drm/xe: Only set VM->asid for platforms that support a ASID

This will help with TLB invalidation as the ASID in TLB invalidate
should be zero for platforms that do not support a ASID.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>

drm/xe: Add TDR for invalidation fence timeout cleanup

Endless fences are not good, add a TDR to cleanup any invalidation
fences which have not received an invalidation message within a timeout
period.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>

drm/xe: Add TLB invalidation fence ftrace

This will help debug issues with TLB invalidation fences.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Kernel doc GT TLB invalidations

Document all exported functions.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Invalidate TLB after unbind is complete

This gets tricky as we can't do the TLB invalidation until the unbind
operation is done on the hardware and we can't signal the unbind as
complete until the TLB invalidation is done. To work around this we
create an unbind fence which does a TLB invalidation after unbind is
done on the hardware, signals on TLB invalidation completion, and this
fence is installed in the BO dma-resv slot and installed in out-syncs
for the unbind operation.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Suggested-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com
Suggested-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Add TLB invalidation fence

Fence will be signaled when TLB invalidation completion.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Suggested-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Move TLB invalidation variable to own sub-structure in GT

TLB invalidations no longer just restricted to USM, move the variables
to own sub-structure.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Break of TLB invalidation into its own file

TLB invalidation is used by more than USM (page faults) so break this
code out into its own file.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Don't process TLB invalidation done in CT fast-path

We can't currently do this due to TLB invalidation done handler
expecting the seqno being received in-order, with the fast-path a TLB
invalidation done could pass one being processed in the slow-path in an
extreme corner case. Remove TLB invalidation done from the fast-path for
now and in a follow up reenable this once the TLB invalidation done
handler can deal with out of order seqno.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/migrate: Update emit_pte to cope with a size level than 4k

emit_pte assumes the size argument is 4k aligned, this may not be true
for the PTEs emitted for CSS as seen by below call stack:

[   56.734228] xe_migrate_copy:585: size=327680, ccs_start=327680, css_size=1280,4096
[   56.734250] xe_migrate_copy:643: size=262144
[   56.734252] emit_pte:404: ptes=64
[   56.734255] emit_pte:418: chunk=64
[   56.734257] xe_migrate_copy:650: size=1024 @ CCS emit PTE
[   56.734259] emit_pte:404: ptes=1
[   56.734261] emit_pte:418: chunk=1
[   56.734339] xe_migrate_copy:643: size=65536
[   56.734342] emit_pte:404: ptes=16
[   56.734344] emit_pte:418: chunk=16
[   56.734346] xe_migrate_copy:650: size=256 # CCS emit PTE
[   56.734348] emit_pte:404: ptes=1
[   56.734350] emit_pte:418: chunk=1
[   56.734352] xe_res_next:174: size=4096, remaining=0

Update emit_pte to handle sizes less than 4k.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/ggtt: fix GGTT scratch usage for DG2

Scratch page is in VRAM, and therefore requires 64K GTT layout. In GGTT
world this just means having 16 consecutive entries, with 64K GTT
alignment for the GTT address of the first entry (also matching physical
alignment). However to keep things simple just dump it into system
memory, like we already do for ppGTT.  While we are here, also give it
known default value.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/ggtt: fix alignment usage for DG2

Spec says we need to use 64K VRAM pages for GGTT on platforms like DG2.
In GGTT this just means aligning the GTT address to 64K and ensuring
that we have 16 consecutive entries each pointing to the respective 4K
entry. We already ensure we have 64K pages underneath, so it's just a
case of forcing the GTT alignment.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/ppgtt: fix scratch page usage on DG2

On DG2 when running the xe_vm IGT, the kernel generates loads of CAT
errors and GT resets (sometimes at least).  On small-bar systems seems
to trigger a lot more easily (maybe due to difference in allocation
strategy). Appears to be related to scratch, since we seem to use the
64K TLB hint on scratch entries, even though the scratch page is a 4K
vram page. Bumping the scratch page size and physical alignment seems
to fix it. Or at least we no longer hit:

[  148.872683] xe 0000:03:00.0: [drm] Engine memory cat error: guc_id=0
[  148.872701] xe 0000:03:00.0: [drm] Engine memory cat error: guc_id=0
[  148.875108] WARNING: CPU: 0 PID: 953 at drivers/gpu/drm/xe/xe_guc_submit.c:797

However to keep things simple, so we don't have to deal with 64K TLB
hints, just move the scratch page into system memory on platforms that
require 64K VRAM pages.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/ppgtt: clear the scratch page

We need to ensure we don't leak the contents to userspace.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/bo: reduce xe_bo_create_pin_map() restrictions

On DGFX this blows up if can call this with a system memory object:

XE_BUG_ON(!mem_type_is_vram(place->mem_type) && place->mem_type != XE_PL_STOLEN);

If we consider dpt it looks like we can already in theory hit this, if
we run out of vram and stolen vram. It at least seems reasonable to
allow calling this on any object which supports CPU access.

Note this also changes the behaviour with stolen VRAM and suspend, such
that we no longer attempt to migrate stolen objects into system memory.
However nothing in stolen should ever need to be restored (same on
integrated), so should be fine. Also on small-bar systems the stolen
portion is pretty much always non-CPU accessible, and currently pinned
objects use plain memcpy when being moved, which doesn't play nicely.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Map initial FB at the same place in GGTT too

I saw a flicker when booting xe, and it's very likely that the original
FB was not mapped at the same place when inheriting, fix it.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Convert memory device refcount to s32

The comparison with < 0 suggests that the memory device access
should be signed to handle underflow. This makes it work more reliably.

