* [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume
@ 2020-10-28 8:30 Evan Quan
2020-10-28 8:30 ` [PATCH 2/5] drm/amd/pm: correct the baco reset sequence for CI ASICs Evan Quan
` (3 more replies)
0 siblings, 4 replies; 6+ messages in thread
From: Evan Quan @ 2020-10-28 8:30 UTC (permalink / raw)
To: amd-gfx; +Cc: Alexander.Deucher, sandy.8925, Evan Quan
This can address the random SDMA hang after pci config reset
seen on Hawaii.
Change-Id: I2d6147600636cbc90d1be7f3d9a011f050708fbd
Signed-off-by: Evan Quan <evan.quan@amd.com>
---
drivers/gpu/drm/amd/amdgpu/cik_sdma.c | 27 ++++++++++++---------------
1 file changed, 12 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
index 20f108818b2b..a3c3fe96515f 100644
--- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c
@@ -1071,22 +1071,19 @@ static int cik_sdma_soft_reset(void *handle)
{
u32 srbm_soft_reset = 0;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
- u32 tmp = RREG32(mmSRBM_STATUS2);
+ u32 tmp;
- if (tmp & SRBM_STATUS2__SDMA_BUSY_MASK) {
- /* sdma0 */
- tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET);
- tmp |= SDMA0_F32_CNTL__HALT_MASK;
- WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp);
- srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA_MASK;
- }
- if (tmp & SRBM_STATUS2__SDMA1_BUSY_MASK) {
- /* sdma1 */
- tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET);
- tmp |= SDMA0_F32_CNTL__HALT_MASK;
- WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp);
- srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA1_MASK;
- }
+ /* sdma0 */
+ tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET);
+ tmp |= SDMA0_F32_CNTL__HALT_MASK;
+ WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp);
+ srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA_MASK;
+
+ /* sdma1 */
+ tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET);
+ tmp |= SDMA0_F32_CNTL__HALT_MASK;
+ WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp);
+ srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA1_MASK;
if (srbm_soft_reset) {
tmp = RREG32(mmSRBM_SOFT_RESET);
--
2.29.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 2/5] drm/amd/pm: correct the baco reset sequence for CI ASICs 2020-10-28 8:30 [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume Evan Quan @ 2020-10-28 8:30 ` Evan Quan 2020-10-28 8:30 ` [PATCH 3/5] drm/amd/pm: enable baco reset for Hawaii Evan Quan ` (2 subsequent siblings) 3 siblings, 0 replies; 6+ messages in thread From: Evan Quan @ 2020-10-28 8:30 UTC (permalink / raw) To: amd-gfx; +Cc: Alexander.Deucher, sandy.8925, Evan Quan Correct some registers bitmasks and add mmBIOS_SCRATCH_7 reset. Change-Id: I416d7bee7e7ddd7b726dd921d0bb442da6ff4b93 Signed-off-by: Evan Quan <evan.quan@amd.com> --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/ci_baco.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ci_baco.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ci_baco.c index 3be40114e63d..45f608838f6e 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ci_baco.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/ci_baco.c @@ -142,12 +142,12 @@ static const struct baco_cmd_entry exit_baco_tbl[] = { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_BCLK_OFF_MASK, BACO_CNTL__BACO_BCLK_OFF__SHIFT, 0, 0x00 }, { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_POWER_OFF_MASK, BACO_CNTL__BACO_POWER_OFF__SHIFT, 0, 0x00 }, { CMD_DELAY_MS, 0, 0, 0, 20, 0 }, - { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__PWRGOOD_BF_MASK, 0, 0xffffffff, 0x20 }, + { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__PWRGOOD_BF_MASK, 0, 0xffffffff, 0x200 }, { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_ISO_DIS_MASK, BACO_CNTL__BACO_ISO_DIS__SHIFT, 0, 0x01 }, - { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__PWRGOOD_MASK, 0, 5, 0x1c }, + { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__PWRGOOD_MASK, 0, 5, 0x1c00 }, { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_ANA_ISO_DIS_MASK, BACO_CNTL__BACO_ANA_ISO_DIS__SHIFT, 0, 0x01 }, { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_RESET_EN_MASK, BACO_CNTL__BACO_RESET_EN__SHIFT, 0, 0x00 }, - { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__RCU_BIF_CONFIG_DONE_MASK, 0, 5, 0x10 }, + { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__RCU_BIF_CONFIG_DONE_MASK, 0, 5, 0x100 }, { CMD_READMODIFYWRITE, mmBACO_CNTL, BACO_CNTL__BACO_EN_MASK, BACO_CNTL__BACO_EN__SHIFT, 0, 0x00 }, { CMD_WAITFOR, mmBACO_CNTL, BACO_CNTL__BACO_MODE_MASK, 0, 0xffffffff, 0x00 } }; @@ -155,6 +155,7 @@ static const struct baco_cmd_entry exit_baco_tbl[] = static const struct baco_cmd_entry clean_baco_tbl[] = { { CMD_WRITE, mmBIOS_SCRATCH_6, 0, 0, 0, 0 }, + { CMD_WRITE, mmBIOS_SCRATCH_7, 0, 0, 0, 0 }, { CMD_WRITE, mmCP_PFP_UCODE_ADDR, 0, 0, 0, 0 } }; -- 2.29.0 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 3/5] drm/amd/pm: enable baco reset for Hawaii 2020-10-28 8:30 [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume Evan Quan 2020-10-28 8:30 ` [PATCH 2/5] drm/amd/pm: correct the baco reset sequence for CI ASICs Evan Quan @ 2020-10-28 8:30 ` Evan Quan 2020-10-28 8:30 ` [PATCH 4/5] drm/amd/pm: perform SMC reset on suspend/hibernation Evan Quan 2020-10-28 8:30 ` [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running Evan Quan 3 siblings, 0 replies; 6+ messages in thread From: Evan Quan @ 2020-10-28 8:30 UTC (permalink / raw) To: amd-gfx; +Cc: Alexander.Deucher, sandy.8925, Evan Quan Which can be used for S4(hibernation) support. Change-Id: I9c90c916bdd6e128b7cf7f5c6c2c6ca5b7cfc0ef Signed-off-by: Evan Quan <evan.quan@amd.com> --- drivers/gpu/drm/amd/amdgpu/cik.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index 03ff8bd1fee8..5442df094102 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -1336,11 +1336,13 @@ cik_asic_reset_method(struct amdgpu_device *adev) switch (adev->asic_type) { case CHIP_BONAIRE: - case CHIP_HAWAII: /* disable baco reset until it works */ /* smu7_asic_get_baco_capability(adev, &baco_reset); */ baco_reset = false; break; + case CHIP_HAWAII: + baco_reset = cik_asic_supports_baco(adev); + break; default: baco_reset = false; break; -- 2.29.0 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 4/5] drm/amd/pm: perform SMC reset on suspend/hibernation 2020-10-28 8:30 [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume Evan Quan 2020-10-28 8:30 ` [PATCH 2/5] drm/amd/pm: correct the baco reset sequence for CI ASICs Evan Quan 2020-10-28 8:30 ` [PATCH 3/5] drm/amd/pm: enable baco reset for Hawaii Evan Quan @ 2020-10-28 8:30 ` Evan Quan 2020-10-28 8:30 ` [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running Evan Quan 3 siblings, 0 replies; 6+ messages in thread From: Evan Quan @ 2020-10-28 8:30 UTC (permalink / raw) To: amd-gfx; +Cc: Alexander.Deucher, sandy.8925, Evan Quan So that the succeeding resume can be performed based on a clean state. Change-Id: I82f16eb2d1a6e389f171784e6e56e41892e1725e Signed-off-by: Evan Quan <evan.quan@amd.com> --- drivers/gpu/drm/amd/pm/inc/hwmgr.h | 1 + drivers/gpu/drm/amd/pm/inc/smumgr.h | 2 ++ .../drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c | 4 ++++ .../drm/amd/pm/powerplay/smumgr/ci_smumgr.c | 24 +++++++++++++++++++ .../gpu/drm/amd/pm/powerplay/smumgr/smumgr.c | 8 +++++++ 5 files changed, 39 