From 80a0e828293389358f7db56adcdcb22b28df5e11 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 11 Mar 2025 18:00:57 -0400 Subject: [PATCH 01/16] drm/amdgpu/discovery: optionally use fw based ip discovery On chips without native IP discovery support, use the fw binary if available, otherwise we can continue without it. Signed-off-by: Alex Deucher Reviewed-by: Flora Cui Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 40 +++++++++++++++---- 1 file changed, 32 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index 4125d2da7041..dc2713ec95a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -2545,6 +2545,38 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) { int r; + switch (adev->asic_type) { + case CHIP_VEGA10: + case CHIP_VEGA12: + case CHIP_RAVEN: + case CHIP_VEGA20: + case CHIP_ARCTURUS: + case CHIP_ALDEBARAN: + /* this is not fatal. We have a fallback below + * if the new firmwares are not present. some of + * this will be overridden below to keep things + * consistent with the current behavior. + */ + r = amdgpu_discovery_reg_base_init(adev); + if (!r) { + amdgpu_discovery_harvest_ip(adev); + amdgpu_discovery_get_gfx_info(adev); + amdgpu_discovery_get_mall_info(adev); + amdgpu_discovery_get_vcn_info(adev); + } + break; + default: + r = amdgpu_discovery_reg_base_init(adev); + if (r) + return -EINVAL; + + amdgpu_discovery_harvest_ip(adev); + amdgpu_discovery_get_gfx_info(adev); + amdgpu_discovery_get_mall_info(adev); + amdgpu_discovery_get_vcn_info(adev); + break; + } + switch (adev->asic_type) { case CHIP_VEGA10: vega10_reg_base_init(adev); @@ -2709,14 +2741,6 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 1, 0); break; default: - r = amdgpu_discovery_reg_base_init(adev); - if (r) - return -EINVAL; - - amdgpu_discovery_harvest_ip(adev); - amdgpu_discovery_get_gfx_info(adev); - amdgpu_discovery_get_mall_info(adev); - amdgpu_discovery_get_vcn_info(adev); break; } -- 2.51.0 From af23d3c9caabc6269dddaed4a8de632484951fda Mon Sep 17 00:00:00 2001 From: Srinivasan Shanmugam Date: Fri, 21 Mar 2025 06:53:47 +0530 Subject: [PATCH 02/16] drm/amdgpu: Add parameter documentation for amdgpu_sync_fence MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit The 'flags' parameter, which specifies memory allocation behavior while creating a sync entry, Fixes the below with gcc W=1: drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c:162: warning: Function parameter or struct member 'flags' not described in 'amdgpu_sync_fence' Cc: Christian König Cc: Alex Deucher Signed-off-by: Srinivasan Shanmugam Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index c5f9db6b32a4..5576ed0b508f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -154,6 +154,7 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f) * * @sync: sync object to add fence to * @f: fence to sync to + * @flags: memory allocation flags to use when allocating sync entry * * Add the fence to the sync object. */ -- 2.51.0 From 7547510d4a915f4f6d9b1262182d8db6763508f4 Mon Sep 17 00:00:00 2001 From: Asad Kamal Date: Mon, 17 Mar 2025 20:07:06 +0800 Subject: [PATCH 03/16] drm/amd/pm: Update feature list for smu_v13_0_6 Update feature list for smu_v13_0_6 to show vcn & smu deep sleep feature enable status Signed-off-by: Asad Kamal Reviewed-by: Yang Wang Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c index f2837dadd897..682646068000 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c @@ -231,7 +231,11 @@ static const struct cmn2asic_mapping smu_v13_0_6_feature_mask_map[SMU_FEATURE_CO