linux-stable.git/drivers/gpu, branch v4.19.66

drm/i915/gvt: fix incorrect cache entry for guest page mapping

2019-08-06T17:06:56+00:00

commit 7366aeb77cd840f3edea02c65065d40affaa7f45 upstream.

GPU hang observed during the guest OCL conformance test which is caused
by THP GTT feature used durning the test.

It was observed the same GFN with different size (4K and 2M) requested
from the guest in GVT. So during the guest page dma map stage, it is
required to unmap first with orginal size and then remap again with
requested size.

Fixes: b901b252b6cf ("drm/i915/gvt: Add 2M huge gtt support")
Cc: stable@vger.kernel.org
Reviewed-by: Zhenyu Wang 
Signed-off-by: Xiaolin Zhang 
Signed-off-by: Zhenyu Wang 
Signed-off-by: Greg Kroah-Hartman

drm/nouveau: fix memory leak in nouveau_conn_reset()

2019-08-06T17:06:53+00:00

[ Upstream commit 09b90e2fe35faeace2488234e2a7728f2ea8ba26 ]

In nouveau_conn_reset(), if connector->state is true,
__drm_atomic_helper_connector_destroy_state() will be called,
but the memory pointed by asyc isn't freed. Memory leak happens
in the following function __drm_atomic_helper_connector_reset(),
where newly allocated asyc->state will be assigned to connector->state.

So using nouveau_conn_atomic_destroy_state() instead of
__drm_atomic_helper_connector_destroy_state to free the "old" asyc.

Here the is the log showing memory leak.

unreferenced object 0xffff8c5480483c80 (size 192):
  comm "kworker/0:2", pid 188, jiffies 4294695279 (age 53.179s)
  hex dump (first 32 bytes):
    00 f0 ba 7b 54 8c ff ff 00 00 00 00 00 00 00 00  ...{T...........
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<000000005005c0d0>] kmem_cache_alloc_trace+0x195/0x2c0
    [<00000000a122baed>] nouveau_conn_reset+0x25/0xc0 [nouveau]
    [<000000004fd189a2>] nouveau_connector_create+0x3a7/0x610 [nouveau]
    [<00000000c73343a8>] nv50_display_create+0x343/0x980 [nouveau]
    [<000000002e2b03c3>] nouveau_display_create+0x51f/0x660 [nouveau]
    [<00000000c924699b>] nouveau_drm_device_init+0x182/0x7f0 [nouveau]
    [<00000000cc029436>] nouveau_drm_probe+0x20c/0x2c0 [nouveau]
    [<000000007e961c3e>] local_pci_probe+0x47/0xa0
    [<00000000da14d569>] work_for_cpu_fn+0x1a/0x30
    [<0000000028da4805>] process_one_work+0x27c/0x660
    [<000000001d415b04>] worker_thread+0x22b/0x3f0
    [<0000000003b69f1f>] kthread+0x12f/0x150
    [<00000000c94c29b7>] ret_from_fork+0x3a/0x50

Signed-off-by: Yongxin Liu 
Signed-off-by: Ben Skeggs 
Signed-off-by: Sasha Levin

drm/msm: Depopulate platform on probe failure

2019-07-31T05:27:01+00:00

[ Upstream commit 4368a1539c6b41ac3cddc06f5a5117952998804c ]

add_display_components() calls of_platform_populate, and we depopluate
on pdev remove, but not when probe fails. So if we get a probe deferral
in one of the components, we won't depopulate the platform. This causes
the core to keep references to devices which should be destroyed, which
causes issues when those same devices try to re-initialize on the next
probe attempt.

I think this is the reason we had issues with the gmu's device-managed
resources on deferral (worked around in commit 94e3a17f33a5).

Reviewed-by: Rob Clark 
Signed-off-by: Sean Paul 
Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-3-sean@poorly.run
Signed-off-by: Sasha Levin

drm/rockchip: Properly adjust to a true clock in adjusted_mode

2019-07-31T05:27:00+00:00

[ Upstream commit 99b9683f2142b20bad78e61f7f829e8714e45685 ]

When fixing up the clock in vop_crtc_mode_fixup() we're not doing it
quite correctly.  Specifically if we've got the true clock 266666667 Hz,
we'll perform this calculation:
   266666667 / 1000 => 266666

Later when we try to set the clock we'll do clk_set_rate(266666 *
1000).  The common clock framework won't actually pick the proper clock
in this case since it always wants clocks <= the specified one.

Let's solve this by using DIV_ROUND_UP.

