We only fetch it once from userland, so let's also only notify the
user once and not on every runtime resume.
As you can notice by the tags chain, more than one user found this
annoying.
Reported-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
Suggested-by: Abel Vesa <abel.vesa@linaro.org>
Suggested-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/637062/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Debugging incorrect UAPI usage tends to be a bit painful, so add a
helper macro to make it easier to add debug logging which can be enabled
at runtime via drm.debug.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/630578/
The call chain on a5xx_gpu_init is such that pdev is not going to be
null, so the null check on pdev can be removed. This also cleans up
a static analysis warning where pdev is dereferenced before the null
check which cannot actually occur.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Patchwork: https://patchwork.freedesktop.org/patch/624180/
Signed-off-by: Rob Clark <robdclark@chromium.org>
The "submit->cmd[i].size" and "submit->cmd[i].offset" variables are u32
values that come from the user via the submit_lookup_cmds() function.
This addition could lead to an integer wrapping bug so use size_add()
to prevent that.
Fixes: 198725337e ("drm/msm: fix cmdstream size check")
Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/624696/
Signed-off-by: Rob Clark <robdclark@chromium.org>
This adds MSM_PARAM_UCHE_TRAP_BASE that will be used by Mesa
implementation for VK_KHR_shader_clock and GL_ARB_shader_clock.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Patchwork: https://patchwork.freedesktop.org/patch/627036/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Now all the DDR bandwidth voting via the GPU Management Unit (GMU)
is in place, declare the Bus Control Modules (BCMs) and the
corresponding parameters in the GPU info struct.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629401/
Signed-off-by: Rob Clark <robdclark@chromium.org>
The Adreno GPU Management Unit (GMU) can also scale the DDR Bandwidth
along the Frequency and Power Domain level, until now we left the OPP
core scale the OPP bandwidth via the interconnect path.
In order to enable bandwidth voting via the GPU Management
Unit (GMU), when an opp is set by devfreq we also look for
the corresponding bandwidth index in the previously generated
bw_table and pass this value along the frequency index to the GMU.
The GMU also takes another vote called AB which is a 16bit quantized
value of the floor bandwidth against the maximum supported bandwidth.
The AB is calculated with a default 25% of the bandwidth like the
downstream implementation too inform the GMU firmware the minimal
quantity of bandwidth we require for this OPP. Only pass the AB
vote starting from A750 GPUs.
Since we now vote for all resources via the GMU, setting the OPP
is no more needed, so we can completely skip calling
dev_pm_opp_set_opp() in this situation.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Patchwork: https://patchwork.freedesktop.org/patch/629397/
Signed-off-by: Rob Clark <robdclark@chromium.org>
The Adreno GPU Management Unit (GMU) can also scale the ddr
bandwidth along the frequency and power domain level, but for
now we statically fill the bw_table with values from the
downstream driver.
Only the first entry is used, which is a disable vote, so we
currently rely on scaling via the linux interconnect paths.
Let's dynamically generate the bw_table with the vote values
previously calculated from the OPPs.
Those entries will then be used by the GMU when passing the
appropriate bandwidth level while voting for a gpu frequency.
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Patchwork: https://patchwork.freedesktop.org/patch/629396/
Signed-off-by: Rob Clark <robdclark@chromium.org>
The Adreno GPU Management Unit (GMU) can also scale DDR Bandwidth along
the Frequency and Power Domain level, but by default we leave the
OPP core scale the interconnect ddr path.
While scaling via the interconnect path was sufficient, newer GPUs
like the A750 requires specific vote paremeters and bandwidth to
achieve full functionality.
In order to calculate vote values used by the GPU Management
Unit (GMU), we need to parse all the possible OPP Bandwidths and
create a vote value to be sent to the appropriate Bus Control
Modules (BCMs) declared in the GPU info struct.
This vote value is called IB, while on the other side the GMU also
takes another vote called AB which is a 16bit quantized value
of the floor bandwidth against the maximum supported bandwidth.
The AB vote will be calculated later when setting the frequency.
The vote array will then be used to dynamically generate the GMU
bw_table sent during the GMU power-up.
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Patchwork: https://patchwork.freedesktop.org/patch/629395/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Even if the code uses ARRAY_SIZE() to fill those tables,
it's still a best practice to not use magic values for
tables in structs.
Suggested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629393/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Log the version for informational purposes, such as for keeping track
of possible GMU fw-related failures in crash / CI logs.