SMU_13_0_6_FEA_MAP(SMU_FEATURE_FW_CTF_BIT, FEATURE_FW_CTF), SMU_13_0_6_FEA_MAP(SMU_FEATURE_THERMAL_BIT, FEATURE_THERMAL), SMU_13_0_6_FEA_MAP(SMU_FEATURE_XGMI_PER_LINK_PWR_DWN_BIT, FEATURE_XGMI_PER_LINK_PWR_DOWN), - SMU_13_0_6_FEA_MAP(SMU_FEATURE_DF_CSTATE_BIT, FEATURE_DF_CSTATE), + SMU_13_0_6_FEA_MAP(SMU_FEATURE_DF_CSTATE_BIT, FEATURE_DF_CSTATE), + SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_VCN_BIT, FEATURE_DS_VCN), + SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MP1CLK_BIT, FEATURE_DS_MP1CLK), + SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MPIOCLK_BIT, FEATURE_DS_MPIOCLK), + SMU_13_0_6_FEA_MAP(SMU_FEATURE_DS_MP0CLK_BIT, FEATURE_DS_MP0CLK), }; #define TABLE_PMSTATUSLOG 0 -- 2.51.0 From f23e9116ebb71b63fe9cec0dcac792aa9af30b0c Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Fri, 21 Mar 2025 13:52:31 +0300 Subject: [PATCH 04/16] drm/amd/pm: Prevent division by zero The user can set any speed value. If speed is greater than UINT_MAX/8, division by zero is possible. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: c05d1c401572 ("drm/amd/swsmu: add aldebaran smu13 ip support (v3)") Signed-off-by: Denis Arefev Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index 0915d6377613..ba5a9012dbd5 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -1226,7 +1226,7 @@ int smu_v13_0_set_fan_speed_rpm(struct smu_context *smu, uint32_t tach_period; int ret; - if (!speed) + if (!speed || speed > UINT_MAX/8) return -EINVAL; ret = smu_v13_0_auto_fan_control(smu, 0); -- 2.51.0 From 7d641c2b83275d3b0424127b2e0d2d0f7dd82aef Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Fri, 21 Mar 2025 13:52:32 +0300 Subject: [PATCH 05/16] drm/amd/pm: Prevent division by zero The user can set any speed value. If speed is greater than UINT_MAX/8, division by zero is possible. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: b64625a303de ("drm/amd/pm: correct the address of Arcturus fan related registers") Signed-off-by: Denis Arefev Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c index 8aa61a9f7778..453952cdc353 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c @@ -1267,6 +1267,9 @@ static int arcturus_set_fan_speed_rpm(struct smu_context *smu, uint32_t crystal_clock_freq = 2500; uint32_t tach_period; + if (!speed || speed > UINT_MAX/8) + return -EINVAL; + tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed); WREG32_SOC15(THM, 0, mmCG_TACH_CTRL_ARCT, REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL_ARCT), -- 2.51.0 From 4e3d9508c056d7e0a56b58d5c81253e2a0d22b6c Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Fri, 21 Mar 2025 13:52:33 +0300 Subject: [PATCH 06/16] drm/amd/pm: Prevent division by zero The user can set any speed value. If speed is greater than UINT_MAX/8, division by zero is possible. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: 031db09017da ("drm/amd/powerplay/vega20: enable fan RPM and pwm settings V2") Signed-off-by: Denis Arefev Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_thermal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_thermal.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_thermal.c index a3331ffb2daf..1b1c88590156 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_thermal.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega20_thermal.c @@ -191,7 +191,7 @@ int vega20_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) uint32_t tach_period, crystal_clock_freq; int result = 0; - if (!speed) + if (!speed || speed > UINT_MAX/8) return -EINVAL; if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) { -- 2.51.0 From 4b8c3c0d17c07f301011e2908fecd2ebdcfe3d1c Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Fri, 21 Mar 2025 14:08:15 +0300 Subject: [PATCH 07/16] drm/amd/pm: Prevent division by zero The user can set any speed value. If speed is greater than UINT_MAX/8, division by zero is possible. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: c52dcf49195d ("drm/amd/pp: Avoid divide-by-zero in fan_ctrl_set_fan_speed_rpm") Signed-off-by: Denis Arefev Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_thermal.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_thermal.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_thermal.c index 379012494da5..56423aedf3fa 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_thermal.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_thermal.c @@ -307,10 +307,10 @@ int vega10_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) int result = 0; if (hwmgr->thermal_controller.fanInfo.bNoFan || - speed == 0 || + (!speed || speed > UINT_MAX/8) || (speed < hwmgr->thermal_controller.fanInfo.ulMinRPM) || (speed > hwmgr->thermal_controller.fanInfo.ulMaxRPM)) - return -1; + return -EINVAL; if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) result = vega10_fan_ctrl_stop_smc_fan_control(hwmgr); -- 2.51.0 From 7c246a05df51c52fe0852ce56ba10c41e6ed1f39 Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Fri, 21 Mar 2025 14:08:16 +0300 Subject: [PATCH 08/16] drm/amd/pm: Prevent division by zero The user can set any speed value. If speed is greater than UINT_MAX/8, division by zero is possible. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: c52dcf49195d ("drm/amd/pp: Avoid divide-by-zero in fan_ctrl_set_fan_speed_rpm") Signed-off-by: Denis Arefev Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_thermal.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_thermal.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_thermal.c index a8fc0fa44db6..ba5c1237fcfe 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_thermal.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_thermal.c @@ -267,10 +267,10 @@ int smu7_fan_ctrl_set_fan_speed_rpm(struct pp_hwmgr *hwmgr, uint32_t speed) if (hwmgr->thermal_controller.fanInfo.bNoFan || (hwmgr->thermal_controller.fanInfo. ucTachometerPulsesPerRevolution == 0) || - speed == 0 || + (!speed || speed > UINT_MAX/8) || (speed < hwmgr->thermal_controller.fanInfo.ulMinRPM) || (speed > hwmgr->thermal_controller.fanInfo.ulMaxRPM)) - return 0; + return -EINVAL; if (PP_CAP(PHM_PlatformCaps_MicrocodeFanControl)) smu7_fan_ctrl_stop_smc_fan_control(hwmgr); -- 2.51.0 From b03f1810db7bf609a64b90704f11da46e3baa050 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Tomasz=20Paku=C5=82a?= Date: Tue, 25 Mar 2025 11:26:52 +0100 Subject: [PATCH 09/16] drm/amd/pm: Add zero RPM enabled OD setting support for SMU14.0.2 MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Hook up zero RPM enable for 9070 and 9070 XT based on RDNA3 (smu 13.0.0 and 13.0.7) code. Tested on 9070 XT Hellhound Signed-off-by: Tomasz Pakuła Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.12.x Signed-off-by: Alex Deucher --- .../drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c | 55 ++++++++++++++++++- 1 file changed, 54 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c index f7cfe1f35cae..82c2db972491 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c @@ -79,6 +79,7 @@ #define PP_OD_FEATURE_FAN_ACOUSTIC_TARGET 8 #define PP_OD_FEATURE_FAN_TARGET_TEMPERATURE 9 #define PP_OD_FEATURE_FAN_MINIMUM_PWM 10 +#define PP_OD_FEATURE_FAN_ZERO_RPM_ENABLE 11 static struct cmn2asic_msg_mapping