As a result, the max refcount is now S32_MAX instead.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/uapi: Add XE_ENGINE_GET_PROPERTY uAPI

This is intended to get some properties that are of interest of UMDs
like the ban state.

Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/uapi: Rename XE_ENGINE_PROPERTY_X to XE_ENGINE_SET_PROPERTY_X

Engine property get uAPI will be added, so to avoid ambiguity here
renaming XE_ENGINE_PROPERTY_X to XE_ENGINE_SET_PROPERTY_X.

No changes in behavior.

Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Use global macros to set PM functions

This aligns with other drivers and fixes build failure when
CONFIG_PM_SLEEP is not set, such as on RISC-V.

Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: skip Kunit tests requiring real hardware when running on UML

Some tests are meant to run only on real hardware. Skip those,
if no device was found.

Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/guc_pc: Fix Meteor Lake registers.

When adding the frequency management, Meteor Lake platform
was left behind. Handling it properly now.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Cc: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Fix Meteor Lake rsa issue on guc loading

[  117.901473] xe 0000:00:02.0: [drm] GuC load failed: status = 0x400000A0
[  117.901506] xe 0000:00:02.0: [drm] GuC load failed: status: Reset = 0, BootROM = 0x50, UKernel = 0x00, MIA = 0x00, Auth = 0x01

Signed-off-by: Philippe Lecluse <philippe.lecluse@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/gt: Enable interrupt while initializing root gt

At present the interrupts are enabled while initializing the last GT.
But this is incorrect for a Multi-GT platform, as root GT initialization
will fail with interrupt disabled. Interrupts are required for
the GuC submission triggered during initialization.
Enable the interrupt during the root GT initialization.

Signed-off-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: KUnit tests depend on CONFIG_DRM_FBDEV_EMULATION

ERROR:root:../drivers/gpu/drm/xe/display/intel_fbdev.c:585:5: error: redefinition of ‘intel_fbdev_init’
  585 | int intel_fbdev_init(struct drm_device *dev)
      |     ^~~~~~~~~~~~~~~~
In file included from ../drivers/gpu/drm/xe/display/intel_fbdev.c:55:
../drivers/gpu/drm/xe/display/intel_fbdev.h:26:19: note: previous definition of ‘intel_fbdev_init’ with type ‘int(struct drm_device *)’
   26 | static inline int intel_fbdev_init(struct drm_device *dev)
      |                   ^~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:626:6: error: redefinition of ‘intel_fbdev_initial_config_async’
  626 | void intel_fbdev_initial_config_async(struct drm_device *dev)
      |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:31:20: note: previous definition of ‘intel_fbdev_initial_config_async’ with type ‘void(struct drm_device *)’
   31 | static inline void intel_fbdev_initial_config_async(struct drm_device *dev)
      |                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:646:6: error: redefinition of ‘intel_fbdev_unregister’
  646 | void intel_fbdev_unregister(struct drm_i915_private *dev_priv)
      |      ^~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:35:20: note: previous definition of ‘intel_fbdev_unregister’ with type ‘void(struct xe_device *)’
   35 | static inline void intel_fbdev_unregister(struct drm_i915_private *dev_priv)
      |                    ^~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:661:6: error: redefinition of ‘intel_fbdev_fini’
  661 | void intel_fbdev_fini(struct drm_i915_private *dev_priv)
      |      ^~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:39:20: note: previous definition of ‘intel_fbdev_fini’ with type ‘void(struct xe_device *)’
   39 | static inline void intel_fbdev_fini(struct drm_i915_private *dev_priv)
      |                    ^~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:692:6: error: redefinition of ‘intel_fbdev_set_suspend’
  692 | void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
      |      ^~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:43:20: note: previous definition of ‘intel_fbdev_set_suspend’ with type ‘void(struct drm_device *, int,  bool)’ {aka ‘void(struct drm_device *, int,  _Bool)’}
   43 | static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
      |                    ^~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:751:6: error: redefinition of ‘intel_fbdev_output_poll_changed’
  751 | void intel_fbdev_output_poll_changed(struct drm_device *dev)
      |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:47:20: note: previous definition of ‘intel_fbdev_output_poll_changed’ with type ‘void(struct drm_device *)’
   47 | static inline void intel_fbdev_output_poll_changed(struct drm_device *dev)
      |                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:770:6: error: redefinition of ‘intel_fbdev_restore_mode’
  770 | void intel_fbdev_restore_mode(struct drm_device *dev)
      |      ^~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:51:20: note: previous definition of ‘intel_fbdev_restore_mode’ with type ‘void(struct drm_device *)’
   51 | static inline void intel_fbdev_restore_mode(struct drm_device *dev)
      |                    ^~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.c:785:27: error: redefinition of ‘intel_fbdev_framebuffer’
  785 | struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev)
      |                           ^~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_fbdev.h:54:41: note: previous definition of ‘intel_fbdev_framebuffer’ with type ‘struct intel_framebuffer *(struct intel_fbdev *)’
   54 | static inline struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev)
      |                                         ^~~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/Kconfig.debug: select DEBUG_FS for KUnit runs

KUnit reuquires debugfs, as otherwise, it won't build:

