drm/amdgpu: refactor RLCG access path part 1
Subsystems (3)
PATCH 1/2 drm/amdgpu: refactor RLCG access path part 1
what changed:
1)provide new implementation interface for the rlcg access path 2)put SQ_CMD/SQ_IND_INDEX/SQ_IND_DATA to GFX9 RLCG path to align with SRIOV RLCG logic
background:
we what to clear the code path for WREG32_RLC, to make it only covered and handled by amdgpu_mm_wreg() routine, this way we can let RLCG to serve the register access even through UMR (via debugfs interface) the current implementation cannot achieve that goal because it can only hardcode everywhere, but UMR only pass "offset" as varable to driver
--- drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h | 2 + drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 80 ++++++++++++++- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 177 +++++++++++++++++++++++++++++++- drivers/gpu/drm/amd/amdgpu/soc15.h | 7 ++ 4 files changed, 264 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h +2 -0
@@ -127,6 +127,8 @@ struct amdgpu_rlc_funcs {void (*reset)(struct amdgpu_device *adev);void (*start)(struct amdgpu_device *adev);void (*update_spm_vmid)(struct amdgpu_device *adev,unsigned vmid); +void (*rlcg_wreg)(struct amdgpu_device *adev,u32 offset,u32 v); +bool (*is_rlcg_access_range)(struct amdgpu_device *adev,uint32_t reg); };struct amdgpu_rlc {
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +79 -1
98 lines not on this page · raw message
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +176 -1
201 lines not on this page · raw message
drivers/gpu/drm/amd/amdgpu/soc15.h +7 -0
@@ -42,6 +42,13 @@ struct soc15_reg_golden {u32 or_mask; }; +struct soc15_reg_rlcg { +u32 hwip; +u32 instance; +u32 segment; +u32 reg; +}; +struct soc15_reg_entry {uint32_t hwip;uint32_t inst;
Signature · 6 lines
2.7.4 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 2/2 drm/amdgpu: refactor RLCG access path part 2
switch to new RLCG access path, and drop the legacy WREG32_RLC macros
--- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 30 +++---- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 ++ drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 8 +- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 104 +++++++++++----------- drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c | 2 +- drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 28 +++--- drivers/gpu/drm/amd/amdgpu/soc15.c | 11 +-- drivers/gpu/drm/amd/amdgpu/soc15_common.h | 57 ------------ 8 files changed, 93 insertions(+), 152 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +15 -15
@@ -105,8 +105,8 @@ void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,lock_srbm (kgd,0 ,0 ,0 , vmid); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_CONFIG), sh_mem_config); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_BASES), sh_mem_bases); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_CONFIG), sh_mem_config); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_BASES), sh_mem_bases);/* APE1 no longer exists on GFX9 */ unlock_srbm (kgd);
@@ -242,13 +242,13 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,for (reg = hqd_base; reg <=SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_HI); reg++) -WREG32_RLC (reg, mqd_hqd[reg - hqd_base]); +WREG32 (reg, mqd_hqd[reg - hqd_base]);/* Activate doorbell logic before triggering WPTR poll. */ data =REG_SET_FIELD (m->cp_hqd_pq_doorbell_control,CP_HQD_PQ_DOORBELL_CONTROL ,DOORBELL_EN ,1 ); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_DOORBELL_CONTROL), data); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_DOORBELL_CONTROL), data);if (wptr) {/* Don't read wptr with get_user because the user
