Evan Quan [Wed, 2 Sep 2020 08:10:10 +0000 (16:10 +0800)]
drm/amd/pm: increase mclk switch threshold to 200 us
To avoid underflow seen on Polaris10 with some 3440x1440
144Hz displays. As the threshold of 190 us cuts too close
to minVBlankTime of 192 us.
Signed-off-by: Evan Quan <evan.quan@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
./drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h:345: warning: Function parameter or member 'force_timing_sync' not described in 'amdgpu_display_manager'
This new struct member was not documented at kernel-doc markup.
Fixes: 3d4e52d0cf24 ("drm/amd/display: Add debugfs for forcing stream timing sync") Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Bernard Zhao [Mon, 12 Oct 2020 11:46:16 +0000 (04:46 -0700)]
drm/amd/display: remove no need return value
Functions (disable_all_writeback_pipes_for_stream &
dc_enable_stereo & dc_post_update_surfaces_to_stream)
always return true, there is no need to keep the return value.
This change is to make the code a bit more readable.
Signed-off-by: Bernard Zhao <bernard@vivo.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fixes: b4a7db71ea06 ("drm/amdgpu: add per device user friendly xgmi events for vega20") CC: Jonathan Kim <jonathan.kim@amd.com> Signed-off-by: kernel test robot <lkp@intel.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Roman Li [Thu, 8 Oct 2020 17:32:47 +0000 (13:32 -0400)]
drm/amd/display: Add green_sardine support to DM
Display Manager support for green_sardine
Signed-off-by: Roman Li <Roman.Li@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Roman Li [Thu, 8 Oct 2020 17:28:41 +0000 (13:28 -0400)]
drm/amd/display: Add green_sardine support to DC
Display Core support for green_sardine
Signed-off-by: Roman Li <Roman.Li@amd.com> Acked-by: Hersen Wu <hersenxs.wu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Wed, 5 Aug 2020 21:59:09 +0000 (17:59 -0400)]
drm/amdgpu: enable jpeg3.0 PG and CG for dimgrey_cavefish
Enable JPEG3.0 PG and CG for dimgrey_cavefish.
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Kenneth Feng <kenneth.feng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
James Zhu [Wed, 5 Aug 2020 21:54:21 +0000 (17:54 -0400)]
drm/amdgpu: enable VCN3.0 PG and CG for dimgrey_cavefish
Enable VCN3.0 PG and CG for dimgrey_cavefish
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Kenneth Feng <kenneth.feng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Rodrigo Siqueira [Fri, 9 Oct 2020 13:36:01 +0000 (09:36 -0400)]
drm/amd/display: Fix module load hangs when connected to an eDP
It was recently introduced a change that enables driver to disable
streams if pixel clock changes. Consequently, the code path executed in
the disable vbios function expanded to an encoder verification part.
The encoder loop is nested inside the pipe count loop, and both loops
share the 'i' variable in control of their flow. This situation may lead
to an infinite loop because the encoder loop constantly updates the `i`
variable, making the first loop always positive. As a result, we can see
a soft hang during the module load (modprobe amdgpu) and a series of
dmesg log that looks like this:
This issue was fixed by introducing a second variable for the internal
loop.
Fixes: 8353d30e747f4e ("drm/amd/display: disable stream if pixel clock changed with link active") Reviewed-by: Roman Li <Roman.Li@amd.com> Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ye Bin [Fri, 9 Oct 2020 07:47:58 +0000 (15:47 +0800)]
drm/amdgpu: Fix inconsistent of format with argument type in amdgpu_xgmi.c
Fix follow warning:
[drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c:249]: (warning) %d in format
string (no. 1) requires 'int' but the argument type is 'unsigned int'.
Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Ye Bin <yebin10@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ye Bin [Fri, 9 Oct 2020 07:42:18 +0000 (15:42 +0800)]
drm/amdgpu: Fix invalid number of character '{' in amdgpu_acpi_init
Fix follow warning:
Checking drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c...
[drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:770]: (error) Invalid number
of character '{' when these macros are defined: ''.
Checking drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c: CONFIG_ACPI...
[drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:770]: (error) Invalid number
of character '{' when these macros are defined: 'CONFIG_ACPI'.
......
Checking drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c: CONFIG_X86...
[drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:770]: (error) Invalid number
of character '{' when these macros are defined: 'CONFIG_X86'.
Checking drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c: _X86_...
[drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:770]: (error) Invalid number
of character '{' when these macros are defined: '_X86_'.
Checking drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c: __linux__...
[drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:770]: (error) Invalid number
of character '{' when these macros are defined: '__linux__'.
