From a8f921a10a8c2e2b209ad9ed1f1b7f48192c30aa Mon Sep 17 00:00:00 2001 From: ganglxie Date: Mon, 24 Feb 2025 15:06:51 +0800 Subject: [PATCH 01/16] drm/amdgpu: Change page/record number calculation based on nps save only one record to save eeprom space,and bad_page_num = pa_rec_num + mca_rec_num*16 Signed-off-by: ganglxie Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 49 +++++++++---------- .../gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c | 17 +++---- .../gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h | 20 +++----- 3 files changed, 35 insertions(+), 51 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index f0349094f8c9..493dd004d6fa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2981,24 +2981,14 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev, /* only new entries are saved */ if (save_count > 0) { - if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA) { + for (i = 0; i < unit_num; i++) { if (amdgpu_ras_eeprom_append(control, - &data->bps[control->ras_num_recs], - save_count)) { + &data->bps[bad_page_num + i * adev->umc.retire_unit], + 1)) { dev_err(adev->dev, "Failed to save EEPROM table data!"); return -EIO; } - } else { - for (i = 0; i < unit_num; i++) { - if (amdgpu_ras_eeprom_append(control, - &data->bps[bad_page_num + i * adev->umc.retire_unit], - 1)) { - dev_err(adev->dev, "Failed to save EEPROM table data!"); - return -EIO; - } - } } - dev_info(adev->dev, "Saved %d pages to EEPROM table.\n", save_count); } @@ -3014,7 +3004,7 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev) struct amdgpu_ras_eeprom_control *control = &adev->psp.ras_context.ras->eeprom_control; struct eeprom_table_record *bps; - int ret; + int ret, i = 0; /* no bad page record, skip eeprom access */ if (control->ras_num_recs == 0 || amdgpu_bad_page_threshold == 0) @@ -3028,13 +3018,23 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev) if (ret) { dev_err(adev->dev, "Failed to load EEPROM table records!"); } else { - if (control->ras_num_recs > 1 && - adev->umc.ras && adev->umc.ras->convert_ras_err_addr) { - if ((bps[0].address == bps[1].address) && - (bps[0].mem_channel == bps[1].mem_channel)) - control->rec_type = AMDGPU_RAS_EEPROM_REC_PA; - else - control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA; + if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr) { + for (i = 0; i < control->ras_num_recs; i++) { + if ((control->ras_num_recs - i) >= adev->umc.retire_unit) { + if ((bps[i].address == bps[i + 1].address) && + (bps[i].mem_channel == bps[i + 1].mem_channel)) { + control->ras_num_pa_recs += adev->umc.retire_unit; + i += (adev->umc.retire_unit - 1); + } else { + control->ras_num_mca_recs += + (control->ras_num_recs - i); + break; + } + } else { + control->ras_num_mca_recs += (control->ras_num_recs - i); + break; + } + } } ret = amdgpu_ras_eeprom_check(control); @@ -3440,12 +3440,7 @@ int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev) return ret; if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr) - control->rec_type = AMDGPU_RAS_EEPROM_REC_PA; - - /* default status is MCA storage */ - if (control->ras_num_recs <= 1 && - adev->umc.ras && adev->umc.ras->convert_ras_err_addr) - control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA; + control->ras_num_pa_recs = control->ras_num_recs; if (control->ras_num_recs) { ret = amdgpu_ras_load_bad_pages(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index 87fcdda3ec61..ab27cecb5519 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -727,11 +727,9 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control, - control->ras_fri) % control->ras_max_record_count; - if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA) - control->ras_num_bad_pages = control->ras_num_recs; - else - control->ras_num_bad_pages = - control->ras_num_recs * adev->umc.retire_unit; + control->ras_num_mca_recs += num; + control->ras_num_bad_pages += num * adev->umc.retire_unit; + Out: kfree(buf); return res; @@ -1396,6 +1394,8 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control) } control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset); + control->ras_num_mca_recs = 0; + control->ras_num_pa_recs = 0; return 0; } @@ -1416,11 +1416,8 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control) if (!__get_eeprom_i2c_addr(adev, control)) return -EINVAL; - if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA) - control->ras_num_bad_pages = control->ras_num_recs; - else - control->ras_num_bad_pages = - control->ras_num_recs * adev->umc.retire_unit; + control->ras_num_bad_pages = control->ras_num_pa_recs + + control->ras_num_mca_recs * adev->umc.retire_unit; if (hdr->header == RAS_TABLE_HDR_VAL) { DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records", diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h index 81d55cb7b397..13f7eda9a696 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h @@ -43,19 +43,6 @@ enum amdgpu_ras_eeprom_err_type { AMDGPU_RAS_EEPROM_ERR_COUNT, }; -/* - * one UMC MCA address could map to multiply physical address (PA), - * such as 1:16, we use eeprom_table_record.address to store MCA - * address and use eeprom_table_record.retired_page to save PA. - * - * AMDGPU_RAS_EEPROM_REC_PA: one record store one PA - * AMDGPU_RAS_EEPROM_REC_MCA: one record store one MCA address - */ -enum amdgpu_ras_eeprom_rec_type { - AMDGPU_RAS_EEPROM_REC_PA, - AMDGPU_RAS_EEPROM_REC_MCA, -}; - struct amdgpu_ras_eeprom_table_header { uint32_t header; uint32_t version; @@ -100,6 +87,12 @@ struct amdgpu_ras_eeprom_control { */ u32 ras_num_bad_pages; + /* Number of records store mca address */ + u32 ras_num_mca_recs; + + /* Number of records store physical address */ + u32 ras_num_pa_recs; + /* First record index to read, 0-based. * Range is [0, num_recs-1]. This is * an absolute index, starting right after @@ -120,7 +113,6 @@ struct amdgpu_ras_eeprom_control { /* Record channel info which occurred bad pages */ u32 bad_channel_bitmap; - enum amdgpu_ras_eeprom_rec_type rec_type; }; /* -- 2.51.0 From 7c62aacc3b452f73a1284198c81551035fac6d71 Mon Sep 17 00:00:00 2001 From: Pierre-Eric Pelloux-Prayer Date: Thu, 20 Feb 2025 14:41:59 +0100 Subject: [PATCH 02/16] drm/amdgpu: init return value in amdgpu_ttm_clear_buffer MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Otherwise an uninitialized value can be returned if amdgpu_res_cleared returns true for all regions. Possibly closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3812 Fixes: a68c7eaa7a8f ("drm/amdgpu: Enable clear page functionality") Signed-off-by: Pierre-Eric Pelloux-Prayer Acked-by: Alex Deucher Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index bcb4bcc4ab75..53b71e9d8076 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -2295,7 +2295,7 @@ int amdgpu_ttm_clear_buffer(struct amdgpu_bo *bo, struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring; struct amdgpu_res_cursor cursor; u64 addr; - int r; + int r = 0; if (!adev->mman.buffer_funcs_enabled) return -EINVAL; -- 2.51.0 From 59f9c2c9f6f87c640d82a9751647d2eb1c4f0095 Mon Sep 17 00:00:00 2001 From: Srinivasan Shanmugam Date: Mon, 24 Feb 2025 17:26:22 +0530 Subject: [PATCH 03/16] drm/amdgpu: Fix correct parameter desc for VCN idle check functions MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Fixes the kdoc for the following VCN idle check functions by updating the parameter description from 'handle' to 'ip_block': - vcn_v4_0_is_idle - vcn_v4_0_3_is_idle - vcn_v4_0_5_is_idle - vcn_v5_0_1_is_idle Fixes the below with gcc W=1: drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c:935: warning: Function parameter or struct member 'ip_block' not described in 'vcn_v5_0_1_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c:935: warning: Excess function parameter 'handle' description in 'vcn_v5_0_1_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c:1972: warning: Function parameter or struct member 'ip_block' not described in 'vcn_v4_0_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c:1972: warning: Excess function parameter 'handle' description in 'vcn_v4_0_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c:1583: warning: Function parameter or struct member 'ip_block' not described in 'vcn_v4_0_3_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c:1583: warning: Excess function parameter 'handle' description in 'vcn_v4_0_3_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c:1200: warning: Function parameter or struct member 'ip_block' not described in 'vcn_v5_0_0_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c:1200: warning: Excess function parameter 'handle' description in 'vcn_v5_0_0_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c:1460: warning: Function parameter or struct member 'ip_block' not described in 'vcn_v4_0_5_is_idle' drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c:1460: warning: Excess function parameter 'handle' description in 'vcn_v4_0_5_is_idle' Cc: Christian König Cc: Alex Deucher Signed-off-by: Srinivasan Shanmugam Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index fe539f7957f0..0dd844243531 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -1964,7 +1964,7 @@ static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev) /** * vcn_v4_0_is_idle - check VCN block is idle * - * @handle: amdgpu_device pointer + * @ip_block: Pointer to the amdgpu_ip_block structure * * Check whether VCN block is idle */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index dda5ee187948..c936bd08963e 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -1575,7 +1575,7 @@ static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev) /** * vcn_v4_0_3_is_idle - check VCN block is idle * - * @handle: amdgpu_device pointer + * @ip_block: Pointer to the amdgpu_ip_block structure * * Check whether VCN block is idle */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index a482658cbf86..5e6ad825cc47 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -1452,7 +1452,7 @@ static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev) /** * vcn_v4_0_5_is_idle - check VCN block is idle * - * @handle: amdgpu_device pointer + * @ip_block: Pointer to the amdgpu_ip_block structure * * Check whether VCN block is idle */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index 721c29f52451..40f36c9a4132 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -927,7 +927,7 @@ static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev) /** * vcn_v5_0_1_is_idle - check VCN block is idle * - * @handle: amdgpu_device pointer + * @ip_block: Pointer to the amdgpu_ip_block structure * * Check whether VCN block is idle */ -- 2.51.0 From dab993bf151359e4e4d045c5469f34333e88e359 Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Tue, 25 Feb 2025 19:18:12 +0800 Subject: [PATCH 04/16] drm/amdgpu: increase AMDGPU_MAX_RINGS Increase it since a cper ring is introduced. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index 52f7a9a79e7b..b4fd1e17205e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -37,7 +37,7 @@ struct amdgpu_job; struct amdgpu_vm; /* max number of rings */ -#define AMDGPU_MAX_RINGS 132 +#define AMDGPU_MAX_RINGS 133 #define AMDGPU_MAX_HWIP_RINGS 64 #define AMDGPU_MAX_GFX_RINGS 2 #define AMDGPU_MAX_SW_GFX_RINGS 2 -- 2.51.0 From dae870019866b247a7796359c4327ade3170faa4 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 24 Feb 2025 11:13:27 -0500 Subject: [PATCH 05/16] drm/amdgpu/vcn2.5: fix VCN stop logic Need to make sure we call amdgpu_dpm_enable_vcn() in vcn_v2_5_stop() at the end if there are errors or DPG is enabled. Fixes: ebc25499de12 ("drm/amdgpu/vcn2.5: split code along instances") Reviewed-by: Boyuan Zhang Suggested-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index a6726afcf89c..e36e2a5676df 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -1414,13 +1414,15 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev, int i) if (adev->vcn.harvest_config & (1 << i)) return 0; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v2_5_stop_dpg_mode(adev, i); + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + r = vcn_v2_5_stop_dpg_mode(adev, i); + goto done; + } /* wait for vcn idle */ r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7); if (r) - return r; + goto done; tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | UVD_LMI_STATUS__READ_CLEAN_MASK | @@ -1428,7 +1430,7 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev, int i) UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); if (r) - return r; + goto done; /* block LMI UMC channel */ tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2); @@ -1439,7 +1441,7 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev, int i) UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); if (r) - return r; + goto done; /* block VCPU register access */ WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), @@ -1465,10 +1467,11 @@ static int vcn_v2_5_stop(struct amdgpu_device *adev, int i) UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK); +done: if (adev->pm.dpm_enabled) amdgpu_dpm_enable_vcn(adev, false, i); - return 0; + return r; } static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev, -- 2.51.0 From d39f1bb57741ff6e34dd3c2937d5cddfed8a499c Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 11:47:49 -0500 Subject: [PATCH 06/16] drm/amdgpu/vcn3.0: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. v2: squash in fix for stop() from Boyuan Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 443 +++++++++++++------------- 1 file changed, 218 insertions(+), 225 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index f75f996d929e..ec42a456c541 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -1134,192 +1134,186 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo return 0; } -static int vcn_v3_0_start(struct amdgpu_device *adev) +static int vcn_v3_0_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t rb_bufsz, tmp; - int i, j, k, r; + int j, k, r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, true, i); - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - /* disable VCN power gating */ - vcn_v3_0_disable_static_power_gating(adev, i); + /* disable VCN power gating */ + vcn_v3_0_disable_static_power_gating(adev, i); - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp); + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp); - /*SW clock gating */ - vcn_v3_0_disable_clock_gating(adev, i); + /* SW clock gating */ + vcn_v3_0_disable_clock_gating(adev, i); - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); - - /* setup mmUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL); - WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - /* setup mmUVD_MPC_CNTL */ - tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL); - tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; - tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; - WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp); - - /* setup UVD_MPC_SET_MUXA0 */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0, - ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); - - /* setup UVD_MPC_SET_MUXB0 */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0, - ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); - - /* setup mmUVD_MPC_SET_MUX */ - WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX, - ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | - (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - - vcn_v3_0_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - for (j = 0; j < 10; ++j) { - uint32_t status; + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, i, mmUVD_STATUS); - if (status & 2) - break; - mdelay(10); - } - r = 0; - if (status & 2) - break; + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - DRM_ERROR("VCN[%d] decode not responding, trying to reset the VCPU!!!\n", i); - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); + + /* setup mmUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL); + WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL, tmp | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup mmUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, i, mmUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, i, mmUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup mmUVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, i, mmUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v3_0_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, i, mmUVD_STATUS); + if (status & 2) + break; mdelay(10); - r = -1; } + r = 0; + if (status & 2) + break; - if (r) { - DRM_ERROR("VCN[%d] decode not responding, giving up!!!\n", i); - return r; - } + DRM_ERROR("VCN[%d] decode not responding, trying to reset the VCPU!!!\n", i); + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + mdelay(10); + r = -1; + } - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + if (r) { + DRM_ERROR("VCN[%d] decode not responding, giving up!!!\n", i); + return r; + } - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - ring = &adev->vcn.inst[i].ring_dec; - /* force RBC into idle state */ - rb_bufsz = order_base_2(ring->ring_size); - tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); - tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->multi_queue.decode_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_VMID, 0); - /* programm the RB_BASE for ring buffer */ - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW, - lower_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH, - upper_32_bits(ring->gpu_addr)); + ring = &adev->vcn.inst[i].ring_dec; + /* force RBC into idle state */ + rb_bufsz = order_base_2(ring->ring_size); + tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1); + tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1); + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_CNTL, tmp); + + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->multi_queue.decode_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); - /* Initialize the ring buffer's read and write pointers */ - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0); + /* programm the RB_BASE for ring buffer */ + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH, + upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_SCRATCH2, 0); - ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR); - WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR, - lower_32_bits(ring->wptr)); - fw_shared->rb.wptr = lower_32_bits(ring->wptr); - fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); - - if (amdgpu_ip_version(adev, UVD_HWIP, 0) != - IP_VERSION(3, 0, 33)) { - fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); - ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4); - fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); - - fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); - ring = &adev->vcn.inst[i].ring_enc[1]; - WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr)); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr); - WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4); - fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); - } + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR, 0); + + WREG32_SOC15(VCN, i, mmUVD_SCRATCH2, 0); + ring->wptr = RREG32_SOC15(VCN, i, mmUVD_RBC_RB_RPTR); + WREG32_SOC15(VCN, i, mmUVD_RBC_RB_WPTR, + lower_32_bits(ring->wptr)); + fw_shared->rb.wptr = lower_32_bits(ring->wptr); + fw_shared->multi_queue.decode_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); + + if (amdgpu_ip_version(adev, UVD_HWIP, 0) != + IP_VERSION(3, 0, 33)) { + fw_shared->multi_queue.encode_generalpurpose_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); + ring = &adev->vcn.inst[i].ring_enc[0]; + WREG32_SOC15(VCN, i, mmUVD_RB_RPTR, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_WPTR, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_RB_SIZE, ring->ring_size / 4); + fw_shared->multi_queue.encode_generalpurpose_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); + + fw_shared->multi_queue.encode_lowlatency_queue_mode |= cpu_to_le32(FW_QUEUE_RING_RESET); + ring = &adev->vcn.inst[i].ring_enc[1]; + WREG32_SOC15(VCN, i, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr)); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_LO2, ring->gpu_addr); + WREG32_SOC15(VCN, i, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, mmUVD_RB_SIZE2, ring->ring_size / 4); + fw_shared->multi_queue.encode_lowlatency_queue_mode &= cpu_to_le32(~FW_QUEUE_RING_RESET); } return 0; @@ -1565,81 +1559,78 @@ static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) return 0; } -static int vcn_v3_0_stop(struct amdgpu_device *adev) +static int vcn_v3_0_stop(struct amdgpu_device *adev, int i) { uint32_t tmp; - int i, r = 0; + int r = 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v3_0_stop_dpg_mode(adev, i); - continue; - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + r = vcn_v3_0_stop_dpg_mode(adev, i); + goto done; + } - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + goto done; - /* disable LMI UMC channel */ - tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK| - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; + + /* disable LMI UMC channel */ + tmp = RREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, i, mmUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK| + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, mmUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - /* apply soft reset */ - tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); + /* apply soft reset */ + tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, i, mmUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, mmUVD_SOFT_RESET, tmp); - /* clear status */ - WREG32_SOC15(VCN, i, mmUVD_STATUS, 0); + /* clear status */ + WREG32_SOC15(VCN, i, mmUVD_STATUS, 0); - /* apply HW clock gating */ - vcn_v3_0_enable_clock_gating(adev, i); + /* apply HW clock gating */ + vcn_v3_0_enable_clock_gating(adev, i); - /* enable VCN power gating */ - vcn_v3_0_enable_static_power_gating(adev, i); - } + /* enable VCN power gating */ + vcn_v3_0_enable_static_power_gating(adev, i); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } +done: + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, false, i); - return 0; + return r; } static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev, @@ -2163,7 +2154,7 @@ static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating @@ -2177,10 +2168,12 @@ static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v3_0_stop(adev); - else - ret = vcn_v3_0_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v3_0_stop(adev, i); + else + ret |= vcn_v3_0_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From f4cd7a85db75d78f7cb907de785b1140e8b8b8aa Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 12:01:44 -0500 Subject: [PATCH 07/16] drm/amdgpu/vcn4.0: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. v2: squash in fix for stop() from Boyuan Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 440 +++++++++++++------------- 1 file changed, 218 insertions(+), 222 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 0dd844243531..c99c121faac0 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -1087,182 +1087,177 @@ static int vcn_v4_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo * vcn_v4_0_start - VCN start * * @adev: amdgpu_device pointer + * @i: instance to start * * Start VCN block */ -static int vcn_v4_0_start(struct amdgpu_device *adev) +static int vcn_v4_0_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t tmp; - int i, j, k, r; + int j, k, r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, true, i); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v4_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v4_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - /* disable VCN power gating */ - vcn_v4_0_disable_static_power_gating(adev, i); - - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); - - /*SW clock gating */ - vcn_v4_0_disable_clock_gating(adev, i); - - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - - /* setup regUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - /* setup regUVD_MPC_CNTL */ - tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL); - tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; - tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; - WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp); - - /* setup UVD_MPC_SET_MUXA0 */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0, - ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); - - /* setup UVD_MPC_SET_MUXB0 */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0, - ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); - - /* setup UVD_MPC_SET_MUX */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX, - ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | - (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - - vcn_v4_0_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - for (j = 0; j < 10; ++j) { - uint32_t status; - - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, i, regUVD_STATUS); - if (status & 2) - break; - mdelay(10); - if (amdgpu_emu_mode == 1) - msleep(1); - } + /* disable VCN power gating */ + vcn_v4_0_disable_static_power_gating(adev, i); + + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); + + /*SW clock gating */ + vcn_v4_0_disable_clock_gating(adev, i); + + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); + + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - if (amdgpu_emu_mode == 1) { - r = -1; - if (status & 2) { - r = 0; - break; - } - } else { + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup regUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup UVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v4_0_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, i, regUVD_STATUS); + if (status & 2) + break; + mdelay(10); + if (amdgpu_emu_mode == 1) + msleep(1); + } + + if (amdgpu_emu_mode == 1) { + r = -1; + if (status & 2) { r = 0; - if (status & 2) - break; - - dev_err(adev->dev, "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - mdelay(10); - r = -1; + break; } + } else { + r = 0; + if (status & 2) + break; + + dev_err(adev->dev, "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + mdelay(10); + r = -1; } + } - if (r) { - dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); - return r; - } + if (r) { + dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); + return r; + } - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, - ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | - VCN_RB1_DB_CTRL__EN_MASK); - - WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; - WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); - - tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); - ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp |= VCN_RB_ENABLE__RB1_EN_MASK; - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); - } + ring = &adev->vcn.inst[i].ring_enc[0]; + WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, + ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); + + WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; + WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); + + tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); + ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB1_EN_MASK; + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); return 0; } @@ -1549,86 +1544,85 @@ static void vcn_v4_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) * vcn_v4_0_stop - VCN stop * * @adev: amdgpu_device pointer + * @i: instance to stop * * Stop VCN block */ -static int vcn_v4_0_stop(struct amdgpu_device *adev) +static int vcn_v4_0_stop(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; uint32_t tmp; - int i, r = 0; + int r = 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + if (adev->vcn.harvest_config & (1 << i)) + return 0; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - vcn_v4_0_stop_dpg_mode(adev, i); - continue; - } + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v4_0_stop_dpg_mode(adev, i); + r = 0; + goto done; + } - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + goto done; - /* disable LMI UMC channel */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; + + /* disable LMI UMC channel */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - /* apply soft reset */ - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + /* apply soft reset */ + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - /* clear status */ - WREG32_SOC15(VCN, i, regUVD_STATUS, 0); + /* clear status */ + WREG32_SOC15(VCN, i, regUVD_STATUS, 0); - /* apply HW clock gating */ - vcn_v4_0_enable_clock_gating(adev, i); + /* apply HW clock gating */ + vcn_v4_0_enable_clock_gating(adev, i); - /* enable VCN power gating */ - vcn_v4_0_enable_static_power_gating(adev, i); - } + /* enable VCN power gating */ + vcn_v4_0_enable_static_power_gating(adev, i); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } +done: + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, false, i); return 0; } @@ -2051,7 +2045,7 @@ static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating @@ -2065,10 +2059,12 @@ static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v4_0_stop(adev); - else - ret = vcn_v4_0_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v4_0_stop(adev, i); + else + ret |= vcn_v4_0_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From 5826d5a5d5c5548a11fbc503d9a0bdee8af310d2 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 12:13:15 -0500 Subject: [PATCH 08/16] drm/amdgpu/vcn4.0.3: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. v2: squash in fix for stop() from Boyuan Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 405 ++++++++++++------------ 1 file changed, 202 insertions(+), 203 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index c936bd08963e..00b94d01cd4c 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -1117,172 +1117,169 @@ static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev) * vcn_v4_0_3_start - VCN start * * @adev: amdgpu_device pointer + * @i: instance to start * * Start VCN block */ -static int vcn_v4_0_3_start(struct amdgpu_device *adev) +static int vcn_v4_0_3_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; struct amdgpu_ring *ring; - int i, j, k, r, vcn_inst; + int j, k, r, vcn_inst; uint32_t tmp; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - vcn_inst = GET_INST(VCN, i); - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | - UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp); - - /*SW clock gating */ - vcn_v4_0_3_disable_clock_gating(adev, i); - - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, - ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - - /* setup regUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL); - WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, - tmp | UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - /* setup regUVD_MPC_CNTL */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL); - tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; - tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; - WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL, tmp); - - /* setup UVD_MPC_SET_MUXA0 */ - WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXA0, - ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); - - /* setup UVD_MPC_SET_MUXB0 */ - WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXB0, - ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); - - /* setup UVD_MPC_SET_MUX */ - WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUX, - ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | - (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - - vcn_v4_0_3_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, vcn_inst, regUVD_GFX8_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + vcn_inst = GET_INST(VCN, i); + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | + UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp); - for (j = 0; j < 10; ++j) { - uint32_t status; + /* SW clock gating */ + vcn_v4_0_3_disable_clock_gating(adev, i); - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, vcn_inst, - regUVD_STATUS); - if (status & 2) - break; - mdelay(10); - } - r = 0; - if (status & 2) - break; + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, + ~UVD_VCPU_CNTL__CLK_EN_MASK); - DRM_DEV_ERROR(adev->dev, - "VCN decode not responding, trying to reset the VCPU!!!\n"); - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, - regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, - regUVD_VCPU_CNTL), - 0, ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, + tmp | UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup regUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup UVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, vcn_inst, regUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v4_0_3_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_GFX8_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, vcn_inst, + regUVD_STATUS); + if (status & 2) + break; mdelay(10); - r = -1; } + r = 0; + if (status & 2) + break; - if (r) { - DRM_DEV_ERROR(adev->dev, "VCN decode not responding, giving up!!!\n"); - return r; - } + DRM_DEV_ERROR(adev->dev, + "VCN decode not responding, trying to reset the VCPU!!!\n"); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, + regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, + regUVD_VCPU_CNTL), + 0, ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + mdelay(10); + r = -1; + } - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + if (r) { + DRM_DEV_ERROR(adev->dev, "VCN decode not responding, giving up!!!\n"); + return r; + } - ring = &adev->vcn.inst[i].ring_enc[0]; - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - /* program the RB_BASE for ring buffer */ - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, - lower_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, - upper_32_bits(ring->gpu_addr)); + ring = &adev->vcn.inst[i].ring_enc[0]; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + + /* program the RB_BASE for ring buffer */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, + lower_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, + upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, - ring->ring_size / sizeof(uint32_t)); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, + ring->ring_size / sizeof(uint32_t)); - /* resetting ring, fw should not check RB ring */ - tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); - tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK); - WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + /* resetting ring, fw should not check RB ring */ + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB_EN_MASK); + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); - /* Initialize the ring buffer's read and write pointers */ - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); + /* Initialize the ring buffer's read and write pointers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); - tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); - tmp |= VCN_RB_ENABLE__RB_EN_MASK; - WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB_EN_MASK; + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); - ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); - fw_shared->sq.queue_mode &= - cpu_to_le32(~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF)); + ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); + fw_shared->sq.queue_mode &= + cpu_to_le32(~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF)); - } return 0; } @@ -1322,81 +1319,81 @@ static int vcn_v4_0_3_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) * vcn_v4_0_3_stop - VCN stop * * @adev: amdgpu_device pointer + * @i: instance to stop * * Stop VCN block */ -static int vcn_v4_0_3_stop(struct amdgpu_device *adev) +static int vcn_v4_0_3_stop(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; - int i, r = 0, vcn_inst; + int r = 0, vcn_inst; uint32_t tmp; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - vcn_inst = GET_INST(VCN, i); + vcn_inst = GET_INST(VCN, i); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - vcn_v4_0_3_stop_dpg_mode(adev, i); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v4_0_3_stop_dpg_mode(adev, i); + goto Done; + } - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, - UVD_STATUS__IDLE, 0x7); - if (r) - goto Done; - - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, - tmp); - if (r) - goto Done; - - /* stall UMC channel */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, - tmp); - if (r) - goto Done; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, + UVD_STATUS__IDLE, 0x7); + if (r) + goto Done; + + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, + tmp); + if (r) + goto Done; + + /* stall UMC channel */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, + tmp); + if (r) + goto Done; - /* Unblock VCPU Register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* Unblock VCPU Register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - /* reset LMI UMC/LMI/VCPU */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + /* reset LMI UMC/LMI/VCPU */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - /* clear VCN status */ - WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); + /* clear VCN status */ + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); + + /* apply HW clock gating */ + vcn_v4_0_3_enable_clock_gating(adev, i); - /* apply HW clock gating */ - vcn_v4_0_3_enable_clock_gating(adev, i); - } Done: return 0; } @@ -1653,7 +1650,7 @@ static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating @@ -1667,10 +1664,12 @@ static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block, if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v4_0_3_stop(adev); - else - ret = vcn_v4_0_3_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v4_0_3_stop(adev, i); + else + ret |= vcn_v4_0_3_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From ecc9ab4e924b7eb9e2c4a668162aaa1d9d60d08c Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 12:21:18 -0500 Subject: [PATCH 09/16] drm/amdgpu/vcn4.0.5: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. v2: squash in fix for stop() from Boyuan Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 444 ++++++++++++------------ 1 file changed, 220 insertions(+), 224 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 5e6ad825cc47..c2bbacb176c1 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -990,183 +990,178 @@ static int vcn_v4_0_5_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, b * vcn_v4_0_5_start - VCN start * * @adev: amdgpu_device pointer + * @i: instance to start * * Start VCN block */ -static int vcn_v4_0_5_start(struct amdgpu_device *adev) +static int vcn_v4_0_5_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t tmp; - int i, j, k, r; + int j, k, r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, true, i); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v4_0_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v4_0_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - /* disable VCN power gating */ - vcn_v4_0_5_disable_static_power_gating(adev, i); - - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); - - /*SW clock gating */ - vcn_v4_0_5_disable_clock_gating(adev, i); - - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - - /* setup regUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - /* setup regUVD_MPC_CNTL */ - tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL); - tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; - tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; - WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp); - - /* setup UVD_MPC_SET_MUXA0 */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0, - ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); - - /* setup UVD_MPC_SET_MUXB0 */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0, - ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | - (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | - (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); - - /* setup UVD_MPC_SET_MUX */ - WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX, - ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | - (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | - (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - - vcn_v4_0_5_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - for (j = 0; j < 10; ++j) { - uint32_t status; - - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, i, regUVD_STATUS); - if (status & 2) - break; - mdelay(10); - if (amdgpu_emu_mode == 1) - msleep(1); - } + /* disable VCN power gating */ + vcn_v4_0_5_disable_static_power_gating(adev, i); + + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); - if (amdgpu_emu_mode == 1) { - r = -1; - if (status & 2) { - r = 0; - break; - } - } else { + /* SW clock gating */ + vcn_v4_0_5_disable_clock_gating(adev, i); + + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); + + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + /* setup regUVD_MPC_CNTL */ + tmp = RREG32_SOC15(VCN, i, regUVD_MPC_CNTL); + tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK; + tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT; + WREG32_SOC15(VCN, i, regUVD_MPC_CNTL, tmp); + + /* setup UVD_MPC_SET_MUXA0 */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXA0, + ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT))); + + /* setup UVD_MPC_SET_MUXB0 */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUXB0, + ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) | + (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) | + (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT))); + + /* setup UVD_MPC_SET_MUX */ + WREG32_SOC15(VCN, i, regUVD_MPC_SET_MUX, + ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) | + (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | + (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); + + vcn_v4_0_5_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, i, regUVD_STATUS); + if (status & 2) + break; + mdelay(10); + if (amdgpu_emu_mode == 1) + msleep(1); + } + + if (amdgpu_emu_mode == 1) { + r = -1; + if (status & 2) { r = 0; - if (status & 2) - break; - - dev_err(adev->dev, - "VCN[%d] is not responding, trying to reset VCPU!!!\n", i); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - mdelay(10); - r = -1; + break; } + } else { + r = 0; + if (status & 2) + break; + + dev_err(adev->dev, + "VCN[%d] is not responding, trying to reset VCPU!!!\n", i); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + mdelay(10); + r = -1; } + } - if (r) { - dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); - return r; - } + if (r) { + dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); + return r; + } - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, - ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | - VCN_RB1_DB_CTRL__EN_MASK); - - WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; - WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); - - tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); - ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp |= VCN_RB_ENABLE__RB1_EN_MASK; - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); - } + ring = &adev->vcn.inst[i].ring_enc[0]; + WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, + ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); + + WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; + WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); + + tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); + ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB1_EN_MASK; + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); return 0; } @@ -1203,88 +1198,87 @@ static void vcn_v4_0_5_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) * vcn_v4_0_5_stop - VCN stop * * @adev: amdgpu_device pointer + * @i: instance to stop * * Stop VCN block */ -static int vcn_v4_0_5_stop(struct amdgpu_device *adev) +static int vcn_v4_0_5_stop(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; uint32_t tmp; - int i, r = 0; + int r = 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + if (adev->vcn.harvest_config & (1 << i)) + return 0; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - vcn_v4_0_5_stop_dpg_mode(adev, i); - continue; - } + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v4_0_5_stop_dpg_mode(adev, i); + r = 0; + goto done; + } - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + goto done; - /* disable LMI UMC channel */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; + + /* disable LMI UMC channel */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - /* apply soft reset */ - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + /* apply soft reset */ + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - /* clear status */ - WREG32_SOC15(VCN, i, regUVD_STATUS, 0); + /* clear status */ + WREG32_SOC15(VCN, i, regUVD_STATUS, 0); - /* apply HW clock gating */ - vcn_v4_0_5_enable_clock_gating(adev, i); + /* apply HW clock gating */ + vcn_v4_0_5_enable_clock_gating(adev, i); - /* enable VCN power gating */ - vcn_v4_0_5_enable_static_power_gating(adev, i); - } + /* enable VCN power gating */ + vcn_v4_0_5_enable_static_power_gating(adev, i); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } +done: + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, false, i); - return 0; + return r; } /** @@ -1539,15 +1533,17 @@ static int vcn_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v4_0_5_stop(adev); - else - ret = vcn_v4_0_5_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v4_0_5_stop(adev, i); + else + ret |= vcn_v4_0_5_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From 0797c54502d6cdb806aaf8d482b3ed445d6e298f Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 12:27:45 -0500 Subject: [PATCH 10/16] drm/amdgpu/vcn5.0.0: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. v2: squash in fix for stop() from Boyuan Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c | 390 ++++++++++++------------ 1 file changed, 193 insertions(+), 197 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index e937ed682abd..188cd6f1dda2 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -767,154 +767,149 @@ static int vcn_v5_0_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, b * vcn_v5_0_0_start - VCN start * * @adev: amdgpu_device pointer + * @i: instance to start * * Start VCN block */ -static int vcn_v5_0_0_start(struct amdgpu_device *adev) +static int vcn_v5_0_0_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn5_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t tmp; - int i, j, k, r; + int j, k, r; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, true, i); - } + if (adev->vcn.harvest_config & (1 << i)) + return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, true, i); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v5_0_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v5_0_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - /* disable VCN power gating */ - vcn_v5_0_0_disable_static_power_gating(adev, i); - - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); - - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - - /* setup regUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - vcn_v5_0_0_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - for (j = 0; j < 10; ++j) { - uint32_t status; - - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, i, regUVD_STATUS); - if (status & 2) - break; - mdelay(10); - if (amdgpu_emu_mode == 1) - msleep(1); - } + /* disable VCN power gating */ + vcn_v5_0_0_disable_static_power_gating(adev, i); + + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, i, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, i, regUVD_STATUS, tmp); + + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); + + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - if (amdgpu_emu_mode == 1) { - r = -1; - if (status & 2) { - r = 0; - break; - } - } else { + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL, tmp | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + vcn_v5_0_0_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, i, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, i, regUVD_STATUS); + if (status & 2) + break; + mdelay(10); + if (amdgpu_emu_mode == 1) + msleep(1); + } + + if (amdgpu_emu_mode == 1) { + r = -1; + if (status & 2) { r = 0; - if (status & 2) - break; - - dev_err(adev->dev, - "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - mdelay(10); - r = -1; + break; } + } else { + r = 0; + if (status & 2) + break; + + dev_err(adev->dev, + "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + mdelay(10); + r = -1; } + } - if (r) { - dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); - return r; - } + if (r) { + dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); + return r; + } - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, - ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | - VCN_RB1_DB_CTRL__EN_MASK); - - WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; - WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); - - tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); - WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); - ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); - - tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); - tmp |= VCN_RB_ENABLE__RB1_EN_MASK; - WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); - } + ring = &adev->vcn.inst[i].ring_enc[0]; + WREG32_SOC15(VCN, i, regVCN_RB1_DB_CTRL, + ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); + + WREG32_SOC15(VCN, i, regUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, i, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, i, regUVD_RB_SIZE, ring->ring_size / 4); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; + WREG32_SOC15(VCN, i, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, 0); + + tmp = RREG32_SOC15(VCN, i, regUVD_RB_RPTR); + WREG32_SOC15(VCN, i, regUVD_RB_WPTR, tmp); + ring->wptr = RREG32_SOC15(VCN, i, regUVD_RB_WPTR); + + tmp = RREG32_SOC15(VCN, i, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB1_EN_MASK; + WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); return 0; } @@ -953,85 +948,84 @@ static void vcn_v5_0_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) * vcn_v5_0_0_stop - VCN stop * * @adev: amdgpu_device pointer + * @i: instance to stop * * Stop VCN block */ -static int vcn_v5_0_0_stop(struct amdgpu_device *adev) +static int vcn_v5_0_0_stop(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn5_fw_shared *fw_shared; uint32_t tmp; - int i, r = 0; + int r = 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->vcn.harvest_config & (1 << i)) + return 0; - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - vcn_v5_0_0_stop_dpg_mode(adev, i); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v5_0_0_stop_dpg_mode(adev, i); + r = 0; + goto done; + } - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + goto done; - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; + + /* disable LMI UMC channel */ + tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); + if (r) + goto done; - /* disable LMI UMC channel */ - tmp = RREG32_SOC15(VCN, i, regUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, i, regUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, i, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - - /* apply soft reset */ - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); - - /* clear status */ - WREG32_SOC15(VCN, i, regUVD_STATUS, 0); - - /* enable VCN power gating */ - vcn_v5_0_0_enable_static_power_gating(adev, i); - } + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->pm.dpm_enabled) - amdgpu_dpm_enable_vcn(adev, false, i); - } + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, i, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - return 0; + /* apply soft reset */ + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, i, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, i, regUVD_SOFT_RESET, tmp); + + /* clear status */ + WREG32_SOC15(VCN, i, regUVD_STATUS, 0); + + /* enable VCN power gating */ + vcn_v5_0_0_enable_static_power_gating(adev, i); + +done: + if (adev->pm.dpm_enabled) + amdgpu_dpm_enable_vcn(adev, false, i); + + return r; } /** @@ -1279,15 +1273,17 @@ static int vcn_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v5_0_0_stop(adev); - else - ret = vcn_v5_0_0_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret |= vcn_v5_0_0_stop(adev, i); + else + ret |= vcn_v5_0_0_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From 94629182f3096fb8beafabc7e7cba097630185b7 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 10 Dec 2024 12:34:54 -0500 Subject: [PATCH 11/16] drm/amdgpu/vcn5.0.1: split code along instances Split the code on a per instance basis. This will allow us to use the per instance functions in the future to handle more things per instance. Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 352 ++++++++++++------------ 1 file changed, 174 insertions(+), 178 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index 40f36c9a4132..bfe48e4a4859 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -568,146 +568,142 @@ static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, b * * Start VCN block */ -static int vcn_v5_0_1_start(struct amdgpu_device *adev) +static int vcn_v5_0_1_start(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t tmp; - int i, j, k, r, vcn_inst; + int j, k, r, vcn_inst; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v5_0_1_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + return vcn_v5_0_1_start_dpg_mode(adev, i, adev->vcn.indirect_sram); - vcn_inst = GET_INST(VCN, i); + vcn_inst = GET_INST(VCN, i); - /* set VCN status busy */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; - WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp); - - /* enable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); - - /* disable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0, - ~UVD_MASTINT_EN__VCPU_EN_MASK); - - /* enable LMI MC and UMC channels */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0, - ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); - - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - - /* setup regUVD_LMI_CTRL */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL); - WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, tmp | - UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | - UVD_LMI_CTRL__MASK_MC_URGENT_MASK | - UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | - UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); - - vcn_v5_0_1_mc_resume(adev, i); - - /* VCN global tiling registers */ - WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG, - adev->gfx.config.gb_addr_config); - - /* unblock VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* release VCPU reset to boot */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - for (j = 0; j < 10; ++j) { - uint32_t status; - - for (k = 0; k < 100; ++k) { - status = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS); - if (status & 2) - break; - mdelay(100); - if (amdgpu_emu_mode == 1) - msleep(20); - } + /* set VCN status busy */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS) | UVD_STATUS__UVD_BUSY; + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, tmp); + + /* enable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK); + + /* disable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), 0, + ~UVD_MASTINT_EN__VCPU_EN_MASK); + + /* enable LMI MC and UMC channels */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_LMI_CTRL2), 0, + ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); + + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + /* setup regUVD_LMI_CTRL */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL); + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL, tmp | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__MASK_MC_URGENT_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK); + + vcn_v5_0_1_mc_resume(adev, i); + + /* VCN global tiling registers */ + WREG32_SOC15(VCN, vcn_inst, regUVD_GFX10_ADDR_CONFIG, + adev->gfx.config.gb_addr_config); + + /* unblock VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), 0, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* release VCPU reset to boot */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + for (j = 0; j < 10; ++j) { + uint32_t status; + + for (k = 0; k < 100; ++k) { + status = RREG32_SOC15(VCN, vcn_inst, regUVD_STATUS); + if (status & 2) + break; + mdelay(100); + if (amdgpu_emu_mode == 1) + msleep(20); + } - if (amdgpu_emu_mode == 1) { - r = -1; - if (status & 2) { - r = 0; - break; - } - } else { + if (amdgpu_emu_mode == 1) { + r = -1; + if (status & 2) { r = 0; - if (status & 2) - break; - - dev_err(adev->dev, - "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - mdelay(10); - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - mdelay(10); - r = -1; + break; } + } else { + r = 0; + if (status & 2) + break; + + dev_err(adev->dev, + "VCN[%d] is not responding, trying to reset the VCPU!!!\n", i); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + mdelay(10); + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + mdelay(10); + r = -1; } + } - if (r) { - dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); - return r; - } + if (r) { + dev_err(adev->dev, "VCN[%d] is not responding, giving up!!!\n", i); + return r; + } - /* enable master interrupt */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), - UVD_MASTINT_EN__VCPU_EN_MASK, - ~UVD_MASTINT_EN__VCPU_EN_MASK); + /* enable master interrupt */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_MASTINT_EN), + UVD_MASTINT_EN__VCPU_EN_MASK, + ~UVD_MASTINT_EN__VCPU_EN_MASK); - /* clear the busy bit of VCN_STATUS */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0, - ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); + /* clear the busy bit of VCN_STATUS */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_STATUS), 0, + ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); - ring = &adev->vcn.inst[i].ring_enc[0]; + ring = &adev->vcn.inst[i].ring_enc[0]; - WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL, - ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | - VCN_RB1_DB_CTRL__EN_MASK); + WREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL, + ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT | + VCN_RB1_DB_CTRL__EN_MASK); - /* Read DB_CTRL to flush the write DB_CTRL command. */ - RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL); + /* Read DB_CTRL to flush the write DB_CTRL command. */ + RREG32_SOC15(VCN, vcn_inst, regVCN_RB1_DB_CTRL); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, ring->gpu_addr); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / 4); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, ring->gpu_addr); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_SIZE, ring->ring_size / 4); - tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); - tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); - WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp &= ~(VCN_RB_ENABLE__RB1_EN_MASK); + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode |= FW_QUEUE_RING_RESET; + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR, 0); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, 0); - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR); - WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp); - ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_RPTR); + WREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR, tmp); + ring->wptr = RREG32_SOC15(VCN, vcn_inst, regUVD_RB_WPTR); - tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); - tmp |= VCN_RB_ENABLE__RB1_EN_MASK; - WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); - fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); - } + tmp = RREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE); + tmp |= VCN_RB_ENABLE__RB1_EN_MASK; + WREG32_SOC15(VCN, vcn_inst, regVCN_RB_ENABLE, tmp); + fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF); return 0; } @@ -747,71 +743,69 @@ static void vcn_v5_0_1_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) * * Stop VCN block */ -static int vcn_v5_0_1_stop(struct amdgpu_device *adev) +static int vcn_v5_0_1_stop(struct amdgpu_device *adev, int i) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; uint32_t tmp; - int i, r = 0, vcn_inst; + int r = 0, vcn_inst; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - vcn_inst = GET_INST(VCN, i); + vcn_inst = GET_INST(VCN, i); - fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; - fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; + fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; + fw_shared->sq.queue_mode |= FW_QUEUE_DPG_HOLD_OFF; - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - vcn_v5_0_1_stop_dpg_mode(adev, i); - continue; - } + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { + vcn_v5_0_1_stop_dpg_mode(adev, i); + return 0; + } - /* wait for vcn idle */ - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); - if (r) - return r; + /* wait for vcn idle */ + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_STATUS, UVD_STATUS__IDLE, 0x7); + if (r) + return r; - tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | - UVD_LMI_STATUS__READ_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_MASK | - UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK | + UVD_LMI_STATUS__READ_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_MASK | + UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp); + if (r) + return r; - /* disable LMI UMC channel */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2); - tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp); - tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | - UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; - r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp); - if (r) - return r; + /* disable LMI UMC channel */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2); + tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_LMI_CTRL2, tmp); + tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK | + UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK; + r = SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_LMI_STATUS, tmp, tmp); + if (r) + return r; - /* block VCPU register access */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), - UVD_RB_ARB_CTRL__VCPU_DIS_MASK, - ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); - - /* reset VCPU */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), - UVD_VCPU_CNTL__BLK_RST_MASK, - ~UVD_VCPU_CNTL__BLK_RST_MASK); - - /* disable VCPU clock */ - WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, - ~(UVD_VCPU_CNTL__CLK_EN_MASK)); - - /* apply soft reset */ - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); - tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; - WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); - - /* clear status */ - WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); - } + /* block VCPU register access */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_RB_ARB_CTRL), + UVD_RB_ARB_CTRL__VCPU_DIS_MASK, + ~UVD_RB_ARB_CTRL__VCPU_DIS_MASK); + + /* reset VCPU */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), + UVD_VCPU_CNTL__BLK_RST_MASK, + ~UVD_VCPU_CNTL__BLK_RST_MASK); + + /* disable VCPU clock */ + WREG32_P(SOC15_REG_OFFSET(VCN, vcn_inst, regUVD_VCPU_CNTL), 0, + ~(UVD_VCPU_CNTL__CLK_EN_MASK)); + + /* apply soft reset */ + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + tmp = RREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET); + tmp |= UVD_SOFT_RESET__LMI_SOFT_RESET_MASK; + WREG32_SOC15(VCN, vcn_inst, regUVD_SOFT_RESET, tmp); + + /* clear status */ + WREG32_SOC15(VCN, vcn_inst, regUVD_STATUS, 0); return 0; } @@ -1004,15 +998,17 @@ static int vcn_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block, enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; - int ret; + int ret = 0, i; if (state == adev->vcn.cur_state) return 0; - if (state == AMD_PG_STATE_GATE) - ret = vcn_v5_0_1_stop(adev); - else - ret = vcn_v5_0_1_start(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + if (state == AMD_PG_STATE_GATE) + ret = vcn_v5_0_1_stop(adev, i); + else + ret = vcn_v5_0_1_start(adev, i); + } if (!ret) adev->vcn.cur_state = state; -- 2.51.0 From bee48570cf8e312faec9b8f465762052555158d0 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 14:43:15 -0500 Subject: [PATCH 12/16] drm/amdgpu/vcn: switch work handler to be per instance Have a separate work handler for each VCN instance. This paves the way for per instance VCN power gating at runtime. v2: index instances directly on vcn1.0 and 2.0 to make it clear that they only support a single instance (Lijo) Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 84 +++++++++++++------------ drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 8 ++- drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 17 ++--- drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 4 +- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 4 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 5 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 4 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 5 +- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c | 5 +- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 4 +- 12 files changed, 80 insertions(+), 64 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 83faf6e6788a..0c3081a198d9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -100,6 +100,9 @@ int amdgpu_vcn_early_init(struct amdgpu_device *adev) amdgpu_ucode_ip_version_decode(adev, UVD_HWIP, ucode_prefix, sizeof(ucode_prefix)); for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + adev->vcn.inst[i].adev = adev; + adev->vcn.inst[i].inst = i; + if (i == 1 && amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 6)) r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw, AMDGPU_UCODE_REQUIRED, @@ -124,12 +127,13 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) unsigned int fw_shared_size, log_offset; int i, r; - INIT_DELAYED_WORK(&adev->vcn.idle_work, amdgpu_vcn_idle_work_handler); - mutex_init(&adev->vcn.vcn_pg_lock); mutex_init(&adev->vcn.vcn1_jpeg1_workaround); - atomic_set(&adev->vcn.total_submission_cnt, 0); - for (i = 0; i < adev->vcn.num_vcn_inst; i++) + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + mutex_init(&adev->vcn.inst[i].vcn_pg_lock); + atomic_set(&adev->vcn.inst[i].total_submission_cnt, 0); + INIT_DELAYED_WORK(&adev->vcn.inst[i].idle_work, amdgpu_vcn_idle_work_handler); atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0); + } if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) @@ -277,10 +281,10 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) amdgpu_ring_fini(&adev->vcn.inst[j].ring_enc[i]); amdgpu_ucode_release(&adev->vcn.inst[j].fw); + mutex_destroy(&adev->vcn.inst[j].vcn_pg_lock); } mutex_destroy(&adev->vcn.vcn1_jpeg1_workaround); - mutex_destroy(&adev->vcn.vcn_pg_lock); return 0; } @@ -331,8 +335,10 @@ int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev) int amdgpu_vcn_suspend(struct amdgpu_device *adev) { bool in_ras_intr = amdgpu_ras_intr_triggered(); + int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); /* err_event_athub will corrupt VCPU buffer, so we need to * restore fw data and clear buffer in amdgpu_vcn_resume() */ @@ -388,46 +394,45 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev) static void amdgpu_vcn_idle_work_handler(struct work_struct *work) { - struct amdgpu_device *adev = - container_of(work, struct amdgpu_device, vcn.idle_work.work); + struct amdgpu_vcn_inst *vcn_inst = + container_of(work, struct amdgpu_vcn_inst, idle_work.work); + struct amdgpu_device *adev = vcn_inst->adev; unsigned int fences = 0, fence[AMDGPU_MAX_VCN_INSTANCES] = {0}; - unsigned int i, j; + unsigned int i = vcn_inst->inst, j; int r = 0; - for (j = 0; j < adev->vcn.num_vcn_inst; ++j) { - if (adev->vcn.harvest_config & (1 << j)) - continue; - - for (i = 0; i < adev->vcn.num_enc_rings; ++i) - fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_enc[i]); + if (adev->vcn.harvest_config & (1 << i)) + return; - /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && - !adev->vcn.using_unified_queue) { - struct dpg_pause_state new_state; + for (j = 0; j < adev->vcn.num_enc_rings; ++j) + fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_enc[j]); - if (fence[j] || - unlikely(atomic_read(&adev->vcn.inst[j].dpg_enc_submission_cnt))) - new_state.fw_based = VCN_DPG_STATE__PAUSE; - else - new_state.fw_based = VCN_DPG_STATE__UNPAUSE; + /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && + !adev->vcn.using_unified_queue) { + struct dpg_pause_state new_state; - adev->vcn.pause_dpg_mode(adev, j, &new_state); - } + if (fence[i] || + unlikely(atomic_read(&vcn_inst->dpg_enc_submission_cnt))) + new_state.fw_based = VCN_DPG_STATE__PAUSE; + else + new_state.fw_based = VCN_DPG_STATE__UNPAUSE; - fence[j] += amdgpu_fence_count_emitted(&adev->vcn.inst[j].ring_dec); - fences += fence[j]; + adev->vcn.pause_dpg_mode(adev, i, &new_state); } - if (!fences && !atomic_read(&adev->vcn.total_submission_cnt)) { + fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_dec); + fences += fence[i]; + + if (!fences && !atomic_read(&vcn_inst->total_submission_cnt)) { amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN, - AMD_PG_STATE_GATE); + AMD_PG_STATE_GATE); r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO, - false); + false); if (r) dev_warn(adev->dev, "(%d) failed to disable video power profile mode\n", r); } else { - schedule_delayed_work(&adev->vcn.idle_work, VCN_IDLE_TIMEOUT); + schedule_delayed_work(&vcn_inst->idle_work, VCN_IDLE_TIMEOUT); } } @@ -436,18 +441,18 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) struct amdgpu_device *adev = ring->adev; int r = 0; - atomic_inc(&adev->vcn.total_submission_cnt); + atomic_inc(&adev->vcn.inst[ring->me].total_submission_cnt); - if (!cancel_delayed_work_sync(&adev->vcn.idle_work)) { + if (!cancel_delayed_work_sync(&adev->vcn.inst[ring->me].idle_work)) { r = amdgpu_dpm_switch_power_profile(adev, PP_SMC_POWER_PROFILE_VIDEO, true); if (r) dev_warn(adev->dev, "(%d) failed to switch to video power profile mode\n", r); } - mutex_lock(&adev->vcn.vcn_pg_lock); + mutex_lock(&adev->vcn.inst[ring->me].vcn_pg_lock); amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN, - AMD_PG_STATE_UNGATE); + AMD_PG_STATE_UNGATE); /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && @@ -472,7 +477,7 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) adev->vcn.pause_dpg_mode(adev, ring->me, &new_state); } - mutex_unlock(&adev->vcn.vcn_pg_lock); + mutex_unlock(&adev->vcn.inst[ring->me].vcn_pg_lock); } void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring) @@ -485,9 +490,10 @@ void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring) !adev->vcn.using_unified_queue) atomic_dec(&ring->adev->vcn.inst[ring->me].dpg_enc_submission_cnt); - atomic_dec(&ring->adev->vcn.total_submission_cnt); + atomic_dec(&ring->adev->vcn.inst[ring->me].total_submission_cnt); - schedule_delayed_work(&ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT); + schedule_delayed_work(&ring->adev->vcn.inst[ring->me].idle_work, + VCN_IDLE_TIMEOUT); } int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index c92f683ee595..f1685283c2f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -295,6 +295,8 @@ struct amdgpu_vcn_fw_shared { }; struct amdgpu_vcn_inst { + struct amdgpu_device *adev; + int inst; struct amdgpu_bo *vcpu_bo; void *cpu_addr; uint64_t gpu_addr; @@ -316,6 +318,9 @@ struct amdgpu_vcn_inst { const struct firmware *fw; /* VCN firmware */ uint8_t vcn_config; uint32_t vcn_codec_disable_mask; + atomic_t total_submission_cnt; + struct mutex vcn_pg_lock; + struct