$ make ARCH=x86_64 O=.kunit --jobs=8
ERROR:root:../drivers/gpu/drm/xe/display/intel_display_debugfs.c:1612:6: error: redefinition of ‘intel_display_debugfs_register’
 1612 | void intel_display_debugfs_register(struct drm_i915_private *i915)
      |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../drivers/gpu/drm/xe/display/intel_display_debugfs.c:18:
../drivers/gpu/drm/xe/display/intel_display_debugfs.h:18:20: note: previous definition of ‘intel_display_debugfs_register’ with type ‘void(struct xe_device *)’
   18 | static inline void intel_display_debugfs_register(struct drm_i915_private *i915) {}
      |                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_display_debugfs.c:1935:6: error: redefinition of ‘intel_connector_debugfs_add’
 1935 | void intel_connector_debugfs_add(struct intel_connector *intel_connector)
      |      ^~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_display_debugfs.h:19:20: note: previous definition of ‘intel_connector_debugfs_add’ with type ‘void(struct intel_connector *)’
   19 | static inline void intel_connector_debugfs_add(struct intel_connector *connector) {}
      |                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_display_debugfs.c:1993:6: error: redefinition of ‘intel_crtc_debugfs_add’
 1993 | void intel_crtc_debugfs_add(struct drm_crtc *crtc)
      |      ^~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/xe/display/intel_display_debugfs.h:20:20: note: previous definition of ‘intel_crtc_debugfs_add’ with type ‘void(struct drm_crtc *)’
   20 | static inline void intel_crtc_debugfs_add(struct drm_crtc *crtc) {}
      |                    ^~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Add min config for kunit integration ARCH=um

Some of the tests may benefit from running with ARCH=um, forgoing any
additional setup on the CI build side. Add min config for that.

Tested with:

./tools/testing/kunit/kunit.py build \
--kunitconfig drivers/gpu/drm/xe/.kunitconfig \
--jobs $(nproc) \
--build_dir build_kunit

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>

drm/xe: Add documentation for mem_type

mem_type field was added in commit d8b52a02cb40 ("drm/xe: Implement
stolen memory.") to designate the TTM memory type for that mgr. Add
kernel-doc with its description.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/gt: Fix min() with u32 and u64

Fix the following error while building for 32b:

In file included from ../drivers/gpu/drm/xe/xe_gt.c:6:
../drivers/gpu/drm/xe/xe_gt.c: In function ‘gt_ttm_mgr_init’:
../include/linux/minmax.h:20:35: error: comparison of distinct pointer types lacks a cast [-Werror]
   20 |         (!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
      |                                   ^~

Cast it to u64 so size of the second operand matches the first one when
building it for 32 bits.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Fix tracepoints on 32b

Leave the types as u64, but cast the pointers to unsigned long before
assigning so the compiler doesn't throw warning about casting a pointer
to integer of different size.

Also, size_t should use %zu, not %ld.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/mmio: Use non-atomic writeq/readq variant for 32b

writeq() and readq() and other functions working on 64 bit variables
are not provided by 32b arch. For that it's needed to choose between
linux/io-64-nonatomic-hi-lo.h and linux/io-64-nonatomic-lo-hi.h,
spliting the read/write in 2 accesses. For xe driver, it doesn't matter
much, so just choose one and include in xe_mmio.h.

This also removes some ifdef CONFIG_64BIT we had around because of the
missing 64bit functions.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Fix some log messages on 32b

Either use the proper format or cast up to 64b depending on the case.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/ggtt: Use BIT_ULL() for 64bit

Make sure it's 64bit value in both 32b and 64b arch.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: fix xe_mmio_total_vram_size

As also cause issue on PVC, moving back
to what we did before stolen was
introduced

Signed-off-by: Philippe Lecluse <philippe.lecluse@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/pcode: fix pcode error check

On DG2 we are now getting:

[  104.456607] xe 0000:03:00.0: [drm] *ERROR* PCODE timeout, retrying with preemption disabled

Looks like we just need to invert the error check for
xe_pcode_try_request(), which returns zero on success.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: enforce GSMBASE for DG1 instead of BAR2

On DG1, BAR2 is not reliable for reporting Vram size, need to use GSMBASE.
Simplify xe_mmio_total_vram_size to report vram size and usable size.

Signed-off-by: Philippe Lecluse <philippe.lecluse@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Fix hidden gotcha regression with bo create

The bo_create ioctl relied on the internal ordering of memory regions to
be the same, make sure we don't allocate stolen instead of VRAM0.

Also remove a debug warning left in.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Philippe Lecluse <philippe.lecluse1@gmail.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/guc: Add support GuC MMIO send / recv

SRIOV has a use case of GuC MMIO send / recv, add a function for it.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Philippe Lecluse <philippe.lecluse1@gmail.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/guc: s/xe_guc_send_mmio/xe_guc_mmio_send

Now aligns with the xe_guc_ct_send naming.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Philippe Lecluse <philippe.lecluse1@gmail.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/guc: Report submission version of GuC firmware

Starting in 70.6.* GuC firmware the CSS header includes the submission
version, pull this from the CSS header. Prior 70.* versions accidentally
omitted this informatio so hard code to the correct values. This
information will be used by VFs when communicating with the PF.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Philippe Lecluse <philippe.lecluse1@gmail.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Fake pulling gt->info.engine_mask from hwconfig blob

The blob doesn't fully support this yet, so fake for now to ensure our
driver load order is correct.

