| Age | Commit message (Collapse) | Author |
|---|
|
Most of the driver code doesn't need to reach in to msm specific fields,
so just use the drm_gpuvm/drm_gpuva types directly. This should
hopefully improve commonality with other drivers and make the code
easier to understand.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Rob Clark <robin.clark@oss.qualcomm.com>
Tested-by: Antonino Maniscalco <antomani103@gmail.com>
Reviewed-by: Antonino Maniscalco <antomani103@gmail.com>
Patchwork: https://patchwork.freedesktop.org/patch/661483/
|
|
Re-aligning naming to better match drm_gpuvm terminology will make
things less confusing at the end of the drm_gpuvm conversion.
This is just rename churn, no functional change.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Rob Clark <robin.clark@oss.qualcomm.com>
Tested-by: Antonino Maniscalco <antomani103@gmail.com>
Reviewed-by: Antonino Maniscalco <antomani103@gmail.com>
Patchwork: https://patchwork.freedesktop.org/patch/661466/
|
|
ACD a.k.a Adaptive Clock Distribution is a feature which helps to reduce
the power consumption. In some chipsets, it is also a requirement to
support higher GPU frequencies. This patch adds support for GPU ACD by
sending necessary data to GMU and AOSS. The feature support for the
chipset is detected based on devicetree data.
Signed-off-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Tested-by: Maya Matuszczyk <maccraft123mc@gmail.com>
Tested-by: Anthony Ruhier <aruhier@mailbox.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Patchwork: https://patchwork.freedesktop.org/patch/649342/
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 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>
|
|
Clang-19 and above sometimes end up with multiple copies of the large
a6xx_hfi_msg_bw_table structure on the stack. The problem is that
a6xx_hfi_send_bw_table() calls a number of device specific functions to
fill the structure, but these create another copy of the structure on
the stack which gets copied to the first.
If the functions get inlined, that busts the warning limit:
drivers/gpu/drm/msm/adreno/a6xx_hfi.c:631:12: error: stack frame size (1032) exceeds limit (1024) in 'a6xx_hfi_send_bw_table' [-Werror,-Wframe-larger-than]
Fix this by kmalloc-ating struct a6xx_hfi_msg_bw_table instead of using
the stack. Also, use this opportunity to skip re-initializing this table
to optimize gpu wake up latency.
Cc: Arnd Bergmann <arnd@kernel.org>
Signed-off-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/621814/
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Totally useless.
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Patchwork: https://patchwork.freedesktop.org/patch/588804/
Link: https://lore.kernel.org/r/20240410-topic-msm_rw-v1-1-e1fede9ffaba@linaro.org
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
|
|
The QMP mailbox expects to be notified of the ACD (Adaptive Clock
Distribution) state. Get a handle to the mailbox at probe time and
poke it at GMU resume.
Since we don't fully support ACD yet, hardcode the message to "val: 0"
(state = disabled).
Tested-by: Neil Armstrong <neil.armstrong@linaro.org> # on SM8550-QRD
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> # sm8450
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/559287/
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
These two will be reused by at least A619_holi in the non-gmu
paths. Turn them non-static them to make it possible.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/542751/
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Sparse reports plenty of warnings against the a6xx code because of
a6xx_gmu::mmio and a6xx_gmu::rscc members. For some reason they were
defined as __iomem pointers rather than pointers to __iomem memory.
Correct the __iomem attribute.
Fixes: 02ef80c54e7c ("drm/msm/a6xx: update pdc/rscc GMU registers for A640/A650")
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202304070550.NrbhJCvP-lkp@intel.com/
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/531583/
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
As per the recommended recovery sequence of adreno gpu, cx gdsc should
collapse at hardware before it is turned back ON. This helps to clear
out the stale states in hardware before it is reinitialized. Use the
genpd notifier along with the newly introduced
dev_pm_genpd_synced_poweroff() api to ensure that cx gdsc has collapsed
before we turn it back ON.