@@ -277,25 +277,25 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id, guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size -1 ); guessed_wptr += (uint64_t )m->cp_hqd_pq_wptr_hi <<32 ; -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_LO), +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_LO),lower_32_bits (guessed_wptr)); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_HI), +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_HI),upper_32_bits (guessed_wptr)); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_POLL_ADDR), +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_POLL_ADDR),lower_32_bits ((uintptr_t )wptr)); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_POLL_ADDR_HI), +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),upper_32_bits ((uintptr_t )wptr));WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_PQ_WPTR_POLL_CNTL1), - (uint32_t )get_queue_mask (adev, pipe_id, queue_id)); +get_queue_mask (adev, pipe_id, queue_id)); }/* Start the EOP fetcher */ -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_EOP_RPTR), +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_EOP_RPTR),REG_SET_FIELD (m->cp_hqd_eop_rptr,CP_HQD_EOP_RPTR ,INIT_FETCHER ,1 )); data =REG_SET_FIELD (m->cp_hqd_active,CP_HQD_ACTIVE ,ACTIVE ,1 ); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_ACTIVE), data); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_ACTIVE), data);release_queue (kgd);
@@ -547,7 +547,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,acquire_queue (kgd, pipe_id, queue_id);if (m->cp_hqd_vmid ==0 ) -WREG32_FIELD15_RLC (GC ,0 ,RLC_CP_SCHEDULERS , scheduler1,0 ); +WREG32_FIELD15 (GC ,0 ,RLC_CP_SCHEDULERS , scheduler1,0 );switch (reset_type) {case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN :
@@ -561,7 +561,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,break ; } -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_DEQUEUE_REQUEST), type); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_DEQUEUE_REQUEST), type); end_jiffies = (utimeout *HZ /1000 ) + jiffies;while (true) {
@@ -656,7 +656,7 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,mutex_lock (&adev->grbm_idx_mutex); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_INDEX, gfx_index_val); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_INDEX, gfx_index_val);WREG32 (SOC15_REG_OFFSET (GC ,0 , mmSQ_CMD), sq_cmd); data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
@@ -666,7 +666,7 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd, data =REG_SET_FIELD (data,GRBM_GFX_INDEX ,SE_BROADCAST_WRITES ,1 ); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_INDEX, data); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_INDEX, data);mutex_unlock (&adev->grbm_idx_mutex);return 0 ;
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +5 -0
@@ -328,6 +328,11 @@ void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,if ((acc_flags &AMDGPU_REGS_KIQ ) || (!(acc_flags &AMDGPU_REGS_NO_KIQ ) &&amdgpu_sriov_runtime (adev)))return amdgpu_kiq_wreg (adev, reg, v); +if (amdgpu_sriov_vf (adev) && adev->gfx.rlc.funcs && + adev->gfx.rlc.funcs->is_rlcg_access_range) +if (adev->gfx.rlc.funcs->is_rlcg_access_range (adev, reg)) +return adev->gfx.rlc.funcs->rlcg_wreg (adev, reg, v); +if ((reg *4 ) < adev->rmmio_size && !(acc_flags &AMDGPU_REGS_IDX ))writel (v, ((void __iomem *)adev->rmmio) + (reg *4 ));else {
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +4 -4
@@ -1835,11 +1835,11 @@ static int gfx_v10_0_init_csb(struct amdgpu_device *adev) adev->gfx.rlc.funcs->get_csb_buffer (adev, adev->gfx.rlc.cs_ptr);/* csib */ -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_ADDR_HI, +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_ADDR_HI, adev->gfx.rlc.clear_state_gpu_addr >>32 ); -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_ADDR_LO, +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_ADDR_LO, adev->gfx.rlc.clear_state_gpu_addr &0xfffffffc ); -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size); +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size);return 0 ; }
@@ -2447,7 +2447,7 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)for (i =0 ; i < adev->gfx.num_gfx_rings; i++) adev->gfx.gfx_ring[i].sched.ready = false; } -WREG32_SOC15_RLC (GC ,0 , mmCP_ME_CNTL, tmp); +WREG32_SOC15 (GC ,0 , mmCP_ME_CNTL, tmp);for (i =0 ; i < adev->usec_timeout; i++) {if (RREG32_SOC15 (GC ,0 , mmCP_STAT) ==0 )
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +52 -52
247 lines not on this page · raw message
drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c +1 -1