Once the blob supports pulling gt->info.engine_mask from the blob, this
patch can be removed.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Ensure VMA not userptr before calling xe_bo_is_stolen

Fix the below splat:

[  142.510525] [IGT] xe_exec_basic: starting subtest once-userptr
[  142.511339] BUG: kernel NULL pointer dereference, address: 0000000000000228
[  142.518311] #PF: supervisor read access in kernel mode
[  142.523458] #PF: error_code(0x0000) - not-present page
[  142.528604] PGD 0 P4D 0
[  142.531153] Oops: 0000 [#1] PREEMPT SMP NOPTI
[  142.535518] CPU: 4 PID: 1199 Comm: kworker/u16:8 Not tainted 6.1.0-rc1-xe+ #1
[  142.542656] Hardware name: Intel Corporation Tiger Lake Client Platform/TigerLake U DDR4 SODIMM RVP, BIOS TGLSFWI1.R00.3243.A01.2006102133 06/10/2020
[  142.556033] Workqueue: events_unbound async_op_work_func [xe]
[  142.561810] RIP: 0010:xe_bo_is_stolen+0x0/0x20 [xe]
[  142.566709] Code: 20 c8 75 05 83 fa 07 74 05 c3 cc cc cc cc 48 8b 87 08 02 00 00 0f b6 80 2c ff ff ff c3 cc cc cc cc 66 0f 1f 84 00 00 00 00 00 <48> 8b 87 28 02 00 00 83 78 10 07 0f 94 c0 c3 cc cc cc cc 66 66 2e
[  142.585447] RSP: 0018:ffffc900019eb888 EFLAGS: 00010246
[  142.590678] RAX: 0000000000000002 RBX: 0000000000000000 RCX: ffff88813f6a2108
[  142.597821] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[  142.604962] RBP: ffffc900019ebbc0 R08: 0000000000000001 R09: 0000000000000000
[  142.612101] R10: 0000000000000001 R11: 0000000000000001 R12: ffff88814107d600
[  142.619242] R13: ffffc900019eba20 R14: ffff888140442000 R15: 0000000000000000
[  142.626378] FS:  0000000000000000(0000) GS:ffff88849fa00000(0000) knlGS:0000000000000000
[  142.634468] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  142.640219] CR2: 0000000000000228 CR3: 000000010a4c0006 CR4: 0000000000770ee0
[  142.647361] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  142.654505] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  142.661639] PKRU: 55555554
[  142.664367] Call Trace:
[  142.666830]  <TASK>
[  142.668947]  __xe_pt_bind_vma+0x1a1/0xa50 [xe]
[  142.673417]  ? unwind_next_frame+0x187/0x770
[  142.677699]  ? __thaw_task+0xc0/0xc0
[  142.681293]  ? __lock_acquire+0x5e4/0x26e0
[  142.685409]  ? lockdep_hardirqs_on+0xbf/0x140
[  142.689779]  ? lock_acquire+0xd2/0x310
[  142.693548]  ? mark_held_locks+0x49/0x80
[  142.697485]  ? xe_vm_bind_vma+0xf1/0x3d0 [xe]
[  142.701866]  xe_vm_bind_vma+0xf1/0x3d0 [xe]
[  142.706082]  xe_vm_bind+0x76/0x140 [xe]
[  142.709944]  vm_bind_ioctl+0x26f/0xb40 [xe]
[  142.714161]  ? async_op_work_func+0x20c/0x450 [xe]
[  142.718974]  async_op_work_func+0x20c/0x450 [xe]
[  142.723620]  process_one_work+0x263/0x580
[  142.727645]  ? process_one_work+0x580/0x580
[  142.731839]  worker_thread+0x4d/0x3b0
[  142.735518]  ? process_one_work+0x580/0x580
[  142.739714]  kthread+0xeb/0x120
[  142.742872]  ? kthread_complete_and_exit+0x20/0x20
[  142.747671]  ret_from_fork+0x1f/0x30
[  142.751264]  </TASK>

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Implement stolen memory.

This adds support for stolen memory, with the same allocator as
vram_mgr. This allows us to skip a whole lot of copy-paste,
by re-using parts of xe_ttm_vram_mgr.

The stolen memory may be bound using VM_BIND, so it performs like any
other memory region.

We should be able to map a stolen BO directly using the physical memory
location instead of through GGTT even on old platforms, but I don't know
what the effects are on coherency.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe: Take memory ref on kernel job creation

When a job is inflight we may access memory to read the hardware seqno.
All user jobs have VM open which has a ref but kernel jobs do not
require VM so it is possible to not have memory ref. To avoid this, take
a memory ref on kernel job creation.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/xe/migrate: Add kerneldoc for the migrate subsystem

Add kerneldoc for structs and external functions.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Mauro Carvalho Chehab <mchehab@kernel.org>

drm/xe: Remove i915_utils dependency from xe_pcode.

Expand xe_mmio_wait32 to accept atomic and then use
that directly when possible, and create own routine to
wait for the pcode status.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Convert xe_mmio_wait32 to us so we can stop using wait_for_us.

Another clean-up towards killing the usage of i915_utils.h

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Let's avoid i915_utils in the xe_force_wake.

We can run the bit operation locally without yet another macro.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Stop using i915_utils in xe_wopcm.

We don't need any macro for a simple check we can do explicitly
and clear.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Remove i915_utils dependency from xe_guc_pc.

To make it simpler, all of the status checks also waits and
times out.

Also, no ktime precision is needed in this case, and we
can use usleep_range because we are not in atomic paths here.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Wait for success on guc done.

Rather than a constant check on proto and wait not busy,
let's wait for the expected success and then check the
protocol afterwards.

