linux.git/drivers/gpu/drm/i915, branch v5.8

drm/i915: Recalculate FBC w/a stride when needed

2020-07-14T17:31:45+00:00

Currently we're failing to recalculate the gen9 FBC w/a stride
unless something more drastic than just the modifier itself has
changed. This often leaves us with FBC enabled with the linear
fbdev framebuffer without the w/a stride enabled. That will cause
an immediate underrun and FBC will get promptly disabled.

Fix the problem by checking if the w/a stride is about to change,
and go through the full dance if so. This part of the FBC code
is still pretty much a disaster and will need lots more work.
But this should at least fix the immediate issue.

v2: Deactivate FBC when the modifier changes since that will
    likely require resetting the w/a CFB stride

Cc: stable@vger.kernel.org
Signed-off-by: Ville Syrjälä 
Link: https://patchwork.freedesktop.org/patch/msgid/20200711080336.13423-1-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza 
(cherry picked from commit 0428ab013fdd39dbfb8f4cd8ad2b60af3776c6b9)
Signed-off-by: Jani Nikula

drm/i915: Move cec_notifier to intel_hdmi_connector_unregister, v2.

2020-07-14T17:31:39+00:00

This fixes the following KASAN splash on module reload:
[  145.136327] ==================================================================
[  145.136502] BUG: KASAN: use-after-free in intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136514] Read of size 8 at addr ffff888216641830 by task kworker/1:1/134

[  145.136535] CPU: 1 PID: 134 Comm: kworker/1:1 Tainted: G     U          T 5.5.0-rc7-valkyria+ #5783
[  145.136539] Hardware name: GIGABYTE GB-BKi3A-7100/MFLP3AP-00, BIOS F1 07/27/2016
[  145.136546] Workqueue: events drm_connector_free_work_fn
[  145.136551] Call Trace:
[  145.136560]  dump_stack+0xa1/0xe0
[  145.136571]  print_address_description.constprop.0+0x1e/0x210
[  145.136639]  ? intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136703]  ? intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136710]  __kasan_report.cold+0x1b/0x37
[  145.136790]  ? intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136863]  ? intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136870]  kasan_report+0x27/0x30
[  145.136881]  __asan_report_load8_noabort+0x1c/0x20
[  145.136946]  intel_hdmi_destroy+0x74/0x80 [i915]
[  145.136954]  drm_connector_free_work_fn+0xd1/0x100
[  145.136967]  process_one_work+0x86e/0x1610
[  145.136987]  ? pwq_dec_nr_in_flight+0x2f0/0x2f0
[  145.137004]  ? move_linked_works+0x128/0x2c0
[  145.137021]  worker_thread+0x63e/0xc90
[  145.137048]  kthread+0x2f6/0x3f0
[  145.137054]  ? calculate_sigpending+0x81/0xa0
[  145.137059]  ? process_one_work+0x1610/0x1610
[  145.137064]  ? kthread_bind+0x40/0x40
[  145.137075]  ret_from_fork+0x24/0x30

[  145.137111] Allocated by task 0:
[  145.137119] (stack is not available)

[  145.137137] Freed by task 5053:
[  145.137147]  save_stack+0x28/0x90
[  145.137152]  __kasan_slab_free+0x136/0x180
[  145.137157]  kasan_slab_free+0x26/0x30
[  145.137161]  kfree+0xe6/0x350
[  145.137242]  intel_ddi_encoder_destroy+0x60/0x80 [i915]
[  145.137252]  drm_mode_config_cleanup+0x11d/0x8f0
[  145.137329]  intel_modeset_driver_remove+0x1f5/0x350 [i915]
[  145.137403]  i915_driver_remove+0xc4/0x130 [i915]
[  145.137482]  i915_pci_remove+0x3e/0x90 [i915]
[  145.137489]  pci_device_remove+0x108/0x2d0
[  145.137494]  device_release_driver_internal+0x1e6/0x4a0
[  145.137499]  driver_detach+0xcb/0x198
[  145.137503]  bus_remove_driver+0xde/0x204
[  145.137508]  driver_unregister+0x6d/0xa0
[  145.137513]  pci_unregister_driver+0x2e/0x230
[  145.137576]  i915_exit+0x1f/0x26 [i915]
[  145.137157]  kasan_slab_free+0x26/0x30
[  145.137161]  kfree+0xe6/0x350
[  145.137242]  intel_ddi_encoder_destroy+0x60/0x80 [i915]
[  145.137252]  drm_mode_config_cleanup+0x11d/0x8f0
[  145.137329]  intel_modeset_driver_remove+0x1f5/0x350 [i915]
[  145.137403]  i915_driver_remove+0xc4/0x130 [i915]
[  145.137482]  i915_pci_remove+0x3e/0x90 [i915]
[  145.137489]  pci_device_remove+0x108/0x2d0
[  145.137494]  device_release_driver_internal+0x1e6/0x4a0
[  145.137499]  driver_detach+0xcb/0x198
[  145.137503]  bus_remove_driver+0xde/0x204
[  145.137508]  driver_unregister+0x6d/0xa0
[  145.137513]  pci_unregister_driver+0x2e/0x230
[  145.137576]  i915_exit+0x1f/0x26 [i915]
[  145.137581]  __x64_sys_delete_module+0x35b/0x470
[  145.137586]  do_syscall_64+0x99/0x4e0
[  145.137591]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