delayed_work idle_work; }; struct amdgpu_vcn_ras { @@ -324,7 +329,6 @@ struct amdgpu_vcn_ras { struct amdgpu_vcn { unsigned fw_version; - struct delayed_work idle_work; unsigned num_enc_rings; enum amd_powergating_state cur_state; bool indirect_sram; @@ -332,9 +336,7 @@ struct amdgpu_vcn { uint8_t num_vcn_inst; struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES]; struct amdgpu_vcn_reg internal; - struct mutex vcn_pg_lock; struct mutex vcn1_jpeg1_workaround; - atomic_t total_submission_cnt; unsigned harvest_config; int (*pause_dpg_mode)(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c index 03b8b7cd5229..8031406e20ff 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c @@ -604,7 +604,7 @@ static void jpeg_v1_0_set_irq_funcs(struct amdgpu_device *adev) static void jpeg_v1_0_ring_begin_use(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.idle_work); + bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); int cnt = 0; mutex_lock(&adev->vcn.vcn1_jpeg1_workaround); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 06e1bbcf1a52..cccf8e5b0037 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -150,7 +150,7 @@ static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block) return r; /* Override the work func */ - adev->vcn.idle_work.work.func = vcn_v1_0_idle_work_handler; + adev->vcn.inst[0].idle_work.work.func = vcn_v1_0_idle_work_handler; amdgpu_vcn_setup_ucode(adev); @@ -277,7 +277,7 @@ static int vcn_v1_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - cancel_delayed_work_sync(&adev->vcn.idle_work); + cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && @@ -301,7 +301,7 @@ static int vcn_v1_0_suspend(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; bool idle_work_unexecuted; - idle_work_unexecuted = cancel_delayed_work_sync(&adev->vcn.idle_work); + idle_work_unexecuted = cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); if (idle_work_unexecuted) { if (adev->pm.dpm_enabled) amdgpu_dpm_enable_vcn(adev, false, 0); @@ -1828,8 +1828,9 @@ static int vcn_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block, static void vcn_v1_0_idle_work_handler(struct work_struct *work) { - struct amdgpu_device *adev = - container_of(work, struct amdgpu_device, vcn.idle_work.work); + struct amdgpu_vcn_inst *vcn_inst = + container_of(work, struct amdgpu_vcn_inst, idle_work.work); + struct amdgpu_device *adev = vcn_inst->adev; unsigned int fences = 0, i; for (i = 0; i < adev->vcn.num_enc_rings; ++i) @@ -1862,14 +1863,14 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work) amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN, AMD_PG_STATE_GATE); } else { - schedule_delayed_work(&adev->vcn.idle_work, VCN_IDLE_TIMEOUT); + schedule_delayed_work(&adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT); } } static void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.idle_work); + bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); mutex_lock(&adev->vcn.vcn1_jpeg1_workaround); @@ -1921,7 +1922,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks) void vcn_v1_0_ring_end_use(struct amdgpu_ring *ring) { - schedule_delayed_work(&ring->adev->vcn.idle_work, VCN_IDLE_TIMEOUT); + schedule_delayed_work(&ring->adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT); mutex_unlock(&ring->adev->vcn.vcn1_jpeg1_workaround); } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index 1e1c8f57d4cd..86d20141f39c 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -313,7 +313,7 @@ static int vcn_v2_0_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - cancel_delayed_work_sync(&adev->vcn.idle_work); + cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index e36e2a5676df..aee236aaa7d0 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -390,12 +390,12 @@ static int vcn_v2_5_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, i, mmUVD_STATUS))) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index ec42a456c541..dd3f3c409105 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -422,12 +422,12 @@ static int vcn_v3_0_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index c99c121faac0..4d20d58802f6 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -359,11 +359,12 @@ static int vcn_v4_0_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; + + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index 00b94d01cd4c..932b65072dbc 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -349,8 +349,10 @@ static int vcn_v4_0_3_hw_init(struct amdgpu_ip_block *ip_block) static int vcn_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); if (adev->vcn.cur_state != AMD_PG_STATE_GATE) vcn_v4_0_3_set_powergating_state(ip_block, AMD_PG_STATE_GATE); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index c2bbacb176c1..83093e5c5cca 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -300,11 +300,12 @@ static int vcn_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; + + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index 188cd6f1dda2..26f788ceeb11 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -280,11 +280,12 @@ static int vcn_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; + + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || (adev->vcn.cur_state != AMD_PG_STATE_GATE && diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index bfe48e4a4859..512e5f6ae7b3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -206,8 +206,10 @@ static int vcn_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block) static int vcn_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - cancel_delayed_work_sync(&adev->vcn.idle_work); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); return 0; } -- 2.51.0 From 9bf9442051c0af835b683411d9ebc6a5036d4cbb Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 13 Nov 2024 15:28:41 -0500 Subject: [PATCH 13/16] drm/amdgpu/vcn: make powergating status per instance Store it per instance so we can track it per instance. v2: index instances directly on vcn1.0 and 2.0 to make it clear that they only support a single instance (Lijo) Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 2 +- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 6 +-- drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 11 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 43 +++++++++++----- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 49 ++++++++++++------ drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 60 ++++++++++++--------- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 63 ++++++++++++++--------- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 48 +++++++++++------ drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c | 48 +++++++++++------ drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 40 +++++++++----- drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c | 8 +-- 11 files changed, 241 insertions(+), 137 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index f1685283c2f1..0bcf4d1c7175 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -320,6 +320,7 @@ struct amdgpu_vcn_inst { uint32_t vcn_codec_disable_mask; atomic_t total_submission_cnt; struct mutex vcn_pg_lock; + enum amd_powergating_state cur_state; struct delayed_work idle_work; }; @@ -330,7 +331,6 @@ struct amdgpu_vcn_ras { struct amdgpu_vcn { unsigned fw_version; unsigned num_enc_rings; - enum amd_powergating_state cur_state; bool indirect_sram; uint8_t num_vcn_inst; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index cccf8e5b0037..291a5866c148 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -280,7 +280,7 @@ static int vcn_v1_0_hw_fini(struct amdgpu_ip_block *ip_block) cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && + (adev->vcn.inst[0].cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, 0, mmUVD_STATUS))) { vcn_v1_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE); } @@ -1813,7 +1813,7 @@ static int vcn_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block, int ret; struct amdgpu_device *adev = ip_block->adev; - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[0].cur_state) return 0; if (state == AMD_PG_STATE_GATE) @@ -1822,7 +1822,7 @@ static int vcn_v1_0_set_powergating_state(struct amdgpu_ip_block *ip_block, ret = vcn_v1_0_start(adev); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[0].cur_state = state; return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index 86d20141f39c..cc2b9dcc8868 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -316,8 +316,8 @@ static int vcn_v2_0_hw_fini(struct amdgpu_ip_block *ip_block) cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && - RREG32_SOC15(VCN, 0, mmUVD_STATUS))) + (adev->vcn.inst[0].cur_state != AMD_PG_STATE_GATE && + RREG32_SOC15(VCN, 0, mmUVD_STATUS))) vcn_v2_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE); return 0; @@ -1810,11 +1810,11 @@ static int vcn_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block, struct amdgpu_device *adev = ip_block->adev; if (amdgpu_sriov_vf(adev)) { - adev->vcn.cur_state = AMD_PG_STATE_UNGATE; + adev->vcn.inst[0].cur_state = AMD_PG_STATE_UNGATE; return 0; } - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[0].cur_state) return 0; if (state == AMD_PG_STATE_GATE) @@ -1823,7 +1823,8 @@ static int vcn_v2_0_set_powergating_state(struct amdgpu_ip_block *ip_block, ret = vcn_v2_0_start(adev); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[0].cur_state = state; + return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index aee236aaa7d0..0d78377f8ffe 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -95,8 +95,11 @@ static const struct amdgpu_hwip_reg_entry vcn_reg_list_2_5[] = { static void vcn_v2_5_set_dec_ring_funcs(struct amdgpu_device *adev); static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev); static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v2_5_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v2_5_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); static int vcn_v2_5_sriov_start(struct amdgpu_device *adev); @@ -397,9 +400,9 @@ static int vcn_v2_5_hw_fini(struct amdgpu_ip_block *ip_block) cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && + (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, i, mmUVD_STATUS))) - vcn_v2_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + vcn_v2_5_set_powergating_state_inst(ip_block, AMD_PG_STATE_GATE, i); if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) amdgpu_irq_put(adev, &adev->vcn.inst[i].ras_poison_irq, 0); @@ -1808,27 +1811,39 @@ static int vcn_v2_5_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) +static int vcn_v2_5_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) { struct amdgpu_device *adev = ip_block->adev; - int ret = 0, i; + int ret; if (amdgpu_sriov_vf(adev)) return 0; - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[i].cur_state) return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v2_5_stop(adev, i); - else - ret |= vcn_v2_5_start(adev, i); - } + if (state == AMD_PG_STATE_GATE) + ret = vcn_v2_5_stop(adev, i); + else + ret = vcn_v2_5_start(adev, i); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[i].cur_state = state; + + return ret; +} + +static int vcn_v2_5_set_powergating_state(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0, i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + ret |= vcn_v2_5_set_powergating_state_inst(ip_block, + state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index dd3f3c409105..acfe85b0c062 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -105,8 +105,11 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev); static void vcn_v3_0_set_dec_ring_funcs(struct amdgpu_device *adev); static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev); static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v3_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); @@ -430,9 +433,11 @@ static int vcn_v3_0_hw_fini(struct amdgpu_ip_block *ip_block) if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && - RREG32_SOC15(VCN, i, mmUVD_STATUS))) { - vcn_v3_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE && + RREG32_SOC15(VCN, i, mmUVD_STATUS))) { + vcn_v3_0_set_powergating_state_inst(ip_block, + AMD_PG_STATE_GATE, + i); } } } @@ -2150,33 +2155,45 @@ static int vcn_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) +static int vcn_v3_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) { struct amdgpu_device *adev = ip_block->adev; - int ret = 0, i; + int ret = 0; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating * guest should avoid touching CGC and PG */ if (amdgpu_sriov_vf(adev)) { - adev->vcn.cur_state = AMD_PG_STATE_UNGATE; + adev->vcn.inst[i].cur_state = AMD_PG_STATE_UNGATE; return 0; } - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[i].cur_state) return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v3_0_stop(adev, i); - else - ret |= vcn_v3_0_start(adev, i); - } + if (state == AMD_PG_STATE_GATE) + ret = vcn_v3_0_stop(adev, i); + else + ret = vcn_v3_0_start(adev, i); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[i].cur_state = state; + + return ret; +} + +static int vcn_v3_0_set_powergating_state(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0, i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + ret |= vcn_v3_0_set_powergating_state_inst(ip_block, + state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 4d20d58802f6..1f67e5973cf8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -96,8 +96,11 @@ static int amdgpu_ih_clientid_vcns[] = { static int vcn_v4_0_start_sriov(struct amdgpu_device *adev); static void vcn_v4_0_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v4_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v4_0_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); static void vcn_v4_0_unified_ring_set_wptr(struct amdgpu_ring *ring); @@ -367,9 +370,9 @@ static int vcn_v4_0_hw_fini(struct amdgpu_ip_block *ip_block) if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && + (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE && RREG32_SOC15(VCN, i, regUVD_STATUS))) { - vcn_v4_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + vcn_v4_0_set_powergating_state_inst(ip_block, AMD_PG_STATE_GATE, i); } } if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) @@ -2034,41 +2037,52 @@ static int vcn_v4_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -/** - * vcn_v4_0_set_powergating_state - set VCN block powergating state - * - * @ip_block: amdgpu_ip_block pointer - * @state: power gating state - * - * Set VCN block powergating state - */ -static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) +static int vcn_v4_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) { struct amdgpu_device *adev = ip_block->adev; - int ret = 0, i; + int ret = 0; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating * guest should avoid touching CGC and PG */ if (amdgpu_sriov_vf(adev)) { - adev->vcn.cur_state = AMD_PG_STATE_UNGATE; + adev->vcn.inst[i].cur_state = AMD_PG_STATE_UNGATE; return 0; } - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[i].cur_state) return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v4_0_stop(adev, i); - else - ret |= vcn_v4_0_start(adev, i); - } + if (state == AMD_PG_STATE_GATE) + ret = vcn_v4_0_stop(adev, i); + else + ret = vcn_v4_0_start(adev, i); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[i].cur_state = state; + + return ret; +} + +/** + * vcn_v4_0_set_powergating_state - set VCN block powergating state + * + * @ip_block: amdgpu_ip_block pointer + * @state: power gating state + * + * Set VCN block powergating state + */ +static int vcn_v4_0_set_powergating_state(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0, i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + ret |= vcn_v4_0_set_powergating_state_inst(ip_block, state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index 932b65072dbc..73b10813d9b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -87,8 +87,11 @@ static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_3[] = { static int vcn_v4_0_3_start_sriov(struct amdgpu_device *adev); static void vcn_v4_0_3_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v4_0_3_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v4_0_3_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v4_0_3_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); static void vcn_v4_0_3_unified_ring_set_wptr(struct amdgpu_ring *ring); @@ -351,11 +354,12 @@ static int vcn_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); - if (adev->vcn.cur_state != AMD_PG_STATE_GATE) - vcn_v4_0_3_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + if (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE) + vcn_v4_0_3_set_powergating_state_inst(ip_block, AMD_PG_STATE_GATE, i); + } return 0; } @@ -1640,41 +1644,52 @@ static int vcn_v4_0_3_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } -/** - * vcn_v4_0_3_set_powergating_state - set VCN block powergating state - * - * @ip_block: amdgpu_ip_block pointer - * @state: power gating state - * - * Set VCN block powergating state - */ -static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) +static int vcn_v4_0_3_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) { struct amdgpu_device *adev = ip_block->adev; - int ret = 0, i; + int ret = 0; /* for SRIOV, guest should not control VCN Power-gating * MMSCH FW should control Power-gating and clock-gating * guest should avoid touching CGC and PG */ if (amdgpu_sriov_vf(adev)) { - adev->vcn.cur_state = AMD_PG_STATE_UNGATE; + adev->vcn.inst[i].cur_state = AMD_PG_STATE_UNGATE; return 0; } - if (state == adev->vcn.cur_state) + if (state == adev->vcn.inst[i].cur_state) return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v4_0_3_stop(adev, i); - else - ret |= vcn_v4_0_3_start(adev, i); - } + if (state == AMD_PG_STATE_GATE) + ret = vcn_v4_0_3_stop(adev, i); + else + ret = vcn_v4_0_3_start(adev, i); if (!ret) - adev->vcn.cur_state = state; + adev->vcn.inst[i].cur_state = state; + + return ret; +} + +/** + * vcn_v4_0_3_set_powergating_state - set VCN block powergating state + * + * @ip_block: amdgpu_ip_block pointer + * @state: power gating state + * + * Set VCN block powergating state + */ +static int vcn_v4_0_3_set_powergating_state(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0, i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + vcn_v4_0_3_set_powergating_state_inst(ip_block, state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 83093e5c5cca..06b5ef4b3e84 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -95,8 +95,11 @@ static int amdgpu_ih_clientid_vcns[] = { static void vcn_v4_0_5_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v4_0_5_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v4_0_5_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v4_0_5_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring); @@ -308,9 +311,9 @@ static int vcn_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block) if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && - RREG32_SOC15(VCN, i, regUVD_STATUS))) { - vcn_v4_0_5_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE && + RREG32_SOC15(VCN, i, regUVD_STATUS))) { + vcn_v4_0_5_set_powergating_state_inst(ip_block, AMD_PG_STATE_GATE, i); } } } @@ -1522,6 +1525,27 @@ static int vcn_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } +static int vcn_v4_0_5_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0; + + if (state == adev->vcn.inst[i].cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = vcn_v4_0_5_stop(adev, i); + else + ret = vcn_v4_0_5_start(adev, i); + + if (!ret) + adev->vcn.inst[i].cur_state = state; + + return ret; +} + /** * vcn_v4_0_5_set_powergating_state - set VCN block powergating state * @@ -1531,23 +1555,13 @@ static int vcn_v4_0_5_set_clockgating_state(struct amdgpu_ip_block *ip_block, * Set VCN block powergating state */ static int vcn_v4_0_5_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) + enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; int ret = 0, i; - if (state == adev->vcn.cur_state) - return 0; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v4_0_5_stop(adev, i); - else - ret |= vcn_v4_0_5_start(adev, i); - } - - if (!ret) - adev->vcn.cur_state = state; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + vcn_v4_0_5_set_powergating_state_inst(ip_block, state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index 26f788ceeb11..8e0bd2cc1002 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -78,8 +78,11 @@ static int amdgpu_ih_clientid_vcns[] = { static void vcn_v5_0_0_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v5_0_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static int vcn_v5_0_0_pause_dpg_mode(struct amdgpu_device *adev, int inst_idx, struct dpg_pause_state *new_state); static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring); @@ -288,9 +291,9 @@ static int vcn_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block) if (!amdgpu_sriov_vf(adev)) { if ((adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) || - (adev->vcn.cur_state != AMD_PG_STATE_GATE && - RREG32_SOC15(VCN, i, regUVD_STATUS))) { - vcn_v5_0_0_set_powergating_state(ip_block, AMD_PG_STATE_GATE); + (adev->vcn.inst[i].cur_state != AMD_PG_STATE_GATE && + RREG32_SOC15(VCN, i, regUVD_STATUS))) { + vcn_v5_0_0_set_powergating_state_inst(ip_block, AMD_PG_STATE_GATE, i); } } } @@ -1262,6 +1265,27 @@ static int vcn_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } +static int vcn_v5_0_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0; + + if (state == adev->vcn.inst[i].cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = vcn_v5_0_0_stop(adev, i); + else + ret = vcn_v5_0_0_start(adev, i); + + if (!ret) + adev->vcn.inst[i].cur_state = state; + + return ret; +} + /** * vcn_v5_0_0_set_powergating_state - set VCN block powergating state * @@ -1271,23 +1295,13 @@ static int vcn_v5_0_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, * Set VCN block powergating state */ static int vcn_v5_0_0_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state) + enum amd_powergating_state state) { struct amdgpu_device *adev = ip_block->adev; int ret = 0, i; - if (state == adev->vcn.cur_state) - return 0; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret |= vcn_v5_0_0_stop(adev, i); - else - ret |= vcn_v5_0_0_start(adev, i); - } - - if (!ret) - adev->vcn.cur_state = state; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + vcn_v5_0_0_set_powergating_state_inst(ip_block, state, i); return ret; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index 512e5f6ae7b3..83e873767a86 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -40,8 +40,11 @@ static void vcn_v5_0_1_set_unified_ring_funcs(struct amdgpu_device *adev); static void vcn_v5_0_1_set_irq_funcs(struct amdgpu_device *adev); +static int vcn_v5_0_1_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i); static int vcn_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block, - enum amd_powergating_state state); + enum amd_powergating_state state); static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring); /** @@ -988,6 +991,27 @@ static int vcn_v5_0_1_set_clockgating_state(struct amdgpu_ip_block *ip_block, return 0; } +static int vcn_v5_0_1_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, + enum amd_powergating_state state, + int i) +{ + struct amdgpu_device *adev = ip_block->adev; + int ret = 0; + + if (state == adev->vcn.inst[i].cur_state) + return 0; + + if (state == AMD_PG_STATE_GATE) + ret = vcn_v5_0_1_stop(adev, i); + else + ret = vcn_v5_0_1_start(adev, i); + + if (!ret) + adev->vcn.inst[i].cur_state = state; + + return ret; +} + /** * vcn_v5_0_1_set_powergating_state - set VCN block powergating state * @@ -1002,18 +1026,8 @@ static int vcn_v5_0_1_set_powergating_state(struct amdgpu_ip_block *ip_block, struct amdgpu_device *adev = ip_block->adev; int ret = 0, i; - if (state == adev->vcn.cur_state) - return 0; - - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (state == AMD_PG_STATE_GATE) - ret = vcn_v5_0_1_stop(adev, i); - else - ret = vcn_v5_0_1_start(adev, i); - } - - if (!ret) - adev->vcn.cur_state = state; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + vcn_v5_0_1_set_powergating_state_inst(ip_block, state, i); return ret; } diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index 37ba84898bd5..8cfb07549f54 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -2041,15 +2041,15 @@ static int smu_hw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev) && !amdgpu_sriov_is_pp_one_vf(adev)) return 0; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { smu_dpm_set_vcn_enable(smu, false, i); + adev->vcn.inst[i].cur_state = AMD_PG_STATE_GATE; + } smu_dpm_set_jpeg_enable(smu, false); + adev->jpeg.cur_state = AMD_PG_STATE_GATE; smu_dpm_set_vpe_enable(smu, false); smu_dpm_set_umsch_mm_enable(smu, false); - adev->vcn.cur_state = AMD_PG_STATE_GATE; - adev->jpeg.cur_state = AMD_PG_STATE_GATE; - if (!smu->pm_enabled) return 0; -- 2.51.0 From cb107271687d569c3b706b974bcae6b9fdcfe9d0 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 15 Nov 2024 16:19:23 -0500 Subject: [PATCH 14/16] drm/amdgpu/vcn: move more instanced data to vcn_instance Move more per instance data into the per instance structure. v2: index instances directly on vcn1.0 and 2.0 to make it clear that they only support a single instance (Lijo) v3: fix typo on vcn 2.5 Reviewed-by: Boyuan Zhang (v2) Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 196 ++++++++++++------------ drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 17 +- drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c | 4 +- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 42 ++--- drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 92 +++++------ drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 51 +++--- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 67 ++++---- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 19 ++- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 14 +- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 18 ++- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c | 16 +- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 8 +- 13 files changed, 279 insertions(+), 267 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 05c73bf7541c..27bfe9c8af06 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -459,7 +459,7 @@ static int amdgpu_hw_ip_info(struct amdgpu_device *adev, if (adev->vcn.harvest_config & (1 << i)) continue; - for (j = 0; j < adev->vcn.num_enc_rings; j++) + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; j++) if (adev->vcn.inst[i].ring_enc[j].sched.ready) ++num_rings; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 0c3081a198d9..9901b8e17f30 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -127,93 +127,88 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) unsigned int fw_shared_size, log_offset; int i, r; - mutex_init(&adev->vcn.vcn1_jpeg1_workaround); for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + mutex_init(&adev->vcn.inst[i].vcn1_jpeg1_workaround); mutex_init(&adev->vcn.inst[i].vcn_pg_lock); atomic_set(&adev->vcn.inst[i].total_submission_cnt, 0); INIT_DELAYED_WORK(&adev->vcn.inst[i].idle_work, amdgpu_vcn_idle_work_handler); atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0); - } - - if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && - (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) - adev->vcn.indirect_sram = true; - - /* - * Some Steam Deck's BIOS versions are incompatible with the - * indirect SRAM mode, leading to amdgpu being unable to get - * properly probed (and even potentially crashing the kernel). - * Hence, check for these versions here - notice this is - * restricted to Vangogh (Deck's APU). - */ - if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 0, 2)) { - const char *bios_ver = dmi_get_system_info(DMI_BIOS_VERSION); - - if (bios_ver && (!strncmp("F7A0113", bios_ver, 7) || - !strncmp("F7A0114", bios_ver, 7))) { - adev->vcn.indirect_sram = false; - dev_info(adev->dev, - "Steam Deck quirk: indirect SRAM disabled on BIOS %s\n", bios_ver); + if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && + (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) + adev->vcn.inst[i].indirect_sram = true; + + /* + * Some Steam Deck's BIOS versions are incompatible with the + * indirect SRAM mode, leading to amdgpu being unable to get + * properly probed (and even potentially crashing the kernel). + * Hence, check for these versions here - notice this is + * restricted to Vangogh (Deck's APU). + */ + if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 0, 2)) { + const char *bios_ver = dmi_get_system_info(DMI_BIOS_VERSION); + + if (bios_ver && (!strncmp("F7A0113", bios_ver, 7) || + !strncmp("F7A0114", bios_ver, 7))) { + adev->vcn.inst[i].indirect_sram = false; + dev_info(adev->dev, + "Steam Deck quirk: indirect SRAM disabled on BIOS %s\n", bios_ver); + } } - } - /* from vcn4 and above, only unified queue is used */ - adev->vcn.using_unified_queue = - amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0); - - hdr = (const struct common_firmware_header *)adev->vcn.inst[0].fw->data; - adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version); - - /* Bit 20-23, it is encode major and non-zero for new naming convention. - * This field is part of version minor and DRM_DISABLED_FLAG in old naming - * convention. Since the l:wq!atest version minor is 0x5B and DRM_DISABLED_FLAG - * is zero in old naming convention, this field is always zero so far. - * These four bits are used to tell which naming convention is present. - */ - fw_check = (le32_to_cpu(hdr->ucode_version) >> 20) & 0xf; - if (fw_check) { - unsigned int dec_ver, enc_major, enc_minor, vep, fw_rev; - - fw_rev = le32_to_cpu(hdr->ucode_version) & 0xfff; - enc_minor = (le32_to_cpu(hdr->ucode_version) >> 12) & 0xff; - enc_major = fw_check; - dec_ver = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xf; - vep = (le32_to_cpu(hdr->ucode_version) >> 28) & 0xf; - DRM_INFO("Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n", - enc_major, enc_minor, dec_ver, vep, fw_rev); - } else { - unsigned int version_major, version_minor, family_id; + /* from vcn4 and above, only unified queue is used */ + adev->vcn.inst[i].using_unified_queue = + amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0); + + hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; + adev->vcn.inst[i].fw_version = le32_to_cpu(hdr->ucode_version); + adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version); + + /* Bit 20-23, it is encode major and non-zero for new naming convention. + * This field is part of version minor and DRM_DISABLED_FLAG in old naming + * convention. Since the l:wq!atest version minor is 0x5B and DRM_DISABLED_FLAG + * is zero in old naming convention, this field is always zero so far. + * These four bits are used to tell which naming convention is present. + */ + fw_check = (le32_to_cpu(hdr->ucode_version) >> 20) & 0xf; + if (fw_check) { + unsigned int dec_ver, enc_major, enc_minor, vep, fw_rev; + + fw_rev = le32_to_cpu(hdr->ucode_version) & 0xfff; + enc_minor = (le32_to_cpu(hdr->ucode_version) >> 12) & 0xff; + enc_major = fw_check; + dec_ver = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xf; + vep = (le32_to_cpu(hdr->ucode_version) >> 28) & 0xf; + DRM_INFO("Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n", + enc_major, enc_minor, dec_ver, vep, fw_rev); + } else { + unsigned int version_major, version_minor, family_id; - family_id = le32_to_cpu(hdr->ucode_version) & 0xff; - version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff; - version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff; - DRM_INFO("Found VCN firmware Version: %u.%u Family ID: %u\n", - version_major, version_minor, family_id); - } + family_id = le32_to_cpu(hdr->ucode_version) & 0xff; + version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff; + version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff; + DRM_INFO("Found VCN firmware Version: %u.%u Family ID: %u\n", + version_major, version_minor, family_id); + } - bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); + bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE; + if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) + bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); - if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(5, 0, 0)) { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)); - log_offset = offsetof(struct amdgpu_vcn5_fw_shared, fw_log); - } else if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0)) { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)); - log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log); - } else { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)); - log_offset = offsetof(struct amdgpu_fw_shared, fw_log); - } - - bo_size += fw_shared_size; + if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(5, 0, 0)) { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)); + log_offset = offsetof(struct amdgpu_vcn5_fw_shared, fw_log); + } else if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0)) { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)); + log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log); + } else { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)); + log_offset = offsetof(struct amdgpu_fw_shared, fw_log); + } - if (amdgpu_vcnfw_log) - bo_size += AMDGPU_VCNFW_LOG_SIZE; + bo_size += fw_shared_size; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (amdgpu_vcnfw_log) + bo_size += AMDGPU_VCNFW_LOG_SIZE; r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM | @@ -239,7 +234,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev) adev->vcn.inst[i].fw_shared.log_offset = log_offset; } - if (adev->vcn.indirect_sram) { + if (adev->vcn.inst[i].indirect_sram) { r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT, @@ -277,15 +272,14 @@ int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) amdgpu_ring_fini(&adev->vcn.inst[j].ring_dec); - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) amdgpu_ring_fini(&adev->vcn.inst[j].ring_enc[i]); amdgpu_ucode_release(&adev->vcn.inst[j].fw); mutex_destroy(&adev->vcn.inst[j].vcn_pg_lock); + mutex_destroy(&adev->vcn.inst[j].vcn1_jpeg1_workaround); } - mutex_destroy(&adev->vcn.vcn1_jpeg1_workaround); - return 0; } @@ -404,12 +398,12 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work) if (adev->vcn.harvest_config & (1 << i)) return; - for (j = 0; j < adev->vcn.num_enc_rings; ++j) + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_enc[j]); /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && - !adev->vcn.using_unified_queue) { + !adev->vcn.inst[i].using_unified_queue) { struct dpg_pause_state new_state; if (fence[i] || @@ -418,7 +412,7 @@ static void amdgpu_vcn_idle_work_handler(struct work_struct *work) else new_state.fw_based = VCN_DPG_STATE__UNPAUSE; - adev->vcn.pause_dpg_mode(adev, i, &new_state); + adev->vcn.inst[i].pause_dpg_mode(adev, i, &new_state); } fence[i] += amdgpu_fence_count_emitted(&vcn_inst->ring_dec); @@ -456,7 +450,7 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && - !adev->vcn.using_unified_queue) { + !adev->vcn.inst[ring->me].using_unified_queue) { struct dpg_pause_state new_state; if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC) { @@ -466,7 +460,7 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) unsigned int fences = 0; unsigned int i; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[ring->me].num_enc_rings; ++i) fences += amdgpu_fence_count_emitted(&adev->vcn.inst[ring->me].ring_enc[i]); if (fences || atomic_read(&adev->vcn.inst[ring->me].dpg_enc_submission_cnt)) @@ -475,7 +469,7 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring) new_state.fw_based = VCN_DPG_STATE__UNPAUSE; } - adev->vcn.pause_dpg_mode(adev, ring->me, &new_state); + adev->vcn.inst[ring->me].pause_dpg_mode(adev, ring->me, &new_state); } mutex_unlock(&adev->vcn.inst[ring->me].vcn_pg_lock); } @@ -487,7 +481,7 @@ void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring) /* Only set DPG pause for VCN3 or below, VCN4 and above will be handled by FW */ if (ring->adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG && ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC && - !adev->vcn.using_unified_queue) + !adev->vcn.inst[ring->me].using_unified_queue) atomic_dec(&ring->adev->vcn.inst[ring->me].dpg_enc_submission_cnt); atomic_dec(&ring->adev->vcn.inst[ring->me].total_submission_cnt); @@ -511,7 +505,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring) r = amdgpu_ring_alloc(ring, 3); if (r) return r; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.scratch9, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.scratch9, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { @@ -576,14 +570,14 @@ static int amdgpu_vcn_dec_send_msg(struct amdgpu_ring *ring, goto err; ib = &job->ibs[0]; - ib->ptr[0] = PACKET0(adev->vcn.internal.data0, 0); + ib->ptr[0] = PACKET0(adev->vcn.inst[ring->me].internal.data0, 0); ib->ptr[1] = addr; - ib->ptr[2] = PACKET0(adev->vcn.internal.data1, 0); + ib->ptr[2] = PACKET0(adev->vcn.inst[ring->me].internal.data1, 0); ib->ptr[3] = addr >> 32; - ib->ptr[4] = PACKET0(adev->vcn.internal.cmd, 0); + ib->ptr[4] = PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0); ib->ptr[5] = 0; for (i = 6; i < 16; i += 2) { - ib->ptr[i] = PACKET0(adev->vcn.internal.nop, 0); + ib->ptr[i] = PACKET0(adev->vcn.inst[ring->me].internal.nop, 0); ib->ptr[i+1] = 0; } ib->length_dw = 16; @@ -746,7 +740,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, uint32_t ib_pack_in_dw; int i, r; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) ib_size_dw += 8; r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, @@ -759,7 +753,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, ib->length_dw = 0; /* single queue headers */ - if (adev->vcn.using_unified_queue) { + if (adev->vcn.inst[ring->me].using_unified_queue) { ib_pack_in_dw = sizeof(struct amdgpu_vcn_decode_buffer) / sizeof(uint32_t) + 4 + 2; /* engine info + decoding ib in dw */ ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, ib_pack_in_dw, false); @@ -778,7 +772,7 @@ static int amdgpu_vcn_dec_sw_send_msg(struct amdgpu_ring *ring, for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, ib_pack_in_dw); r = amdgpu_job_submit_direct(job, ring, &f); @@ -876,7 +870,7 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand uint64_t addr; int i, r; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) ib_size_dw += 8; r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, @@ -890,7 +884,7 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand ib->length_dw = 0; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, 0x11, true); ib->ptr[ib->length_dw++] = 0x00000018; @@ -912,7 +906,7 @@ static int amdgpu_vcn_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t hand for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, 0x11); r = amdgpu_job_submit_direct(job, ring, &f); @@ -943,7 +937,7 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han uint64_t addr; int i, r; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) ib_size_dw += 8; r = amdgpu_job_alloc_with_ib(ring->adev, NULL, NULL, @@ -957,7 +951,7 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han ib->length_dw = 0; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) ib_checksum = amdgpu_vcn_unified_ring_ib_header(ib, 0x11, true); ib->ptr[ib->length_dw++] = 0x00000018; @@ -979,7 +973,7 @@ static int amdgpu_vcn_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t han for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - if (adev->vcn.using_unified_queue) + if (adev->vcn.inst[ring->me].using_unified_queue) amdgpu_vcn_unified_ring_ib_checksum(&ib_checksum, 0x11); r = amdgpu_job_submit_direct(job, ring, &f); @@ -1396,7 +1390,7 @@ void amdgpu_debugfs_vcn_sched_mask_init(struct amdgpu_device *adev) struct dentry *root = minor->debugfs_root; char name[32]; - if (adev->vcn.num_vcn_inst <= 1 || !adev->vcn.using_unified_queue) + if (adev->vcn.num_vcn_inst <= 1 || !adev->vcn.inst[0].using_unified_queue) return; sprintf(name, "amdgpu_vcn_sched_mask"); debugfs_create_file(name, 0600, root, adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index 0bcf4d1c7175..3b059f51894b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -322,6 +322,15 @@ struct amdgpu_vcn_inst { struct mutex vcn_pg_lock; enum amd_powergating_state cur_state; struct delayed_work idle_work; + unsigned fw_version; + unsigned num_enc_rings; + bool indirect_sram; + struct amdgpu_vcn_reg internal; + struct mutex vcn1_jpeg1_workaround; + int (*pause_dpg_mode)(struct amdgpu_device *adev, + int inst_idx, + struct dpg_pause_state *new_state); + bool using_unified_queue; }; struct amdgpu_vcn_ras { @@ -330,24 +339,16 @@ struct amdgpu_vcn_ras { struct amdgpu_vcn { unsigned fw_version; - unsigned num_enc_rings; - bool indirect_sram; - uint8_t num_vcn_inst; struct amdgpu_vcn_inst inst[AMDGPU_MAX_VCN_INSTANCES]; - struct amdgpu_vcn_reg internal; - struct mutex vcn1_jpeg1_workaround; unsigned harvest_config; - int (*pause_dpg_mode)(struct amdgpu_device *adev, - int inst_idx, struct dpg_pause_state *new_state); struct ras_common_if *ras_if; struct amdgpu_vcn_ras *ras; uint16_t inst_mask; uint8_t num_inst_per_aid; - bool using_unified_queue; /* IP reg dump */ uint32_t *ip_dump; diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c index 8031406e20ff..9e428e669ada 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c @@ -607,12 +607,12 @@ static void jpeg_v1_0_ring_begin_use(struct amdgpu_ring *ring) bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); int cnt = 0; - mutex_lock(&adev->vcn.vcn1_jpeg1_workaround); + mutex_lock(&adev->vcn.inst[0].vcn1_jpeg1_workaround); if (amdgpu_fence_wait_empty(&adev->vcn.inst->ring_dec)) DRM_ERROR("JPEG dec: vcn dec ring may not be empty\n"); - for (cnt = 0; cnt < adev->vcn.num_enc_rings; cnt++) { + for (cnt = 0; cnt < adev->vcn.inst[0].num_enc_rings; cnt++) { if (amdgpu_fence_wait_empty(&adev->vcn.inst->ring_enc[cnt])) DRM_ERROR("JPEG dec: vcn enc ring[%d] may not be empty\n", cnt); } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 291a5866c148..edbcd0819d17 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -105,7 +105,7 @@ static int vcn_v1_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - adev->vcn.num_enc_rings = 2; + adev->vcn.inst[0].num_enc_rings = 2; vcn_v1_0_set_dec_ring_funcs(adev); vcn_v1_0_set_enc_ring_funcs(adev); @@ -138,7 +138,7 @@ static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block) return r; /* VCN ENC TRAP */ - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, i + VCN_1_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq); if (r) @@ -166,18 +166,18 @@ static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - adev->vcn.internal.scratch9 = adev->vcn.inst->external.scratch9 = + adev->vcn.inst[0].internal.scratch9 = adev->vcn.inst->external.scratch9 = SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9); - adev->vcn.internal.data0 = adev->vcn.inst->external.data0 = + adev->vcn.inst[0].internal.data0 = adev->vcn.inst->external.data0 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0); - adev->vcn.internal.data1 = adev->vcn.inst->external.data1 = + adev->vcn.inst[0].internal.data1 = adev->vcn.inst->external.data1 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1); - adev->vcn.internal.cmd = adev->vcn.inst->external.cmd = + adev->vcn.inst[0].internal.cmd = adev->vcn.inst->external.cmd = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD); - adev->vcn.internal.nop = adev->vcn.inst->external.nop = + adev->vcn.inst[0].internal.nop = adev->vcn.inst->external.nop = SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP); - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); ring = &adev->vcn.inst->ring_enc[i]; @@ -189,7 +189,7 @@ static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - adev->vcn.pause_dpg_mode = vcn_v1_0_pause_dpg_mode; + adev->vcn.inst[0].pause_dpg_mode = vcn_v1_0_pause_dpg_mode; if (amdgpu_vcnfw_log) { volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst->fw_shared.cpu_addr; @@ -253,7 +253,7 @@ static int vcn_v1_0_hw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { ring = &adev->vcn.inst->ring_enc[i]; r = amdgpu_ring_test_helper(ring); if (r) @@ -1833,7 +1833,7 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work) struct amdgpu_device *adev = vcn_inst->adev; unsigned int fences = 0, i; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { @@ -1849,7 +1849,7 @@ static void vcn_v1_0_idle_work_handler(struct work_struct *work) else new_state.jpeg = VCN_DPG_STATE__UNPAUSE; - adev->vcn.pause_dpg_mode(adev, 0, &new_state); + adev->vcn.inst[0].pause_dpg_mode(adev, 0, &new_state); } fences += amdgpu_fence_count_emitted(adev->jpeg.inst->ring_dec); @@ -1872,7 +1872,7 @@ static void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring) struct amdgpu_device *adev = ring->adev; bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.inst[0].idle_work); - mutex_lock(&adev->vcn.vcn1_jpeg1_workaround); + mutex_lock(&adev->vcn.inst[0].vcn1_jpeg1_workaround); if (amdgpu_fence_wait_empty(ring->adev->jpeg.inst->ring_dec)) DRM_ERROR("VCN dec: jpeg dec ring may not be empty\n"); @@ -1898,7 +1898,7 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks) struct dpg_pause_state new_state; unsigned int fences = 0, i; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]); if (fences) @@ -1916,14 +1916,14 @@ void vcn_v1_0_set_pg_for_begin_use(struct amdgpu_ring *ring, bool set_clocks) else if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG) new_state.jpeg = VCN_DPG_STATE__PAUSE; - adev->vcn.pause_dpg_mode(adev, 0, &new_state); + adev->vcn.inst[0].pause_dpg_mode(adev, 0, &new_state); } } void vcn_v1_0_ring_end_use(struct amdgpu_ring *ring) { schedule_delayed_work(&ring->adev->vcn.inst[0].idle_work, VCN_IDLE_TIMEOUT); - mutex_unlock(&ring->adev->vcn.vcn1_jpeg1_workaround); + mutex_unlock(&ring->adev->vcn.inst[0].vcn1_jpeg1_workaround); } static void vcn_v1_0_print_ip_state(struct amdgpu_ip_block *ip_block, struct drm_printer *p) @@ -2057,11 +2057,11 @@ static int vcn_v1_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p, uint32_t reg = amdgpu_ib_get_value(ib, i); uint32_t val = amdgpu_ib_get_value(ib, i + 1); - if (reg == PACKET0(p->adev->vcn.internal.data0, 0)) { + if (reg == PACKET0(p->adev->vcn.inst[0].internal.data0, 0)) { msg_lo = val; - } else if (reg == PACKET0(p->adev->vcn.internal.data1, 0)) { + } else if (reg == PACKET0(p->adev->vcn.inst[0].internal.data1, 0)) { msg_hi = val; - } else if (reg == PACKET0(p->adev->vcn.internal.cmd, 0)) { + } else if (reg == PACKET0(p->adev->vcn.inst[0].internal.cmd, 0)) { r = vcn_v1_0_validate_bo(p, job, ((u64)msg_hi) << 32 | msg_lo); if (r) @@ -2146,7 +2146,7 @@ static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev) { int i; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) adev->vcn.inst->ring_enc[i].funcs = &vcn_v1_0_enc_ring_vm_funcs; } @@ -2157,7 +2157,7 @@ static const struct amdgpu_irq_src_funcs vcn_v1_0_irq_funcs = { static void vcn_v1_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->vcn.inst->irq.num_types = adev->vcn.num_enc_rings + 2; + adev->vcn.inst->irq.num_types = adev->vcn.inst[0].num_enc_rings + 2; adev->vcn.inst->irq.funcs = &vcn_v1_0_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index cc2b9dcc8868..d94fd075a910 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -110,9 +110,9 @@ static int vcn_v2_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; if (amdgpu_sriov_vf(adev)) - adev->vcn.num_enc_rings = 1; + adev->vcn.inst[0].num_enc_rings = 1; else - adev->vcn.num_enc_rings = 2; + adev->vcn.inst[0].num_enc_rings = 2; vcn_v2_0_set_dec_ring_funcs(adev); vcn_v2_0_set_enc_ring_funcs(adev); @@ -145,7 +145,7 @@ static int vcn_v2_0_sw_init(struct amdgpu_ip_block *ip_block) return r; /* VCN ENC TRAP */ - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, i + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq); @@ -175,25 +175,25 @@ static int vcn_v2_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; - adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; - adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; - adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; - adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; adev->vcn.inst->external.scratch9 = SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9); - adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; adev->vcn.inst->external.data0 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0); - adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; adev->vcn.inst->external.data1 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1); - adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; adev->vcn.inst->external.cmd = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD); - adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; + adev->vcn.inst[0].