio_debug as described in commit
0a9b26306d6a ("drm/i915: split out uncore_mmio_debug") was to be a
singleton structure that could be shared between multiple GTs' uncore
objects in a multi-tile system.  Somehow we went off track and
started allocating separate instances of this structure for each GT,
which defeats that original goal.

But in reality, there isn't even a need to share the mmio_debug between
multiple GTs; on all modern platforms (i.e., everything after gen7)
unclaimed register accesses are something that can only be detected for
display registers.  There's no point in grabbing the debug spinlock and
checking for unclaimed accesses on an uncore used by an xehpsdv or pvc
remote tile GT, or the uncore used by a mtl standalone media GT since
all of the display accesses go through the primary intel_uncore.

The simplest solution is to simply leave uncore->debug NULL on all
intel_uncore instances except for the primary one.  This will allow us
to avoid the pointless debug spinlock acquisition we've been doing on
MMIO accesses coming in through these intel_uncores.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-3-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Move locking and unclaimed check into mmio_debug_{suspend, resume}

Moving the locking for MMIO debug (and the final check for unclaimed
accesses when resuming debug after a userspace-initiated forcewake) will
make it simpler to completely skip MMIO debug handling on uncores that
don't support it in future patches.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220906234934.3655440-2-matthew.d.roper@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gt: Fix perf limit reasons bit positions

Perf limit reasons bit positions were off by one.

Fixes: fa68bff7cf27 ("drm/i915/gt: Add sysfs throttle frequency interfaces")
Cc: stable@vger.kernel.org # v5.18+
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Sujaritha Sundaresan <sujaritha.sundaresan@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220908155821.1662110-1-ashutosh.dixit@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Revert "drm/i915/dg2: extend Wa_1409120013 to DG2"

This reverts commit 487970e8bb776c989013bb59d6cbb22e45b9afc6.

Updated bspec and workaround database note Wa_1409120013 is not needed
for DG2 (or any Xe_LPD) platform. Simply check by display version 12.

v2: Simplify condition check to display version (Matt Roper)

Cc: Matt Atwood <matthew.s.atwood@intel.com>
Cc: Clint Taylor <Clinton.A.Taylor@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220907232541.1720966-1-lucas.demarchi@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/guc: Cancel GuC engine busyness worker synchronously

The worker is canceled in gt_park path, but earlier it was assumed that
gt_park path cannot sleep and the cancel is asynchronous. This caused a
race with suspend flow where the worker runs after suspend and causes an
unclaimed register access warning. Cancel the worker synchronously since
the gt_park is indeed allowed to sleep.

v2: Fix author name and sign-off mismatch

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4419
Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu")
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/20220827002135.139349-1-umesh.nerlige.ramappa@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gsc: allocate extended operational memory in LMEM

GSC requires more operational memory than available on chip.
Reserve 4M of LMEM for GSC operation. The memory is provided to the
GSC as struct resource to the auxiliary data of the child device.

Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Alan Previn <alan.previn.teres.alexis@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220907215113.1596567-16-tomas.winkler@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: debugfs: add pxp mode to devstate in debugfs

Add pxp mode devstate to debugfs to monitor pxp state machine progress.
This is useful to debug issues in scenarios in which the pxp state
needs to be re-initialized, like during power transitions such as
suspend/resume. With this debugfs the state could be monitored
to ensure that pxp is in the ready state.

CC: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-15-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: drop ready bits check after start

The check that hardware and host ready bits are set after start
is redundant and may fail and disable driver if there is
back-to-back link reset issued right after start.
This happens during pxp mode transitions when firmware
undergo reset. Remove these checks to eliminate such failures.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-14-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: gsc: add transition to PXP mode in resume flow

Added transition to PXP mode in resume flow.

CC: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-13-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: gsc: setup gsc extended operational memory

1. Retrieve extended operational memory physical pointers from the
   auxiliary device info.
2. Setup memory registers.
3. Notify firmware that the memory is ready by sending the memory
   ready command.
4. Disable PXP device if GSC is not in PXP mode.

CC: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-12-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: mkhi: add memory ready command

Add GSC memory ready command.
The command indicates to the firmware that extend operation
memory was setup and the firmware may enter PXP mode.

CC: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-11-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: bus: export common mkhi definitions into a separate header

Exported common mkhi definitions from bus-fixup.c into a separate
header file mkhi.h for other driver usage.