@@ -114,7 +114,7 @@ static void gfx_v9_4_select_se_sh(struct amdgpu_device *adev, u32 se_num,else data =REG_SET_FIELD (data,GRBM_GFX_INDEX ,SH_INDEX , sh_num); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_INDEX, data); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_INDEX, data); }static const struct soc15_ras_field_entry gfx_v9_4_ras_fields[] = {
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +14 -14
@@ -71,13 +71,13 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)uint64_t value;/* Program the AGP BAR */ -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_BASE,0 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_BOT, adev->gmc.agp_start >>24 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_TOP, adev->gmc.agp_end >>24 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_BASE,0 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_BOT, adev->gmc.agp_start >>24 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_TOP, adev->gmc.agp_end >>24 );if (!amdgpu_sriov_vf (adev) || adev->asic_type <=CHIP_VEGA10 ) {/* Program the system aperture low logical page number. */ -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_SYSTEM_APERTURE_LOW_ADDR, +WREG32_SOC15 (GC ,0 , mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,min (adev->gmc.fb_start, adev->gmc.agp_start) >>18 );if (adev->asic_type ==CHIP_RAVEN && adev->rev_id >=0x8 )
@@ -88,12 +88,12 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)* aperture high address (add 1) to get rid of the VM * fault and hardware hang. */ -WREG32_SOC15_RLC (GC ,0 , +WREG32_SOC15 (GC ,0 , mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,max ((adev->gmc.fb_end >>18 ) +0x1 , adev->gmc.agp_end >>18 )); else -WREG32_SOC15_RLC ( +WREG32_SOC15 ( GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
@@ -134,7 +134,7 @@ static void gfxhub_v1_0_init_tlb_regs(struct amdgpu_device *adev) MTYPE, MTYPE_UC);/* XXX for emulation. */ tmp =REG_SET_FIELD (tmp,MC_VM_MX_L1_TLB_CNTL ,ATC_EN ,1 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_MX_L1_TLB_CNTL, tmp); +WREG32_SOC15 (GC ,0 , mmMC_VM_MX_L1_TLB_CNTL, tmp); }static void gfxhub_v1_0_init_cache_regs (struct amdgpu_device *adev)
@@ -151,12 +151,12 @@ static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev) tmp =REG_SET_FIELD (tmp,VM_L2_CNTL ,PDE_FAULT_CLASSIFICATION ,0 ); tmp =REG_SET_FIELD (tmp,VM_L2_CNTL ,CONTEXT1_IDENTITY_ACCESS_MODE ,1 ); tmp =REG_SET_FIELD (tmp,VM_L2_CNTL ,IDENTITY_MODE_FRAGMENT_SIZE ,0 ); -WREG32_SOC15_RLC (GC ,0 , mmVM_L2_CNTL, tmp); +WREG32_SOC15 (GC ,0 , mmVM_L2_CNTL, tmp); tmp =RREG32_SOC15 (GC ,0 , mmVM_L2_CNTL2); tmp =REG_SET_FIELD (tmp,VM_L2_CNTL2 ,INVALIDATE_ALL_L1_TLBS ,1 ); tmp =REG_SET_FIELD (tmp,VM_L2_CNTL2 ,INVALIDATE_L2_CACHE ,1 ); -WREG32_SOC15_RLC (GC ,0 , mmVM_L2_CNTL2, tmp); +WREG32_SOC15 (GC ,0 , mmVM_L2_CNTL2, tmp); tmp = mmVM_L2_CNTL3_DEFAULT;if (adev->gmc.translate_further) {
@@ -168,12 +168,12 @@ static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev) tmp =REG_SET_FIELD (tmp,VM_L2_CNTL3 ,L2_CACHE_BIGK_FRAGMENT_SIZE ,6 ); } -WREG32_SOC15_RLC (GC ,0 , mmVM_L2_CNTL3, tmp); +WREG32_SOC15 (GC ,0 , mmVM_L2_CNTL3, tmp); tmp = mmVM_L2_CNTL4_DEFAULT; tmp =REG_SET_FIELD (tmp,VM_L2_CNTL4 ,VMC_TAP_PDE_REQUEST_PHYSICAL ,0 ); tmp =REG_SET_FIELD (tmp,VM_L2_CNTL4 ,VMC_TAP_PTE_REQUEST_PHYSICAL ,0 ); -WREG32_SOC15_RLC (GC ,0 , mmVM_L2_CNTL4, tmp); +WREG32_SOC15 (GC ,0 , mmVM_L2_CNTL4, tmp); }static void gfxhub_v1_0_enable_system_domain (struct amdgpu_device *adev)
@@ -275,9 +275,9 @@ int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev)* VF copy registers so vbios post doesn't program them, for * SRIOV driver need to program them */ -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_FB_LOCATION_BASE, +WREG32_SOC15 (GC ,0 , mmMC_VM_FB_LOCATION_BASE, adev->gmc.vram_start >>24 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_FB_LOCATION_TOP, +WREG32_SOC15 (GC ,0 , mmMC_VM_FB_LOCATION_TOP, adev->gmc.vram_end >>24 ); }
