linux-stable.git/drivers/gpu, branch v5.10.44 Linux kernel stable tree drm/msm/a6xx: avoid shadow NULL reference in failure path 2021-06-16T10:01:44+00:00 Jonathan Marek jonathan@marek.ca 2021-05-13T17:14:00+00:00 fd681a8c7ac8f649a0718f6cbf2fe75d0587c9a2 commit ce86c239e4d218ae6040bec18e6d19a58edb8b7c upstream. If a6xx_hw_init() fails before creating the shadow_bo, the a6xx_pm_suspend code referencing it will crash. Change the condition to one that avoids this problem (note: creation of shadow_bo is behind this same condition) Fixes: e8b0b994c3a5 ("drm/msm/a6xx: Clear shadow on suspend") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org> Link: https://lore.kernel.org/r/20210513171431.18632-6-jonathan@marek.ca Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit ce86c239e4d218ae6040bec18e6d19a58edb8b7c upstream.

If a6xx_hw_init() fails before creating the shadow_bo, the a6xx_pm_suspend
code referencing it will crash. Change the condition to one that avoids
this problem (note: creation of shadow_bo is behind this same condition)

Fixes: e8b0b994c3a5 ("drm/msm/a6xx: Clear shadow on suspend")
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org>
Link: https://lore.kernel.org/r/20210513171431.18632-6-jonathan@marek.ca
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/msm/a6xx: update/fix CP_PROTECT initialization 2021-06-16T10:01:44+00:00 Jonathan Marek jonathan@marek.ca 2021-05-13T17:13:59+00:00 0bc79f4b7ac60f9fd4dfd1c65ad17e5a2369e2ca commit 408434036958699a7f50ddec984f7ba33e11a8f5 upstream. Update CP_PROTECT register programming based on downstream. A6XX_PROTECT_RW is renamed to A6XX_PROTECT_NORDWR to make things aligned and also be more clear about what it does. Note that this required switching to use the CP_ALWAYS_ON_COUNTER as the GMU counter is not accessible from the cmdstream. Which also means using the CPU counter for the msm_gpu_submit_flush() tracepoint (as catapult depends on being able to compare this to the start/end values captured in cmdstream). This may need to be revisited when IFPC is enabled. Also, compared to downstream, this opens up CP_PERFCTR_CP_SEL as the userspace performance tooling (fdperf and pps-producer) expect to be able to configure the CP counters. Fixes: 4b565ca5a2cb ("drm/msm: Add A6XX device support") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org> Link: https://lore.kernel.org/r/20210513171431.18632-5-jonathan@marek.ca [switch to CP_ALWAYS_ON_COUNTER, open up CP_PERFCNTR_CP_SEL, and spiff up commit msg] Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 408434036958699a7f50ddec984f7ba33e11a8f5 upstream.

Update CP_PROTECT register programming based on downstream.

A6XX_PROTECT_RW is renamed to A6XX_PROTECT_NORDWR to make things aligned
and also be more clear about what it does.

Note that this required switching to use the CP_ALWAYS_ON_COUNTER as the
GMU counter is not accessible from the cmdstream.  Which also means
using the CPU counter for the msm_gpu_submit_flush() tracepoint (as
catapult depends on being able to compare this to the start/end values
captured in cmdstream).  This may need to be revisited when IFPC is
enabled.

Also, compared to downstream, this opens up CP_PERFCTR_CP_SEL as the
userspace performance tooling (fdperf and pps-producer) expect to be
able to configure the CP counters.

Fixes: 4b565ca5a2cb ("drm/msm: Add A6XX device support")
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org>
Link: https://lore.kernel.org/r/20210513171431.18632-5-jonathan@marek.ca
[switch to CP_ALWAYS_ON_COUNTER, open up CP_PERFCNTR_CP_SEL, and spiff
 up commit msg]
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/msm/a6xx: fix incorrectly set uavflagprd_inv field for A650 2021-06-16T10:01:44+00:00 Jonathan Marek jonathan@marek.ca 2021-05-13T17:13:58+00:00 5b7dc8329d7a51bab34452d132230fd76330a22b commit b4387eaf3821a4c4241ac3a556e13244eb1fdaa5 upstream. Value was shifted in the wrong direction, resulting in the field always being zero, which is incorrect for A650. Fixes: d0bac4e9cd66 ("drm/msm/a6xx: set ubwc config for A640 and A650") Signed-off-by: Jonathan Marek <jonathan@marek.ca> Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org> Link: https://lore.kernel.org/r/20210513171431.18632-4-jonathan@marek.ca Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit b4387eaf3821a4c4241ac3a556e13244eb1fdaa5 upstream.

