[Why]
Unplug mst hub will cause warning. That's because
dm_helpers_construct_old_payload() is changed to be called after
payload removement from dc link.
In dm_helpers_construct_old_payload(), We refer to the vcpi in
payload allocation table of dc link to construct the old payload
and payload is no longer in the table when we call the function
now.
[How]
Refer to the mst_state to construct the number of time slot for old
payload now. Note that dm_helpers_construct_old_payload() is just
a quick workaround before and we are going to abandon it soon.
Fixes: 5aa1dfcdf0 ("drm/mst: Refactor the flow for payload allocation/removement")
Reviewed-by: Jerry Zuo <jerry.zuo@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231005080405.169841-1-Wayne.Lin@amd.com
Some BO's might be mapped onto physical memory chunkwise and on demand,
like Panfrost's tiler heap. In this case, even though the
drm_gem_shmem_object page array might already be allocated, only a very
small fraction of the BO is currently backed by system memory, but
drm_show_memory_stats will then proceed to add its entire virtual size to
the file's total resident size regardless.
This led to very unrealistic RSS sizes being reckoned for Panfrost, where
said tiler heap buffer is initially allocated with a virtual size of 128
MiB, but only a small part of it will eventually be backed by system memory
after successive GPU page faults.
Provide a new DRM object generic function that would allow drivers to
return a more accurate RSS and purgeable sizes for their BOs.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230929181616.2769345-5-adrian.larumbe@collabora.com
The drm-stats fdinfo tags made available to user space are drm-engine,
drm-cycles, drm-max-freq and drm-curfreq, one per job slot.
This deviates from standard practice in other DRM drivers, where a single
set of key:value pairs is provided for the whole render engine. However,
Panfrost has separate queues for fragment and vertex/tiler jobs, so a
decision was made to calculate bus cycles and workload times separately.
Maximum operating frequency is calculated at devfreq initialisation time.
Current frequency is made available to user space because nvtop uses it
when performing engine usage calculations.
It is important to bear in mind that both GPU cycle and kernel time numbers
provided are at best rough estimations, and always reported in excess from
the actual figure because of two reasons:
- Excess time because of the delay between the end of a job processing,
the subsequent job IRQ and the actual time of the sample.
- Time spent in the engine queue waiting for the GPU to pick up the next
job.
To avoid race conditions during enablement/disabling, a reference counting
mechanism was introduced, and a job flag that tells us whether a given job
increased the refcount. This is necessary, because user space can toggle
cycle counting through a debugfs file, and a given job might have been in
flight by the time cycle counting was disabled.
The main goal of the debugfs cycle counter knob is letting tools like nvtop
or IGT's gputop switch it at any time, to avoid power waste in case no
engine usage measuring is necessary.
Also add a documentation file explaining the possible values for fdinfo's
engine keystrings and Panfrost-specific drm-curfreq-<keystr> pairs.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230929181616.2769345-3-adrian.larumbe@collabora.com
After a recent change, two variables are only used in an #ifdef:
drivers/gpu/drm/nouveau/dispnv50/disp.c: In function 'nv50_sor_atomic_disable':
drivers/gpu/drm/nouveau/dispnv50/disp.c:1569:13: error: unused variable 'ret' [-Werror=unused-variable]
1569 | int ret;
| ^~~
drivers/gpu/drm/nouveau/dispnv50/disp.c:1568:28: error: unused variable 'aux' [-Werror=unused-variable]
1568 | struct drm_dp_aux *aux = &nv_connector->aux;
| ^~~
Move them into the same conditional block, along with the nv_connector variable
that becomes unused during that fix.
Fixes: 757033808c ("drm/nouveau/kms/nv50-: fixup sink D3 before tearing down link")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Danilo Krummrich <dakr@redhat.com>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230925155930.677620-1-arnd@kernel.org
Rename struct drm_gpuva_manager to struct drm_gpuvm including
corresponding functions. This way the GPUVA manager's structures align
very well with the documentation of VM_BIND [1] and VM_BIND locking [2].
It also provides a better foundation for the naming of data structures
and functions introduced for implementing a common dma-resv per GPU-VM
including tracking of external and evicted objects in subsequent
patches.
[1] Documentation/gpu/drm-vm-bind-async.rst
[2] Documentation/gpu/drm-vm-bind-locking.rst
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230920144343.64830-2-dakr@redhat.com
Multiple power domains need to be handled explicitly in each driver. The
driver core can not handle it automatically since it is not aware of
power sequencing requirements the hardware might have. This is not a
problem for simpledrm since everything is expected to be powered on by
the bootloader. simpledrm has just ensure it remains powered on during
its lifetime.
This is required on Apple silicon M2 and M2 Pro/Max/Ultra desktop
systems. The HDMI output initialized by the bootloader requires keeping
the display controller and a DP phy power domain on.
Signed-off-by: Janne Grunau <j@jannau.net>
Reviewed-by: Eric Curtin <ecurtin@redhat.com>
Reviewed-by: Neal Gompa <neal@gompa.dev>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20230912-simpledrm-multiple-power-domains-v2-1-01b66bfb1980@jannau.net
The driver uses a naming convention where functions for struct drm_*_funcs
callbacks are named ssd130x_$object_$operation, while the callbacks for
struct drm_*_helper_funcs are named ssd130x_$object_helper_$operation.