Signed-off-by: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-10-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: extend timeouts on slow devices

Parametrize operational timeouts in order
to support slow firmware on some graphics devices.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-9-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: gsc: wait for reset thread on stop

Wait for reset work to complete before initiating
stop reset flow sequence.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-8-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: gsc: use polling instead of interrupts

A work-around for a HW issue in XEHPSDV that manifests itself when SW reads
a gsc register when gsc is sending an interrupt. The work-around is
to disable interrupts and to use polling instead.

Cc: James Ausmus <james.ausmus@intel.com>
Signed-off-by: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-7-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gsc: add GSC XeHP SDV platform definition

Define GSC on XeHP SDV (Intel(R) dGPU without display)

XeHP SDV uses the same hardware settings as DG1, but uses polling
instead of interrupts and runs the firmware in slow pace due to
hardware limitations.

Signed-off-by: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-6-tomas.winkler@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gsc: add slow_firmware flag to the gsc device definition

Add slow_firmware flag to the gsc device definition
and pass it to mei auxiliary device, this instructs
the driver to use longer operation timeouts.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-5-tomas.winkler@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: add slow_firmware flag to the mei auxiliary device

Add slow_firmware flag to the mei auxiliary device info
to inform the mei driver about slow underlying firmware.
Such firmware will require to use larger operation timeouts.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-4-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

mei: add kdoc for struct mei_aux_device

struct mei_aux_device is an interface structure
requires proper documenation.

Signed-off-by: Tomas Winkler <tomas.winkler@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/20220907215113.1596567-3-tomas.winkler@intel.com
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/gsc: skip irq initialization if using polling

Some platforms require the host to poll on the
GSC registers instead of relaying on the interrupts.
For those platforms, irq initialization should be skipped

Signed-off-by: Vitaly Lubart <vitaly.lubart@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@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/20220907215113.1596567-2-tomas.winkler@intel.com
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Set correct domains values at _i915_vma_move_to_active

Fix regression introduced by commit:
"drm/i915: Individualize fences before adding to dma_resv obj"
which sets obj->read_domains to 0 for both read and write paths.
Also set obj->write_domain to 0 on read path which was removed by
the commit.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/6639
Fixes: 420a07b841d0 ("drm/i915: Individualize fences before adding to dma_resv obj")
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Cc: <stable@vger.kernel.org> # v5.16+
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Andrzej Hajda <andrzej.hajda@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/20220907172641.12555-1-nirmoy.das@intel.com

drm/i915: Rename ggtt_view as gtt_view

So far, different views (normal, partial, rotated and remapped)
into the same object are only supported for GGTT mappings.
But with the upcoming VM_BIND feature, PPGTT will also use the
partial view mapping. Hence rename ggtt_view to more generic
gtt_view.

Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220901183854.3446-1-niranjana.vishwanathapura@intel.com

drm/i915/uc: Add patch level version number support

With the move to un-versioned filenames, it becomes more difficult to
know exactly what version of a given firmware is being used. So add
the patch level version number to the debugfs output.

Also, support matching by patch level when selecting code paths for
firmware compatibility. While a patch level change cannot be backwards
breaking, it is potentially possible that a new feature only works
from a given patch level onwards (even though it was theoretically
added in an earlier version that bumped the major or minor version).

Signed-off-by: John Harrison <John.C.Harrison@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/20220906230147.479945-2-daniele.ceraolospurio@intel.com

drm/i915/uc: Support for version reduced and multiple firmware files

There was a misunderstanding in how firmware file compatibility should
be managed within i915. This has been clarified as:
  i915 must support all existing firmware releases forever
  new minor firmware releases should replace prior versions
  only backwards compatibility breaking releases should be a new file

This patch cleans up the single fallback file support that was added
as a quick fix emergency effort. That is now removed in preference to
supporting arbitrary numbers of firmware files per platform.

The patch also adds support for having GuC firmware files that are
named by major version only (because the major version indicates
backwards breaking changes that affect the KMD) and for having HuC
firmware files with no version number at all (because the KMD has no
interface requirements with the HuC).

For GuC, the driver will report via dmesg if the found file is older than
expected. For HuC, the KMD will no longer require updating for any new
HuC release so will not be able to report what the latest expected
version is.