Value was shifted in the wrong direction, resulting in the field always
being zero, which is incorrect for A650.

Fixes: d0bac4e9cd66 ("drm/msm/a6xx: set ubwc config for A640 and A650")
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Akhil P Oommen <akhilpo@codeaurora.org>
Link: https://lore.kernel.org/r/20210513171431.18632-4-jonathan@marek.ca
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/mcde: Fix off by 10^3 in calculation 2021-06-16T10:01:44+00:00 Linus Walleij linus.walleij@linaro.org 2021-06-08T21:33:18+00:00 5a61f69da3b8d735b01dddee72fee4671510d907 commit c8a570443943304cac2e4186dbce6989b6c2b8b5 upstream. The calclulation of how many bytes we stuff into the DSI pipeline for video mode panels is off by three orders of magnitude because we did not account for the fact that the DRM mode clock is in kilohertz rather than hertz. This used to be: drm_mode_vrefresh(mode) * mode->htotal * mode->vtotal which would become for example for s6e63m0: 60 x 514 x 831 = 25628040 Hz, but mode->clock is 25628 as it is in kHz. This affects only the Samsung GT-I8190 "Golden" phone right now since it is the only MCDE device with a video mode display. Curiously some specimen work with this code and wild settings in the EOL and empty packets at the end of the display, but I have noticed an eeire flicker until now. Others were not so lucky and got black screens. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reported-by: Stephan Gerhold <stephan@gerhold.net> Fixes: 920dd1b1425b ("drm/mcde: Use mode->clock instead of reverse calculating it from the vrefresh") Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Stephan Gerhold <stephan@gerhold.net> Reviewed-by: Stephan Gerhold <stephan@gerhold.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210608213318.3897858-1-linus.walleij@linaro.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit c8a570443943304cac2e4186dbce6989b6c2b8b5 upstream.

The calclulation of how many bytes we stuff into the
DSI pipeline for video mode panels is off by three
orders of magnitude because we did not account for the
fact that the DRM mode clock is in kilohertz rather
than hertz.

This used to be:
drm_mode_vrefresh(mode) * mode->htotal * mode->vtotal
which would become for example for s6e63m0:
60 x 514 x 831 = 25628040 Hz, but mode->clock is
25628 as it is in kHz.

This affects only the Samsung GT-I8190 "Golden" phone
right now since it is the only MCDE device with a video
mode display.

Curiously some specimen work with this code and wild
settings in the EOL and empty packets at the end of the
display, but I have noticed an eeire flicker until now.
Others were not so lucky and got black screens.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reported-by: Stephan Gerhold <stephan@gerhold.net>
Fixes: 920dd1b1425b ("drm/mcde: Use mode->clock instead of reverse calculating it from the vrefresh")
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Tested-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20210608213318.3897858-1-linus.walleij@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm: Lock pointer access in drm_master_release() 2021-06-16T10:01:39+00:00 Desmond Cheong Zhi Xi desmondcheongzx@gmail.com 2021-06-09T09:21:19+00:00 aa8591a58cbd2986090709e4202881f18e8ae30e commit c336a5ee984708db4826ef9e47d184e638e29717 upstream. This patch eliminates the following smatch warning: drivers/gpu/drm/drm_auth.c:320 drm_master_release() warn: unlocked access 'master' (line 318) expected lock '&dev->master_mutex' The 'file_priv->master' field should be protected by the mutex lock to '&dev->master_mutex'. This is because other processes can concurrently modify this field and free the current 'file_priv->master' pointer. This could result in a use-after-free error when 'master' is dereferenced in subsequent function calls to 'drm_legacy_lock_master_cleanup()' or to 'drm_lease_revoke()'. An example of a scenario that would produce this error can be seen from a similar bug in 'drm_getunique()' that was reported by Syzbot: https://syzkaller.appspot.com/bug?id=148d2f1dfac64af52ffd27b661981a540724f803 In the Syzbot report, another process concurrently acquired the device's master mutex in 'drm_setmaster_ioctl()', then overwrote 'fpriv->master' in 'drm_new_set_master()'. The old value of 'fpriv->master' was subsequently freed before the mutex was unlocked. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com> Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20210609092119.173590-1-desmondcheongzx@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit c336a5ee984708db4826ef9e47d184e638e29717 upstream.

