Despite having different IP offsets the computed address of the register(s)
are the same between gfx7..gfx10. This patch fixes the offset relative
to the GC block on gfx10.
(v2): SQ_DEBUG_STS_GLOBAL2 is 0x10 higher ...
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
in amdgpu_display_crtc_set_config, the call to pm_runtime_get_sync
increments the counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
in amdgpu_drm_ioctl the call to pm_runtime_get_sync increments the
counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
in amdgpu_driver_open_kms the call to pm_runtime_get_sync increments the
counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
The memcpy() function copies n bytes from memory area src to memory area
dest. So specify the firmware size in bytes.
[How]
Correct the calculation.
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Lei Guo <raykwok1150@163.com>
Reviewed-by: Junwei Zhang <zjunweihit@163.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add rename the gpu busy percentage for consistency and
add the mem busy percentage documentation.
Reviewed-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Nirmoy Das <nirmoy.das@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
kobject_init_and_add() takes reference even when it fails.
If this function returns an error, kobject_put() must be called to
properly clean up the memory associated with the object.
Signed-off-by: Qiushi Wu <wu000273@umn.edu>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Use WARN to print messages with backtrace when evictions are triggered.
This can help determine the root cause of evictions and help spot driver
bugs triggering evictions unintentionally, or help with performance tuning
by avoiding conditions that cause evictions in a specific workload.
The messages are controlled by a new module parameter that can be changed
at runtime:
echo Y > /sys/module/amdgpu/parameters/debug_evictions
echo N > /sys/module/amdgpu/parameters/debug_evictions
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In function fill_iram_v_2, the ram_table->bright_neg_gain`s
first element [0][0] seems to be missing. This change is just
to make the code a bit readable.
Signed-off-by: Bernard Zhao <bernard@vivo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
SDMA ULV can benefit low-power modes, but can sometimes cause
latency increases in small SDMA transfers. Server SKUs have a
different trade-off space in this domain, so this configures
the server SKUs' ULV hysteresis times differently than consumer
SKUs'.
Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The existing code used the major version number of the DRM driver
instead of the device major number of the DRM subsystem for
validating access for a devices cgroup.
This meant that accesses allowed by the devices cgroup weren't
permitted and certain accesses denied by the devices cgroup were
permitted (if they matched the wrong major device number).
Signed-off-by: Lorenz Brun <lorenz@brun.one>
Fixes: 6b855f7b83 ("drm/amdkfd: Check against device cgroup")
Reviewed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Header Changes]
- Add new initialization bits for driver to check
firmware status
- Add command for HW locking via DMUB
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
dmub FW running abnormal after resume from S0i3 due
to data aliagnment issue.
[How]
Before having a solution for this issue, temparory
not doing data pack.
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Sung Lee <Sung.Lee@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
during S0i3, set power state is toggled a few times,
and dmub uC will restart with current reset/hw_init.
[How]
Remove reset in set power state, and before doing hw_init,
check if dmub is enabled, and doing FW autoload check only
if dmub is already enabled.
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
For some special timing with border, like DMT 640*480 72Hz,
pipe split can't handle well. Thus, it will be black screen
for these special timing.
[How]
Disable pipe split for these timing with borders as W/A.
Signed-off-by: Dale Zhao <dale.zhao@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Multi-adapter calculate regamma table at the same time.
Two thread used the same global variable cause race
condition.
[How]
Change global buffer to local buffer
Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DC is very fast at link training and stream enablement
which causes issues such as blackscreens for non-compliant
monitors.
[How]
After debugging with scaler vendors we implement the
minimum delays at the necessary locations to ensure
the monitor does not hang. Delays are generic due to
lack of IEEE OUI information on the failing displays.
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Acked-by: Tony Cheng <Tony.Cheng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
During the debugging process related to a hot-plug
problem with 4k display, we realized that we had
some issues related to the global state validation.
This problem was not explicitly highlighted in the
dmesg log, for this reason, this commit adds a function
that converts `enum dc_status` to a human-readable
string and appends the proper warning message in case
of failure.
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why]
Dummy pstate latency actually varies between different
UCLK frequencies, when calculating watermark C, if DAL
always assumes worst case, then it can lead to dummy
pstate not supported scenarios.