With this, we can now use the regular xe_mmio_wait32
and kill this local need for the wait_for.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Convert guc_ready to regular xe_mmio_wait32

Possible now that the wait function returns the last read value.

So we can remove the users of i915's wait_for one by one...

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Let's return last value read on xe_mmio_wait32.

This is already useful because it avoids some extra reads
where registers might have changed after the timeout decision.

But also, it will be important to end the kill of i915's wait_for.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Stop using i915's range_overflows_t macro.

Let's do it directly.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Implement a local xe_mmio_wait32

Then, move the i915_utils.h include to its user.

The overall goal is to kill all the usages of the i915_utils
stuff.

Yes, wait_for also depends on <linux/delay.h>, so they go
together to where it is needed. It will be likely needed
anyway directly for udelay or usleep_range.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>

drm/xe: Introduce a new DRM driver for Intel GPUs

Xe, is a new driver for Intel GPUs that supports both integrated and
discrete platforms starting with Tiger Lake (first Intel Xe Architecture).

The code is at a stage where it is already functional and has experimental
support for multiple platforms starting from Tiger Lake, with initial
support implemented in Mesa (for Iris and Anv, our OpenGL and Vulkan
drivers), as well as in NEO (for OpenCL and Level0).

The new Xe driver leverages a lot from i915.

As for display, the intent is to share the display code with the i915
driver so that there is maximum reuse there. But it is not added
in this patch.

This initial work is a collaboration of many people and unfortunately
the big squashed patch won't fully honor the proper credits. But let's
get some git quick stats so we can at least try to preserve some of the
credits:

Co-developed-by: Matthew Brost <matthew.brost@intel.com>
Co-developed-by: Matthew Auld <matthew.auld@intel.com>
Co-developed-by: Matt Roper <matthew.d.roper@intel.com>
Co-developed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Co-developed-by: Francois Dugast <francois.dugast@intel.com>
Co-developed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Co-developed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Co-developed-by: Philippe Lecluse <philippe.lecluse@intel.com>
Co-developed-by: Nirmoy Das <nirmoy.das@intel.com>
Co-developed-by: Jani Nikula <jani.nikula@intel.com>
Co-developed-by: José Roberto de Souza <jose.souza@intel.com>
Co-developed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Co-developed-by: Dave Airlie <airlied@redhat.com>
Co-developed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Co-developed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Co-developed-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>

Merge tag 'drm-misc-next-2023-12-07' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for 6.8:

UAPI Changes:
  - Remove Userspace Mode-Setting ioctls
  - v3d: New uapi to handle jobs involving the CPU

Cross-subsystem Changes:

Core Changes:
  - atomic: Add support for FB-less planes which got reverted a bit
    later for lack of IGT tests and userspace code, Dump private objects
    state in drm_state_dump.
  - dma-buf: Add fence deadline support
  - encoder: Create per-encoder debugfs directory, move the bridge chain
    file to that directory

Driver Changes:
  - Include drm_auth.h in driver that use it but don't include it, Drop
    drm_plane_helper.h from drivers that include it but don't use it
  - imagination: Plenty of small fixes
  - panfrost: Improve interrupt handling at poweroff
  - qaic: Convert to persistent DRM devices
  - tidss: Support for the AM62A7, a few probe improvements, some cleanups
  - v3d: Support for jobs involving the CPU

  - bridge:
    - Create transparent aux-bridge for DP/USB-C
    - lt8912b: Add suspend/resume support and power regulator support

  - panel:
    - himax-hx8394: Drop prepare, unprepare and shutdown logic, Support
      panel rotation
    - New panels: BOE BP101WX1-100, Powkiddy X55, Ampire AM8001280G,
      Evervision VGG644804, SDC ATNA45AF01

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <mripard@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/yu5heqaufyeo4nlowzieu4s5unwqrqyx4jixbfjmzdon677rpk@t53vceua2dao

Merge tag 'drm-intel-next-2023-12-07' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

- Improve display debug msgs and other general clean-ups (Ville, Rahuul)
- PSR fixes and improvements around selective fetch (Jouni, Ville)
- Remove FBC restrictions for Xe2LPD displays (Vinod)
- Skip some timing checks on BXT/GLK DSI transcoders (Ville)
- DP MST Fixes (Ville)
- Correct the input parameter on _intel_dsb_commit (heminhong)
- Fix IP version of the display WAs (Bala)
- DGFX uses direct VBT pin mapping (Clint)
- Proper handling of bool on PIPE_CONF_CHECK macros (Jani)
- Skip state verification with TBT-ALT mod (Mika Kahona)
- General organization of display code for reusage with Xe
  (Jouni, Luca, Jani, Maarten)
- Squelch a sparse warning (Jani)
- Don't use "proxy" headers (Andy Shevchenko)
- Use devm_gpiod_get() for all GPIOs (Hans)
- Fix ADL+ tiled plane stride (Ville)
- Use octal permissions in display debugfs (Jani)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ZXIWG6bRYaUw0w6-@intel.com

drm/i915/display: Add intel_fb_bo_framebuffer_fini

Xe needs intel_fb_bo_framebuffer_fini for taking care of unpinning the fb
and taking reference.  In i915 this can be empty.