@@ -313,7 +313,7 @@ void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)MC_VM_MX_L1_TLB_CNTL ,ENABLE_ADVANCED_DRIVER_MODEL , 0); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_MX_L1_TLB_CNTL, tmp); +WREG32_SOC15 (GC ,0 , mmMC_VM_MX_L1_TLB_CNTL, tmp);/* Setup L2 cache */ WREG32_FIELD15 (GC ,0 ,VM_L2_CNTL ,ENABLE_L2_CACHE ,0 );
drivers/gpu/drm/amd/amdgpu/soc15.c +2 -9
@@ -297,7 +297,7 @@ void soc15_grbm_select(struct amdgpu_device *adev, grbm_gfx_cntl =REG_SET_FIELD (grbm_gfx_cntl,GRBM_GFX_CNTL ,VMID , vmid); grbm_gfx_cntl =REG_SET_FIELD (grbm_gfx_cntl,GRBM_GFX_CNTL ,QUEUEID , queue); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_CNTL, grbm_gfx_cntl); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_CNTL, grbm_gfx_cntl); }static void soc15_vga_set_state (struct amdgpu_device *adev,bool state)
@@ -459,13 +459,7 @@ void soc15_program_register_sequence(struct amdgpu_device *adev, tmp |= (entry->or_mask & entry->and_mask); } -if (reg ==SOC15_REG_OFFSET (GC ,0 , mmPA_SC_BINNER_EVENT_CNTL_3) || - reg ==SOC15_REG_OFFSET (GC ,0 , mmPA_SC_ENHANCE) || - reg ==SOC15_REG_OFFSET (GC ,0 , mmPA_SC_ENHANCE_1) || - reg ==SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_CONFIG)) -WREG32_RLC (reg, tmp); -else -WREG32 (reg, tmp); +WREG32 (reg, tmp); }
@@ -991,7 +985,6 @@ static uint64_t soc15_get_pcie_replay_count(struct amdgpu_device *adev)/* Add the total number of NAKs, i.e the number of replays */ return (nak_r + nak_g); } -static const struct amdgpu_asic_funcs soc15_asic_funcs = { .read_disabled_bios = &soc15_read_disabled_bios,
drivers/gpu/drm/amd/amdgpu/soc15_common.h +0 -57
61 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
RE: [refactor RLCG wreg path 2/2] drm/amdgpu: refactor RLCG access path part 2
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of Monk Liu
Sent: 2020年3月11日 13:58
To: [email protected]
Cc: Liu, Monk <[email protected]>
Subject: [refactor RLCG wreg path 2/2] drm/amdgpu: refactor RLCG access path part 2
switch to new RLCG access path, and drop the legacy WREG32_RLC macros
--- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c | 30 +++---- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 ++ drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 8 +- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 104 +++++++++++----------- drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c | 2 +- drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 28 +++--- drivers/gpu/drm/amd/amdgpu/soc15.c | 11 +-- drivers/gpu/drm/amd/amdgpu/soc15_common.h | 57 ------------ 8 files changed, 93 insertions(+), 152 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +4 -4
@@ -105,8 +105,8 @@ void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,lock_srbm (kgd,0 ,0 ,0 , vmid); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_CONFIG), sh_mem_config); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_BASES), sh_mem_bases); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_CONFIG), sh_mem_config); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmSH_MEM_BASES), sh_mem_bases);/* APE1 no longer exists on GFX9 */ unlock_srbm (kgd);
@@ -242,13 +242,13 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,for (reg = hqd_base; reg <=SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_WPTR_HI); reg++) -WREG32_RLC (reg, mqd_hqd[reg - hqd_base]); +WREG32 (reg, mqd_hqd[reg - hqd_base]);/* Activate doorbell logic before triggering WPTR poll. */ data =REG_SET_FIELD (m->cp_hqd_pq_doorbell_control,CP_HQD_PQ_DOORBELL_CONTROL ,DOORBELL_EN ,1 ); -WREG32_RLC (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_DOORBELL_CONTROL), data); +WREG32 (SOC15_REG_OFFSET (GC ,0 , mmCP_HQD_PQ_DOORBELL_CONTROL), data);if (wptr) {/* Don't read wptr with get_user because the user @@ -277,25 +277,25 @@ int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1); guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
lower_32_bits(guessed_wptr));
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
upper_32_bits(guessed_wptr));
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
lower_32_bits((uintptr_t)wptr));
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
upper_32_bits((uintptr_t)wptr));
WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1),