smu_v14_0_2_message_map[SMU_MSG_MAX_COUNT] = { MSG_MAP(TestMessage, PPSMC_MSG_TestMessage, 1), @@ -1052,6 +1053,10 @@ static void smu_v14_0_2_get_od_setting_limits(struct smu_context *smu, od_min_setting = overdrive_lowerlimits->FanMinimumPwm; od_max_setting = overdrive_upperlimits->FanMinimumPwm; break; + case PP_OD_FEATURE_FAN_ZERO_RPM_ENABLE: + od_min_setting = overdrive_lowerlimits->FanZeroRpmEnable; + od_max_setting = overdrive_upperlimits->FanZeroRpmEnable; + break; default: od_min_setting = od_max_setting = INT_MAX; break; @@ -1330,6 +1335,24 @@ static int smu_v14_0_2_print_clk_levels(struct smu_context *smu, min_value, max_value); break; + case SMU_OD_FAN_ZERO_RPM_ENABLE: + if (!smu_v14_0_2_is_od_feature_supported(smu, + PP_OD_FEATURE_ZERO_FAN_BIT)) + break; + + size += sysfs_emit_at(buf, size, "FAN_ZERO_RPM_ENABLE:\n"); + size += sysfs_emit_at(buf, size, "%d\n", + (int)od_table->OverDriveTable.FanZeroRpmEnable); + + size += sysfs_emit_at(buf, size, "%s:\n", "OD_RANGE"); + smu_v14_0_2_get_od_setting_limits(smu, + PP_OD_FEATURE_FAN_ZERO_RPM_ENABLE, + &min_value, + &max_value); + size += sysfs_emit_at(buf, size, "ZERO_RPM_ENABLE: %u %u\n", + min_value, max_value); + break; + case SMU_OD_RANGE: if (!smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_GFXCLK_BIT) && !smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_UCLK_BIT) && @@ -2270,7 +2293,9 @@ static void smu_v14_0_2_set_supported_od_feature_mask(struct smu_context *smu) OD_OPS_SUPPORT_FAN_TARGET_TEMPERATURE_RETRIEVE | OD_OPS_SUPPORT_FAN_TARGET_TEMPERATURE_SET | OD_OPS_SUPPORT_FAN_MINIMUM_PWM_RETRIEVE | - OD_OPS_SUPPORT_FAN_MINIMUM_PWM_SET; + OD_OPS_SUPPORT_FAN_MINIMUM_PWM_SET | + OD_OPS_SUPPORT_FAN_ZERO_RPM_ENABLE_RETRIEVE | + OD_OPS_SUPPORT_FAN_ZERO_RPM_ENABLE_SET; } static int smu_v14_0_2_get_overdrive_table(struct smu_context *smu, @@ -2349,6 +2374,8 @@ static int smu_v14_0_2_set_default_od_settings(struct smu_context *smu) user_od_table_bak.OverDriveTable.FanTargetTemperature; user_od_table->OverDriveTable.FanMinimumPwm = user_od_table_bak.OverDriveTable.FanMinimumPwm; + user_od_table->OverDriveTable.FanZeroRpmEnable = + user_od_table_bak.OverDriveTable.FanZeroRpmEnable; } smu_v14_0_2_set_supported_od_feature_mask(smu); @@ -2396,6 +2423,11 @@ static int smu_v14_0_2_od_restore_table_single(struct smu_context *smu, long inp od_table->OverDriveTable.FanMode = FAN_MODE_AUTO; od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT); break; + case PP_OD_EDIT_FAN_ZERO_RPM_ENABLE: + od_table->OverDriveTable.FanZeroRpmEnable = + boot_overdrive_table->OverDriveTable.FanZeroRpmEnable; + od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_ZERO_FAN_BIT); + break; case PP_OD_EDIT_ACOUSTIC_LIMIT: od_table->OverDriveTable.AcousticLimitRpmThreshold = boot_overdrive_table->OverDriveTable.AcousticLimitRpmThreshold; @@ -2678,6 +2710,27 @@ static int smu_v14_0_2_od_edit_dpm_table(struct smu_context *smu, od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_FAN_CURVE_BIT); break; + case PP_OD_EDIT_FAN_ZERO_RPM_ENABLE: + if (!smu_v14_0_2_is_od_feature_supported(smu, PP_OD_FEATURE_ZERO_FAN_BIT)) { + dev_warn(adev->dev, "Zero RPM setting not supported!\n"); + return -ENOTSUPP; + } + + smu_v14_0_2_get_od_setting_limits(smu, + PP_OD_FEATURE_FAN_ZERO_RPM_ENABLE, + &minimum, + &maximum); + if (input[0] < minimum || + input[0] > maximum) { + dev_info(adev->dev, "zero RPM enable setting(%ld) must be within [%d, %d]!