Also move intel_frontbuffer_get to be done after
intel_fb_bo_framebuffer_init to have reasonable sequences:

intel_fb_bo_framebuffer_init
intel_frontbuffer_get
...
intel_frontbuffer_put
intel_fb_bo_framebuffer_fini

v2: Empty function instead of define

Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231207083451.2184562-1-jouni.hogander@intel.com

drm/i915/display: Use i915_gem_object_get_dma_address to get dma address

Works better for xe like that. obj is no longer const.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204134946.16219-1-maarten.lankhorst@linux.intel.com
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>

drm/i915: use octal permissions in display debugfs

Octal permissions are preferred over the symbolics ones.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205134143.2427661-3-jani.nikula@intel.com

drm/i915: pass struct intel_connector to connector debugfs fops

Prefer struct intel_connector over struct drm_connector, and unify the
declarations in the fops.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205134143.2427661-2-jani.nikula@intel.com

drm/i915: use intel_connector in intel_connector_debugfs_add()

Prefer struct intel_connector over struct drm_connector.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205134143.2427661-1-jani.nikula@intel.com

drm/i915/rpm: add rpm_to_i915() helper around container_of()

Reduce the duplication.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205121545.2338665-1-jani.nikula@intel.com

drm/i915: handle uncore spinlock when not available

The uncore code may not always be available (e.g. when we build the
display code with Xe), so we can't always rely on having the uncore's
spinlock.

To handle this, split the spin_lock/unlock_irqsave/restore() into
spin_lock/unlock() followed by a call to local_irq_save/restore() and
create wrapper functions for locking and unlocking the uncore's
spinlock.  In these functions, we have a condition check and only
actually try to lock/unlock the spinlock when I915 is defined, and
thus uncore is available.

This keeps the ifdefs contained in these new functions and all such
logic inside the display code.

Cc: Tvrtko Ursulin <tvrto.ursulin@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231201100032.1367589-1-luciano.coelho@intel.com

drm/mipi-dsi: Fix detach call without attach

It's been reported that DSI host driver's detach can be called without
the attach ever happening:

https://lore.kernel.org/all/20230412073954.20601-1-tony@atomide.com/

After reading the code, I think this is what happens:

We have a DSI host defined in the device tree and a DSI peripheral under
that host (i.e. an i2c device using the DSI as data bus doesn't exhibit
this behavior).

The host driver calls mipi_dsi_host_register(), which causes (via a few
functions) mipi_dsi_device_add() to be called for the DSI peripheral. So
now we have a DSI device under the host, but attach hasn't been called.

Normally the probing of the devices continues, and eventually the DSI
peripheral's driver will call mipi_dsi_attach(), attaching the
peripheral.

However, if the host driver's probe encounters an error after calling
mipi_dsi_host_register(), and before the peripheral has called
mipi_dsi_attach(), the host driver will do cleanups and return an error
from its probe function. The cleanups include calling
mipi_dsi_host_unregister().

mipi_dsi_host_unregister() will call two functions for all its DSI
peripheral devices: mipi_dsi_detach() and mipi_dsi_device_unregister().
The latter makes sense, as the device exists, but the former may be
wrong as attach has not necessarily been done.

To fix this, track the attached state of the peripheral, and only detach
from mipi_dsi_host_unregister() if the peripheral was attached.

Note that I have only tested this with a board with an i2c DSI
peripheral, not with a "pure" DSI peripheral.

However, slightly related, the unregister machinery still seems broken.
E.g. if the DSI host driver is unbound, it'll detach and unregister the
DSI peripherals. After that, when the DSI peripheral driver unbound
it'll call detach either directly or using the devm variant, leading to
a crash. And probably the driver will crash if it happens, for some
reason, to try to send a message via the DSI bus.

But that's another topic.

Tested-by: H. Nikolaus Schaller <hns@goldelico.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230921-dsi-detach-fix-v1-1-d0de2d1621d9@ideasonboard.com

drm/bridge: tc358767: Fix return value on error case

If the hpd_pin is invalid, the driver returns 'ret'. But 'ret' contains
0, instead of an error value.

Return -EINVAL instead.

Fixes: f25ee5017e4f ("drm/bridge: tc358767: add IRQ and HPD support")
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231103-uninit-fixes-v2-4-c22b2444f5f5@ideasonboard.com

drm/bridge: cdns-mhdp8546: Fix use of uninitialized variable

'ret' could be uninitialized at the end of the function, although it's
not clear if that can happen in practice.

Fixes: 6a3608eae6d3 ("drm: bridge: cdns-mhdp8546: Enable HDCP")
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231103-uninit-fixes-v2-3-c22b2444f5f5@ideasonboard.com

drm/framebuffer: Fix use of uninitialized variable

smatch reports:

drivers/gpu/drm/drm_framebuffer.c:654 drm_mode_getfb2_ioctl() error: uninitialized symbol 'ret'.

'ret' is possibly not set when there are no errors, causing the error
above. I can't say if that ever happens in real-life, but in any case I
think it is good to initialize 'ret' to 0.

Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231103-uninit-fixes-v2-2-c22b2444f5f5@ideasonboard.com

drm/drm_file: fix use of uninitialized variable

smatch reports:

drivers/gpu/drm/drm_file.c:967 drm_show_memory_stats() error: uninitialized symbol 'supported_status'.

'supported_status' is only set in one code path. I'm not familiar with
the code to say if that path will always be ran in real life, but
whether that is the case or not, I think it is good to initialize
'supported_status' to 0 to silence the warning (and possibly fix a bug).

Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231103-uninit-fixes-v2-1-c22b2444f5f5@ideasonboard.com

drm/bridge: nxp-ptn3460: simplify some error checking

The i2c_master_send/recv() functions return negative error codes or
they return "len" on success.  So the error handling here can be written
as just normal checks for "if (ret < 0) return ret;".  No need to
complicate things.