- (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+ get_queue_mask(adev, pipe_id, queue_id));
}
/* Start the EOP fetcher */
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
REG_SET_FIELD(m->cp_hqd_eop_rptr,
CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);
release_queue(kgd);
@@ -547,7 +547,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
acquire_queue(kgd, pipe_id, queue_id);
if (m->cp_hqd_vmid == 0)
- WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+ WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
switch (reset_type) {
case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
@@ -561,7 +561,7 @@ int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
break;
}
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);
end_jiffies = (utimeout * HZ / 1000) + jiffies;
while (true) {
@@ -656,7 +656,7 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
mutex_lock(&adev->grbm_idx_mutex);
- WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
+ WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd);
data = REG_SET_FIELD(data, GRBM_GFX_INDEX, @@ -666,7 +666,7 @@ int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd, data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
SE_BROADCAST_WRITES, 1);
- WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data);
+ WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, data);
mutex_unlock(&adev->grbm_idx_mutex);
return 0;
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +5 -0
@@ -328,6 +328,11 @@ void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,if ((acc_flags &AMDGPU_REGS_KIQ ) || (!(acc_flags &AMDGPU_REGS_NO_KIQ ) &&amdgpu_sriov_runtime (adev)))return amdgpu_kiq_wreg (adev, reg, v); +if (amdgpu_sriov_vf (adev) && adev->gfx.rlc.funcs && + adev->gfx.rlc.funcs->is_rlcg_access_range) +if (adev->gfx.rlc.funcs->is_rlcg_access_range (adev, reg)) +return adev->gfx.rlc.funcs->rlcg_wreg (adev, reg, v); +if ((reg *4 ) < adev->rmmio_size && !(acc_flags &AMDGPU_REGS_IDX ))writel (v, ((void __iomem *)adev->rmmio) + (reg *4 ));else {
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +4 -3
@@ -1835,11 +1835,11 @@ static int gfx_v10_0_init_csb(struct amdgpu_device *adev) adev->gfx.rlc.funcs->get_csb_buffer (adev, adev->gfx.rlc.cs_ptr);/* csib */ -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_ADDR_HI, +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_ADDR_HI, adev->gfx.rlc.clear_state_gpu_addr >>32 ); -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_ADDR_LO, +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_ADDR_LO, adev->gfx.rlc.clear_state_gpu_addr &0xfffffffc ); -WREG32_SOC15_RLC (GC ,0 , mmRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size); +WREG32_SOC15 (GC ,0 , mmRLC_CSIB_LENGTH, +adev->gfx.rlc.clear_state_size);return 0 ;
}
@@ -2447,7 +2447,7 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
for (i = 0; i < adev->gfx.num_gfx_rings; i++)
adev->gfx.gfx_ring[i].sched.ready = false;
}
- WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp);
+ WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp);
for (i = 0; i < adev->usec_timeout; i++) {
if (RREG32_SOC15(GC, 0, mmCP_STAT) == 0) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 1fc430d..57d370a 100644
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +1 -1
@@ -2500,7 +2500,7 @@ static void gfx_v9_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 shelse data =REG_SET_FIELD (data,GRBM_GFX_INDEX ,SH_INDEX , sh_num); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_INDEX, data); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_INDEX, data); } static u32 gfx_v9_0_get_rb_active_bitmap(struct amdgpu_device *adev) @@ -2568,8 +2568,8 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
for (i = FIRST_COMPUTE_VMID; i < LAST_COMPUTE_VMID; i++) {
soc15_grbm_select(adev, 0, 0, 0, i);
/* CP and shaders */
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
+ WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
+ WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
}
soc15_grbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex);
@@ -2623,7 +2623,7 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