Fixes: 97d798b276e9 ("drm/amdgpu: simplify ATIF backlight handling") Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Ye Bin <yebin10@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Wed, 30 Sep 2020 03:54:01 +0000 (11:54 +0800)]
drm/amd/pm: setup APU dpm clock table in SMU HW initialization
As the dpm clock table is needed during DC HW initialization.
And that (DC HW initialization) comes before smu_late_init()
where current APU dpm clock table setup is performed. So, NULL
pointer dereference will be triggered. By moving APU dpm clock
table setup to smu_hw_init(), this can be avoided.
Fixes: 02cf91c113ea ("drm/amd/powerplay: postpone operations not required for hw setup to late_init") Signed-off-by: Evan Quan <evan.quan@amd.com> Reported-by: Dirk Gouders <dirk@gouders.net> Acked-by: Nirmoy Das <nirmoy.das@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Thu, 8 Oct 2020 14:23:05 +0000 (09:23 -0500)]
drm/amdgpu: Use struct_size() helper in kmalloc()
Make use of the new struct_size() helper instead of the offsetof() idiom.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:10:44 +0000 (11:10 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_ppt_v1_pcie_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_ppt_v1_pcie_table, instead of a one-element array, and use
the struct_size() helper to calculate the size for the allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7db0bc.7Xivn4K83f7XW0ug%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:10:16 +0000 (11:10 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_ppt_v1_voltage_lookup_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_ppt_v1_voltage_lookup_table, instead of a one-element array,
and use the struct_size() helper to calculate the size for the allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7d61df.jWrFfnjxGbjSkPOp%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:09:34 +0000 (11:09 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_ppt_v1_mm_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_ppt_v1_mm_clock_voltage_dependency_table, instead of a
one-element array, and use the struct_size() helper to calculate the
size for the allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7d61e2.qiTVTyG2pVoG8bb0%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:08:53 +0000 (11:08 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_ppt_v1_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_ppt_v1_clock_voltage_dependency_table, instead of a one-element
array, and use the struct_size() helper to calculate the size for the
allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c433c.TTk9rnA+F58kyDUy%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:08:08 +0000 (11:08 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_samu_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_samu_clock_voltage_dependency_table, instead of a one-element array,
and use the struct_size() helper to calculate the size for the allocation.
Also, save some heap space as the original code is multiplying
table->numEntries by sizeof(struct phm_samu_clock_voltage_dependency_table)
when it should have been multiplied it by
sizeof(struct phm_samu_clock_voltage_dependency_record) instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c5d3a.ryM4GmZr3e0JeZy+%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:07:38 +0000 (11:07 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_cac_leakage_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_cac_leakage_table, instead of a one-element array,
and use the struct_size() helper to calculate the size for the allocation.
Also, save some heap space as the original code is multiplying
table->ucNumEntries by sizeof(struct phm_cac_leakage_table) when it
should have been multiplied it by sizeof(struct phm_cac_leakage_record)
instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c5d38.iT%2FQTjN+659XUDo5%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:06:48 +0000 (11:06 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_vce_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_vce_clock_voltage_dependency_table, instead of a one-element array,
and use the struct_size() helper to calculate the size for the allocation.
Also, save some heap space as the original code is multiplying
table->numEntries by sizeof(struct phm_vce_clock_voltage_dependency_table)
when it should have multiplied it by sizeof(struct phm_vce_clock_voltage_dependency_record)
instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c5d35.pJToGs3H9khZK6ws%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:06:14 +0000 (11:06 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_phase_shedding_limits_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_phase_shedding_limits_table, instead of a one-element array,
and use the struct_size() helper to calculate the size for the allocation.
Also, save some heap space as the original code is multiplying
ptable->ucNumEntries by sizeof(struct phm_phase_shedding_limits_table)
when it should have multiplied it by sizeof(struct phm_phase_shedding_limits_record)
instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c5d36.6PStUZp2HRxAz7IM%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:05:22 +0000 (11:05 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_acp_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_acp_clock_voltage_dependency_table, instead of a one-element
array, and use the struct_size() helper to calculate the size for the
allocation.
Also, save some heap space as the original code is multiplying
table->numEntries by sizeof(struct phm_acp_clock_voltage_dependency_table)
when it should have multiplied it by sizeof(phm_acp_clock_voltage_dependency_record)
instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c5d3c.TyfOhg%2FA6JycL6ZN%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:04:49 +0000 (11:04 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_uvd_clock_voltage_dependency_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_uvd_clock_voltage_dependency_table, instead of a one-element
array, and use the struct_size() helper to calculate the size for the
allocation.