Fixes: b59b8de31497 ("drm/rockchip: return a true clock rate to adjusted_mode")
Signed-off-by: Douglas Anderson 
Signed-off-by: Sean Paul 
Reviewed-by: Yakir Yang 
Signed-off-by: Heiko Stuebner 
Link: https://patchwork.freedesktop.org/patch/msgid/20190614224730.98622-1-dianders@chromium.org
Signed-off-by: Sasha Levin

drm/amd/display: fix compilation error

2019-07-31T05:27:00+00:00

[ Upstream commit 88099f53cc3717437f5fc9cf84205c5b65118377 ]

this patch fixes below compilation error

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn10/dcn10_hw_sequencer.c: In
function ‘dcn10_apply_ctx_for_surface’:
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn10/dcn10_hw_sequencer.c:2378:3:
error: implicit declaration of function ‘udelay’
[-Werror=implicit-function-declaration]
   udelay(underflow_check_delay_us);

Signed-off-by: Hariprasad Kelam 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/virtio: Add memory barriers for capset cache.

2019-07-31T05:27:00+00:00

[ Upstream commit 9ff3a5c88e1f1ab17a31402b96d45abe14aab9d7 ]

After data is copied to the cache entry, atomic_set is used indicate
that the data is the entry is valid without appropriate memory barriers.
Similarly the read side was missing the corresponding memory barriers.

Signed-off-by: David Riley 
Link: http://patchwork.freedesktop.org/patch/msgid/20190610211810.253227-5-davidriley@chromium.org
Signed-off-by: Gerd Hoffmann 
Signed-off-by: Sasha Levin

drm/amd/display: Always allocate initial connector state state

2019-07-31T05:27:00+00:00

[ Upstream commit f04bee34d6e35df26cbb2d65e801adfd0d8fe20d ]

[Why]
Unlike our regular connectors, MST connectors don't start off with
an initial connector state. This causes a NULL pointer dereference to
occur when attaching the bpc property since it tries to modify the
connector state.

We need an initial connector state on the connector to avoid the crash.

[How]
Use our reset helper to allocate an initial state and reset the values
to their defaults. We were already doing this before, just not for
MST connectors.

Signed-off-by: Nicholas Kazlauskas 
Reviewed-by: Leo Li 
Signed-off-by: Alex Deucher 
Signed-off-by: Sasha Levin

drm/crc-debugfs: Also sprinkle irqrestore over early exits

2019-07-31T05:26:59+00:00

[ Upstream commit d99004d7201aa653658ff2390d6e516567c96ebc ]

I. was. blind.

Caught with vkms, which has some really slow crc computation function.

Fixes: 1882018a70e0 ("drm/crc-debugfs: User irqsafe spinlock in drm_crtc_add_crc_entry")
Cc: Rodrigo Siqueira 
Cc: Tomeu Vizoso 
Cc: Emil Velikov 
Cc: Benjamin Gaignard 
Cc: Ville Syrjälä 
Reviewed-by: Emil Velikov 
Reviewed-by: Benjamin Gaignard 
Signed-off-by: Daniel Vetter 
Link: https://patchwork.freedesktop.org/patch/msgid/20190606211544.5389-1-daniel.vetter@ffwll.ch
Signed-off-by: Sasha Levin

drm/crc-debugfs: User irqsafe spinlock in drm_crtc_add_crc_entry

2019-07-31T05:26:59+00:00

[ Upstream commit 1882018a70e06376234133e69ede9dd743b4dbd9 ]

We can be called from any context, we need to be prepared.

Noticed this while hacking on vkms, which calls this function from a
normal worker. Which really upsets lockdep.

Cc: Rodrigo Siqueira 
Cc: Tomeu Vizoso 
Cc: Emil Velikov 
Cc: Benjamin Gaignard 
Reviewed-by: Benjamin Gaignard 
Reviewed-by: Ville Syrjälä 
Signed-off-by: Daniel Vetter 
Link: https://patchwork.freedesktop.org/patch/msgid/20190605194556.16744-1-daniel.vetter@ffwll.ch
Signed-off-by: Sasha Levin

gpu: host1x: Increase maximum DMA segment size

2019-07-31T05:26:59+00:00

[ Upstream commit 1e390478cfb527e34c9ab89ba57212cb05c33c51 ]

Recent versions of the DMA API debug code have started to warn about
violations of the maximum DMA segment size. This is because the segment
size defaults to 64 KiB, which can easily be exceeded in large buffer
allocations such as used in DRM/KMS for framebuffers.

Technically the Tegra SMMU and ARM SMMU don't have a maximum segment
size (they map individual pages irrespective of whether they are
contiguous or not), so the choice of 4 MiB is a bit arbitrary here. The
maximum segment size is a 32-bit unsigned integer, though, so we can't
set it to the correct maximum size, which would be the size of the
aperture.

Signed-off-by: Thierry Reding 
Signed-off-by: Sasha Levin