This patch eliminates the following smatch warning:
drivers/gpu/drm/drm_auth.c:320 drm_master_release() warn: unlocked access 'master' (line 318) expected lock '&dev->master_mutex'

The 'file_priv->master' field should be protected by the mutex lock to
'&dev->master_mutex'. This is because other processes can concurrently
modify this field and free the current 'file_priv->master'
pointer. This could result in a use-after-free error when 'master' is
dereferenced in subsequent function calls to
'drm_legacy_lock_master_cleanup()' or to 'drm_lease_revoke()'.

An example of a scenario that would produce this error can be seen
from a similar bug in 'drm_getunique()' that was reported by Syzbot:
https://syzkaller.appspot.com/bug?id=148d2f1dfac64af52ffd27b661981a540724f803

In the Syzbot report, another process concurrently acquired the
device's master mutex in 'drm_setmaster_ioctl()', then overwrote
'fpriv->master' in 'drm_new_set_master()'. The old value of
'fpriv->master' was subsequently freed before the mutex was unlocked.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210609092119.173590-1-desmondcheongzx@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm: Fix use-after-free read in drm_getunique() 2021-06-16T10:01:39+00:00 Desmond Cheong Zhi Xi desmondcheongzx@gmail.com 2021-06-08T11:04:36+00:00 491d52e0078860b33b6c14f0a7ac74ca1b603bd6 commit b436acd1cf7fac0ba987abd22955d98025c80c2b upstream. There is a time-of-check-to-time-of-use error in drm_getunique() due to retrieving file_priv->master prior to locking the device's master mutex. An example can be seen in the crash report of the use-after-free error found by Syzbot: https://syzkaller.appspot.com/bug?id=148d2f1dfac64af52ffd27b661981a540724f803 In the report, the master pointer was used after being freed. This is because another process had acquired the device's master mutex in drm_setmaster_ioctl(), then overwrote fpriv->master in drm_new_set_master(). The old value of fpriv->master was subsequently freed before the mutex was unlocked. To fix this, we lock the device's master mutex before retrieving the pointer from from fpriv->master. This patch passes the Syzbot reproducer test. Reported-by: syzbot+c3a706cec1ea99e1c693@syzkaller.appspotmail.com Signed-off-by: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com> Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20210608110436.239583-1-desmondcheongzx@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit b436acd1cf7fac0ba987abd22955d98025c80c2b upstream.

There is a time-of-check-to-time-of-use error in drm_getunique() due
to retrieving file_priv->master prior to locking the device's master
mutex.

An example can be seen in the crash report of the use-after-free error
found by Syzbot:
https://syzkaller.appspot.com/bug?id=148d2f1dfac64af52ffd27b661981a540724f803

In the report, the master pointer was used after being freed. This is
because another process had acquired the device's master mutex in
drm_setmaster_ioctl(), then overwrote fpriv->master in
drm_new_set_master(). The old value of fpriv->master was subsequently
freed before the mutex was unlocked.

To fix this, we lock the device's master mutex before retrieving the
pointer from from fpriv->master. This patch passes the Syzbot
reproducer test.