Signed-off-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/516472/
Link: https://lore.kernel.org/r/20230102161757.v5.5.I9e10545c6a448d5eb1b734839b871d1b3146dac3@changeid
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
When a device has multiple power domains, dev->power_domain is left
empty during probe. That didn't cause any issue so far because we are
freeloading on smmu driver's vote on cx gdsc. Instead of that, create
a device_link between cx genpd device and gmu device to keep a vote from
gpu driver.
Before this patch:
localhost ~ # cat /sys/kernel/debug/pm_genpd/pm_genpd_summary
gx_gdsc on 0
/devices/genpd:1:3d6a000.gmu active 0
cx_gdsc on 0
/devices/platform/soc@0/3da0000.iommu active 0
After this patch:
localhost ~ # cat /sys/kernel/debug/pm_genpd/pm_genpd_summary
gx_gdsc on 0
/devices/genpd:1:3d6a000.gmu active 0
cx_gdsc on 0
/devices/platform/soc@0/3da0000.iommu active 0
/devices/genpd:0:3d6a000.gmu active 0
Signed-off-by: Akhil P Oommen <quic_akhilpo@quicinc.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/516468/
Link: https://lore.kernel.org/r/20230102161757.v5.3.I7f545d8494dcdbe6e96a15fbe8aaf5bb0c003d50@changeid
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
With sparse ("make C=2"), lots of
error: return expression in void function
messages are seen.
Fix this by removing the return statements to propagate void return
values.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Patchwork: https://patchwork.freedesktop.org/patch/492529/
Link: https://lore.kernel.org/r/0083bc7e23753c19902580b902582ae499b44dbf.1657113388.git.geert@linux-m68k.org
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
I've seen some crashes in our crash reporting that *look* like multiple
threads stomping on each other while communicating with GMU. So wrap
all those paths in a lock.
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
This patch adds support for the gpu found in the Snapdragon 7c Gen 3
compute platform. This gpu is similar to the exisiting a660 gpu with
minor delta in the programing sequence. As the Adreno GPUs are moving
away from a numeric chipid based naming scheme to a string, it was
decided to use 0x06030500 as the chip id of this gpu to communicate
to the userspace driver.
Signed-off-by: Akhil P Oommen <akhilpo@codeaurora.org>
Link: https://lore.kernel.org/r/20210730011945.v4.3.I610377db0934b6b7deda532ec2bf786a02c38c01@changeid
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Now that the bug is fixed in the minimal way for stable, go make the
code table-driven.
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
We were using the same force-poweron bit in the two codepaths, so they
could race to have one of them lose GPU power early.
freedreno CI was seeing intermittent errors like:
[drm:_a6xx_gmu_set_oob] *ERROR* Timeout waiting for GMU OOB set GPU_SET: 0x0
and this issue could have contributed to it.
Signed-off-by: Eric Anholt <eric@anholt.net>
Fixes: 4b565ca5a2cb ("drm/msm: Add A6XX device support")
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
A650 has a separate RSCC region, so dump RSCC registers separately, reading
them from the RSCC base. Without this change a GPU hang will cause a system
reset if CONFIG_DEV_COREDUMP is enabled.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
This is required for a650 to work.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Update the gmu_pdc registers for A640 and A650.
Some of the RSCC registers on A650 are in a separate region.
Note this also changes the address of these registers:
RSCC_TCS1_DRV0_STATUS
RSCC_TCS2_DRV0_STATUS
RSCC_TCS3_DRV0_STATUS
Based on the values in msm-4.14 and msm-4.19 kernels.
v3: replaced adreno_is_a650 around ->rscc with checks for "rscc" resource
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Newer GPUs have different GMU firmware path.
v3: updated a6xx_gmu_fw_load based on feedback, including gmu_write_bulk,
and removed extra whitespace change
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Add HFI v2 code paths required by Adreno 640 and 650 GPUs.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
This gives more fine-grained control over how memory is allocated over the
DMA api. In particular, it allows using an address range or pinning to
a fixed address.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
The GMU has very few memory allocations and uses a flat memory space so
there is no good reason to go out of our way to bypass the DMA APIs which
were basically designed for this exact scenario.
v7: Check return value of dma_set_mask_and_coherent
v4: Use dma_alloc_wc()
v3: Set the dma mask correctly and use dma_addr_t for the iova type
v2: Pass force_dma false to of_dma_configure to require that the DMA
region be set up and return error from of_dma_configure to fail probe.