Also, save some heap space as the original code is multiplying
table->numEntries by sizeof(struct phm_uvd_clock_voltage_dependency_table)
when it should have multiplied it by sizeof(phm_uvd_clock_voltage_dependency_record)
instead.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c433e.pXkC6KsN6HN%2FLdhj%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:04:12 +0000 (11:04 -0500)]
drm/amd/pm: Replace one-element array with flexible-array in struct phm_clock_array
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_clock_array, instead of a one-element array, and use the
struct_size() helper to calculate the size for the allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c433f.ZyMD+YUIVAwiHGVe%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:03:27 +0000 (11:03 -0500)]
drm/amd/pm: Replace one-element array with flexible-array member in struct vi_dpm_table
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Use a flexible-array member in struct vi_dpm_table instead of a
one-element array.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c433c.TTk9rnA+F58kyDUy%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gustavo A. R. Silva [Wed, 7 Oct 2020 16:02:26 +0000 (11:02 -0500)]
drm/amd/pm: Replace one-element array with flexible-array member
There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].
Refactor the code according to the use of a flexible-array member in
struct phm_clock_voltage_dependency_table, instead of a one-element
array, and use the struct_size() helper to calculate the size for the
allocation.
Build-tested-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/lkml/5f7c295c.8iqp1Ifc6oiVDq%2F%2F%25lkp@intel.com/ Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 6 Oct 2020 13:10:24 +0000 (09:10 -0400)]
drm/amdgpu: prevent spurious warning
The default auto setting for kcq should not generate
a warning.
Fixes: a300de40f66b ("drm/amdgpu: introduce a new parameter to configure how many KCQ we want(v5)") Reviewed-by: Kent Russell <kent.russell@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 6 Oct 2020 13:20:47 +0000 (09:20 -0400)]
drm/amdgpu/swsmu: fix ARC build errors
We want to use the dev_* functions here rather than the pr_* variants.
Switch to using dev_warn() which mirrors what we do on other asics.
Fixes the following build errors on ARC:
../drivers/gpu/drm/amd/amdgpu/../powerplay/navi10_ppt.c: In function 'navi10_fill_i2c_req':
../arch/arc/include/asm/bug.h:24:2: error: implicit declaration of function 'pr_warn'; did you mean 'drm_warn'? [-Werror=implicit-function-declaration]
../drivers/gpu/drm/amd/amdgpu/../powerplay/sienna_cichlid_ppt.c: In function 'sienna_cichlid_fill_i2c_req':
../arch/arc/include/asm/bug.h:24:2: error: implicit declaration of function 'pr_warn'; did you mean 'drm_warn'? [-Werror=implicit-function-declaration]
Reported-by: kernel test robot <lkp@intel.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: Evan Quan <evan.quan@amd.com> Cc: Vineet Gupta <vgupta@synopsys.com> Cc: linux-snps-arc@lists.infradead.org Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jonathan Kim [Wed, 2 Sep 2020 03:59:21 +0000 (23:59 -0400)]
drm/amdgpu: add xgmi perfmons for arcturus
Add xgmi perfmons for Arcturus.
v4: Align with patch 2 with struct improvements
v3: Align with patch 2 streamlining perf types versus event config types.
v2: Resend for re-review with alignment for v3 in patch 2.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Tested-by: Chris Freehill <chris.freehill@amd.com> Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jonathan Kim [Wed, 2 Sep 2020 00:21:00 +0000 (20:21 -0400)]
drm/amdgpu: add per device user friendly xgmi events for vega20
Non-outbound data metrics are non useful so mark them as legacy.
Bucket new perf counters into device and not device ip.
Bind events to chip instead of IP.
Report available event counters and not number of hw counter banks.
Move DF public macros to private since not needed outside of IP version.
v5: cleanup by moving per chip configs into structs
v4: After more discussion, replace *_LEGACY references with IP references
to indicate concept of pmu-typed versus event-config-typed event
registration.
v3: attr groups const array is global but attr groups are allocated per
device which doesn't work and causes problems on memory allocation and
de-allocation for pmu unregister. Switch to building const attr groups
per pmu instead to simplify solution.
v2: add comments on sysfs structure and formatting.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jonathan Kim [Thu, 18 Jun 2020 14:40:14 +0000 (10:40 -0400)]
drm/amdgpu: fix xgmi perfmon a-b-a problem
Mapping hw counters per event config will cause ABA problems so map per
event instead.
v2: Discontinue starting perf counters if add fails. Make it clear what's
happening with pmc_start.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
projects / users / jedix / linux-maple.git / log