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/inc/hwmgr.h b/drivers/gpu/drm/amd/pm/inc/hwmgr.h index 0e4707a76f6d..7e29ec2902dc 100644 --- a/drivers/gpu/drm/amd/pm/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/pm/inc/hwmgr.h @@ -229,6 +229,7 @@ struct pp_smumgr_func { bool (*is_hw_avfs_present)(struct pp_hwmgr *hwmgr); int (*update_dpm_settings)(struct pp_hwmgr *hwmgr, void *profile_setting); int (*smc_table_manager)(struct pp_hwmgr *hwmgr, uint8_t *table, uint16_t table_id, bool rw); /*rw: true for read, false for write */ + int (*stop_smc)(struct pp_hwmgr *hwmgr); }; struct pp_hwmgr_func { diff --git a/drivers/gpu/drm/amd/pm/inc/smumgr.h b/drivers/gpu/drm/amd/pm/inc/smumgr.h index ad100b533d04..5f46f1a4f38e 100644 --- a/drivers/gpu/drm/amd/pm/inc/smumgr.h +++ b/drivers/gpu/drm/amd/pm/inc/smumgr.h @@ -113,4 +113,6 @@ extern int smum_update_dpm_settings(struct pp_hwmgr *hwmgr, void *profile_settin extern int smum_smc_table_manager(struct pp_hwmgr *hwmgr, uint8_t *table, uint16_t table_id, bool rw); +extern int smum_stop_smc(struct pp_hwmgr *hwmgr); + #endif diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c index 5937150e6b37..0879cd7df59f 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c @@ -1667,6 +1667,10 @@ static int smu7_disable_dpm_tasks(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((tmp_result == 0), "Failed to reset to default!", result = tmp_result); + tmp_result = smum_stop_smc(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to stop smc!", result = tmp_result); + tmp_result = smu7_force_switch_to_arbf0(hwmgr); PP_ASSERT_WITH_CODE((tmp_result == 0), "Failed to force to switch arbf0!", result = tmp_result); diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c index e4d1f3d66ef4..09128122b493 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c @@ -2939,6 +2939,29 @@ static int ci_update_smc_table(struct pp_hwmgr *hwmgr, uint32_t type) return 0; } +static void ci_reset_smc(struct pp_hwmgr *hwmgr) +{ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SMC_SYSCON_RESET_CNTL, + rst_reg, 1); +} + + +static void ci_stop_smc_clock(struct pp_hwmgr *hwmgr) +{ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SMC_SYSCON_CLOCK_CNTL_0, + ck_disable, 1); +} + +static int ci_stop_smc(struct pp_hwmgr *hwmgr) +{ + ci_reset_smc(hwmgr); + ci_stop_smc_clock(hwmgr); + + return 0; +} + const struct pp_smumgr_func ci_smu_funcs = { .name = "ci_smu", .smu_init = ci_smu_init, @@ -2964,4 +2987,5 @@ const struct pp_smumgr_func ci_smu_funcs = { .is_dpm_running = ci_is_dpm_running, .update_dpm_settings = ci_update_dpm_settings, .update_smc_table = ci_update_smc_table, + .stop_smc = ci_stop_smc, }; diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smumgr.c index b6fb48066841..b6921db3c130 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/smumgr.c @@ -245,3 +245,11 @@ int smum_smc_table_manager(struct pp_hwmgr *hwmgr, uint8_t *table, uint16_t tabl return -EINVAL; } + +int smum_stop_smc(struct pp_hwmgr *hwmgr) +{ + if (hwmgr->smumgr_funcs->stop_smc) + return hwmgr->smumgr_funcs->stop_smc(hwmgr); + + return 0; +} -- 2.29.0 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running 2020-10-28 8:30 [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume Evan Quan ` (2 preceding siblings ...) 