From 08fffa74d9772d9538338be3f304006c94dde6f0 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Thu, 27 Jul 2023 10:22:20 -0500 Subject: [PATCH 01/14] drm/amd: Disable S/G for APUs when 64GB or more host memory Users report a white flickering screen on multiple systems that is tied to having 64GB or more memory. When S/G is enabled pages will get pinned to both VRAM carve out and system RAM leading to this. Until it can be fixed properly, disable S/G when 64GB of memory or more is detected. This will force pages to be pinned into VRAM. This should fix white screen flickers but if VRAM pressure is encountered may lead to black screens. It's a trade-off for now. Fixes: 81d0bcf99009 ("drm/amdgpu: make display pinning more flexible (v2)") Cc: Hamza Mahfooz Cc: Roman Li Cc: # 6.1.y: bf0207e172703 ("drm/amdgpu: add S/G display parameter") Cc: # 6.4.y Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2735 Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2354 Signed-off-by: Mario Limonciello Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 26 +++++++++++++++++++ .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 5 ++-- 3 files changed, 29 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index a3b86b86dc47..6dc950c1b689 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1296,6 +1296,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, void amdgpu_device_pci_config_reset(struct amdgpu_device *adev); int amdgpu_device_pci_reset(struct amdgpu_device *adev); bool amdgpu_device_need_post(struct amdgpu_device *adev); +bool amdgpu_sg_display_supported(struct amdgpu_device *adev); bool amdgpu_device_pcie_dynamic_switching_supported(void); bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev); bool amdgpu_device_aspm_support_quirk(void); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index a2cdde0ca0a7..45e9d737e5b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -1458,6 +1458,32 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev) return true; } +/* + * On APUs with >= 64GB white flickering has been observed w/ SG enabled. + * Disable S/G on such systems until we have a proper fix. + * https://gitlab.freedesktop.org/drm/amd/-/issues/2354 + * https://gitlab.freedesktop.org/drm/amd/-/issues/2735 + */ +bool amdgpu_sg_display_supported(struct amdgpu_device *adev) +{ + switch (amdgpu_sg_display) { + case -1: + break; + case 0: + return false; + case 1: + return true; + default: + return false; + } + if ((totalram_pages() << (PAGE_SHIFT - 10)) + + (adev->gmc.real_vram_size / 1024) >= 64000000) { + DRM_WARN("Disabling S/G due to >=64GB RAM\n"); + return false; + } + return true; +} + /* * Intel hosts such as Raptor Lake and Sapphire Rapids don't support dynamic * speed switching. Until we have confirmation from Intel that a specific host 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 0fa739fd6a9c..e5554a36e8c8 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -1638,9 +1638,8 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) } break; } - if (init_data.flags.gpu_vm_support && - (amdgpu_sg_display == 0)) - init_data.flags.gpu_vm_support = false; + if (init_data.flags.gpu_vm_support) + init_data.flags.gpu_vm_support = amdgpu_sg_display_supported(adev); if (init_data.flags.gpu_vm_support) adev->mode_info.gpu_vm_support = true; From 730d44e1fa306a20746ad4a85da550662aed9daa Mon Sep 17 00:00:00 2001 From: Tim Huang Date: Thu, 27 Jul 2023 09:59:45 +0800 Subject: [PATCH 02/14] drm/amd/pm: skip the RLC stop when S0i3 suspend for SMU v13.0.4/11 For SMU v13.0.4/11, driver does not need to stop RLC for S0i3, the firmwares will handle that properly. Signed-off-by: Tim Huang Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index ce41a8309582..222af2fae745 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -1581,9 +1581,9 @@ static int smu_disable_dpms(struct smu_context *smu) /* * For SMU 13.0.4/11, PMFW will handle the features disablement properly - * for gpu reset case. Driver involvement is unnecessary. + * for gpu reset and S0i3 cases. Driver involvement is unnecessary. */ - if (amdgpu_in_reset(adev)) { + if (amdgpu_in_reset(adev) || adev->in_s0ix) { switch (adev->ip_versions[MP1_HWIP][0]) { case IP_VERSION(13, 0, 4): case IP_VERSION(13, 0, 11): From d3de41ee5febe5c2d9989fe9810bce2bb54a3a8e Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 25 Jul 2023 19:11:54 +0530 Subject: [PATCH 03/14] drm/amdgpu: Match against exact bootloader status On PSP v13.x ASICs, boot loader will set only the MSB to 1 and clear the least significant bits for any command submission. Hence match against the exact register value, otherwise a register value of all 0xFFs also could falsely indicate that boot loader is ready. Also, from PSP v13.0.6 and newer, bits[7:0] will be used to indicate command error status. Signed-off-by: Lijo Lazar Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/psp_v13_0.