kernel/linux.git/drivers/gpu/drm/amd/include, branch v6.1.2

drm/amdgpu: Fix type of second parameter in odn_edit_dpm_table() callback

2022-12-31T12:33:01+00:00

[ Upstream commit e4d0ef752081e7aa6ffb7ccac11c499c732a2e05 ] With clang's kernel control flow integrity (kCFI, CONFIG_CFI_CLANG), indirect call targets are validated against the expected function pointer prototype to make sure the call target is valid to help mitigate ROP attacks. If they are not identical, there is a failure at run time, which manifests as either a kernel panic or thread getting killed. A proposed warning in clang aims to catch these at compile time, which reveals: drivers/gpu/drm/amd/amdgpu/../pm/swsmu/amdgpu_smu.c:3008:29: error: incompatible function pointer types initializing 'int (*)(void *, uint32_t, long *, uint32_t)' (aka 'int (*)(void *, unsigned int, long *, unsigned int)') with an expression of type 'int (void *, enum PP_OD_DPM_TABLE_COMMAND, long *, uint32_t)' (aka 'int (void *, enum PP_OD_DPM_TABLE_COMMAND, long *, unsigned int)') [-Werror,-Wincompatible-function-pointer-types-strict] .odn_edit_dpm_table = smu_od_edit_dpm_table, ^~~~~~~~~~~~~~~~~~~~~ 1 error generated. There are only two implementations of ->odn_edit_dpm_table() in 'struct amd_pm_funcs': smu_od_edit_dpm_table() and pp_odn_edit_dpm_table(). One has a second parameter type of 'enum PP_OD_DPM_TABLE_COMMAND' and the other uses 'u32'. Ultimately, smu_od_edit_dpm_table() calls ->od_edit_dpm_table() from 'struct pptable_funcs' and pp_odn_edit_dpm_table() calls ->odn_edit_dpm_table() from 'struct pp_hwmgr_func', which both have a second parameter type of 'enum PP_OD_DPM_TABLE_COMMAND'. Update the type parameter in both the prototype in 'struct amd_pm_funcs' and pp_odn_edit_dpm_table() to 'enum PP_OD_DPM_TABLE_COMMAND', which cleans up the warning. Link: https://github.com/ClangBuiltLinux/linux/issues/1750 Reported-by: Sami Tolvanen Reviewed-by: Kees Cook Signed-off-by: Nathan Chancellor Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin

drm/amdkfd: Fix type of reset_type parameter in hqd_destroy() callback

2022-10-19T02:13:12+00:00

When booting a kernel compiled with CONFIG_CFI_CLANG on a machine with an RX 6700 XT, there is a CFI failure in kfd_destroy_mqd_cp(): [ 12.894543] CFI failure at kfd_destroy_mqd_cp+0x2a/0x40 [amdgpu] (target: hqd_destroy_v10_3+0x0/0x260 [amdgpu]; expected type: 0x8594d794) Clang's kernel Control Flow Integrity (kCFI) makes sure that all indirect call targets have a type that exactly matches the function pointer prototype. In this case, hqd_destroy()'s third parameter, reset_type, should have a type of 'uint32_t' but every implementation of this callback has a third parameter type of 'enum kfd_preempt_type'. Update the function pointer prototype to match reality so that there is no more CFI violation. Link: https://github.com/ClangBuiltLinux/linux/issues/1738 Signed-off-by: Nathan Chancellor Signed-off-by: Alex Deucher

drm/amdgpu: Update umc v8_10_0 headers

2022-10-11T15:05:35+00:00

Add GeccCtrl offset and mask to umc v8_10_0 headers. Signed-off-by: Candice Li Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: pass queue size and is_aql_queue to MES

2022-09-29T13:41:44+00:00

Update mes_v11_api_def.h add_queue API with is_aql_queue parameter. Also re-use gds_size for the queue size (unused for KFD). MES requires the queue size in order to compute the actual wptr offset within the queue RB since it increases monotonically for AQL queues. v2: Make is_aql_queue assign clearer Signed-off-by: Graham Sider Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher

drm/amd/amdgpu: Add missing XGMI hive registers for mmhub 9.4.1

2022-09-29T13:41:43+00:00

These are used by umr to sort the hive nodes since the kernel initializes the nodes in order of bus enumeration not XGMI hive enumeration. Signed-off-by: Tom St Denis Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amd/pm:add new gpu_metrics_v2_3 to acquire average temperature info

2022-09-14T16:38:41+00:00

Add new gpu_metrics_v2_3 to acquire average temperature info from SMU metrics. To acquire average temp info from gpu_metrics interface, but gpu_metrics_v2_2 only has members to show current temp info. --- v1: Only add average_temperature_gfx in gpu_metrics_v2_3. v2: Add average temp members for soc, core and l3 in gpu_metrics_v2_3 and put these new members at the end of gpu_metrics_v2_3. Add operation to read average temp info from metrics table. v3: Merge v1 and v2 and rename the patch. v4: Merge v3. Add firmware version judgment in vangogh_common_get_gpu_metrics to maintain backward compatibility and rename the patch. "return ret" on error scenario in smu_cmn_get_smc_version. Signed-off-by: Li Ma Reviewed-by: Evan Quan Signed-off-by: Alex Deucher

drm/amd/amdgpu: update GC 10.3.0 pwrdec

2022-09-13T16:54:23+00:00

The 10.3 GC headers were missing most of the pwrdec block. This patch adds the registers and bits present in the 10.1 header but based on the contents of the 10.3 specs. Signed-off-by: Tom St Denis Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amd/amdgpu: Add missing CGTS*TCC_DISABLE to 10.3 headers

2022-09-08T02:28:42+00:00

The TCC_DISABLE registers were not included in the 10.3 headers and instead just placed directly in the gfx_v10_0.c source. This patch adds them to the headers so tools like umr can scan them and support them. Signed-off-by: Tom St Denis Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher

drm/amdgpu: add gc v11_0_3 ip headers

2022-08-30T20:36:42+00:00

Add gc v11_0_3 register offset and shift masks header files v2: update registers (Alex) Signed-off-by: Frank Min Signed-off-by: Hawking Zhang Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher

drm/amdgpu: Update mes_v11_api_def.h

2022-08-29T21:43:50+00:00

New GFX11 MES FW adds the trap_en bit. For now hardcode to 1 (traps enabled). Signed-off-by: Graham Sider Acked-by: Felix Kuehling Reviewed-by: Harish Kasiviswanathan Signed-off-by: Alex Deucher