[how]
Rather than statically calculating dummy pstate using
worst case, we store the entire table of UCLK to dummy
pstate relationships. On a per mode basis, we calculate
the actual UCLK lower limit, and use the dynamic worst
case dummy pstate latency. This prevents the situation
where we don't support full p-state (which will force
high DPM), but still use low DPM dummy pstate latency.
Signed-off-by: Jun Lei <jun.lei@amd.com>
Reviewed-by: Joshua Aberback <Joshua.Aberback@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
Add emul specific hw function to dmub, in case of
emulator is created, we can runtime switch between
dmub emulator or dmub uC via is_virtual flag in dmub.
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Some sprcified monitor scalar cannot recognize timing
change on demand. Once the link phy disable and enable
during a short period then the Sink protection mechanism
could keep the screen in blank and cannot be recoverred.
[How]
To add 100ms delay between enable link phy and link training.
Signed-off-by: Martin Tsai <martin.tsai@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DML failures occur for 420 modes with dynamic pipe
splitting enabled because the ChromaViewport exceeds
the ChromaSurfaceWidth.
This is caused by adding the viewport_width instead
of the viewport_width_c.
This similarly occurs for rotated modes due to the
use of viewport_height instead of viewport_height_c.
[How]
Correct the calculations.
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DSC calculations fail because the u16 bits_per_pixel from
the DRM struct is being casted to the u8 drm_bpp parameters
and locals. Integer wraparound is happening because this
value is greater than 255.
[How]
Use u16 to match what's in the structure instead of u8.
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When we want to use float point operation on Linux
we need to use within special kernel protection
(`kernel_fpu_{begin,end}()`.), otherwise the kernel
can clobber userspace FPU register state. For detecting
these issues we use a tool named objtool (with -Ffa
flags) to highlight the FPU problems, all warnings can
be summed up as follows:
./tools/objtool/objtool check -Ffa
drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.o
[..] dc/dsc/rc_calc.o: warning: objtool: get_qp_set()+0x2f8:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc.o: warning: objtool: dsc_roundf()+0x5:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc.o: warning: objtool: dsc_ceil()+0x5:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc.o: warning: objtool: get_ofs_set()+0x3eb:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc.o: warning: objtool: calc_rc_params()+0x3c:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/dc_dsc.o: warning: objtool:
get_dsc_bandwidth_range.isra.0()+0x8d:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/dc_dsc.o: warning: objtool: setup_dsc_config()+0x2ef:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc_dpi.o: warning: objtool:copy_pps_fields()+0xbb:
FPU instruction outside of kernel_fpu_{begin,end}()
[..] dc/dsc/rc_calc_dpi.o: warning: objtool:
dscc_compute_dsc_parameters()+0x7b:
FPU instruction outside of kernel_fpu_{begin,end}()
This commit fixes the above issues by rework DSC as described:
1. Isolate all FPU operations in a single file;
2. Use FPU flags only in the file that handles FPU operations;
3. Isolate all functions that require float point operation in static
functions;
4. Add a mid-layer function that does not use any float point operation,
and that could be safely invoked in other parts of the code.
5. Keep float point operation under DC_FP_{START/END} macro.
CC: Christian König <christian.koenig@amd.com>
CC: Alexander Deucher <Alexander.Deucher@amd.com>
CC: Peter Zijlstra <peterz@infradead.org>
CC: Tony Cheng <tony.cheng@amd.com>
CC: Harry Wentland <hwentlan@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Reviewed-by: Mikita Lipski <Mikita.Lipski@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why]
for following new format, no alpha
SURFACE_PIXEL_FORMAT_GRPH_RGB111110_FLOAT/_FIX:
SURFACE_PIXEL_FORMAT_GRPH_BGR101111_FLOAT/_FIX
same as case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555:
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Even though they are technically MMIO registers I put the bits with the sqind block
for organizational purposes.
Requested for UMR debugging.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Guest VM issue the PSP clear_vf_fw command at 2 points:
1.On VF driver loading, after VF message PSP to setup rings,
the next command is “clear_vf_fw”
2.On VF driver unload before VF message to
destroy rings
Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Ack-by: Monk.liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fixes: c1cf79ca5c ("drm/amdgpu: use IP discovery table for renoir")
This nullptr issue should be specific on the Renoir series during try access the PWR_MISC_CNTL_STATUS
when PWR IP not been detected by discovery table. Moreover the PWR IP not existing in Renoir series is
expected therefore just avoid access PWR register in Renoir series.
Signed-off-by: Prike.Liang <Prike.Liang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