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c index e1a392bcea70..af5685f4cb34 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c @@ -137,14 +137,15 @@ static int psp_v13_0_wait_for_bootloader(struct psp_context *psp) int ret; int retry_loop; + /* Wait for bootloader to signify that it is ready having bit 31 of + * C2PMSG_35 set to 1. All other bits are expected to be cleared. + * If there is an error in processing command, bits[7:0] will be set. + * This is applicable for PSP v13.0.6 and newer. + */ for (retry_loop = 0; retry_loop < 10; retry_loop++) { - /* Wait for bootloader to signify that is - ready having bit 31 of C2PMSG_35 set to 1 */ - ret = psp_wait_for(psp, - SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35), - 0x80000000, - 0x80000000, - false); + ret = psp_wait_for( + psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35), + 0x80000000, 0xffffffff, false); if (ret == 0) return 0; From 7ad1dfc144cbf62702fd07838da8fd8a77921083 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Mon, 31 Jul 2023 09:22:05 -0500 Subject: [PATCH 04/14] drm/amd/display: Don't show stack trace for missing eDP Some systems are only connected by HDMI or DP, so warning related to missing eDP is unnecessary. Downgrade to debug instead. Cc: Hamza Mahfooz Fixes: 6d9b6dceaa51 ("drm/amd/display: only warn once in dce110_edp_wait_for_hpd_ready()") Reported-by: Mastan.Katragadda@amd.com Signed-off-by: Mario Limonciello Reviewed-by: Hamza Mahfooz Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c index 20d4d08a6a2f..6966420dfbac 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c @@ -777,7 +777,8 @@ void dce110_edp_wait_for_hpd_ready( dal_gpio_destroy_irq(&hpd); /* ensure that the panel is detected */ - ASSERT(edp_hpd_high); + if (!edp_hpd_high) + DC_LOG_DC("%s: wait timed out!\n", __func__); } void dce110_edp_power_control( From bd60e2eafd8fb053948b6e23e8167baf7a159750 Mon Sep 17 00:00:00 2001 From: Kenneth Feng Date: Thu, 27 Jul 2023 19:37:31 +0800 Subject: [PATCH 05/14] drm/amd/pm: correct the pcie width for smu 13.0.0 correct the pcie width value in pp_dpm_pcie for smu 13.0.0 Signed-off-by: Kenneth Feng Reviewed-by: Harish Kasiviswanathan Acked-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index 3d188616ba24..e191bbe9e994 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -1140,7 +1140,6 @@ static int smu_v13_0_0_print_clk_levels(struct smu_context *smu, (OverDriveTableExternal_t *)smu->smu_table.overdrive_table; struct smu_13_0_dpm_table *single_dpm_table; struct smu_13_0_pcie_table *pcie_table; - const int link_width[] = {0, 1, 2, 4, 8, 12, 16}; uint32_t gen_speed, lane_width; int i, curr_freq, size = 0; int32_t min_value, max_value; @@ -1256,7 +1255,7 @@ static int smu_v13_0_0_print_clk_levels(struct smu_context *smu, (pcie_table->pcie_lane[i] == 6) ? "x16" : "", pcie_table->clk_freq[i], (gen_speed == DECODE_GEN_SPEED(pcie_table->pcie_gen[i])) && - (lane_width == DECODE_LANE_WIDTH(link_width[pcie_table->pcie_lane[i]])) ? + (lane_width == DECODE_LANE_WIDTH(pcie_table->pcie_lane[i])) ? "*" : ""); break; From 61319b8e3b58a7167cf146313fd4523fe72586bc Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 21 Jul 2023 19:18:00 +0800 Subject: [PATCH 06/14] drm/amd/pm: disable the SMU13 OD feature support temporarily The existing OD interface cannot support the growing demand for more OD features. We are in the transition to a new OD mechanism. So, disable the SMU13 OD feature support temporarily. And this should be reverted when the new OD mechanism online. Signed-off-by: Evan Quan Reviewed-by: Guchun Chen Signed-off-by: Alex Deucher --- .../drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 17 ++++++++++++++--- .../drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c | 12 +++++++++--- 2 files changed, 23 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index