\n", + input[0], minimum, maximum); + return -EINVAL; + } + + od_table->OverDriveTable.FanZeroRpmEnable = input[0]; + od_table->OverDriveTable.FeatureCtrlMask |= BIT(PP_OD_FEATURE_ZERO_FAN_BIT); + break; + case PP_OD_RESTORE_DEFAULT_TABLE: if (size == 1) { ret = smu_v14_0_2_od_restore_table_single(smu, input[0]); -- 2.51.0 From 5f054ddead33c1622ea9c0c0aaf07c6843fc7ab0 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Wed, 19 Mar 2025 11:58:31 -0500 Subject: [PATCH 10/16] drm/amd: Handle being compiled without SI or CIK support better If compiled without SI or CIK support but amdgpu tries to load it will run into failures with uninitialized callbacks. Show a nicer message in this case and fail probe instead. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4050 Signed-off-by: Mario Limonciello Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 44 ++++++++++++++----------- 1 file changed, 24 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 23cfce5aa1fc..26bf896f1444 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -1809,7 +1809,6 @@ static const u16 amdgpu_unsupported_pciidlist[] = { }; static const struct pci_device_id pciidlist[] = { -#ifdef CONFIG_DRM_AMDGPU_SI {0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, {0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, {0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI}, @@ -1882,8 +1881,6 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x6665, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY}, {0x1002, 0x6667, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY}, {0x1002, 0x666F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY}, -#endif -#ifdef CONFIG_DRM_AMDGPU_CIK /* Kaveri */ {0x1002, 0x1304, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_MOBILITY|AMD_IS_APU}, {0x1002, 0x1305, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_APU}, @@ -1966,7 +1963,6 @@ static const struct pci_device_id pciidlist[] = { {0x1002, 0x985D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU}, {0x1002, 0x985E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU}, {0x1002, 0x985F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MULLINS|AMD_IS_MOBILITY|AMD_IS_APU}, -#endif /* topaz */ {0x1002, 0x6900, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ}, {0x1002, 0x6901, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TOPAZ}, @@ -2313,14 +2309,14 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, return -ENOTSUPP; } + switch (flags & AMD_ASIC_MASK) { + case CHIP_TAHITI: + case CHIP_PITCAIRN: + case CHIP_VERDE: + case CHIP_OLAND: + case CHIP_HAINAN: #ifdef CONFIG_DRM_AMDGPU_SI - if (!amdgpu_si_support) { - switch (flags & AMD_ASIC_MASK) { - case CHIP_TAHITI: - case CHIP_PITCAIRN: - case CHIP_VERDE: - case CHIP_OLAND: - case CHIP_HAINAN: + if (!amdgpu_si_support) { dev_info(&pdev->dev, "SI support provided by radeon.\n"); dev_info(&pdev->dev, @@ -2328,16 +2324,18 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, ); return -ENODEV; } - } + break; +#else + dev_info(&pdev->dev, "amdgpu is built without SI support.\n"); + return -ENODEV; #endif + case CHIP_KAVERI: + case CHIP_BONAIRE: + case CHIP_HAWAII: + case CHIP_KABINI: + case CHIP_MULLINS: #ifdef CONFIG_DRM_AMDGPU_CIK - if (!amdgpu_cik_support) { - switch (flags & AMD_ASIC_MASK) { - case CHIP_KAVERI: - case CHIP_BONAIRE: - case CHIP_HAWAII: - case CHIP_KABINI: - case CHIP_MULLINS: + if (!amdgpu_cik_support) { dev_info(&pdev->dev, "CIK support provided by radeon.\n"); dev_info(&pdev->dev, @@ -2345,8 +2343,14 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, ); return -ENODEV; } - } + break; +#else + dev_info(&pdev->dev, "amdgpu is built without CIK support.