Intentionally not implemented on the if (gmu->legacy) codepath, as
these registers seem not to be used on there.
Downstream additionally warns if the firmware version is too old for
a given GPU, but we already pair the binary to a given GPU, so let's
not go there at the moment.
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/629934/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Add a register that contains the GMU core firmware version on non-
legacy (non-sdm845-family) SoCs.
The name is guesstimated based on what it does downstream, but it'll
do.
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/629932/
Signed-off-by: Rob Clark <robdclark@chromium.org>
The CWB mux has its own registers and set of operations. Add dpu_hw_cwb
abstraction to allow driver to configure the CWB mux.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629254/
Link: https://lore.kernel.org/r/20241216-concurrent-wb-v4-12-fe220297a7f0@quicinc.com
[DB: added #include <linux/bitfield.h>]
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
The struct dpu_rm_requirements was used to wrap display topology and
hw resources, which meant INTF indices. As of commit ef58e0ad34
("drm/msm/dpu: get INTF blocks directly rather than through RM") the hw
resources struct was removed, leaving struct dpu_rm_requirements
containing a single field (topology). Remove the useless wrapper.
Signed-off-by: Jessica Zhang <quic_jesszhan@quicinc.com>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
[DB: dropped stray msm_drv.h inclusion]
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629259/
Link: https://lore.kernel.org/r/20241216-concurrent-wb-v4-5-fe220297a7f0@quicinc.com
MSM display drivers provide kms structure allocated during probe().
Don't clean up priv->kms field in case of an error. Otherwise probe
functions might fail after KMS probe deferral.
Fixes: a2ab5d5bb6 ("drm/msm: allow passing struct msm_kms to msm_drv_probe()")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Fixes: 506efcba31 ("drm/msm: carve out KMS code from msm_drv.c")
Patchwork: https://patchwork.freedesktop.org/patch/590411/
Link: https://lore.kernel.org/r/20240420-mdp4-fixes-v1-1-96a70f64fa85@linaro.org
Because the of_dma_configure() will returns '-EPROBE_DEFER' if the probe
procedure of the specific platform IOMMU driver is not finished yet. It
can also return other error code for various reasons.
Stop pretending that it will always suceess, quit if it fail.
Signed-off-by: Sui Jingfeng <sui.jingfeng@linux.dev>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Fixes: 29ac8979cd ("drm/msm/a6xx: use msm_gem for GMU memory objects")
Fixes: 5a903a44a9 ("drm/msm/a6xx: Introduce GMU wrapper support")
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/622782/
Link: https://lore.kernel.org/r/20241104090738.529848-1-sui.jingfeng@linux.dev
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Passing a variable string as a printf style format is potentially
dangerous that -Wformat-security can warn about if enabled. A new
instance just got added:
drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c: In function 'dpu_kms_mdp_snapshot':
drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c:1046:49: error: format not a string literal and no format arguments [-Werror=format-security]
1046 | vbif->name);
| ~~~~^~~~~~
Fix this one and the preexisting -Wformat-security warnings the in the
DRM code for snapdragon.
Fixes: 1a40bb31fc ("drm/msm/dpu: Add VBIF to DPU snapshot") # and others
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/629126/
Link: https://lore.kernel.org/r/20241216083319.1838449-1-arnd@kernel.org
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
If a driver calls dev_pm_opp_find_bw_ceil/floor() the retrieve bandwidth
from the OPP table but the bandwidth table was not created because the
interconnect properties were missing in the OPP consumer node, the
kernel will crash with:
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000004
...
pc : _read_bw+0x8/0x10
lr : _opp_table_find_key+0x9c/0x174
...
Call trace:
_read_bw+0x8/0x10 (P)
_opp_table_find_key+0x9c/0x174 (L)
_find_key+0x98/0x168
dev_pm_opp_find_bw_ceil+0x50/0x88
...
In order to fix the crash, create an assert function to check
if the bandwidth table was created before trying to get a
bandwidth with _read_bw().
Fixes: add1dc094a ("OPP: Use generic key finding helpers for bandwidth key")
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Pass the freq index to the assert function to make sure
we do not read a freq out of the opp->rates[] table when called
from the indexed variants:
dev_pm_opp_find_freq_exact_indexed() or
dev_pm_opp_find_freq_ceil/floor_indexed().