[  145.137606] The buggy address belongs to the object at ffff888216640000
                which belongs to the cache kmalloc-8k of size 8192
[  145.137618] The buggy address is located 6192 bytes inside of
                8192-byte region [ffff888216640000, ffff888216642000)
[  145.137630] The buggy address belongs to the page:
[  145.137640] page:ffffea0008599000 refcount:1 mapcount:0 mapping:ffff888107c02a80 index:0xffff888216644000 compound_mapcount: 0
[  145.137647] raw: 0200000000010200 0000000000000000 0000000100000001 ffff888107c02a80
[  145.137652] raw: ffff888216644000 0000000080020001 00000001ffffffff 0000000000000000
[  145.137656] page dumped because: kasan: bad access detected

[  145.137668] Memory state around the buggy address:
[  145.137678]  ffff888216641700: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  145.137687]  ffff888216641780: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  145.137697] >ffff888216641800: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  145.137706]                                      ^
[  145.137715]  ffff888216641880: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  145.137724]  ffff888216641900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  145.137733] ==================================================================
[  145.137742] Disabling lock debugging due to kernel taint

Changes since v1:
- Add fixes tags.
- Use early unregister.

Signed-off-by: Maarten Lankhorst 
Fixes: 9c229127aee2 ("drm/i915: hdmi: add CEC notifier to intel_hdmi")
Cc:  # v4.19+
Reviewed-by: Daniel Vetter 
Link: https://patchwork.freedesktop.org/patch/msgid/20200212135445.1469133-1-maarten.lankhorst@linux.intel.com
(cherry picked from commit a581483b1e5466d28fc50ff623fba31cea2cccb6)
Signed-off-by: Jani Nikula

drm/i915/gt: Only swap to a random sibling once upon creation

2020-07-14T17:31:23+00:00

The danger in switching at random upon intel_context_pin is that the
context may still actually be inflight, as it will not be scheduled out
until a context switch after it is complete -- that may be a long time
after we do a final intel_context_unpin.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2118
Fixes: 6d06779e8672 ("drm/i915: Load balancing across a virtual engine")
Signed-off-by: Chris Wilson 
Cc: Tvrtko Ursulin 
Cc:  # v5.3+
Reviewed-by: Tvrtko Ursulin 
Link: https://patchwork.freedesktop.org/patch/msgid/20200713160549.17344-1-chris@chris-wilson.co.uk
(cherry picked from commit 90a987205c6cf74116a102ed446d22d92cdaf915)
Signed-off-by: Jani Nikula

drm/i915/gt: Ignore irq enabling on the virtual engines

2020-07-14T17:31:14+00:00

We do not use the virtual engines for interrupts (they have physical
components), but we do use them to decouple the fence signaling during
submission. Currently, when we submit a completed request, we try to
enable the interrupt handler for the virtual engine, but we never disarm
it. A quick fix is then to mark the irq as enabled, and it will then
remain enabled -- and this prevents us from waking the device and never
letting it sleep again.

Fixes: f8db4d051b5e ("drm/i915: Initialise breadcrumb lists on the virtual engine")
Signed-off-by: Chris Wilson 
Cc: Mika Kuoppala 
Cc: Tvrtko Ursulin 
Cc:  # v5.5+
Reviewed-by: Tvrtko Ursulin 
Link: https://patchwork.freedesktop.org/patch/msgid/20200711203236.12330-1-chris@chris-wilson.co.uk
(cherry picked from commit 4fe6abb8f51355224808ab02a9febf65d184c40b)
Signed-off-by: Jani Nikula

drm/i915/perf: Use GTT when saving/restoring engine GPR

2020-07-14T17:31:05+00:00

MI_STORE_REGISTER_MEM and MI_LOAD_REGISTER_MEM need to know which
translation to use when saving restoring the engine general purpose
registers to and from the GT scratch. Since GT scratch is mapped to
ggtt, we need to set an additional bit in the command to use GTT.

Fixes: daed3e44396d17 ("drm/i915/perf: implement active wait for noa configurations")
Suggested-by: Prathap Kumar Valsan 
Signed-off-by: Umesh Nerlige Ramappa 
Reviewed-by: Lionel Landwerlin 
Reviewed-by: Chris Wilson 
Link: https://patchwork.freedesktop.org/patch/msgid/20200709224504.11345-1-chris@chris-wilson.co.uk
Signed-off-by: Chris Wilson 
(cherry picked from commit e43ff99c8deda85234e6233e0f4af6cb09566a37)
Signed-off-by: Jani Nikula

drm/i915/selftests: Fix compare functions provided for sorting

2020-07-14T17:30:58+00:00

Both cmp_u32 and cmp_u64 are comparing the pointers instead of the value
at those pointers. This will result in incorrect/unsorted list. Fix it
by deferencing the pointers before comparison.