u32 tmp;
int i;
- WREG32_FIELD15_RLC(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
+ WREG32_FIELD15(GC, 0, GRBM_CNTL, READ_TIMEOUT, 0xff);
gfx_v9_0_tiling_mode_table_init(adev);
@@ -2642,19 +2642,19 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
SH_MEM_ALIGNMENT_MODE_UNALIGNED);
tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
!!amdgpu_noretry);
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, tmp);
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, 0);
+ WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, tmp);
+ WREG32_SOC15(GC, 0, mmSH_MEM_BASES, 0);
} else {
tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
SH_MEM_ALIGNMENT_MODE_UNALIGNED);
tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
!!amdgpu_noretry);
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, tmp);
+ WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, tmp);
tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE,
(adev->gmc.private_aperture_start >> 48));
tmp = REG_SET_FIELD(tmp, SH_MEM_BASES, SHARED_BASE,
(adev->gmc.shared_aperture_start >> 48));
- WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, tmp);
+ WREG32_SOC15(GC, 0, mmSH_MEM_BASES, tmp);
}
}
soc15_grbm_select(adev, 0, 0, 0, 0);
@@ -2721,11 +2721,11 @@ static void gfx_v9_0_init_csb(struct amdgpu_device *adev) {
adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr); /* csib */
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI),
adev->gfx.rlc.clear_state_gpu_addr >> 32);
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO),
adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc);
- WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH),
+ WREG32(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH),
adev->gfx.rlc.clear_state_size);
}
@@ -3197,7 +3197,7 @@ static void gfx_v9_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
for (i = 0; i < adev->gfx.num_gfx_rings; i++)
adev->gfx.gfx_ring[i].sched.ready = false;
}
- WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp);
+ WREG32_SOC15(GC, 0, mmCP_ME_CNTL, tmp);
udelay(50);
}
@@ -3395,9 +3395,9 @@ static void gfx_v9_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
int i;
if (enable) {
- WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL, 0);
+ WREG32_SOC15(GC, 0, mmCP_MEC_CNTL, 0);
} else {
- WREG32_SOC15_RLC(GC, 0, mmCP_MEC_CNTL,
+ WREG32_SOC15(GC, 0, mmCP_MEC_CNTL,
(CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK));
for (i = 0; i < adev->gfx.num_compute_rings; i++)
adev->gfx.compute_ring[i].sched.ready = false; @@ -3458,9 +3458,9 @@ static void gfx_v9_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
tmp |= 0x80;
- WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
+ WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp);
}
static void gfx_v9_0_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd) @@ -3637,67 +3637,67 @@ static int gfx_v9_0_kiq_init_register(struct amdgpu_ring *ring)
/* disable wptr polling */
WREG32_FIELD15(GC, 0, CP_PQ_WPTR_POLL_CNTL, EN, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
+ WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR,
mqd->cp_hqd_eop_base_addr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI,
mqd->cp_hqd_eop_base_addr_hi);
/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_EOP_CONTROL,
+ WREG32_SOC15(GC, 0, mmCP_HQD_EOP_CONTROL,
mqd->cp_hqd_eop_control);
/* enable doorbell? */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
mqd->cp_hqd_pq_doorbell_control);
/* disable the queue if it's active */
if (RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1) {
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
+ WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
for (j = 0; j < adev->usec_timeout; j++) {
if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1))
break;
udelay(1);
}
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
+ WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
mqd->cp_hqd_dequeue_request);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR,
mqd->cp_hqd_pq_rptr);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO,