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Previously, if the freq were overriden (ie. via sysfs), it would get
reset to max on resume.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
|
|
The driver checks for gmu->mmio as a sign that the device has been
initialized, however there are failures in probe below the mmio init.
If one of those is hit, mmio will be non-null but freed.
In that case, a6xx_gmu_probe will return an error to a6xx_gpu_init which
will in turn call a6xx_gmu_remove which checks gmu->mmio and tries to free
resources for a second time. This causes a great boom.
Fix this by adding an initialized member to gmu which is set on
successful probe and cleared on removal.
Changes in v2:
- None
Cc: Jordan Crouse <jcrouse@codeaurora.org>
Reviewed-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190523171653.138678-1-sean@poorly.run
|
|
The HFI tasklet was removed in df0dff1 ("drm/msm/a6xx: Poll for HFI
responses") but the tasklet_struct was accidentally left behind.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Now that the GX domain is sorted we can wire up a working GMU reset.
IF a GMU hang was detected then try to forcefully shut down the GMU
in the power down sequence which should ensure that it can recover
normally on the next power up.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
99.999% of the time during normal operation the GMU is responsible
for power and clock control on the GX domain and the CPU remains
blissfully unaware. However, there is one situation where the CPU
needs to get involved:
The power sequencing rules dictate that the GX needs to be turned
off before the CX so that the CX can be turned on before the GX
during power up. During normal operation when the CPU is taking
down the CX domain a stop command is sent to the GMU which turns
off the GX domain and then the CPU handles the CX domain.
But if the GMU happened to be unresponsive while the GX domain was
left then the CPU will need to step in and turn off the GX domain
before resetting the CX and rebooting the GMU. This unfortunately
means that the CPU needs to be marginally aware of the GX domain
even though it is expected to usually keep its hands off.
To support this we create a semi-disabled GX power domain that
does nothing to the hardware on power up but tries to shut it
down normally on power down. In this method the reference counting
is correct and we can step in with the pm_runtime_put() at the right
time during the failure path.
This patch sets up the connection to the GX power domain and does
the magic to "enable" and disable it at the right points.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
The GMU code currently has some misguided code to try to work around
a hardware quirk that requires the power domains on the GPU be
collapsed in a certain order. Upcoming patches will do this the
right way so get rid of the unused and unwanted regulator
code.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@chromium.org>
|
|
Add support for gathering and dumping the a6xx GPU state including
registers, GMU registers, indexed registers, shader blocks,
context clusters and debugbus.
v2: Fix bugs discovered by Sharat Masetty
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
|
Implement routines to estimate GPU busy time and fetching the
current frequency for the polling interval. This is required by
the devfreq framework which recommends a frequency change if needed.
The driver code then tries to set this new frequency on the GPU by
sending an Out Of Band(OOB) request to the GMU.
Signed-off-by: Sharat Masetty <smasetty@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
|
Add a simple function to read 64 registers in the GMU domain
Signed-off-by: Sharat Masetty <smasetty@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
|
The only HFI communication with the GMU on sdm845 happens
during initialization and all commands are synchronous. A fancy
interrupt tasklet and associated infrastructure is entirely
not eeded and puts us at the mercy of the scheduler.
Instead poll for the message signal and handle the response
immediately and go on our way.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
|
The current design greedily takes a big chunk of the PDC
register space instead of just the GPU specific sections
which conflicts with other drivers and generally makes
a mess of things.
Furthermore we only need to map the GPU PDC sections
just once during init so map the memory inside the function
that uses it and adjust the pointers and register offsets
accordingly.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
|
Add support for the A6XX family of Adreno GPUs. The biggest addition
is the GMU (Graphics Management Unit) which takes over most of the
power management of the GPU itself but in a ironic twist of fate
needs a goodly amount of management itself. Add support for the
A6XX core code, the GMU and the HFI (hardware firmware interface)
queue that the CPU uses to communicate with the GMU.
Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
|