\n"); + return -ENODEV; #endif + default: + break; + } adev = devm_drm_dev_alloc(&pdev->dev, &amdgpu_kms_driver, typeof(*adev), ddev); if (IS_ERR(adev)) -- 2.51.0 From f21e6d149b49c92f9e68aa0c76033e1e13d9f5da Mon Sep 17 00:00:00 2001 From: Leo Li Date: Mon, 10 Mar 2025 12:20:39 -0400 Subject: [PATCH 11/16] drm/amd/display: Increase vblank offdelay for PSR panels [Why] Depending on when the HW latching event (vupdate) of double-buffered registers happen relative to the PSR SDP (signals panel psr enter/exit) deadline, and how bad the Panel clock has drifted since the last ALPM off event, there can be up to 3 frames of delay between sending the PSR exit cmd to DMUB fw, and when the panel starts displaying live frames. This can manifest as micro-stuttering when userspace commit patterns cause rapid toggling of the DRM vblank counter, since PSR enter/exit is hooked up to DRM vblank disable/enable respectively. In the ideal world, the panel should present the live frame immediately on PSR exit cmd. But due to HW design and PSR limitations, immediate exit can only happen by chance, when: 1. PSR exit cmd is ack'd by FW before HW latching (vupdate) event, and 2. Panel's SDP deadline -- determined by it's PSR Start Delay in DPCD 71h -- is after the vupdate event. The PSR exit SDP can then be sent immediately after HW latches. Otherwise, we have to wait 1 frame. And 3. There is negligible drift between the panel's clock and source clock. Otherwise, there can be up to 1 frame of drift. Note that this delay is not expected with Panel Replay. [How] Since PSR power savings can be quite substantial, and there are a lot of systems in the wild with PSR panels, It'll be nice to have a middle ground that balances user experience with power savings. A simple way to achieve this is by extending the vblank offdelay, such that additional PSR exit delays will be less perceivable. We can set: 20/100 * offdelay_ms = 3_frames_ms => offdelay_ms = 5 * 3_frames_ms This ensures that `3_frames_ms` will only be experienced as a 20% delay on top how long the panel has been static, and thus make the delay less perceivable. If this ends up being too high of a percentage, it can be dropped further in a future change. Fixes: 537ef0f88897 ("drm/amd/display: use new vblank enable policy for DCN35+") Reviewed-by: Harry Wentland Signed-off-by: Leo Li Signed-off-by: Aurabindo Pillai Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 41 +++++++++++++++---- 1 file changed, 33 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index bae83a129b5f..c75fcf726a44 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -8707,14 +8707,39 @@ static void manage_dm_interrupts(struct amdgpu_device *adev, int offdelay; if (acrtc_state) { - if (amdgpu_ip_version(adev, DCE_HWIP, 0) < - IP_VERSION(3, 5, 0) || - acrtc_state->stream->link->psr_settings.psr_version < - DC_PSR_VERSION_UNSUPPORTED || - !(adev->flags & AMD_IS_APU)) { - timing = &acrtc_state->stream->timing; - - /* at least 2 frames */ + timing = &acrtc_state->stream->timing; + + /* + * Depending on when the HW latching event of double-buffered + * registers happen relative to the PSR SDP deadline, and how + * bad the Panel clock has drifted since the last ALPM off + * event, there can be up to 3 frames of delay between sending + * the PSR exit cmd to DMUB fw, and when the panel starts + * displaying live frames. + * + * We can set: + * + * 20/100 * offdelay_ms = 3_frames_ms + * => offdelay_ms = 5 * 3_frames_ms + * + * This ensures that `3_frames_ms` will only be experienced as a + * 20% delay on top how long the display has been static, and + * thus make the delay less perceivable. + */ + if (acrtc_state->stream->link->psr_settings.psr_version < + DC_PSR_VERSION_UNSUPPORTED) { + offdelay = DIV64_U64_ROUND_UP((u64)5 * 3 * 10 * + timing->v_total * + timing->h_total, + timing->pix_clk_100hz); + config.offdelay_ms = offdelay ?: 30; + } else if (amdgpu_ip_version(adev, DCE_HWIP, 0) < + IP_VERSION(3, 5, 0) || + !