The idea is that this helper_ prefix in the function names denote that are
for struct drm_*_helper_funcs callbacks. This convention was copied from
other drivers when ssd130x was written, but Maxime pointed out that is the
exception rather than the norm.
So let's get rid of the _helper prefixes from the function handlers names.
Suggested-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20230914195138.1518065-1-javierm@redhat.com
There are some weird EDIDs floating around that have the sync
pulse extending beyond the end of the blanking period.
On the currently problemtic machine (HP Omni 120) EDID reports
the following mode:
"1600x900": 60 108000 1600 1780 1860 1800 900 910 913 1000 0x40 0x5
which is then "corrected" to have htotal=1861 by the current drm_edid.c
code.
The fixup code was originally added in commit 7064fef563 ("drm: work
around EDIDs with bad htotal/vtotal values"). Googling around we end up in
https://bugs.launchpad.net/ubuntu/hardy/+source/xserver-xorg-video-intel/+bug/297245
where we find an EDID for a Dell Studio 15, which reports:
(II) VESA(0): clock: 65.0 MHz Image Size: 331 x 207 mm
(II) VESA(0): h_active: 1280 h_sync: 1328 h_sync_end 1360 h_blank_end 1337 h_border: 0
(II) VESA(0): v_active: 800 v_sync: 803 v_sync_end 809 v_blanking: 810 v_border: 0
Note that if we use the hblank size (as opposed of the hsync_end)
from the DTD to determine htotal we get exactly 60Hz refresh rate in
both cases, whereas using hsync_end to determine htotal we get a
slightly lower refresh rates. This makes me believe the using the
hblank size is what was intended even in those cases.
Also note that in case of the HP Onmi 120 the VBIOS boots with these:
crtc timings: 108000 1600 1780 1860 1800 900 910 913 1000, type: 0x40 flags: 0x5
ie. it just blindly stuffs the bogus hsync_end and htotal from the DTD
into the transcoder timing registers, and the display works. I believe
the (at least more modern) hardware will automagically terminate the hsync
pulse when the timing generator reaches htotal, which again points that we
should use the hblank size to determine htotal. Unfortunatley the old bug
reports for the Dell machines are extremely lacking in useful details so
we have no idea what kind of timings the VBIOS programmed into the
hardware :(
Let's just flip this quirk around and reduce the length of the sync
pulse instead of extending the blanking period. This at least seems
to be the correct thing to do on more modern hardware. And if any
issues crop up on older hardware we need to debug them properly.
v2: Add debug message breadcrumbs (Jani)
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/8895
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230920211934.14920-1-ville.syrjala@linux.intel.com
Based on grepping through the source code this driver appears to be
missing a call to drm_atomic_helper_shutdown() at system shutdown time
and at driver unbind time. Among other things, this means that if a
panel is in use that it won't be cleanly powered off at system
shutdown time.
The fact that we should call drm_atomic_helper_shutdown() in the case
of OS shutdown/restart and at driver remove (or unbind) time comes
straight out of the kernel doc "driver instance overview" in
drm_drv.c.
I have attempted to put this in the right place at unbind time. In
most other DRM drivers the call is made right after the call to
drm_kms_helper_poll_fini(), so I've put it there. That means that this
call will also be made in the case that we hit errors in bind, since
kirin_drm_kms_cleanup() is called both in the bind error path and in
unbind. I believe this is harmless even though it's not needed in the
bind error path.
For handling shutdown, we rely on the common technique of seeing if
the drvdata is NULL to know whether we need to call
drm_atomic_helper_shutdown(). This makes it important to make sure
that the drvdata is NULL if bind failed or if unbind was called. We
don't need the actual check for NULL and we'll rely on the patch
("drm/atomic-helper: drm_atomic_helper_shutdown(NULL) should be a
noop").
Suggested-by: Maxime Ripard <mripard@kernel.org>
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20230901163944.RFT.6.I21e0916bbd276033f7d31979c0da171458dedd4d@changeid
Based on grepping through the source code these drivers appear to be
missing a call to drm_atomic_helper_shutdown() at system shutdown time
and at driver remove (or unbind) time. Among other things, this means
that if a panel is in use that it won't be cleanly powered off at
system shutdown time.
The fact that we should call drm_atomic_helper_shutdown() in the case
of OS shutdown/restart and at driver remove (or unbind) time comes
straight out of the kernel doc "driver instance overview" in
drm_drv.c.
A few notes about these fixes:
- I confirmed that these drivers were all DRIVER_MODESET type drivers,
which I believe makes this relevant.
- I confirmed that these drivers were all DRIVER_ATOMIC.
- When adding drm_atomic_helper_shutdown() to the remove/unbind path,
I added it after drm_kms_helper_poll_fini() when the driver had
it. This seemed to be what other drivers did. If
drm_kms_helper_poll_fini() wasn't there I added it straight after
drm_dev_unregister().
- This patch deals with drivers using the component model in similar
ways as the patch ("drm: Call drm_atomic_helper_shutdown() at
shutdown time for misc drivers")
- These fixes rely on the patch ("drm/atomic-helper:
drm_atomic_helper_shutdown(NULL) should be a noop") to simplify
shutdown.
Suggested-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> # tilcdc
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20230901163944.RFT.5.I771eb4bd03d8772b19e7dcfaef3e2c167bce5846@changeid