e191bbe9e994..fddcd834bcec 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -331,11 +331,13 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu) struct smu_13_0_0_powerplay_table *powerplay_table = table_context->power_play_table; struct smu_baco_context *smu_baco = &smu->smu_baco; +#if 0 PPTable_t *pptable = smu->smu_table.driver_pptable; const OverDriveLimits_t * const overdrive_upperlimits = &pptable->SkuTable.OverDriveLimitsBasicMax; const OverDriveLimits_t * const overdrive_lowerlimits = &pptable->SkuTable.OverDriveLimitsMin; +#endif if (powerplay_table->platform_caps & SMU_13_0_0_PP_PLATFORM_CAP_HARDWAREDC) smu->dc_controlled_by_gpio = true; @@ -347,18 +349,27 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu) if (powerplay_table->platform_caps & SMU_13_0_0_PP_PLATFORM_CAP_MACO) smu_baco->maco_support = true; + /* + * We are in the transition to a new OD mechanism. + * Disable the OD feature support for SMU13 temporarily. + * TODO: get this reverted when new OD mechanism online + */ +#if 0 if (!overdrive_lowerlimits->FeatureCtrlMask || !overdrive_upperlimits->FeatureCtrlMask) smu->od_enabled = false; - table_context->thermal_controller_type = - powerplay_table->thermal_controller_type; - /* * Instead of having its own buffer space and get overdrive_table copied, * smu->od_settings just points to the actual overdrive_table */ smu->od_settings = &powerplay_table->overdrive_table; +#else + smu->od_enabled = false; +#endif + + table_context->thermal_controller_type = + powerplay_table->thermal_controller_type; return 0; } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index b1f0937ccade..62f2886ab4df 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -323,10 +323,12 @@ static int smu_v13_0_7_check_powerplay_table(struct smu_context *smu) struct smu_baco_context *smu_baco = &smu->smu_baco; PPTable_t *smc_pptable = table_context->driver_pptable; BoardTable_t *BoardTable = &smc_pptable->BoardTable; +#if 0 const OverDriveLimits_t * const overdrive_upperlimits = &smc_pptable->SkuTable.OverDriveLimitsBasicMax; const OverDriveLimits_t * const overdrive_lowerlimits = &smc_pptable->SkuTable.OverDriveLimitsMin; +#endif if (powerplay_table->platform_caps & SMU_13_0_7_PP_PLATFORM_CAP_HARDWAREDC) smu->dc_controlled_by_gpio = true; @@ -338,18 +340,22 @@ static int smu_v13_0_7_check_powerplay_table(struct smu_context *smu) if (smu_baco->platform_support && (BoardTable->HsrEnabled || BoardTable->VddqOffEnabled)) smu_baco->maco_support = true; +#if 0 if (!overdrive_lowerlimits->FeatureCtrlMask || !overdrive_upperlimits->FeatureCtrlMask) smu->od_enabled = false; - table_context->thermal_controller_type = - powerplay_table->thermal_controller_type; - /* * Instead of having its own buffer space and get overdrive_table copied, * smu->od_settings just points to the actual overdrive_table */ smu->od_settings = &powerplay_table->overdrive_table; +#else + smu->od_enabled = false; +#endif + + table_context->thermal_controller_type = + powerplay_table->thermal_controller_type; return 0; } From 96b020e2163fb2197266b2f71b1007495206e6bb Mon Sep 17 00:00:00 2001 From: Melissa Wen Date: Mon, 31 Jul 2023 07:35:05 -0100 Subject: [PATCH 07/14] drm/amd/display: check attr flag before set cursor degamma on DCN3+ Don't set predefined degamma curve to cursor plane if the cursor attribute flag is not set. Applying a degamma curve to the cursor by default breaks userspace expectation. Checking the flag before performing any color transformation prevents too dark cursor gamma in DCN3+ on many Linux desktop environment (KDE Plasma, GNOME, wlroots-based, etc.) as reported at: - https://gitlab.freedesktop.org/drm/amd/-/issues/1513 This is the same approach followed by DCN2 drivers where the issue is not present. Fixes: 03f54d7d3448 ("drm/amd/display: Add DCN3 DPP") Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1513 Signed-off-by: Melissa Wen Reviewed-by: Harry Wentland Tested-by: Alex Hung Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp.c index e5b7ef7422b8..50dc83404644 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dpp.c @@ -357,8 +357,11 @@ void dpp3_set_cursor_attributes( int cur_rom_en = 0; if (color_format == CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA || - color_format == CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA) - cur_rom_en = 1; + color_format == CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA) { + if (cursor_attributes->attribute_flags.bits.ENABLE_CURSOR_DEGAMMA) { + cur_rom_en = 1; + } + } REG_UPDATE_3(CURSOR0_CONTROL, CUR0_MODE, color_format, From a73ea79a0c94bacfab4df23a1043644d14f56591 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Wed, 2 Aug 2023 10:24:36 +0530 Subject: [PATCH 08/14] drm/amd/pm: Fix SMU v13.0.6 energy reporting Energy counter should be reported in units of 15.259 uJ. Don't apply any conversion. Signed-off-by: Lijo Lazar Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) 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 1ac552142763..fe4ee2daa5d8 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 @@ -1993,9 +1993,8 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table gpu_metrics->average_socket_power = SMUQ10_TO_UINT(metrics->SocketPower); - /* Energy is reported in 15.625mJ units */ - gpu_metrics->energy_accumulator = - SMUQ10_TO_UINT(metrics->SocketEnergyAcc); + /* Energy counter reported in 15.259uJ (2^-16) units */ + gpu_metrics->energy_accumulator = metrics->SocketEnergyAcc; gpu_metrics->current_gfxclk = SMUQ10_TO_UINT(metrics->GfxclkFrequency[xcc0]); From 90e065677e0362a777b9db97ea21d43a39211399 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 28 Jul 2023 11:14:05 -0400 Subject: [PATCH 09/14] drm/amdgpu: fix possible UAF in amdgpu_cs_pass1() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since the gang_size check is outside of chunk parsing loop, we need to reset i before we free the chunk data. Suggested by Ye Zhang (@VAR10CK) of Baidu Security. Reviewed-by: Guchun Chen Reviewed-by: Christian König Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 040f4cb6ab2d..fb78a8f47587 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -295,7 +295,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, if (!p->gang_size) { ret = -EINVAL; - goto free_partial_kdata; + goto free_all_kdata; } for (i = 0; i < p->gang_size; ++i) { From 3bb575572bf498a9d39e9d1ca5c06cc3152928a1 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Fri, 28 Jul 2023 17:04:01 -0500 Subject: [PATCH 10/14] drm/amd/display: Fix a regression on Polaris cards DCE products don't define a `remove_stream_from_ctx` like DCN ones do. This means that when compute_mst_dsc_configs_for_state() is called it always returns -EINVAL which causes MST to fail to setup. Cc: stable@vger.kernel.org # 6.4.y Cc: Harry Wentland Reported-by: Klaus.Kusche@computerix.info Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/2671 Fixes: efa4c4df864e ("drm/amd/display: call remove_stream_from_ctx from res_pool funcs") Signed-off-by: Mario Limonciello Reviewed-by: Harry Wentland Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 9bc86deac9e8..b885c39bd16b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -1320,7 +1320,7 @@ int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, if (computed_streams[i]) continue; - if (!res_pool->funcs->remove_stream_from_ctx || + if (res_pool->funcs->remove_stream_from_ctx && res_pool->funcs->remove_stream_from_ctx(stream->ctx->dc, dc_state, stream) != DC_OK) return -EINVAL; From 2e91e731f24817bc55f9c9acc95a8939c4077b05 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 28 Jul 2023 16:53:49 -0400 Subject: [PATCH 11/14] drm/amdgpu/gfx11: only enable CP GFX shadowing on SR-IOV MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is only required for SR-IOV world switches, but it adds additional latency leading to reduced performance in some benchmarks. Disable for now on bare metal. Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 3a7af59e83ca..0451533ddde4 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -471,8 +471,12 @@ static void gfx_v11_0_check_fw_cp_gfx_shadow(struct amdgpu_device *adev) case IP_VERSION(11, 0, 3): if ((adev->gfx.me_fw_version >= 1505) && (adev->gfx.pfp_fw_version >= 1600) && - (adev->gfx.mec_fw_version >= 512)) - adev->gfx.cp_gfx_shadow = true; + (adev->gfx.mec_fw_version >= 512)) { + if (amdgpu_sriov_vf(adev)) + adev->gfx.cp_gfx_shadow = true; + else + adev->gfx.cp_gfx_shadow = false; + } break; default: adev->gfx.cp_gfx_shadow = false; From a6dea2d64ff92851e68cd4e20a35f6534286e016 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 28 Jul 2023 11:45:53 -0400 Subject: [PATCH 12/14] drm/amdkfd: ignore crat by default MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We are dropping the IOMMUv2 path, so no need to enable this. It's often buggy on consumer platforms anyway. Reviewed-by: Felix Kuehling Acked-by: Christian König Tested-by: Mike Lothian Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c index 49f40d9f16e8..f5a6f562e2a8 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c @@ -1543,11 +1543,7 @@ static bool kfd_ignore_crat(void) if (ignore_crat) return true; -#ifndef KFD_SUPPORT_IOMMU_V2 ret = true; -#else - ret = false; -#endif return ret; } From 616f92d188ee7142a95a52068efdbea82645f859 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 28 Jul 2023 11:34:59 -0400 Subject: [PATCH 13/14] drm/amdkfd: disable IOMMUv2 support for KV/CZ MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use the dGPU path instead. There were a lot of platform issues with IOMMU in general on these chips due to windows not enabling IOMMU at the time. The dGPU path has been used for a long time with newer APUs and works fine. This also paves the way to simplify the driver significantly. v2: use the dGPU queue manager functions Reviewed-by: Felix Kuehling Acked-by: Christian König Tested-by: Mike Lothian Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 6 ------ drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 8 +------- 2 files changed, 1 insertion(+), 13 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 0b3dc754e06b..7ae44d68ddc9 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -233,10 +233,6 @@ static void kfd_device_info_init(struct kfd_dev *kfd, asic_type != CHIP_TONGA) kfd->device_info.supports_cwsr = true; - if (asic_type == CHIP_KAVERI || - asic_type == CHIP_CARRIZO) - kfd->device_info.needs_iommu_device = true; - if (asic_type != CHIP_HAWAII && !vf) kfd->device_info.needs_pci_atomics = true; } @@ -249,7 +245,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) uint32_t gfx_target_version = 0; switch (adev->asic_type) { -#ifdef KFD_SUPPORT_IOMMU_V2 #ifdef CONFIG_DRM_AMDGPU_CIK case CHIP_KAVERI: gfx_target_version = 70000; @@ -262,7 +257,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) if (!vf) f2g = &gfx_v8_kfd2kgd; break; -#endif #ifdef CONFIG_DRM_AMDGPU_CIK case CHIP_HAWAII: gfx_target_version = 70001; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index 2df153828ff4..01192f5abe46 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -2538,18 +2538,12 @@ struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev) } switch (dev->adev->asic_type) { - case CHIP_CARRIZO: - device_queue_manager_init_vi(&dqm->asic_ops); - break; - case CHIP_KAVERI: - device_queue_manager_init_cik(&dqm->asic_ops); - break; - case CHIP_HAWAII: device_queue_manager_init_cik_hawaii(&dqm->asic_ops); break; + case CHIP_CARRIZO: case CHIP_TONGA: case CHIP_FIJI: case CHIP_POLARIS10: From 091ae5473f96ced844af6ba39b94757359b12348 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 28 Jul 2023 11:38:02 -0400 Subject: [PATCH 14/14] drm/amdkfd: disable IOMMUv2 support for Raven MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use the dGPU path instead. There were a lot of platform issues with IOMMU in general on these chips due to windows not enabling IOMMU at the time. The dGPU path has been used for a long time with newer APUs and works fine. This also paves the way to simplify the driver significantly. Reviewed-by: Felix Kuehling Acked-by: Christian König Tested-by: Mike Lothian Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 7ae44d68ddc9..a53e0757fe64 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -194,11 +194,6 @@ static void kfd_device_info_init(struct kfd_dev *kfd, kfd_device_info_set_event_interrupt_class(kfd); - /* Raven */ - if (gc_version == IP_VERSION(9, 1, 0) || - gc_version == IP_VERSION(9, 2, 2)) - kfd->device_info.needs_iommu_device = true; - if (gc_version < IP_VERSION(11, 0, 0)) { /* Navi2x+, Navi1x+ */ if (gc_version == IP_VERSION(10, 3, 6)) @@ -292,7 +287,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) gfx_target_version = 90000; f2g = &gfx_v9_kfd2kgd; break; -#ifdef KFD_SUPPORT_IOMMU_V2 /* Raven */ case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 2): @@ -300,7 +294,6 @@ struct kfd_dev *kgd2kfd_probe(struct amdgpu_device *adev, bool vf) if (!vf) f2g = &gfx_v9_kfd2kgd; break; -#endif /* Vega12 */ case IP_VERSION(9, 2, 1): gfx_target_version = 90004;