(adev->flags & AMD_IS_APU)) { + /* + * Older HW and DGPU have issues with instant off; + * use a 2 frame offdelay. + */ offdelay = DIV64_U64_ROUND_UP((u64)20 * timing->v_total * timing->h_total, -- 2.51.0 From 0389f2a3a23c9a2d4cf58d2ab0d3b283f19e8630 Mon Sep 17 00:00:00 2001 From: Charlene Liu Date: Thu, 13 Mar 2025 20:07:13 -0400 Subject: [PATCH 12/16] Revert "drm/amd/display: dml2 soc dscclk use DPM table clk setting" [why] this dscclk use DCN defined per DPM level will cause a DCFCLK increase. needs to follow up. This reverts commit 15b959534a39530a21d378190557cc8d1eab7b09 Reviewed-by: Yihan Zhu Reviewed-by: Alvin Lee Signed-off-by: Charlene Liu Signed-off-by: Aurabindo Pillai Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c index 70c39df62533..2061d43b92e1 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c @@ -590,11 +590,11 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc, p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz; p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz; + p->out_states->state_array[i].dscclk_mhz = max_dispclk_mhz / 3.0; p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz; p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz; /* Dependent states. */ - p->out_states->state_array[i].dscclk_mhz = p->in_states->state_array[i].dscclk_mhz; p->out_states->state_array[i].dram_speed_mts = p->in_states->state_array[i].dram_speed_mts; p->out_states->state_array[i].fabricclk_mhz = p->in_states->state_array[i].fabricclk_mhz; p->out_states->state_array[i].socclk_mhz = p->in_states->state_array[i].socclk_mhz; -- 2.51.0 From 8058061ed9d6bc259d1e678607b07d259342c08f Mon Sep 17 00:00:00 2001 From: Brendan Tam Date: Fri, 14 Mar 2025 13:09:13 -0400 Subject: [PATCH 13/16] drm/amd/display: prevent hang on link training fail [Why] When link training fails, the phy clock will be disabled. However, in enable_streams, it is assumed that link training succeeded and the mux selects the phy clock, causing a hang when a register write is made. [How] When enable_stream is hit, check if link training failed. If it did, fall back to the ref clock to avoid a hang and keep the system in a recoverable state. Reviewed-by: Dillon Varone Signed-off-by: Brendan Tam Signed-off-by: Aurabindo Pillai Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c | 6 +++++- drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 7 +++++-- 2 files changed, 10 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c index 926c08e790c1..846c9c51f2d9 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c @@ -3033,7 +3033,11 @@ void dcn20_enable_stream(struct pipe_ctx *pipe_ctx) dccg->funcs->set_dpstreamclk(dccg, DTBCLK0, tg->inst, dp_hpo_inst); phyd32clk = get_phyd32clk_src(link); - dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); + if (link->cur_link_settings.link_rate == LINK_RATE_UNKNOWN) { + dccg->funcs->disable_symclk32_se(dccg, dp_hpo_inst); + } else { + dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); + } } else { if (dccg->funcs->enable_symclk_se) dccg->funcs->enable_symclk_se(dccg, stream_enc->stream_enc_inst, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c index 8f5da0ded850..5489f3d431f6 