Btw, in this code the "len" parameter can never be zero, but even if
it were, then I feel like this would still be the best way to write it.

Fixes: 914437992876 ("drm/bridge: nxp-ptn3460: fix i2c_master_send() error checking")
Suggested-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Robert Foss <rfoss@kernel.org>
Signed-off-by: Robert Foss <rfoss@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/04242630-42d8-4920-8c67-24ac9db6b3c9@moroto.mountain

drm/imagination: move update_logtype() into ifdef section

This function is only used when debugfs is enabled, and otherwise
causes a build warning:

drivers/gpu/drm/imagination/pvr_fw_trace.c:135:1: error: 'update_logtype' defined but not used [-Werror=unused-function]

Move the #ifdef check to include this function as well.

Fixes: cb56cd610866 ("drm/imagination: Add firmware trace to debugfs")
Acked-by: Frank Binns <frank.binns@imgtec.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231204073231.1164163-1-arnd@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>

drm/imagination: Removed unused functions in pvr_fw_trace

Fixing the warning below due to an unused file level vtable. Removing
only this causes additional warnings for the now unused functions, so
I've removed those too.

>> drivers/gpu/drm/imagination/pvr_fw_trace.c:205:37: warning: 'pvr_fw_trace_group_mask_fops' defined but not used [-Wunused-const-variable=]
     205 | static const struct file_operations pvr_fw_trace_group_mask_fops = {
         |                                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~

Changes since v1:
- Corrected hash in Fixes tag.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202311302054.MVYPxFCE-lkp@intel.com/
Fixes: cb56cd610866 ("drm/imagination: Add firmware trace to debugfs")
Signed-off-by: Donald Robson <donald.robson@imgtec.com>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204151337.60930-1-donald.robson@imgtec.com

drm/xlnx: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-9-tzimmermann@suse.de

drm/simpledrm: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Sui Jingfeng <suijingfeng@loongson.cn>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-8-tzimmermann@suse.de

drm/ofdrm: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Sui Jingfeng <suijingfeng@loongson.cn>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-7-tzimmermann@suse.de

drm/solomon: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-6-tzimmermann@suse.de

drm/shmobile: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-5-tzimmermann@suse.de

drm/loongson: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Sui Jingfeng <suijingfeng@loongson.cn>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Sui Jingfeng <suijingfeng@loongson.cn>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-4-tzimmermann@suse.de

drm/amdgpu: Do not include <drm/drm_plane_helper.h>

Remove unnecessary include statements for <drm/drm_plane_helper.h>.
The file contains helpers for non-atomic code and should not be
required by most drivers. No functional changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-3-tzimmermann@suse.de

drm/plane-helper: Move drm_plane_helper_atomic_check() into udl

The udl driver is the only caller of drm_plane_helper_atomic_check().
Move the function into the driver. No functional changes.

v2:
* fix documenation (Sui)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Sui Jingfeng <suijingfeng@loongson.cn>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204090852.1650-2-tzimmermann@suse.de

drm: Remove Kconfig option for legacy support (CONFIG_DRM_LEGACY)

Remove CONFIG_DRM_LEGACY from Kconfig. Nothing depends on the option.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-15-tzimmermann@suse.de

char/agp: Remove frontend code

The AGP subsystem supports a user-space interface via /dev/agpgart. It
is only enabled with DRM support for mode setting in user space. (i.e.,
CONFIG_DRM_LEGACY). All of that DRM code has been removed and the option
will go away. Hence remove the AGP frontend.

Modern DRM drivers with kernel mode setting handle AGP support internally.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-14-tzimmermann@suse.de

drm: Remove source code for non-KMS drivers

Remove all remaining source code for non-KMS drivers. These drivers
have been removed in v6.3 and won't comeback.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-13-tzimmermann@suse.de

drm: Remove locking for legacy ioctls and DRM_UNLOCKED

Modern DRM drivers acquire ioctl locks by themselves. Legacy ioctls
for user-space mode setting used to acquire drm_global_mutex. After
removing the ioctl entry points, also remove the locking code. The only
legacy ioctl without global locking was VBLANK_WAIT, which has been
removed as well. Hence remove the related DRM_UNLOCKED flag.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-12-tzimmermann@suse.de

drm: Remove support for legacy drivers

Remove all hooks and calls into code for user-space mode setting from
the DRM core. Without the drivers and ioctl entry points, none of this
is required any longer.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-11-tzimmermann@suse.de

drm: Remove the legacy DRM_IOCTL_MODESET_CTL ioctl

DRM drivers with user-space mode setting have been removed in Linux
v6.3. [1] Now remove the ioctl entry points for these drivers. Invoking
any of the ioctl ops will unconditionally return -EINVAL to user space.