Signed-off-by: John Harrison <John.C.Harrison@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/20220906230147.479945-1-daniele.ceraolospurio@intel.com

drm/i915: Don't try to disable host RPS when this was never enabled.

Specially in GT reset case this could be triggered and try
to disable things that had never been enabled. Let's add
some protection here.

Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Acked-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220902095126.373036-1-rodrigo.vivi@intel.com

drm/i915: consider HAS_FLAT_CCS() in needs_ccs_pages

Just move the HAS_FLAT_CCS() check into needs_ccs_pages. This also then
fixes i915_ttm_memcpy_allowed() which was incorrectly reporting true on
DG1, even though it doesn't have small-BAR or flat-CCS.

References: https://gitlab.freedesktop.org/drm/intel/-/issues/6605
Fixes: efeb3caf4341 ("drm/i915/ttm: disallow CPU fallback mode for ccs pages")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Nirmoy Das <nirmoy.das@intel.com>
Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220905105329.41455-1-matthew.auld@intel.com

drm/i915/ttm: Abort suspend on i915_ttm_backup failure

On system suspend when system memory is low then i915_gem_obj_copy_ttm()
could fail trying to backup a lmem obj. GEM_WARN_ON() is not enough,
suspend shouldn't continue if i915_ttm_backup() throws an error.

v2: Keep the fdo issue till we have a igt test(Matt).
v3: Use %pe(Andrzej)

References: https://gitlab.freedesktop.org/drm/intel/-/issues/6529
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Suggested-by: Chris P Wilson <chris.p.wilson@intel.com>
Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220901172217.18392-1-nirmoy.das@intel.com

drm/i915/slpc: Let's fix the PCODE min freq table setup for SLPC

We need to inform PCODE of a desired ring frequencies so PCODE update
the memory frequencies to us. rps->min_freq and rps->max_freq are the
frequencies used in that request. However they were unset when SLPC was
enabled and PCODE never updated the memory freq.

v2 (as Suggested by Ashutosh): if SLPC is in use, let's pick the right
   frequencies from the get_ia_constants instead of the fake init of
   rps' min and max.

v3: don't forget the max <= min return

v4: Move all the freq conversion to intel_rps.c. And the max <= min
    check to where it belongs.

v5: (Ashutosh) Fix old comment s/50 HZ/50 MHz and add a doc explaining
    the "raw format"

Fixes: 7ba79a671568 ("drm/i915/guc/slpc: Gate Host RPS when SLPC is enabled")
Cc: <stable@vger.kernel.org> # v5.15+
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Tested-by: Sushma Venkatesh Reddy <sushma.venkatesh.reddy@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220831214538.143950-1-rodrigo.vivi@intel.com

drm/i915/slpc: Fix inconsistent locked return

Fix for intel_guc_slpc_set_min_freq() warn:
inconsistent returns '&slpc->lock'.

v2: Avoid with_intel_runtime_pm with the
    internal goto/return. (Ashutosh)
    Also standardize the 'ret' if this came from
    the efficient setup. And avoid the 'unlikely'.

Fixes: 95ccf312a1e4 ("drm/i915/guc/slpc: Allow SLPC to use efficient frequency")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220830193537.52201-1-rodrigo.vivi@intel.com

drm/i915/ats-m: Add thread execution tuning setting

On client DG2 platforms, optimal performance is achieved with the
hardware's default "age based" thread execution setting.  However on
ATS-M, switching this to "round robin after dependencies" provides
better performance.  We'll add a new "tuning" feature flag to the ATS-M
device info to enable/disable this setting.

Bspec: 68331
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Atwood <matthew.s.atwood@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220826212718.409948-1-matthew.d.roper@intel.com

Revert "drm/i915/dg2: Add preemption changes for Wa_14015141709"

This reverts commit ca6920811aa5428270dd78af0a7a36b10119065a.

The intent of Wa_14015141709 was to inform us that userspace can no
longer control object-level preemption as it has on past platforms
(i.e., by twiddling register bit CS_CHICKEN1[0]).  The description of
the workaround in the spec wasn't terribly well-written, and when we
requested clarification from the hardware teams we were told that on the
kernel side we should also probably stop setting
FF_SLICE_CS_CHICKEN1[14], which is the register bit that directs the
hardware to honor the settings in per-context register CS_CHICKEN1.  It
turns out that this guidance about FF_SLICE_CS_CHICKEN1[14] was a
mistake; even though CS_CHICKEN1[0] is non-operational and useless to
userspace, there are other bits in the register that do still work and
might need to be adjusted by userspace in the future (e.g., to implement
other workarounds that show up).  If we don't set
FF_SLICE_CS_CHICKEN1[14] in i915, then those future workarounds would
not take effect.