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; adev->vcn.inst->external.nop = SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP); - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); ring = &adev->vcn.inst->ring_enc[i]; @@ -210,7 +210,7 @@ static int vcn_v2_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - adev->vcn.pause_dpg_mode = vcn_v2_0_pause_dpg_mode; + adev->vcn.inst[0].pause_dpg_mode = vcn_v2_0_pause_dpg_mode; r = amdgpu_virt_alloc_mm_table(adev); if (r) @@ -292,7 +292,7 @@ static int vcn_v2_0_hw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) ring->sched.ready = false; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { ring = &adev->vcn.inst->ring_enc[i]; r = amdgpu_ring_test_helper(ring); if (r) @@ -981,7 +981,7 @@ static int vcn_v2_0_start(struct amdgpu_device *adev) amdgpu_dpm_enable_vcn(adev, true, 0); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v2_0_start_dpg_mode(adev, adev->vcn.indirect_sram); + return vcn_v2_0_start_dpg_mode(adev, adev->vcn.inst->indirect_sram); vcn_v2_0_disable_static_power_gating(adev); @@ -1421,9 +1421,9 @@ void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0)); amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_START << 1)); } @@ -1438,7 +1438,7 @@ void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[0].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_END << 1)); } @@ -1458,7 +1458,7 @@ void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) WARN_ON(ring->wptr % 2 || count % 2); for (i = 0; i < count / 2; i++) { - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.nop, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.nop, 0)); amdgpu_ring_write(ring, 0); } } @@ -1479,25 +1479,25 @@ void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, struct amdgpu_device *adev = ring->adev; WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.context_id, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.context_id, 0)); amdgpu_ring_write(ring, seq); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0)); amdgpu_ring_write(ring, addr & 0xffffffff); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0)); amdgpu_ring_write(ring, upper_32_bits(addr) & 0xff); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_FENCE << 1)); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0)); amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0)); amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_TRAP << 1)); } @@ -1520,14 +1520,14 @@ void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_device *adev = ring->adev; unsigned vmid = AMDGPU_JOB_GET_VMID(job); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.ib_vmid, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.ib_vmid, 0)); amdgpu_ring_write(ring, vmid); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.ib_bar_low, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.ib_bar_low, 0)); amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.ib_bar_high, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.ib_bar_high, 0)); amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr)); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.ib_size, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.ib_size, 0)); amdgpu_ring_write(ring, ib->length_dw); } @@ -1536,16 +1536,16 @@ void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg, { struct amdgpu_device *adev = ring->adev; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0)); amdgpu_ring_write(ring, reg << 2); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0)); amdgpu_ring_write(ring, val); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.gp_scratch8, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.gp_scratch8, 0)); amdgpu_ring_write(ring, mask); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_REG_READ_COND_WAIT << 1)); } @@ -1570,13 +1570,13 @@ void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring, { struct amdgpu_device *adev = ring->adev; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data0, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data0, 0)); amdgpu_ring_write(ring, reg << 2); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.data1, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.data1, 0)); amdgpu_ring_write(ring, val); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_WRITE_REG << 1)); } @@ -1777,9 +1777,9 @@ int vcn_v2_0_dec_ring_test_ring(struct amdgpu_ring *ring) r = amdgpu_ring_alloc(ring, 4); if (r) return r; - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.cmd, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.cmd, 0)); amdgpu_ring_write(ring, VCN_DEC_KMD_CMD | (VCN_DEC_CMD_PACKET_START << 1)); - amdgpu_ring_write(ring, PACKET0(adev->vcn.internal.scratch9, 0)); + amdgpu_ring_write(ring, PACKET0(adev->vcn.inst[ring->me].internal.scratch9, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { @@ -1863,7 +1863,7 @@ static int vcn_v2_0_start_mmsch(struct amdgpu_device *adev, adev->vcn.inst->ring_dec.wptr_old = 0; vcn_v2_0_dec_ring_set_wptr(&adev->vcn.inst->ring_dec); - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) { adev->vcn.inst->ring_enc[i].wptr = 0; adev->vcn.inst->ring_enc[i].wptr_old = 0; vcn_v2_0_enc_ring_set_wptr(&adev->vcn.inst->ring_enc[i]); @@ -1989,7 +1989,7 @@ static int vcn_v2_0_start_sriov(struct amdgpu_device *adev) SOC15_REG_OFFSET(UVD, i, mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE); - for (r = 0; r < adev->vcn.num_enc_rings; ++r) { + for (r = 0; r < adev->vcn.inst[0].num_enc_rings; ++r) { ring = &adev->vcn.inst->ring_enc[r]; ring->wptr = 0; MMSCH_V2_0_INSERT_DIRECT_WT( @@ -2178,7 +2178,7 @@ static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev) { int i; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) + for (i = 0; i < adev->vcn.inst[0].num_enc_rings; ++i) adev->vcn.inst->ring_enc[i].funcs = &vcn_v2_0_enc_ring_vm_funcs; } @@ -2189,7 +2189,7 @@ static const struct amdgpu_irq_src_funcs vcn_v2_0_irq_funcs = { static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev) { - adev->vcn.inst->irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst->irq.num_types = adev->vcn.inst[0].num_enc_rings + 1; adev->vcn.inst->irq.funcs = &vcn_v2_0_irq_funcs; } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index 0d78377f8ffe..c20619c027e7 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -121,11 +121,13 @@ static int amdgpu_ih_clientid_vcns[] = { static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; if (amdgpu_sriov_vf(adev)) { adev->vcn.num_vcn_inst = 2; adev->vcn.harvest_config = 0; - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) + adev->vcn.inst[i].num_enc_rings = 1; } else { u32 harvest; int i; @@ -134,13 +136,12 @@ static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block) harvest = RREG32_SOC15(VCN, i, mmCC_UVD_HARVESTING); if (harvest & CC_UVD_HARVESTING__UVD_DISABLE_MASK) adev->vcn.harvest_config |= 1 << i; + adev->vcn.inst[i].num_enc_rings = 2; } if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 | AMDGPU_VCN_HARVEST_VCN1)) /* both instances are harvested, disable the block */ return -ENOENT; - - adev->vcn.num_enc_rings = 2; } vcn_v2_5_set_dec_ring_funcs(adev); @@ -176,7 +177,7 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) return r; /* VCN ENC TRAP */ - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) { r = amdgpu_irq_add_id(adev, amdgpu_ih_clientid_vcns[j], i + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst[j].irq); if (r) @@ -205,22 +206,22 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) if (adev->vcn.harvest_config & (1 << j)) continue; - adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; - adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; - adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; - adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; - - adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; + + adev->vcn.inst[j].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; adev->vcn.inst[j].external.scratch9 = SOC15_REG_OFFSET(VCN, j, mmUVD_SCRATCH9); - adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; adev->vcn.inst[j].external.data0 = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_DATA0); - adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; adev->vcn.inst[j].external.data1 = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_DATA1); - adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; adev->vcn.inst[j].external.cmd = SOC15_REG_OFFSET(VCN, j, mmUVD_GPCOM_VCPU_CMD); - adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; + adev->vcn.inst[j].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; adev->vcn.inst[j].external.nop = SOC15_REG_OFFSET(VCN, j, mmUVD_NO_OP); ring = &adev->vcn.inst[j].ring_dec; @@ -240,7 +241,7 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) { enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(i); ring = &adev->vcn.inst[j].ring_enc[i]; @@ -268,6 +269,9 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_vcnfw_log) amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[j].pause_dpg_mode = vcn_v2_5_pause_dpg_mode; } if (amdgpu_sriov_vf(adev)) { @@ -276,9 +280,6 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) return r; } - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v2_5_pause_dpg_mode; - r = amdgpu_vcn_ras_sw_init(adev); if (r) return r; @@ -369,7 +370,7 @@ static int vcn_v2_5_hw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) { ring = &adev->vcn.inst[j].ring_enc[i]; r = amdgpu_ring_test_helper(ring); if (r) @@ -1018,7 +1019,7 @@ static int vcn_v2_5_start(struct amdgpu_device *adev, int i) amdgpu_dpm_enable_vcn(adev, true, i); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v2_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v2_5_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); /* disable register anti-hang mechanism */ WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_POWER_STATUS), 0, @@ -1749,7 +1750,7 @@ static void vcn_v2_5_set_enc_ring_funcs(struct amdgpu_device *adev) for (j = 0; j < adev->vcn.num_vcn_inst; ++j) { if (adev->vcn.harvest_config & (1 << j)) continue; - for (i = 0; i < adev->vcn.num_enc_rings; ++i) { + for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) { adev->vcn.inst[j].ring_enc[i].funcs = &vcn_v2_5_enc_ring_vm_funcs; adev->vcn.inst[j].ring_enc[i].me = j; } @@ -1920,10 +1921,10 @@ static void vcn_v2_5_set_irq_funcs(struct amdgpu_device *adev) for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v2_5_irq_funcs; - adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v2_6_ras_irq_funcs; } } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index acfe85b0c062..282a186cb0c5 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -127,11 +127,13 @@ static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v3_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; if (amdgpu_sriov_vf(adev)) { adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; adev->vcn.harvest_config = 0; - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) + adev->vcn.inst[i].num_enc_rings = 1; } else { if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 | @@ -139,11 +141,13 @@ static int vcn_v3_0_early_init(struct amdgpu_ip_block *ip_block) /* both instances are harvested, disable the block */ return -ENOENT; - if (amdgpu_ip_version(adev, UVD_HWIP, 0) == - IP_VERSION(3, 0, 33)) - adev->vcn.num_enc_rings = 0; - else - adev->vcn.num_enc_rings = 2; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + if (amdgpu_ip_version(adev, UVD_HWIP, 0) == + IP_VERSION(3, 0, 33)) + adev->vcn.inst[i].num_enc_rings = 0; + else + adev->vcn.inst[i].num_enc_rings = 2; + } } vcn_v3_0_set_dec_ring_funcs(adev); @@ -198,22 +202,22 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; - adev->vcn.internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; - adev->vcn.internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; - adev->vcn.internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; - adev->vcn.internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.ib_bar_high = mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.ib_size = mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.gp_scratch8 = mmUVD_GP_SCRATCH8_INTERNAL_OFFSET; - adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET; adev->vcn.inst[i].external.scratch9 = SOC15_REG_OFFSET(VCN, i, mmUVD_SCRATCH9); - adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET; adev->vcn.inst[i].external.data0 = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_DATA0); - adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET; adev->vcn.inst[i].external.data1 = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_DATA1); - adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET; adev->vcn.inst[i].external.cmd = SOC15_REG_OFFSET(VCN, i, mmUVD_GPCOM_VCPU_CMD); - adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; + adev->vcn.inst[i].internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET; adev->vcn.inst[i].external.nop = SOC15_REG_OFFSET(VCN, i, mmUVD_NO_OP); /* VCN DEC TRAP */ @@ -227,7 +231,7 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) ring = &adev->vcn.inst[i].ring_dec; ring->use_doorbell = true; if (amdgpu_sriov_vf(adev)) { - ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1); + ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.inst[i].num_enc_rings + 1); } else { ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 8 * i; } @@ -239,7 +243,7 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) { enum amdgpu_ring_priority_level hw_prio = amdgpu_vcn_get_enc_ring_prio(j); /* VCN ENC TRAP */ @@ -251,7 +255,7 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) ring = &adev->vcn.inst[i].ring_enc[j]; ring->use_doorbell = true; if (amdgpu_sriov_vf(adev)) { - ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1) + 1 + j; + ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.inst[i].num_enc_rings + 1) + 1 + j; } else { ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + j + 8 * i; } @@ -277,6 +281,9 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_vcnfw_log) amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[i].pause_dpg_mode = vcn_v3_0_pause_dpg_mode; } if (amdgpu_sriov_vf(adev)) { @@ -284,8 +291,6 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) if (r) return r; } - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v3_0_pause_dpg_mode; /* Allocate memory for VCN IP Dump buffer */ ptr = kcalloc(adev->vcn.num_vcn_inst * reg_count, sizeof(uint32_t), GFP_KERNEL); @@ -373,7 +378,7 @@ static int vcn_v3_0_hw_init(struct amdgpu_ip_block *ip_block) ring->sched.ready = true; } - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) { ring = &adev->vcn.inst[i].ring_enc[j]; if (amdgpu_vcn_is_disabled_vcn(adev, VCN_ENCODE_RING, i)) { ring->sched.ready = false; @@ -401,7 +406,7 @@ static int vcn_v3_0_hw_init(struct amdgpu_ip_block *ip_block) if (r) return r; - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) { ring = &adev->vcn.inst[i].ring_enc[j]; r = amdgpu_ring_test_helper(ring); if (r) @@ -1153,7 +1158,7 @@ static int vcn_v3_0_start(struct amdgpu_device *adev, int i) amdgpu_dpm_enable_vcn(adev, true, i); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); /* disable VCN power gating */ vcn_v3_0_disable_static_power_gating(adev, i); @@ -1433,7 +1438,7 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) mmUVD_VCPU_CACHE_SIZE2), AMDGPU_VCN_CONTEXT_SIZE); - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) { ring = &adev->vcn.inst[i].ring_enc[j]; ring->wptr = 0; rb_addr = ring->gpu_addr; @@ -1924,11 +1929,11 @@ static int vcn_v3_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p, uint32_t reg = amdgpu_ib_get_value(ib, i); uint32_t val = amdgpu_ib_get_value(ib, i + 1); - if (reg == PACKET0(p->adev->vcn.internal.data0, 0)) { + if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.data0, 0)) { msg_lo = val; - } else if (reg == PACKET0(p->adev->vcn.internal.data1, 0)) { + } else if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.data1, 0)) { msg_hi = val; - } else if (reg == PACKET0(p->adev->vcn.internal.cmd, 0) && + } else if (reg == PACKET0(p->adev->vcn.inst[ring->me].internal.cmd, 0) && val == 0) { r = vcn_v3_0_dec_msg(p, job, ((u64)msg_hi) << 32 | msg_lo); @@ -2092,7 +2097,7 @@ static void vcn_v3_0_set_enc_ring_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; - for (j = 0; j < adev->vcn.num_enc_rings; ++j) { + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) { adev->vcn.inst[i].ring_enc[j].funcs = &vcn_v3_0_enc_ring_vm_funcs; adev->vcn.inst[i].ring_enc[j].me = i; } @@ -2258,7 +2263,7 @@ static void vcn_v3_0_set_irq_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v3_0_irq_funcs; } } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index 1f67e5973cf8..f5bb241c1af1 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -129,8 +129,9 @@ static int vcn_v4_0_early_init(struct amdgpu_ip_block *ip_block) } } - /* re-use enc ring as unified ring */ - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + /* re-use enc ring as unified ring */ + adev->vcn.inst[i].num_enc_rings = 1; vcn_v4_0_set_unified_ring_funcs(adev); vcn_v4_0_set_irq_funcs(adev); @@ -214,7 +215,8 @@ static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block) ring = &adev->vcn.inst[i].ring_enc[0]; ring->use_doorbell = true; if (amdgpu_sriov_vf(adev)) - ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + i * (adev->vcn.num_enc_rings + 1) + 1; + ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + i * + (adev->vcn.inst[i].num_enc_rings + 1) + 1; else ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i; ring->vm_hub = AMDGPU_MMHUB0(0); @@ -226,6 +228,9 @@ static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block) return r; vcn_v4_0_fw_shared_init(adev, i); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_pause_dpg_mode; } /* TODO: Add queue reset mask when FW fully supports it */ @@ -238,8 +243,6 @@ static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v4_0_pause_dpg_mode; r = amdgpu_vcn_ras_sw_init(adev); if (r) @@ -1111,7 +1114,7 @@ static int vcn_v4_0_start(struct amdgpu_device *adev, int i) fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v4_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v4_0_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); /* disable VCN power gating */ vcn_v4_0_disable_static_power_gating(adev, i); @@ -2174,10 +2177,10 @@ static void vcn_v4_0_set_irq_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v4_0_irq_funcs; - adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].ras_poison_irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].ras_poison_irq.funcs = &vcn_v4_0_ras_irq_funcs; } } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index 73b10813d9b8..fbe599c539c8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -114,9 +114,11 @@ static inline bool vcn_v4_0_3_normalizn_reqd(struct amdgpu_device *adev) static int vcn_v4_0_3_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - /* re-use enc ring as unified ring */ - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + /* re-use enc ring as unified ring */ + adev->vcn.inst[i].num_enc_rings = 1; vcn_v4_0_3_set_unified_ring_funcs(adev); vcn_v4_0_3_set_irq_funcs(adev); @@ -194,6 +196,9 @@ static int vcn_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) return r; vcn_v4_0_3_fw_shared_init(adev, i); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_3_pause_dpg_mode; } /* TODO: Add queue reset mask when FW fully supports it */ @@ -206,9 +211,6 @@ static int vcn_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) return r; } - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v4_0_3_pause_dpg_mode; - if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__VCN)) { r = amdgpu_vcn_ras_sw_init(adev); if (r) { @@ -1135,7 +1137,7 @@ static int vcn_v4_0_3_start(struct amdgpu_device *adev, int i) uint32_t tmp; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v4_0_3_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); vcn_inst = GET_INST(VCN, i); /* set VCN status busy */ diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 06b5ef4b3e84..2513929b2d43 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -115,9 +115,11 @@ static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v4_0_5_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - /* re-use enc ring as unified ring */ - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + /* re-use enc ring as unified ring */ + adev->vcn.inst[i].num_enc_rings = 1; vcn_v4_0_5_set_unified_ring_funcs(adev); vcn_v4_0_5_set_irq_funcs(adev); @@ -173,7 +175,7 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) ring->use_doorbell = true; if (amdgpu_sriov_vf(adev)) ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + - i * (adev->vcn.num_enc_rings + 1) + 1; + i * (adev->vcn.inst[i].num_enc_rings + 1) + 1; else ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i; @@ -198,6 +200,9 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_vcnfw_log) amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[i].pause_dpg_mode = vcn_v4_0_5_pause_dpg_mode; } if (amdgpu_sriov_vf(adev)) { @@ -206,9 +211,6 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) return r; } - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v4_0_5_pause_dpg_mode; - /* Allocate memory for VCN IP Dump buffer */ ptr = kcalloc(adev->vcn.num_vcn_inst * reg_count, sizeof(uint32_t), GFP_KERNEL); if (!ptr) { @@ -1014,7 +1016,7 @@ static int vcn_v4_0_5_start(struct amdgpu_device *adev, int i) fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v4_0_5_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v4_0_5_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); /* disable VCN power gating */ vcn_v4_0_5_disable_static_power_gating(adev, i); @@ -1629,7 +1631,7 @@ static void vcn_v4_0_5_set_irq_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v4_0_5_irq_funcs; } } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index 8e0bd2cc1002..50e0928c3a20 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -98,9 +98,11 @@ static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v5_0_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - /* re-use enc ring as unified ring */ - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + /* re-use enc ring as unified ring */ + adev->vcn.inst[i].num_enc_rings = 1; vcn_v5_0_0_set_unified_ring_funcs(adev); vcn_v5_0_0_set_irq_funcs(adev); @@ -184,15 +186,15 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block) if (amdgpu_vcnfw_log) amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]); + + if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) + adev->vcn.inst[i].pause_dpg_mode = vcn_v5_0_0_pause_dpg_mode; } /* TODO: Add queue reset mask when FW fully supports it */ adev->vcn.supported_reset = amdgpu_get_soft_full_reset_mask(&adev->vcn.inst[0].ring_enc[0]); - if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - adev->vcn.pause_dpg_mode = vcn_v5_0_0_pause_dpg_mode; - vcn_v5_0_0_alloc_ip_dump(adev); r = amdgpu_vcn_sysfs_reset_mask_init(adev); @@ -791,7 +793,7 @@ static int vcn_v5_0_0_start(struct amdgpu_device *adev, int i) fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v5_0_0_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v5_0_0_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); /* disable VCN power gating */ vcn_v5_0_0_disable_static_power_gating(adev, i); @@ -1369,7 +1371,7 @@ static void vcn_v5_0_0_set_irq_funcs(struct amdgpu_device *adev) if (adev->vcn.harvest_config & (1 << i)) continue; - adev->vcn.inst[i].irq.num_types = adev->vcn.num_enc_rings + 1; + adev->vcn.inst[i].irq.num_types = adev->vcn.inst[i].num_enc_rings + 1; adev->vcn.inst[i].irq.funcs = &vcn_v5_0_0_irq_funcs; } } diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index 83e873767a86..ebe977507bc3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -58,9 +58,11 @@ static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; + int i; - /* re-use enc ring as unified ring */ - adev->vcn.num_enc_rings = 1; + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) + /* re-use enc ring as unified ring */ + adev->vcn.inst[i].num_enc_rings = 1; vcn_v5_0_1_set_unified_ring_funcs(adev); vcn_v5_0_1_set_irq_funcs(adev); @@ -583,7 +585,7 @@ static int vcn_v5_0_1_start(struct amdgpu_device *adev, int i) fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v5_0_1_start_dpg_mode(adev, i, adev->vcn.indirect_sram); + return vcn_v5_0_1_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); vcn_inst = GET_INST(VCN, i); -- 2.51.0 From f98675638f0a818a2eb802103b1e140b091358c4 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 15 Nov 2024 17:44:01 -0500 Subject: [PATCH 15/16] drm/amdgpu/vcn: switch vcn helpers to be instance based Pass the instance to the helpers. Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 437 ++++++++++++------------ drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h | 12 +- drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 16 +- drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c | 16 +- drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c | 71 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 65 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c | 69 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 70 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 63 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c | 67 ++-- drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 64 ++-- 11 files changed, 541 insertions(+), 409 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c index 9901b8e17f30..010cc64a950a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c @@ -93,192 +93,186 @@ MODULE_FIRMWARE(FIRMWARE_VCN5_0_1); static void amdgpu_vcn_idle_work_handler(struct work_struct *work); -int amdgpu_vcn_early_init(struct amdgpu_device *adev) +int amdgpu_vcn_early_init(struct amdgpu_device *adev, int i) { char ucode_prefix[25]; - int r, i; + int r; + adev->vcn.inst[i].adev = adev; + adev->vcn.inst[i].inst = i; amdgpu_ucode_ip_version_decode(adev, UVD_HWIP, ucode_prefix, sizeof(ucode_prefix)); - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - adev->vcn.inst[i].adev = adev; - adev->vcn.inst[i].inst = i; - - if (i == 1 && amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 6)) - r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw, - AMDGPU_UCODE_REQUIRED, - "amdgpu/%s_%d.bin", ucode_prefix, i); - else - r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw, - AMDGPU_UCODE_REQUIRED, - "amdgpu/%s.bin", ucode_prefix); - if (r) { - amdgpu_ucode_release(&adev->vcn.inst[i].fw); - return r; - } + + if (i == 1 && amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 6)) + r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw, + AMDGPU_UCODE_REQUIRED, + "amdgpu/%s_%d.bin", ucode_prefix, i); + else + r = amdgpu_ucode_request(adev, &adev->vcn.inst[i].fw, + AMDGPU_UCODE_REQUIRED, + "amdgpu/%s.bin", ucode_prefix); + if (r) { + amdgpu_ucode_release(&adev->vcn.inst[i].fw); + return r; } return r; } -int amdgpu_vcn_sw_init(struct amdgpu_device *adev) +int amdgpu_vcn_sw_init(struct amdgpu_device *adev, int i) { unsigned long bo_size; const struct common_firmware_header *hdr; unsigned char fw_check; unsigned int fw_shared_size, log_offset; - int i, r; + int r; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - mutex_init(&adev->vcn.inst[i].vcn1_jpeg1_workaround); - mutex_init(&adev->vcn.inst[i].vcn_pg_lock); - atomic_set(&adev->vcn.inst[i].total_submission_cnt, 0); - INIT_DELAYED_WORK(&adev->vcn.inst[i].idle_work, amdgpu_vcn_idle_work_handler); - atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0); - if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && - (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) - adev->vcn.inst[i].indirect_sram = true; - - /* - * Some Steam Deck's BIOS versions are incompatible with the - * indirect SRAM mode, leading to amdgpu being unable to get - * properly probed (and even potentially crashing the kernel). - * Hence, check for these versions here - notice this is - * restricted to Vangogh (Deck's APU). - */ - if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 0, 2)) { - const char *bios_ver = dmi_get_system_info(DMI_BIOS_VERSION); - - if (bios_ver && (!strncmp("F7A0113", bios_ver, 7) || - !strncmp("F7A0114", bios_ver, 7))) { - adev->vcn.inst[i].indirect_sram = false; - dev_info(adev->dev, - "Steam Deck quirk: indirect SRAM disabled on BIOS %s\n", bios_ver); - } + mutex_init(&adev->vcn.inst[i].vcn1_jpeg1_workaround); + mutex_init(&adev->vcn.inst[i].vcn_pg_lock); + atomic_set(&adev->vcn.inst[i].total_submission_cnt, 0); + INIT_DELAYED_WORK(&adev->vcn.inst[i].idle_work, amdgpu_vcn_idle_work_handler); + atomic_set(&adev->vcn.inst[i].dpg_enc_submission_cnt, 0); + if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) && + (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)) + adev->vcn.inst[i].indirect_sram = true; + + /* + * Some Steam Deck's BIOS versions are incompatible with the + * indirect SRAM mode, leading to amdgpu being unable to get + * properly probed (and even potentially crashing the kernel). + * Hence, check for these versions here - notice this is + * restricted to Vangogh (Deck's APU). + */ + if (amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(3, 0, 2)) { + const char *bios_ver = dmi_get_system_info(DMI_BIOS_VERSION); + + if (bios_ver && (!strncmp("F7A0113", bios_ver, 7) || + !strncmp("F7A0114", bios_ver, 7))) { + adev->vcn.inst[i].indirect_sram = false; + dev_info(adev->dev, + "Steam Deck quirk: indirect SRAM disabled on BIOS %s\n", bios_ver); } + } - /* from vcn4 and above, only unified queue is used */ - adev->vcn.inst[i].using_unified_queue = - amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0); + /* from vcn4 and above, only unified queue is used */ + adev->vcn.inst[i].using_unified_queue = + amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0); + + hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; + adev->vcn.inst[i].fw_version = le32_to_cpu(hdr->ucode_version); + adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version); + + /* Bit 20-23, it is encode major and non-zero for new naming convention. + * This field is part of version minor and DRM_DISABLED_FLAG in old naming + * convention. Since the l:wq!atest version minor is 0x5B and DRM_DISABLED_FLAG + * is zero in old naming convention, this field is always zero so far. + * These four bits are used to tell which naming convention is present. + */ + fw_check = (le32_to_cpu(hdr->ucode_version) >> 20) & 0xf; + if (fw_check) { + unsigned int dec_ver, enc_major, enc_minor, vep, fw_rev; + + fw_rev = le32_to_cpu(hdr->ucode_version) & 0xfff; + enc_minor = (le32_to_cpu(hdr->ucode_version) >> 12) & 0xff; + enc_major = fw_check; + dec_ver = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xf; + vep = (le32_to_cpu(hdr->ucode_version) >> 28) & 0xf; + DRM_INFO("Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n", + enc_major, enc_minor, dec_ver, vep, fw_rev); + } else { + unsigned int version_major, version_minor, family_id; - hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; - adev->vcn.inst[i].fw_version = le32_to_cpu(hdr->ucode_version); - adev->vcn.fw_version = le32_to_cpu(hdr->ucode_version); - - /* Bit 20-23, it is encode major and non-zero for new naming convention. - * This field is part of version minor and DRM_DISABLED_FLAG in old naming - * convention. Since the l:wq!atest version minor is 0x5B and DRM_DISABLED_FLAG - * is zero in old naming convention, this field is always zero so far. - * These four bits are used to tell which naming convention is present. - */ - fw_check = (le32_to_cpu(hdr->ucode_version) >> 20) & 0xf; - if (fw_check) { - unsigned int dec_ver, enc_major, enc_minor, vep, fw_rev; - - fw_rev = le32_to_cpu(hdr->ucode_version) & 0xfff; - enc_minor = (le32_to_cpu(hdr->ucode_version) >> 12) & 0xff; - enc_major = fw_check; - dec_ver = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xf; - vep = (le32_to_cpu(hdr->ucode_version) >> 28) & 0xf; - DRM_INFO("Found VCN firmware Version ENC: %u.%u DEC: %u VEP: %u Revision: %u\n", - enc_major, enc_minor, dec_ver, vep, fw_rev); - } else { - unsigned int version_major, version_minor, family_id; + family_id = le32_to_cpu(hdr->ucode_version) & 0xff; + version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff; + version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff; + DRM_INFO("Found VCN firmware Version: %u.%u Family ID: %u\n", + version_major, version_minor, family_id); + } - family_id = le32_to_cpu(hdr->ucode_version) & 0xff; - version_major = (le32_to_cpu(hdr->ucode_version) >> 24) & 0xff; - version_minor = (le32_to_cpu(hdr->ucode_version) >> 8) & 0xff; - DRM_INFO("Found VCN firmware Version: %u.%u Family ID: %u\n", - version_major, version_minor, family_id); - } + bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE; + if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) + bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); - bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8); + if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(5, 0, 0)) { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)); + log_offset = offsetof(struct amdgpu_vcn5_fw_shared, fw_log); + } else if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0)) { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)); + log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log); + } else { + fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)); + log_offset = offsetof(struct amdgpu_fw_shared, fw_log); + } - if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(5, 0, 0)) { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn5_fw_shared)); - log_offset = offsetof(struct amdgpu_vcn5_fw_shared, fw_log); - } else if (amdgpu_ip_version(adev, UVD_HWIP, 0) >= IP_VERSION(4, 0, 0)) { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_vcn4_fw_shared)); - log_offset = offsetof(struct amdgpu_vcn4_fw_shared, fw_log); - } else { - fw_shared_size = AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared)); - log_offset = offsetof(struct amdgpu_fw_shared, fw_log); - } + bo_size += fw_shared_size; + + if (amdgpu_vcnfw_log) + bo_size += AMDGPU_VCNFW_LOG_SIZE; + + r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT, + &adev->vcn.inst[i].vcpu_bo, + &adev->vcn.inst[i].gpu_addr, + &adev->vcn.inst[i].cpu_addr); + if (r) { + dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r); + return r; + } - bo_size += fw_shared_size; + adev->vcn.inst[i].fw_shared.cpu_addr = adev->vcn.inst[i].cpu_addr + + bo_size - fw_shared_size; + adev->vcn.inst[i].fw_shared.gpu_addr = adev->vcn.inst[i].gpu_addr + + bo_size - fw_shared_size; - if (amdgpu_vcnfw_log) - bo_size += AMDGPU_VCNFW_LOG_SIZE; + adev->vcn.inst[i].fw_shared.mem_size = fw_shared_size; - r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE, + if (amdgpu_vcnfw_log) { + adev->vcn.inst[i].fw_shared.cpu_addr -= AMDGPU_VCNFW_LOG_SIZE; + adev->vcn.inst[i].fw_shared.gpu_addr -= AMDGPU_VCNFW_LOG_SIZE; + adev->vcn.inst[i].fw_shared.log_offset = log_offset; + } + + if (adev->vcn.inst[i].indirect_sram) { + r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT, - &adev->vcn.inst[i].vcpu_bo, - &adev->vcn.inst[i].gpu_addr, - &adev->vcn.inst[i].cpu_addr); + &adev->vcn.inst[i].dpg_sram_bo, + &adev->vcn.inst[i].dpg_sram_gpu_addr, + &adev->vcn.inst[i].dpg_sram_cpu_addr); if (r) { - dev_err(adev->dev, "(%d) failed to allocate vcn bo\n", r); + dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r); return r; } - - adev->vcn.inst[i].fw_shared.cpu_addr = adev->vcn.inst[i].cpu_addr + - bo_size - fw_shared_size; - adev->vcn.inst[i].fw_shared.gpu_addr = adev->vcn.inst[i].gpu_addr + - bo_size - fw_shared_size; - - adev->vcn.inst[i].fw_shared.mem_size = fw_shared_size; - - if (amdgpu_vcnfw_log) { - adev->vcn.inst[i].fw_shared.cpu_addr -= AMDGPU_VCNFW_LOG_SIZE; - adev->vcn.inst[i].fw_shared.gpu_addr -= AMDGPU_VCNFW_LOG_SIZE; - adev->vcn.inst[i].fw_shared.log_offset = log_offset; - } - - if (adev->vcn.inst[i].indirect_sram) { - r = amdgpu_bo_create_kernel(adev, 64 * 2 * 4, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM | - AMDGPU_GEM_DOMAIN_GTT, - &adev->vcn.inst[i].dpg_sram_bo, - &adev->vcn.inst[i].dpg_sram_gpu_addr, - &adev->vcn.inst[i].dpg_sram_cpu_addr); - if (r) { - dev_err(adev->dev, "VCN %d (%d) failed to allocate DPG bo\n", i, r); - return r; - } - } } return 0; } -int amdgpu_vcn_sw_fini(struct amdgpu_device *adev) +int amdgpu_vcn_sw_fini(struct amdgpu_device *adev, int i) { - int i, j; + int j; - for (j = 0; j < adev->vcn.num_vcn_inst; ++j) { - if (adev->vcn.harvest_config & (1 << j)) - continue; + if (adev->vcn.harvest_config & (1 << i)) + return 0; - amdgpu_bo_free_kernel( - &adev->vcn.inst[j].dpg_sram_bo, - &adev->vcn.inst[j].dpg_sram_gpu_addr, - (void **)&adev->vcn.inst[j].dpg_sram_cpu_addr); + amdgpu_bo_free_kernel( + &adev->vcn.inst[i].dpg_sram_bo, + &adev->vcn.inst[i].dpg_sram_gpu_addr, + (void **)&adev->vcn.inst[i].dpg_sram_cpu_addr); - kvfree(adev->vcn.inst[j].saved_bo); + kvfree(adev->vcn.inst[i].saved_bo); - amdgpu_bo_free_kernel(&adev->vcn.inst[j].vcpu_bo, - &adev->vcn.inst[j].gpu_addr, - (void **)&adev->vcn.inst[j].cpu_addr); + amdgpu_bo_free_kernel(&adev->vcn.inst[i].vcpu_bo, + &adev->vcn.inst[i].gpu_addr, + (void **)&adev->vcn.inst[i].cpu_addr); - amdgpu_ring_fini(&adev->vcn.inst[j].ring_dec); + amdgpu_ring_fini(&adev->vcn.inst[i].ring_dec); - for (i = 0; i < adev->vcn.inst[j].num_enc_rings; ++i) - amdgpu_ring_fini(&adev->vcn.inst[j].ring_enc[i]); + for (j = 0; j < adev->vcn.inst[i].num_enc_rings; ++j) + amdgpu_ring_fini(&adev->vcn.inst[i].ring_enc[j]); - amdgpu_ucode_release(&adev->vcn.inst[j].fw); - mutex_destroy(&adev->vcn.inst[j].vcn_pg_lock); - mutex_destroy(&adev->vcn.inst[j].vcn1_jpeg1_workaround); - } + amdgpu_ucode_release(&adev->vcn.inst[i].fw); + mutex_destroy(&adev->vcn.inst[i].vcn_pg_lock); + mutex_destroy(&adev->vcn.inst[i].vcn1_jpeg1_workaround); return 0; } @@ -298,91 +292,102 @@ bool amdgpu_vcn_is_disabled_vcn(struct amdgpu_device *adev, enum vcn_ring_type t return ret; } -int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev) +static int amdgpu_vcn_save_vcpu_bo_inst(struct amdgpu_device *adev, int i) { unsigned int size; void *ptr; - int i, idx; + int idx; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - if (adev->vcn.inst[i].vcpu_bo == NULL) - return 0; + if (adev->vcn.harvest_config & (1 << i)) + return 0; + if (adev->vcn.inst[i].vcpu_bo == NULL) + return 0; - size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo); - ptr = adev->vcn.inst[i].cpu_addr; + size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo); + ptr = adev->vcn.inst[i].cpu_addr; - adev->vcn.inst[i].saved_bo = kvmalloc(size, GFP_KERNEL); - if (!adev->vcn.inst[i].saved_bo) - return -ENOMEM; + adev->vcn.inst[i].saved_bo = kvmalloc(size, GFP_KERNEL); + if (!adev->vcn.inst[i].saved_bo) + return -ENOMEM; - if (drm_dev_enter(adev_to_drm(adev), &idx)) { - memcpy_fromio(adev->vcn.inst[i].saved_bo, ptr, size); - drm_dev_exit(idx); - } + if (drm_dev_enter(adev_to_drm(adev), &idx)) { + memcpy_fromio(adev->vcn.inst[i].saved_bo, ptr, size); + drm_dev_exit(idx); + } + + return 0; +} + +int amdgpu_vcn_save_vcpu_bo(struct amdgpu_device *adev) +{ + int ret, i; + + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + ret = amdgpu_vcn_save_vcpu_bo_inst(adev, i); + if (ret) + return ret; } return 0; } -int amdgpu_vcn_suspend(struct amdgpu_device *adev) +int amdgpu_vcn_suspend(struct amdgpu_device *adev, int i) { bool in_ras_intr = amdgpu_ras_intr_triggered(); - int i; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) - cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); + if (adev->vcn.harvest_config & (1 << i)) + return 0; + + cancel_delayed_work_sync(&adev->vcn.inst[i].idle_work); /* err_event_athub will corrupt VCPU buffer, so we need to * restore fw data and clear buffer in amdgpu_vcn_resume() */ if (in_ras_intr) return 0; - return amdgpu_vcn_save_vcpu_bo(adev); + return amdgpu_vcn_save_vcpu_bo_inst(adev, i); } -int amdgpu_vcn_resume(struct amdgpu_device *adev) +int amdgpu_vcn_resume(struct amdgpu_device *adev, int i) { unsigned int size; void *ptr; - int i, idx; + int idx; - for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { - if (adev->vcn.harvest_config & (1 << i)) - continue; - if (adev->vcn.inst[i].vcpu_bo == NULL) - return -EINVAL; + if (adev->vcn.harvest_config & (1 << i)) + return 0; + if (adev->vcn.inst[i].vcpu_bo == NULL) + return -EINVAL; + + size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo); + ptr = adev->vcn.inst[i].cpu_addr; - size = amdgpu_bo_size(adev->vcn.inst[i].vcpu_bo); - ptr = adev->vcn.inst[i].cpu_addr; + if (adev->vcn.inst[i].saved_bo != NULL) { + if (drm_dev_enter(adev_to_drm(adev), &idx)) { + memcpy_toio(ptr, adev->vcn.inst[i].saved_bo, size); + drm_dev_exit(idx); + } + kvfree(adev->vcn.inst[i].saved_bo); + adev->vcn.inst[i].saved_bo = NULL; + } else { + const struct common_firmware_header *hdr; + unsigned int offset; - if (adev->vcn.inst[i].saved_bo != NULL) { + hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; + if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { + offset = le32_to_cpu(hdr->ucode_array_offset_bytes); if (drm_dev_enter(adev_to_drm(adev), &idx)) { - memcpy_toio(ptr, adev->vcn.inst[i].saved_bo, size); + memcpy_toio(adev->vcn.inst[i].cpu_addr, + adev->vcn.inst[i].fw->data + offset, + le32_to_cpu(hdr->ucode_size_bytes)); drm_dev_exit(idx); } - kvfree(adev->vcn.inst[i].saved_bo); - adev->vcn.inst[i].saved_bo = NULL; - } else { - const struct common_firmware_header *hdr; - unsigned int offset; - - hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { - offset = le32_to_cpu(hdr->ucode_array_offset_bytes); - if (drm_dev_enter(adev_to_drm(adev), &idx)) { - memcpy_toio(adev->vcn.inst[i].cpu_addr, - adev->vcn.inst[i].fw->data + offset, - le32_to_cpu(hdr->ucode_size_bytes)); - drm_dev_exit(idx); - } - size -= le32_to_cpu(hdr->ucode_size_bytes); - ptr += le32_to_cpu(hdr->ucode_size_bytes); - } - memset_io(ptr, 0, size); + size -= le32_to_cpu(hdr->ucode_size_bytes); + ptr += le32_to_cpu(hdr->ucode_size_bytes); } + memset_io(ptr, 0, size); } + return 0; } @@ -1058,36 +1063,32 @@ enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring) } } -void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev) +void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev, int i) { - int i; unsigned int idx; if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) { const struct common_firmware_header *hdr; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { - if (adev->vcn.harvest_config & (1 << i)) - continue; + if (adev->vcn.harvest_config & (1 << i)) + return; - hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; - /* currently only support 2 FW instances */ - if (i >= 2) { - dev_info(adev->dev, "More then 2 VCN FW instances!\n"); - break; - } - idx = AMDGPU_UCODE_ID_VCN + i; - adev->firmware.ucode[idx].ucode_id = idx; - adev->firmware.ucode[idx].fw = adev->vcn.inst[i].fw; - adev->firmware.fw_size += - ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); - - if (amdgpu_ip_version(adev, UVD_HWIP, 0) == - IP_VERSION(4, 0, 3) || - amdgpu_ip_version(adev, UVD_HWIP, 0) == - IP_VERSION(5, 0, 1)) - break; + if ((amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(4, 0, 3) || + amdgpu_ip_version(adev, UVD_HWIP, 0) == IP_VERSION(5, 0, 1)) + && (i > 0)) + return; + + hdr = (const struct common_firmware_header *)adev->vcn.inst[i].fw->data; + /* currently only support 2 FW instances */ + if (i >= 2) { + dev_info(adev->dev, "More then 2 VCN FW instances!\n"); + return; } + idx = AMDGPU_UCODE_ID_VCN + i; + adev->firmware.ucode[idx].ucode_id = idx; + adev->firmware.ucode[idx].fw = adev->vcn.inst[i].fw; + adev->firmware.fw_size += + ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE); } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h index 3b059f51894b..5f92cf28cc75 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h @@ -503,11 +503,11 @@ enum vcn_ring_type { VCN_UNIFIED_RING, }; -int amdgpu_vcn_early_init(struct amdgpu_device *adev); -int amdgpu_vcn_sw_init(struct amdgpu_device *adev); -int amdgpu_vcn_sw_fini(struct amdgpu_device *adev); -int amdgpu_vcn_suspend(struct amdgpu_device *adev); -int amdgpu_vcn_resume(struct amdgpu_device *adev); +int amdgpu_vcn_early_init(struct amdgpu_device *adev, int i); +int amdgpu_vcn_sw_init(struct amdgpu_device *adev, int i); +int amdgpu_vcn_sw_fini(struct amdgpu_device *adev, int i); +int amdgpu_vcn_suspend(struct amdgpu_device *adev, int i); +int amdgpu_vcn_resume(struct amdgpu_device *adev, int i); void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring); void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring); @@ -525,7 +525,7 @@ int amdgpu_vcn_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout); enum amdgpu_ring_priority_level amdgpu_vcn_get_enc_ring_prio(int ring); -void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev); +void amdgpu_vcn_setup_ucode(struct amdgpu_device *adev, int i); void amdgpu_vcn_fwlog_init(struct amdgpu_vcn_inst *vcn); void amdgpu_debugfs_vcn_fwlog_init(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index edbcd0819d17..49542d947727 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -113,7 +113,7 @@ static int vcn_v1_0_early_init(struct amdgpu_ip_block *ip_block) jpeg_v1_0_early_init(ip_block); - return amdgpu_vcn_early_init(adev); + return amdgpu_vcn_early_init(adev, 0); } /** @@ -145,16 +145,16 @@ static int vcn_v1_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - r = amdgpu_vcn_sw_init(adev); + r = amdgpu_vcn_sw_init(adev, 0); if (r) return r; /* Override the work func */ adev->vcn.inst[0].idle_work.work.func = vcn_v1_0_idle_work_handler; - amdgpu_vcn_setup_ucode(adev); + amdgpu_vcn_setup_ucode(adev, 0); - r = amdgpu_vcn_resume(adev); + r = amdgpu_vcn_resume(adev, 0); if (r) return r; @@ -223,13 +223,13 @@ static int vcn_v1_0_sw_fini(struct amdgpu_ip_block *ip_block) int r; struct amdgpu_device *adev = ip_block->adev; - r = amdgpu_vcn_suspend(adev); + r = amdgpu_vcn_suspend(adev, 0); if (r) return r; jpeg_v1_0_sw_fini(ip_block); - r = amdgpu_vcn_sw_fini(adev); + r = amdgpu_vcn_sw_fini(adev, 0); kfree(adev->vcn.ip_dump); @@ -311,7 +311,7 @@ static int vcn_v1_0_suspend(struct amdgpu_ip_block *ip_block) if (r) return r; - r = amdgpu_vcn_suspend(adev); + r = amdgpu_vcn_suspend(adev, 0); return r; } @@ -327,7 +327,7 @@ static int vcn_v1_0_resume(struct amdgpu_ip_block *ip_block) { int r; - r = amdgpu_vcn_resume(ip_block->adev); + r = amdgpu_vcn_resume(ip_block->adev, 0); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c index d94fd075a910..5f537d2a5618 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c @@ -118,7 +118,7 @@ static int vcn_v2_0_early_init(struct amdgpu_ip_block *ip_block) vcn_v2_0_set_enc_ring_funcs(adev); vcn_v2_0_set_irq_funcs(adev); - return amdgpu_vcn_early_init(adev); + return amdgpu_vcn_early_init(adev, 0); } /** @@ -153,13 +153,13 @@ static int vcn_v2_0_sw_init(struct amdgpu_ip_block *ip_block) return r; } - r = amdgpu_vcn_sw_init(adev); + r = amdgpu_vcn_sw_init(adev, 0); if (r) return r; - amdgpu_vcn_setup_ucode(adev); + amdgpu_vcn_setup_ucode(adev, 0); - r = amdgpu_vcn_resume(adev); + r = amdgpu_vcn_resume(adev, 0); if (r) return r; @@ -254,11 +254,11 @@ static int vcn_v2_0_sw_fini(struct amdgpu_ip_block *ip_block) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); + r = amdgpu_vcn_suspend(adev, 0); if (r) return r; - r = amdgpu_vcn_sw_fini(adev); + r = amdgpu_vcn_sw_fini(adev, 0); kfree(adev->vcn.ip_dump); @@ -338,7 +338,7 @@ static int vcn_v2_0_suspend(struct amdgpu_ip_block *ip_block) if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + r = amdgpu_vcn_suspend(ip_block->adev, 0); return r; } @@ -354,7 +354,7 @@ static int vcn_v2_0_resume(struct amdgpu_ip_block *ip_block) { int r; - r = amdgpu_vcn_resume(ip_block->adev); + r = amdgpu_vcn_resume(ip_block->adev, 0); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c index c20619c027e7..3ee00dc6d533 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c @@ -121,7 +121,7 @@ static int amdgpu_ih_clientid_vcns[] = { static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; if (amdgpu_sriov_vf(adev)) { adev->vcn.num_vcn_inst = 2; @@ -139,7 +139,7 @@ static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block) adev->vcn.inst[i].num_enc_rings = 2; } if (adev->vcn.harvest_config == (AMDGPU_VCN_HARVEST_VCN0 | - AMDGPU_VCN_HARVEST_VCN1)) + AMDGPU_VCN_HARVEST_VCN1)) /* both instances are harvested, disable the block */ return -ENOENT; } @@ -149,7 +149,13 @@ static int vcn_v2_5_early_init(struct amdgpu_ip_block *ip_block) vcn_v2_5_set_irq_funcs(adev); vcn_v2_5_set_ras_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } /** @@ -168,6 +174,8 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; for (j = 0; j < adev->vcn.num_vcn_inst; j++) { + volatile struct amdgpu_fw_shared *fw_shared; + if (adev->vcn.harvest_config & (1 << j)) continue; /* VCN DEC TRAP */ @@ -189,23 +197,17 @@ static int vcn_v2_5_sw_init(struct amdgpu_ip_block *ip_block) VCN_2_6__SRCID_UVD_POISON, &adev->vcn.inst[j].ras_poison_irq); if (r) return r; - } - - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - amdgpu_vcn_setup_ucode(adev); + r = amdgpu_vcn_sw_init(adev, j); + if (r) + return r; - r = amdgpu_vcn_resume(adev); - if (r) - return r; + amdgpu_vcn_setup_ucode(adev, j); - for (j = 0; j < adev->vcn.num_vcn_inst; j++) { - volatile struct amdgpu_fw_shared *fw_shared; + r = amdgpu_vcn_resume(adev, j); + if (r) + return r; - if (adev->vcn.harvest_config & (1 << j)) - continue; adev->vcn.inst[j].