Invoking DRM_IOCTL_MODESET_CTL is different from the other legacy
ioctl ops as it returns 0 even without CONFIG_DRM_LEGACY set. From the
original commit 29935554b384 ("drm: Disallow DRM_IOCTL_MODESET_CTL for
KMS drivers") it is not apparent how or why the operation differs from
the others. It is likely just an oversight in commit 61ae227032e7
("drm: allow removal of legacy codepaths (v4.1)"), which allowed
disabling leagacy ioctls in the first place. Still keep this removal
separate from the other ioctls to allow an easy revert, if necessary.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/series/111602/
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-10-tzimmermann@suse.de

drm: Remove entry points for legacy ioctls

DRM drivers with user-space mode setting have been removed in Linux
v6.3. [1] Now remove the ioctl entry points for these drivers. Invoking
any of the ioctl ops will unconditionally return -EINVAL to user space.
This has already been the behavior for kernels without CONFIG_DRM_LEGACY
set.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/series/111602/
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-9-tzimmermann@suse.de

drm/radeon: Do not include <drm/drm_legacy.h>

Including <drm/drm_legacy.h> is not required by radeon.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: "Pan, Xinhui" <Xinhui.Pan@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-8-tzimmermann@suse.de

drm: Include <drm/drm_device.h>

Include <drm/drm_device.h> in drm_ioc32.c. Resolves a depenency
on <drm/drm_legacy.h>, which will be removed.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-7-tzimmermann@suse.de

accel: Include <drm/drm_auth.h>

One of the source files includes <drm/drm_auth.h> via <drm/drm_legacy.h>,
which will be removed. Include drm_auth.h directly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-6-tzimmermann@suse.de

drm/i915: Include <drm/drm_auth.h>

One of the source files includes <drm/drm_auth.h> via <drm/drm_legacy.h>,
which will be removed. Include drm_auth.h directly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
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: intel-gfx@lists.freedesktop.org
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-5-tzimmermann@suse.de

drm: Include <drm/drm_auth.h>

One of the source files includes <drm/drm_auth.h> via <drm/drm_legacy.h>,
which will be removed. Include drm_auth.h directly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-4-tzimmermann@suse.de

drm: Fix TODO list mentioning non-KMS drivers

Non-KMS drivers have been removed from DRM. Update the TODO list
accordingly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: a276afc19eec ("drm: Remove some obsolete drm pciids(tdfx, mga, i810, savage, r128, sis, via)")
Cc: Cai Huoqing <cai.huoqing@linux.dev>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@gmail.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: dri-devel@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v6.3+
Cc: linux-doc@vger.kernel.org
Reviewed-by: David Airlie <airlied@gmail.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-3-tzimmermann@suse.de

drm/edp-panel: Sort the panel entries

Move the order of CMN 0x14e5 to make the list sorted.

Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231205123630.988663-3-treapking@chromium.org

drm/panel-edp: Add SDC ATNA45AF01

Add support for the SDC ATNA45AF01 panel.

Signed-off-by: Abel Vesa <abel.vesa@linaro.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231201-x1e80100-drm-panel-edp-v2-1-b0173484631a@linaro.org

drm/i915: Fix ADL+ tiled plane stride when the POT stride is smaller than the original

plane_view_scanout_stride() currently assumes that we had to pad the
mapping stride with dummy pages in order to align it. But that is not
the case if the original fb stride exceeds the aligned stride used
to populate the remapped view, which is calculated from the user
specified framebuffer width rather than the user specified framebuffer
stride.

Ignore the original fb stride in this case and just stick to the POT
aligned stride. Getting this wrong will cause the plane to fetch the
wrong data, and can lead to fault errors if the page tables at the
bogus location aren't even populated.

TODO: figure out if this is OK for CCS, or if we should instead increase
the width of the view to cover the entire user specified fb stride
instead...

Cc: Imre Deak <imre.deak@intel.com>
Cc: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204202443.31247-1-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Juha-Pekka Heikkila <juhapekka.heikkila@gmail.com>

drm/atomic: add private obj state to state dump

The drm_atomic_print_new_state() already prints private object state via
drm_atomic_private_obj_print_state(). Add private object state dumping
to __drm_state_dump(), so that it is also included into drm_state_dump()
output and into debugfs/dri/N/state file.

Reviewed-by: Rob Clark <robdclark@gmail.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231203000532.1290480-2-dmitry.baryshkov@linaro.org

drm/gpuvm: Let drm_gpuvm_bo_put() report when the vm_bo object is destroyed

Some users need to release resources attached to the vm_bo object when
it's destroyed. In Panthor's case, we need to release the pin ref so
BO pages can be returned to the system when all GPU mappings are gone.

This could be done through a custom drm_gpuvm::vm_bo_free() hook, but
this has all sort of locking implications that would force us to expose
a drm_gem_shmem_unpin_locked() helper, not to mention the fact that
having a ::vm_bo_free() implementation without a ::vm_bo_alloc() one
seems odd. So let's keep things simple, and extend drm_gpuvm_bo_put()
to report when the object is destroyed.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Danilo Krummrich <dakr@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204151406.1977285-1-boris.brezillon@collabora.com

drm/panfrost: Synchronize and disable interrupts before powering off

To make sure that we don't unintentionally perform any unclocked and/or
unpowered R/W operation on GPU registers, before turning off clocks and
regulators we must make sure that no GPU, JOB or MMU ISR execution is
pending: doing that requires to add a mechanism to synchronize the
interrupts on suspend.

Add functions panfrost_{gpu,job,mmu}_suspend_irq() which will perform
interrupts masking and ISR execution synchronization, and then call
those in the panfrost_device_runtime_suspend() handler in the exact
sequence of job (may require mmu!) -> mmu -> gpu.

As a side note, JOB and MMU suspend_irq functions needed some special
treatment: as their interrupt handlers will unmask interrupts, it was
necessary to add an `is_suspended` bitmap which is used to address the
possible corner case of unintentional IRQ unmasking because of ISR
execution after a call to synchronize_irq().

At resume, clear each is_suspended bit in the reset path of JOB/MMU
to allow unmasking the interrupts.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231204114215.54575-4-angelogioacchino.delregno@collabora.com