Reported-by: syzbot+c3a706cec1ea99e1c693@syzkaller.appspotmail.com
Signed-off-by: Desmond Cheong Zhi Xi <desmondcheongzx@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210608110436.239583-1-desmondcheongzx@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/msm/dpu: always use mdp device to scale bandwidth 2021-06-10T11:39:28+00:00 Dmitry Baryshkov dmitry.baryshkov@linaro.org 2021-04-01T02:05:33+00:00 05e41f6f1c4e8c42edb9715b6629d9ab2af61064 commit a670ff578f1fb855fedc7931fa5bbc06b567af22 upstream. Currently DPU driver scales bandwidth and core clock for sc7180 only, while the rest of chips get static bandwidth votes. Make all chipsets scale bandwidth and clock per composition requirements like sc7180 does. Drop old voting path completely. Tested on RB3 (SDM845) and RB5 (SM8250). Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Link: https://lore.kernel.org/r/20210401020533.3956787-2-dmitry.baryshkov@linaro.org Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Amit Pundir <amit.pundir@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit a670ff578f1fb855fedc7931fa5bbc06b567af22 upstream.

Currently DPU driver scales bandwidth and core clock for sc7180 only,
while the rest of chips get static bandwidth votes. Make all chipsets
scale bandwidth and clock per composition requirements like sc7180 does.
Drop old voting path completely.

Tested on RB3 (SDM845) and RB5 (SM8250).

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20210401020533.3956787-2-dmitry.baryshkov@linaro.org
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/amdgpu: make sure we unpin the UVD BO 2021-06-10T11:39:27+00:00 Nirmoy Das nirmoy.das@amd.com 2021-05-28T14:54:16+00:00 3a6b69221f96f87c680bbc9fba01db3415b18f27 commit 07438603a07e52f1c6aa731842bd298d2725b7be upstream. Releasing pinned BOs is illegal now. UVD 6 was missing from: commit 2f40801dc553 ("drm/amdgpu: make sure we unpin the UVD BO") Fixes: 2f40801dc553 ("drm/amdgpu: make sure we unpin the UVD BO") Cc: stable@vger.kernel.org Signed-off-by: Nirmoy Das <nirmoy.das@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 07438603a07e52f1c6aa731842bd298d2725b7be upstream.

Releasing pinned BOs is illegal now. UVD 6 was missing from:
commit 2f40801dc553 ("drm/amdgpu: make sure we unpin the UVD BO")

Fixes: 2f40801dc553 ("drm/amdgpu: make sure we unpin the UVD BO")
Cc: stable@vger.kernel.org
Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/amdgpu: Don't query CE and UE errors 2021-06-10T11:39:27+00:00 Luben Tuikov luben.tuikov@amd.com 2021-05-12T16:33:23+00:00 58da0b509e4b8f4a3a4b1b2e23871d108f81338a commit dce3d8e1d070900e0feeb06787a319ff9379212c upstream. On QUERY2 IOCTL don't query counts of correctable and uncorrectable errors, since when RAS is enabled and supported on Vega20 server boards, this takes insurmountably long time, in O(n^3), which slows the system down to the point of it being unusable when we have GUI up. Fixes: ae363a212b14 ("drm/amdgpu: Add a new flag to AMDGPU_CTX_OP_QUERY_STATE2") Cc: Alexander Deucher <Alexander.Deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit dce3d8e1d070900e0feeb06787a319ff9379212c upstream.

On QUERY2 IOCTL don't query counts of correctable
and uncorrectable errors, since when RAS is
enabled and supported on Vega20 server boards,
this takes insurmountably long time, in O(n^3),
which slows the system down to the point of it
being unusable when we have GUI up.

Fixes: ae363a212b14 ("drm/amdgpu: Add a new flag to AMDGPU_CTX_OP_QUERY_STATE2")
Cc: Alexander Deucher <Alexander.Deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drm/amdgpu/jpeg3: add cancel_delayed_work_sync before power gate 2021-06-10T11:39:23+00:00 James Zhu James.Zhu@amd.com 2021-05-19T16:08:20+00:00 7fa8ee00b5fab915856d03b50113be8fef1337dc [ Upstream commit 20ebbfd22f8115a1e4f60d3d289f66be4d47f1ec ] Add cancel_delayed_work_sync before set power gating state to avoid race condition issue when power gating. Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 20ebbfd22f8115a1e4f60d3d289f66be4d47f1ec ]

Add cancel_delayed_work_sync before set power gating state
to avoid race condition issue when power gating.

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>