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c @@ -936,8 +936,11 @@ void dcn401_enable_stream(struct pipe_ctx *pipe_ctx) if (dc_is_dp_signal(pipe_ctx->stream->signal) || dc_is_virtual_signal(pipe_ctx->stream->signal)) { if (dc->link_srv->dp_is_128b_132b_signal(pipe_ctx)) { dccg->funcs->set_dpstreamclk(dccg, DPREFCLK, tg->inst, dp_hpo_inst); - - dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); + if (link->cur_link_settings.link_rate == LINK_RATE_UNKNOWN) { + dccg->funcs->disable_symclk32_se(dccg, dp_hpo_inst); + } else { + dccg->funcs->enable_symclk32_se(dccg, dp_hpo_inst, phyd32clk); + } } else { dccg->funcs->enable_symclk_se(dccg, stream_enc->stream_enc_inst, link_enc->transmitter - TRANSMITTER_UNIPHY_A); -- 2.51.0 From 704bc361e3a4ead1c0eb40acc255b636b788dc89 Mon Sep 17 00:00:00 2001 From: Leo Li Date: Tue, 11 Mar 2025 09:43:38 -0400 Subject: [PATCH 14/16] drm/amd/display: Actually do immediate vblank disable [Why] The `vblank_config.offdelay` field follows the same semantics as the `drm_vblank_offdelay` parameter. Setting it to 0 will never disable vblank. [How] Set `offdelay` to a positive number. Fixes: e45b6716de4b ("drm/amd/display: use a more lax vblank enable policy for DCN35+") Reviewed-by: Harry Wentland Signed-off-by: Leo Li Signed-off-by: Aurabindo Pillai Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index c75fcf726a44..d0d8ad5368c3 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -8747,6 +8747,8 @@ static void manage_dm_interrupts(struct amdgpu_device *adev, config.offdelay_ms = offdelay ?: 30; } else { + /* offdelay_ms = 0 will never disable vblank */ + config.offdelay_ms = 1; config.disable_immediate = true; } -- 2.51.0 From f3f05a0ec584855c53f2d95024e23259f3ee101d Mon Sep 17 00:00:00 2001 From: Xiang Liu Date: Fri, 21 Mar 2025 20:47:23 +0800 Subject: [PATCH 15/16] drm/amdgpu: Use correct gfx deferred error count In the case of parsing GFX deferred error from SMU corrected error channel, the error count should be set to 1 instead of parsing from MISC0 register, which is 0. Signed-off-by: Xiang Liu Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 736398b0d16d..3caac4a1564a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -872,9 +872,10 @@ static int gfx_v9_4_3_aca_bank_parser(struct aca_handle *handle, break; case ACA_SMU_TYPE_CE: bank->aca_err_type = ACA_BANK_ERR_CE_DE_DECODE(bank); - ret = aca_error_cache_log_bank_error(handle, &info, - bank->aca_err_type, - ACA_REG__MISC0__ERRCNT(misc0)); + ret = aca_error_cache_log_bank_error(handle, &info, bank->aca_err_type, + (bank->aca_err_type == ACA_ERROR_TYPE_CE) ? + ACA_REG__MISC0__ERRCNT(misc0) : + 1); break; default: return -EINVAL; -- 2.51.0 From ee97326fb96505f20b17f38f54df731cd6496bc0 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Mon, 24 Mar 2025 13:26:26 +0530 Subject: [PATCH 16/16] drm/amdgpu: Add NPS2 to DPX compatible mode Compute partition DPX is possible in NPS2 mode. Update the compatible modes for DPX. Signed-off-by: Lijo Lazar Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c index 3c07517be09a..ae071985f26e 100644 --- a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c +++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram.c @@ -473,7 +473,8 @@ static int aqua_vanjaram_get_xcp_res_info(struct amdgpu_xcp_mgr *xcp_mgr, break; case AMDGPU_DPX_PARTITION_MODE: num_xcp = 2; - nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE); + nps_modes = BIT(AMDGPU_NPS1_PARTITION_MODE) | + BIT(AMDGPU_NPS2_PARTITION_MODE); break; case AMDGPU_TPX_PARTITION_MODE: num_xcp = 3; -- 2.51.0