Add a secondary parameter to the assert function, unused
for assert_single_clk() then add assert_clk_index() which
will check for the clock index when called from the _indexed()
find functions.
Fixes: 142e17c1c2 ("OPP: Introduce dev_pm_opp_find_freq_{ceil/floor}_indexed() APIs")
Fixes: a5893928bb ("OPP: Add dev_pm_opp_find_freq_exact_indexed()")
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
The "opp->bandwidth" array has "opp->opp_table->path_count" number of
elements. It's allocated in _opp_allocate(). So this > needs to be >=
to prevent an out of bounds access.
Fixes: d78653dcd8bf ("opp: core: implement dev_pm_opp_get_bw")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Add and implement dev_pm_opp_get_bw() to retrieve the OPP's
bandwidth in the same way as the dev_pm_opp_get_voltage() helper.
Retrieving bandwidth is required in the case of the Adreno GPU
where the GPU Management Unit can handle the Bandwidth scaling.
The helper can get the peak or average bandwidth for any of
the interconnect path.
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
[ Viresh: Fixed commit log and a comment in code ]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
All other submodules pass arguments directly. Drop struct
msm_dp_panel_in that is used to wrap dp_panel's submodule args and pass
all data to msm_dp_panel_get() directly.
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Tested-by: Stephen Boyd <swboyd@chromium.org> # sc7180-trogdor
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629055/
Link: https://lore.kernel.org/r/20241216-fd-dp-audio-fixup-v4-4-f8d1961cf22f@linaro.org
Use msm_dp_utils_pack_sdp_header() and call msm_dp_write_link() directly
to program audio packet data. Use 0 as Packet ID, as it was not
programmed earlier.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Tested-by: Stephen Boyd <swboyd@chromium.org> # sc7180-trogdor
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/629051/
Link: https://lore.kernel.org/r/20241216-fd-dp-audio-fixup-v4-2-f8d1961cf22f@linaro.org
The msm_dp_panel_dump_regs() and msm_dp_catalog_dump_regs() are not
called anywhere. If there is a necessity to dump registers, the
snapshotting should be used instead. Drop these two functions.
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Tested-by: Stephen Boyd <swboyd@chromium.org> # sc7180-trogdor
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629049/
Link: https://lore.kernel.org/r/20241216-fd-dp-audio-fixup-v4-1-f8d1961cf22f@linaro.org
Virtual wide planes give high amount of flexibility, but it is not
always enough:
In parallel multirect case only the half of the usual width is supported
for tiled formats. Thus the whole width of two tiled multirect
rectangles can not be greater than max_linewidth, which is not enough
for some platforms/compositors.
Another example is as simple as wide YUV plane. YUV planes can not use
multirect, so currently they are limited to max_linewidth too.
Now that the planes are fully virtualized, add support for allocating
two SSPP blocks to drive a single DRM plane. This fixes both mentioned
cases and allows all planes to go up to 2*max_linewidth (at the cost of
making some of the planes unavailable to the user).
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/629026/
Link: https://lore.kernel.org/r/20241215-dpu-virtual-wide-v8-2-65221f213ce1@linaro.org
Only several SSPP blocks support such features as YUV output or scaling,
thus different DRM planes have different features. Properly utilizing
all planes requires the attention of the compositor, who should
prefer simpler planes to YUV-supporting ones. Otherwise it is very easy
to end up in a situation when all featureful planes are already
allocated for simple windows, leaving no spare plane for YUV playback.
To solve this problem make all planes virtual. Each plane is registered
as if it supports all possible features, but then at the runtime during
the atomic_check phase the driver selects backing SSPP block for each
plane.
As the planes are attached to the CRTC and not the encoder, the SSPP
blocks are also allocated per CRTC ID (all other resources are currently
allocated per encoder ID). This also matches the hardware requirement,
where both rectangles of a single SSPP can only be used with the LM
pair.
Note, this does not provide support for using two different SSPP blocks
for a single plane or using two rectangles of an SSPP to drive two
planes. Each plane still gets its own SSPP and can utilize either a solo
rectangle or both multirect rectangles depending on the resolution.
Note #2: By default support for virtual planes is turned off and the
driver still uses old code path with preallocated SSPP block for each
plane. To enable virtual planes, pass 'msm.dpu_use_virtual_planes=1'
kernel parameter.
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/629022/
Link: https://lore.kernel.org/r/20241215-dpu-virtual-wide-v8-1-65221f213ce1@linaro.org