This miscommunication came to light because another workaround
(Wa_16013994831) has now shown up that requires userspace to adjust the
value of CS_CHICKEN[10] in certain circumstances.  To ensure userspace's
updates to this chicken bit are handled properly by the hardware, we
need to make sure that FF_SLICE_CS_CHICKEN1[14] is once again set by the
kernel.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Atwood <matthew.s.atwood@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220826210233.406482-1-matthew.d.roper@intel.com

drm/i915/selftests: allow misaligned_pin test work with unmappable memory

In case of Small BAR configurations stolen local memory can be unmappable.
Since the test does not touch the memory, passing I915_BO_ALLOC_GPU_ONLY
flag to i915_gem_object_create_region, will prevent -ENOSPC error from
_i915_gem_object_stolen_init.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6565
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/20220825154239.52343-1-andrzej.hajda@intel.com

drm/i915/guc: Remove log size module parameters

Remove the module parameters for configuring GuC log size.

We should instead rely on tuning the defaults to be usable for
reporting bugs.

v2:
- Use correct 1M unit

Fixes: 8ad0152afb1b ("drm/i915/guc: Make GuC log sizes runtime configurable")
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Alan Previn <alan.previn.teres.alexis@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220826092343.184568-1-joonas.lahtinen@linux.intel.com

drm/i915/dg2: Incorporate Wa_16014892111 into DRAW_WATERMARK tuning

Although register tuning settings are generally implemented via the
workaround infrastructure, it turns out that the DRAW_WATERMARK register
is not properly saved/restored by hardware around power events (i.e.,
RC6 entry) so updates to the value cannot be applied in the usual
manner.  New workaround Wa_16014892111 informs us that any tuning
updates to this register must instead be applied via an INDIRECT_CTX
batch buffer.  This will ensure that the necessary value is re-applied
when a context begins running, even if an RC6 entry had wiped the
register back to hardware defaults since the last context ran.

Fixes: 6dc85721df74 ("drm/i915/dg2: Add additional tuning settings")
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6642
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/20220823202449.83727-1-matthew.d.roper@intel.com

drm/i915/pxp: don't start pxp without mei_pxp bind

pxp will not start correctly until after mei_pxp bind completes and
intel_pxp_init_hw() is called.
Wait for the bind to complete before proceeding with startup.

This fixes a race condition during bootup where we observed a small
window for pxp commands to be sent, starting pxp before mei_pxp bind
completed.

Changes since v2:
- wait for pxp_component to bind instead of returning -EAGAIN (Daniele)

Changes since v1:
- check pxp_component instead of pxp_component_added (Daniele)
- pxp_component needs tee_mutex (Daniele)
- return -EAGAIN so caller knows to retry (Daniele)

Signed-off-by: Juston Li <justonli@chromium.org>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220818174205.2412730-1-justonli@chromium.org

drm/i915/mtl: Don't mask off CCS according to DSS fusing

Unlike the Xe_HP platforms, MTL only has a single CCS engine; the
quad-based engine masking logic does not apply to this platform (or
presumably any future platforms that only have 0 or 1 CCS).

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220818234202.451742-5-radhakrishna.sripada@intel.com

drm/i915/mtl: MMIO range is now 4MB

Previously only dgfx platforms had a 4MB MMIO range, but starting with
MTL we now use the larger range for all platforms.

Bspec: 63834, 63830
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220818234202.451742-4-radhakrishna.sripada@intel.com

drm/i915: Skip Bit12 fw domain reset for gen12+

Bit12 of the Forcewake request register should not be cleared post
gen12. Do not touch this bit while clearing during fw domain reset.

v2: Tweak the comment to drop older platforms(MattR)

Bspec: 52542

Signed-off-by: Sushma Venkatesh Reddy <sushma.venkatesh.reddy@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/20220817224304.255767-1-radhakrishna.sripada@intel.com