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; adev->vcn.inst[j].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; adev->vcn.inst[j].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; @@ -323,15 +325,18 @@ static int vcn_v2_5_sw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); - if (r) - return r; - - r = amdgpu_vcn_sw_fini(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -421,15 +426,20 @@ static int vcn_v2_5_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v2_5_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v2_5_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } - return r; + return 0; } /** @@ -441,11 +451,14 @@ static int vcn_v2_5_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v2_5_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v2_5_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 282a186cb0c5..6052e963fcab 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -127,7 +127,7 @@ static void vcn_v3_0_enc_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v3_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; if (amdgpu_sriov_vf(adev)) { adev->vcn.num_vcn_inst = VCN_INSTANCES_SIENNA_CICHLID; @@ -154,7 +154,12 @@ static int vcn_v3_0_early_init(struct amdgpu_ip_block *ip_block) vcn_v3_0_set_enc_ring_funcs(adev); vcn_v3_0_set_irq_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + return 0; } /** @@ -173,16 +178,6 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) uint32_t *ptr; struct amdgpu_device *adev = ip_block->adev; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; - /* * Note: doorbell assignment is fixed for SRIOV multiple VCN engines * Formula: @@ -202,6 +197,16 @@ static int vcn_v3_0_sw_init(struct amdgpu_ip_block *ip_block) if (adev->vcn.harvest_config & (1 << i)) continue; + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + adev->vcn.inst[i].internal.context_id = mmUVD_CONTEXT_ID_INTERNAL_OFFSET; adev->vcn.inst[i].internal.ib_vmid = mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET; adev->vcn.inst[i].internal.ib_bar_low = mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET; @@ -333,14 +338,18 @@ static int vcn_v3_0_sw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; - r = amdgpu_vcn_sw_fini(adev); + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -459,15 +468,20 @@ static int vcn_v3_0_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v3_0_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v3_0_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } - return r; + return 0; } /** @@ -479,11 +493,14 @@ static int vcn_v3_0_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v3_0_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v3_0_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c index f5bb241c1af1..8c7566960a3b 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c @@ -117,7 +117,7 @@ static void vcn_v4_0_set_ras_funcs(struct amdgpu_device *adev); static int vcn_v4_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; if (amdgpu_sriov_vf(adev)) { adev->vcn.harvest_config = VCN_HARVEST_MMSCH; @@ -137,7 +137,13 @@ static int vcn_v4_0_early_init(struct amdgpu_ip_block *ip_block) vcn_v4_0_set_irq_funcs(adev); vcn_v4_0_set_ras_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } static int vcn_v4_0_fw_shared_init(struct amdgpu_device *adev, int inst_idx) @@ -180,20 +186,20 @@ static int vcn_v4_0_sw_init(struct amdgpu_ip_block *ip_block) uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_4_0); uint32_t *ptr; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { if (adev->vcn.harvest_config & (1 << i)) continue; + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + /* Init instance 0 sched_score to 1, so it's scheduled after other instances */ if (i == 0) atomic_set(&adev->vcn.inst[i].sched_score, 1); @@ -294,16 +300,23 @@ static int vcn_v4_0_sw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + } amdgpu_vcn_sysfs_reset_mask_fini(adev); - r = amdgpu_vcn_sw_fini(adev); + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -394,15 +407,20 @@ static int vcn_v4_0_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v4_0_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } - return r; + return 0; } /** @@ -414,11 +432,14 @@ static int vcn_v4_0_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v4_0_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c index fbe599c539c8..a0b848ad097b 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c @@ -114,7 +114,7 @@ static inline bool vcn_v4_0_3_normalizn_reqd(struct amdgpu_device *adev) static int vcn_v4_0_3_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) /* re-use enc ring as unified ring */ @@ -124,7 +124,13 @@ static int vcn_v4_0_3_early_init(struct amdgpu_ip_block *ip_block) vcn_v4_0_3_set_irq_funcs(adev); vcn_v4_0_3_set_ras_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } static int vcn_v4_0_3_fw_shared_init(struct amdgpu_device *adev, int inst_idx) @@ -156,16 +162,6 @@ static int vcn_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_4_0_3); uint32_t *ptr; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; - /* VCN DEC TRAP */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, VCN_4_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq); @@ -173,6 +169,17 @@ static int vcn_v4_0_3_sw_init(struct amdgpu_ip_block *ip_block) return r; for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + vcn_inst = GET_INST(VCN, i); ring = &adev->vcn.inst[i].ring_enc[0]; @@ -261,16 +268,23 @@ static int vcn_v4_0_3_sw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + } amdgpu_vcn_sysfs_reset_mask_fini(adev); - r = amdgpu_vcn_sw_fini(adev); + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -375,15 +389,20 @@ static int vcn_v4_0_3_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_3_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v4_0_3_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + } - return r; + return 0; } /** @@ -395,11 +414,14 @@ static int vcn_v4_0_3_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_3_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v4_0_3_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c index 2513929b2d43..68ed7950e3d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c @@ -115,7 +115,7 @@ static void vcn_v4_0_5_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v4_0_5_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) /* re-use enc ring as unified ring */ @@ -123,7 +123,13 @@ static int vcn_v4_0_5_early_init(struct amdgpu_ip_block *ip_block) vcn_v4_0_5_set_unified_ring_funcs(adev); vcn_v4_0_5_set_irq_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } /** @@ -141,15 +147,6 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) uint32_t reg_count = ARRAY_SIZE(vcn_reg_list_4_0_5); uint32_t *ptr; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; for (i = 0; i < adev->vcn.num_vcn_inst; i++) { volatile struct amdgpu_vcn4_fw_shared *fw_shared; @@ -157,6 +154,16 @@ static int vcn_v4_0_5_sw_init(struct amdgpu_ip_block *ip_block) if (adev->vcn.harvest_config & (1 << i)) continue; + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + atomic_set(&adev->vcn.inst[i].sched_score, 0); /* VCN UNIFIED TRAP */ @@ -252,15 +259,19 @@ static int vcn_v4_0_5_sw_fini(struct amdgpu_ip_block *ip_block) if (amdgpu_sriov_vf(adev)) amdgpu_virt_free_mm_table(adev); - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; - r = amdgpu_vcn_sw_fini(adev); + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -332,13 +343,18 @@ static int vcn_v4_0_5_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_5_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v4_0_5_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } return r; } @@ -352,11 +368,14 @@ static int vcn_v4_0_5_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v4_0_5_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v4_0_5_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c index 50e0928c3a20..d773766ebbe2 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c @@ -98,7 +98,7 @@ static void vcn_v5_0_0_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v5_0_0_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) /* re-use enc ring as unified ring */ @@ -107,7 +107,13 @@ static int vcn_v5_0_0_early_init(struct amdgpu_ip_block *ip_block) vcn_v5_0_0_set_unified_ring_funcs(adev); vcn_v5_0_0_set_irq_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } void vcn_v5_0_0_alloc_ip_dump(struct amdgpu_device *adev) @@ -138,22 +144,22 @@ static int vcn_v5_0_0_sw_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int i, r; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; - for (i = 0; i < adev->vcn.num_vcn_inst; i++) { volatile struct amdgpu_vcn5_fw_shared *fw_shared; if (adev->vcn.harvest_config & (1 << i)) continue; + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + atomic_set(&adev->vcn.inst[i].sched_score, 0); /* VCN UNIFIED TRAP */ @@ -231,16 +237,23 @@ static int vcn_v5_0_0_sw_fini(struct amdgpu_ip_block *ip_block) drm_dev_exit(idx); } - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + } amdgpu_vcn_sysfs_reset_mask_fini(adev); - r = amdgpu_vcn_sw_fini(adev); + + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -312,13 +325,18 @@ static int vcn_v5_0_0_hw_fini(struct amdgpu_ip_block *ip_block) */ static int vcn_v5_0_0_suspend(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; r = vcn_v5_0_0_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(ip_block->adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } return r; } @@ -332,11 +350,14 @@ static int vcn_v5_0_0_suspend(struct amdgpu_ip_block *ip_block) */ static int vcn_v5_0_0_resume(struct amdgpu_ip_block *ip_block) { - int r; + struct amdgpu_device *adev = ip_block->adev; + int r, i; - r = amdgpu_vcn_resume(ip_block->adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v5_0_0_hw_init(ip_block); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c index ebe977507bc3..afa8f1f0b572 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c @@ -58,7 +58,7 @@ static void vcn_v5_0_1_unified_ring_set_wptr(struct amdgpu_ring *ring); static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int i; + int i, r; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) /* re-use enc ring as unified ring */ @@ -67,7 +67,13 @@ static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block) vcn_v5_0_1_set_unified_ring_funcs(adev); vcn_v5_0_1_set_irq_funcs(adev); - return amdgpu_vcn_early_init(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + r = amdgpu_vcn_early_init(adev, i); + if (r) + return r; + } + + return 0; } /** @@ -83,16 +89,6 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) struct amdgpu_ring *ring; int i, r, vcn_inst; - r = amdgpu_vcn_sw_init(adev); - if (r) - return r; - - amdgpu_vcn_setup_ucode(adev); - - r = amdgpu_vcn_resume(adev); - if (r) - return r; - /* VCN UNIFIED TRAP */ r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN, VCN_5_0__SRCID__UVD_ENC_GENERAL_PURPOSE, &adev->vcn.inst->irq); @@ -104,6 +100,16 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block) vcn_inst = GET_INST(VCN, i); + r = amdgpu_vcn_sw_init(adev, i); + if (r) + return r; + + amdgpu_vcn_setup_ucode(adev, i); + + r = amdgpu_vcn_resume(adev, i); + if (r) + return r; + ring = &adev->vcn.inst[i].ring_enc[0]; ring->use_doorbell = true; ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 9 * vcn_inst; @@ -157,17 +163,23 @@ static int vcn_v5_0_1_sw_fini(struct amdgpu_ip_block *ip_block) drm_dev_exit(idx); } - r = amdgpu_vcn_suspend(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(adev, i); + if (r) + return r; + } - r = amdgpu_vcn_sw_fini(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_sw_fini(adev, i); + if (r) + return r; + } amdgpu_vcn_sysfs_reset_mask_fini(adev); kfree(adev->vcn.ip_dump); - return r; + return 0; } /** @@ -229,13 +241,17 @@ static int vcn_v5_0_1_hw_fini(struct amdgpu_ip_block *ip_block) static int vcn_v5_0_1_suspend(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int r; + int r, i; r = vcn_v5_0_1_hw_fini(ip_block); if (r) return r; - r = amdgpu_vcn_suspend(adev); + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_suspend(ip_block->adev, i); + if (r) + return r; + } return r; } @@ -250,11 +266,13 @@ static int vcn_v5_0_1_suspend(struct amdgpu_ip_block *ip_block) static int vcn_v5_0_1_resume(struct amdgpu_ip_block *ip_block) { struct amdgpu_device *adev = ip_block->adev; - int r; + int r, i; - r = amdgpu_vcn_resume(adev); - if (r) - return r; + for (i = 0; i < adev->vcn.num_vcn_inst; i++) { + r = amdgpu_vcn_resume(ip_block->adev, i); + if (r) + return r; + } r = vcn_v5_0_1_hw_init(ip_block); -- 2.51.0 From 710151263ca5f2ac3d120e3d727d2c75be905d7f Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 19 Nov 2024 16:10:46 -0500 Subject: [PATCH 16/16] drm/amdgpu/vcn3.0: convert internal functions to use vcn_inst Pass the vcn instance structure to these functions rather than adev and the instance number. TODO: clean up the function internals to use the vinst state directly rather than accessing it indirectly via adev->vcn.inst[]. Reviewed-by: Boyuan Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c | 79 ++++++++++++++++++--------- 1 file changed, 52 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c index 6052e963fcab..2fd58d548b19 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c @@ -510,13 +510,14 @@ static int vcn_v3_0_resume(struct amdgpu_ip_block *ip_block) /** * vcn_v3_0_mc_resume - memory controller programming * - * @adev: amdgpu_device pointer - * @inst: instance number + * @vinst: VCN instance * * Let the VCN memory controller know it's offsets */ -static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst) +static void vcn_v3_0_mc_resume(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst = vinst->inst; uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst].fw->size + 4); uint32_t offset; @@ -565,8 +566,11 @@ static void vcn_v3_0_mc_resume(struct amdgpu_device *adev, int inst) AMDGPU_GPU_PAGE_ALIGN(sizeof(struct amdgpu_fw_shared))); } -static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect) +static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst, + bool indirect) { + struct amdgpu_device *adev = vinst->adev; + int inst_idx = vinst->inst; uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.inst[inst_idx].fw->size + 4); uint32_t offset; @@ -661,8 +665,10 @@ static void vcn_v3_0_mc_resume_dpg_mode(struct amdgpu_device *adev, int inst_idx UVD, inst_idx, mmUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect); } -static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int inst) +static void vcn_v3_0_disable_static_power_gating(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst = vinst->inst; uint32_t data = 0; if (adev->pg_flags & AMD_PG_SUPPORT_VCN) { @@ -712,8 +718,10 @@ static void vcn_v3_0_disable_static_power_gating(struct amdgpu_device *adev, int WREG32_SOC15(VCN, inst, mmUVD_POWER_STATUS, data); } -static void vcn_v3_0_enable_static_power_gating(struct amdgpu_device *adev, int inst) +static void vcn_v3_0_enable_static_power_gating(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst = vinst->inst; uint32_t data; if (adev->pg_flags & AMD_PG_SUPPORT_VCN) { @@ -765,8 +773,10 @@ static void vcn_v3_0_enable_static_power_gating(struct amdgpu_device *adev, int * * Disable clock gating for VCN block */ -static void vcn_v3_0_disable_clock_gating(struct amdgpu_device *adev, int inst) +static void vcn_v3_0_disable_clock_gating(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst = vinst->inst; uint32_t data; /* VCN disable CGC */ @@ -893,9 +903,12 @@ static void vcn_v3_0_disable_clock_gating(struct amdgpu_device *adev, int inst) WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data); } -static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev, - uint8_t sram_sel, int inst_idx, uint8_t indirect) +static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_vcn_inst *vinst, + uint8_t sram_sel, + uint8_t indirect) { + struct amdgpu_device *adev = vinst->adev; + int inst_idx = vinst->inst; uint32_t reg_data = 0; /* enable sw clock gating control */ @@ -949,8 +962,10 @@ static void vcn_v3_0_clock_gating_dpg_mode(struct amdgpu_device *adev, * * Enable clock gating for VCN block */ -static void vcn_v3_0_enable_clock_gating(struct amdgpu_device *adev, int inst) +static void vcn_v3_0_enable_clock_gating(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst = vinst->inst; uint32_t data; /* enable VCN CGC */ @@ -1009,8 +1024,10 @@ static void vcn_v3_0_enable_clock_gating(struct amdgpu_device *adev, int inst) WREG32_SOC15(VCN, inst, mmUVD_SUVD_CGC_CTRL, data); } -static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, bool indirect) +static int vcn_v3_0_start_dpg_mode(struct amdgpu_vcn_inst *vinst, bool indirect) { + struct amdgpu_device *adev = vinst->adev; + int inst_idx = vinst->inst; volatile struct amdgpu_fw_shared *fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr; struct amdgpu_ring *ring; uint32_t rb_bufsz, tmp; @@ -1028,7 +1045,7 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo adev->vcn.inst[inst_idx].dpg_sram_curr_addr = (uint32_t *)adev->vcn.inst[inst_idx].dpg_sram_cpu_addr; /* enable clock gating */ - vcn_v3_0_clock_gating_dpg_mode(adev, 0, inst_idx, indirect); + vcn_v3_0_clock_gating_dpg_mode(vinst, 0, indirect); /* enable VCPU clock */ tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT); @@ -1077,7 +1094,7 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0, indirect); - vcn_v3_0_mc_resume_dpg_mode(adev, inst_idx, indirect); + vcn_v3_0_mc_resume_dpg_mode(vinst, indirect); WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET( VCN, inst_idx, mmUVD_REG_XX_MASK), 0x10, 0, indirect); @@ -1161,8 +1178,10 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo return 0; } -static int vcn_v3_0_start(struct amdgpu_device *adev, int i) +static int vcn_v3_0_start(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int i = vinst->inst; volatile struct amdgpu_fw_shared *fw_shared; struct amdgpu_ring *ring; uint32_t rb_bufsz, tmp; @@ -1175,17 +1194,17 @@ static int vcn_v3_0_start(struct amdgpu_device *adev, int i) amdgpu_dpm_enable_vcn(adev, true, i); if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) - return vcn_v3_0_start_dpg_mode(adev, i, adev->vcn.inst[i].indirect_sram); + return vcn_v3_0_start_dpg_mode(vinst, vinst->indirect_sram); /* disable VCN power gating */ - vcn_v3_0_disable_static_power_gating(adev, i); + vcn_v3_0_disable_static_power_gating(vinst); /* set VCN status busy */ tmp = RREG32_SOC15(VCN, i, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY; WREG32_SOC15(VCN, i, mmUVD_STATUS, tmp); /* SW clock gating */ - vcn_v3_0_disable_clock_gating(adev, i); + vcn_v3_0_disable_clock_gating(vinst); /* enable VCPU clock */ WREG32_P(SOC15_REG_OFFSET(VCN, i, mmUVD_VCPU_CNTL), @@ -1238,7 +1257,7 @@ static int vcn_v3_0_start(struct amdgpu_device *adev, int i) (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) | (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT))); - vcn_v3_0_mc_resume(adev, i); + vcn_v3_0_mc_resume(vinst); /* VCN global tiling registers */ WREG32_SOC15(VCN, i, mmUVD_GFX10_ADDR_CONFIG, @@ -1555,8 +1574,10 @@ static int vcn_v3_0_start_sriov(struct amdgpu_device *adev) return 0; } -static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) +static int vcn_v3_0_stop_dpg_mode(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int inst_idx = vinst->inst; struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE}; uint32_t tmp; @@ -1586,8 +1607,10 @@ static int vcn_v3_0_stop_dpg_mode(struct amdgpu_device *adev, int inst_idx) return 0; } -static int vcn_v3_0_stop(struct amdgpu_device *adev, int i) +static int vcn_v3_0_stop(struct amdgpu_vcn_inst *vinst) { + struct amdgpu_device *adev = vinst->adev; + int i = vinst->inst; uint32_t tmp; int r = 0; @@ -1595,7 +1618,7 @@ static int vcn_v3_0_stop(struct amdgpu_device *adev, int i) return 0; if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) { - r = vcn_v3_0_stop_dpg_mode(adev, i); + r = vcn_v3_0_stop_dpg_mode(vinst); goto done; } @@ -1648,10 +1671,10 @@ static int vcn_v3_0_stop(struct amdgpu_device *adev, int i) WREG32_SOC15(VCN, i, mmUVD_STATUS, 0); /* apply HW clock gating */ - vcn_v3_0_enable_clock_gating(adev, i); + vcn_v3_0_enable_clock_gating(vinst); /* enable VCN power gating */ - vcn_v3_0_enable_static_power_gating(adev, i); + vcn_v3_0_enable_static_power_gating(vinst); done: if (adev->pm.dpm_enabled) @@ -2162,15 +2185,16 @@ static int vcn_v3_0_set_clockgating_state(struct amdgpu_ip_block *ip_block, int i; for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i]; if (adev->vcn.harvest_config & (1 << i)) continue; if (enable) { if (RREG32_SOC15(VCN, i, mmUVD_STATUS) != UVD_STATUS__IDLE) return -EBUSY; - vcn_v3_0_enable_clock_gating(adev, i); + vcn_v3_0_enable_clock_gating(vinst); } else { - vcn_v3_0_disable_clock_gating(adev, i); + vcn_v3_0_disable_clock_gating(vinst); } } @@ -2182,6 +2206,7 @@ static int vcn_v3_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, int i) { struct amdgpu_device *adev = ip_block->adev; + struct amdgpu_vcn_inst *vinst = &adev->vcn.inst[i]; int ret = 0; /* for SRIOV, guest should not control VCN Power-gating @@ -2197,9 +2222,9 @@ static int vcn_v3_0_set_powergating_state_inst(struct amdgpu_ip_block *ip_block, return 0; if (state == AMD_PG_STATE_GATE) - ret = vcn_v3_0_stop(adev, i); + ret = vcn_v3_0_stop(vinst); else - ret = vcn_v3_0_start(adev, i); + ret = vcn_v3_0_start(vinst); if (!ret) adev->vcn.inst[i].cur_state = state; -- 2.51.0