2020-10-28 8:30 ` [PATCH 4/5] drm/amd/pm: perform SMC reset on suspend/hibernation Evan Quan @ 2020-10-28 8:30 ` Evan Quan 2020-10-28 13:07 ` Deucher, Alexander 3 siblings, 1 reply; 6+ messages in thread From: Evan Quan @ 2020-10-28 8:30 UTC (permalink / raw) To: amd-gfx; +Cc: Alexander.Deucher, sandy.8925, Evan Quan This reverts commit f87812284172a9809820d10143b573d833cd3f75 "drm/amdgpu: Fix bug where DPM is not enabled after hibernate and resume". It was intended to fix Hawaii S4(hibernation) issue but break S3. As ixFEATURE_STATUS is filled with garbage data on resume which can be only cleared by reloading smc firmware(but that will involve many changes). So, we will revert this S4 fix and seek a new way. Change-Id: If9eed2f5a9c1168fb20be92057b583d854ad779e Signed-off-by: Evan Quan <evan.quan@amd.com> --- drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c index 09128122b493..329bf4d44bbc 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c @@ -2726,10 +2726,7 @@ static int ci_initialize_mc_reg_table(struct pp_hwmgr *hwmgr) static bool ci_is_dpm_running(struct pp_hwmgr *hwmgr) { - return (1 == PHM_READ_INDIRECT_FIELD(hwmgr->device, - CGS_IND_REG__SMC, FEATURE_STATUS, - VOLTAGE_CONTROLLER_ON)) - ? true : false; + return ci_is_smc_ram_running(hwmgr); } static int ci_smu_init(struct pp_hwmgr *hwmgr) -- 2.29.0 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running 2020-10-28 8:30 ` [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running Evan Quan @ 2020-10-28 13:07 ` Deucher, Alexander 0 siblings, 0 replies; 6+ messages in thread From: Deucher, Alexander @ 2020-10-28 13:07 UTC (permalink / raw) To: Quan, Evan, amd-gfx; +Cc: sandy.8925 [-- Attachment #1.1: Type: text/plain, Size: 1901 bytes --] [AMD Public Use] Series is: Reviewed-by: Alex Deucher <alexander.deucher@amd.com> ________________________________ From: Quan, Evan <Evan.Quan@amd.com> Sent: Wednesday, October 28, 2020 4:30 AM To: amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org> Cc: Deucher, Alexander <Alexander.Deucher@amd.com>; sandy.8925@gmail.com <sandy.8925@gmail.com>; Quan, Evan <Evan.Quan@amd.com> Subject: [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running This reverts commit f87812284172a9809820d10143b573d833cd3f75 "drm/amdgpu: Fix bug where DPM is not enabled after hibernate and resume". It was intended to fix Hawaii S4(hibernation) issue but break S3. As ixFEATURE_STATUS is filled with garbage data on resume which can be only cleared by reloading smc firmware(but that will involve many changes). So, we will revert this S4 fix and seek a new way. Change-Id: If9eed2f5a9c1168fb20be92057b583d854ad779e Signed-off-by: Evan Quan <evan.quan@amd.com> --- drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c index 09128122b493..329bf4d44bbc 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c @@ -2726,10 +2726,7 @@ static int ci_initialize_mc_reg_table(struct pp_hwmgr *hwmgr) static bool ci_is_dpm_running(struct pp_hwmgr *hwmgr) { - return (1 == PHM_READ_INDIRECT_FIELD(hwmgr->device, - CGS_IND_REG__SMC, FEATURE_STATUS, - VOLTAGE_CONTROLLER_ON)) - ? true : false; + return ci_is_smc_ram_running(hwmgr); } static int ci_smu_init(struct pp_hwmgr *hwmgr) -- 2.29.0 [-- Attachment #1.2: Type: text/html, Size: 3624 bytes --] [-- Attachment #2: Type: text/plain, Size: 154 bytes --] _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-10-28 13:08 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-10-28 8:30 [PATCH 1/5] drm/amdgpu: perform srbm soft reset always on SDMA resume Evan Quan 2020-10-28 8:30 ` [PATCH 2/5] drm/amd/pm: correct the baco reset sequence for CI ASICs Evan Quan 2020-10-28 8:30 ` [PATCH 3/5] drm/amd/pm: enable baco reset for Hawaii Evan Quan 2020-10-28 8:30 ` [PATCH 4/5] drm/amd/pm: perform SMC reset on suspend/hibernation Evan Quan 2020-10-28 8:30 ` [PATCH 5/5] drm/amd/pm: do not use ixFEATURE_STATUS for checking smc running Evan Quan 2020-10-28 13:07 ` Deucher, Alexander
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).