Fixes: 4ba74e53ada3 ("drm/i915/selftests: Verify frequency scaling with RPS")
Fixes: 8757797ff9c9 ("drm/i915/selftests: Repeat the rps clock frequency measurement")
Cc: Chris Wilson 
Cc: Mika Kuoppala 
Signed-off-by: Sudeep Holla 
Reviewed-by: Chris Wilson 
Signed-off-by: Chris Wilson 
Link: https://patchwork.freedesktop.org/patch/msgid/20200709154931.23310-1-sudeep.holla@arm.com
(cherry picked from commit 2196dfea896f7027b43bae848890ce4aec5c8724)
Signed-off-by: Jani Nikula

drm/i915/gt: Pin the rings before marking active

2020-07-08T01:00:15+00:00

On eviction, we acquire the vm->mutex and then wait on the vma->active.
Therefore when binding and pinning the vma, we must follow the same
sequence, lock/pin the vma then mark it active. Otherwise, we mark the
vma as active, then wait for the vm->mutex, and meanwhile the evictor
holding the mutex waits upon us to complete our activity.

Fixes: 8ccfc20a7d56 ("drm/i915/gt: Mark ring->vma as active while pinned")
Signed-off-by: Chris Wilson 
Cc: Tvrtko Ursulin 
Cc:  # v5.6+
Reviewed-by: Matthew Auld 
Link: https://patchwork.freedesktop.org/patch/msgid/20200706170138.8993-1-chris@chris-wilson.co.uk
(cherry picked from commit 8567774e87e23a57155e5102f81208729b992ae6)
Signed-off-by: Rodrigo Vivi

drm/i915: Also drop vm.ref along error paths for vma construction

2020-07-07T00:16:03+00:00

Not only do we need to release the vm.ref we acquired for the vma on the
duplicate insert branch, but also for the normal error paths, so roll
them all into one.

Reported-by: Andi Shyti 
Suggested-by: Andi Shyti 
Fixes: 2850748ef876 ("drm/i915: Pull i915_vma_pin under the vm->mutex")
Signed-off-by: Chris Wilson 
Cc: Andi Shyti 
Cc:  # v5.5+
Reviewed-by: Andi Shyti 
Link: https://patchwork.freedesktop.org/patch/msgid/20200702211015.29604-1-chris@chris-wilson.co.uk
(cherry picked from commit 03fca66b7a36b52da8915341eee388267f6d5b73)
Signed-off-by: Rodrigo Vivi

drm/i915: Drop vm.ref for duplicate vma on construction

2020-07-07T00:16:00+00:00

As we allow for parallel threads to create the same vma instance
concurrently, and we only filter out the duplicates upon reacquiring the
spinlock for the rbtree, we have to free the loser of the constructors'
race. When freeing, we should also drop any resource references acquired
for the redundant vma.

Fixes: 2850748ef876 ("drm/i915: Pull i915_vma_pin under the vm->mutex")
Signed-off-by: Chris Wilson 
Cc: Tvrtko Ursulin 
Cc:  # v5.5+
Reviewed-by: Tvrtko Ursulin 
Link: https://patchwork.freedesktop.org/patch/msgid/20200702083225.20044-1-chris@chris-wilson.co.uk
(cherry picked from commit 2377427cdd2b7514eb4c40241cf5c4dec63c1bec)
Signed-off-by: Rodrigo Vivi

drm/i915/fbc: Fix fence_y_offset handling

2020-07-07T00:15:57+00:00

The current fence_y_offset calculation is broken. I think it more or
less used to do the right thing, but then I changed the plane code
to put the final x/y source offsets back into the src rectangle so
now it's just subtraacting the same value from itself. The code would
never have worked if we allowed the framebuffer to have a non-zero
offset.

Let's do this in a better way by just calculating the fence_y_offset
from the final plane surface offset. Note that we don't align the
plane surface address to fence rows so with horizontal panning there's
often a horizontal offset from the fence start to the surface address
as well. We have no way to tell the hardware about that so we just
ignore it. Based on some quick tests the invlidation still happens
correctly. I presume due to the invalidation nuking at least the full
line (or a segment of multiple lines).

Fixes: 54d4d719fa11 ("drm/i915: Overcome display engine stride limits via GTT remapping")
Signed-off-by: Ville Syrjälä 
Link: https://patchwork.freedesktop.org/patch/msgid/20200429101034.8208-4-ville.syrjala@linux.intel.com
Reviewed-by: Matt Roper 
(cherry picked from commit 5331889b5ffb11d6257953e418291a9f04c02bed)
Signed-off-by: Rodrigo Vivi