mqd->cp_hqd_pq_wptr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI,
mqd->cp_hqd_pq_wptr_hi);
}
/* set the pointer to the MQD */
- WREG32_SOC15_RLC(GC, 0, mmCP_MQD_BASE_ADDR,
+ WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR,
mqd->cp_mqd_base_addr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_MQD_BASE_ADDR_HI,
+ WREG32_SOC15(GC, 0, mmCP_MQD_BASE_ADDR_HI,
mqd->cp_mqd_base_addr_hi);
/* set MQD vmid to 0 */
- WREG32_SOC15_RLC(GC, 0, mmCP_MQD_CONTROL,
+ WREG32_SOC15(GC, 0, mmCP_MQD_CONTROL,
mqd->cp_mqd_control);
/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_BASE,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE,
mqd->cp_hqd_pq_base_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_BASE_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI,
mqd->cp_hqd_pq_base_hi);
/* set up the HQD, this is similar to CP_RB0_CNTL */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_CONTROL,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_CONTROL,
mqd->cp_hqd_pq_control);
/* set the wb address whether it's enabled or not */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR,
mqd->cp_hqd_pq_rptr_report_addr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR_REPORT_ADDR_HI,
mqd->cp_hqd_pq_rptr_report_addr_hi);
/* only used if CP_PQ_WPTR_POLL_CNTL.CP_PQ_WPTR_POLL_CNTL__EN_MASK=1 */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
mqd->cp_hqd_pq_wptr_poll_addr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
mqd->cp_hqd_pq_wptr_poll_addr_hi);
/* enable the doorbell if requested */ @@ -3708,23 +3708,23 @@ static int gfx_v9_0_kiq_init_register(struct amdgpu_ring *ring)
(adev->doorbell_index.userqueue_end * 2) << 2);
}
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL,
mqd->cp_hqd_pq_doorbell_control);
/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO,
mqd->cp_hqd_pq_wptr_lo);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI,
mqd->cp_hqd_pq_wptr_hi);
/* set the vmid for the queue */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_VMID, mqd->cp_hqd_vmid);
+ WREG32_SOC15(GC, 0, mmCP_HQD_VMID, mqd->cp_hqd_vmid);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PERSISTENT_STATE,
+ WREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE,
mqd->cp_hqd_persistent_state);
/* activate the queue */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_ACTIVE,
+ WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE,
mqd->cp_hqd_active);
if (ring->use_doorbell)
@@ -3741,7 +3741,7 @@ static int gfx_v9_0_kiq_fini_register(struct amdgpu_ring *ring)
/* disable the queue if it's active */
if (RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1) {
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
+ WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, 1);
for (j = 0; j < adev->usec_timeout; j++) {
if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1)) @@ -3753,21 +3753,21 @@ static int gfx_v9_0_kiq_fini_register(struct amdgpu_ring *ring) DRM_DEBUG("KIQ dequeue request failed.\n");
/* Manual disable if dequeue request times out */
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_ACTIVE, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, 0);
}
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
+ WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST,
0);
}
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_IQ_TIMER, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_IB_CONTROL, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PERSISTENT_STATE, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_RPTR, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_HI, 0);
- WREG32_SOC15_RLC(GC, 0, mmCP_HQD_PQ_WPTR_LO, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_IQ_TIMER, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_IB_CONTROL, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PERSISTENT_STATE, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0x40000000);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_RPTR, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI, 0);
+ WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO, 0);
return 0;
}
drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c +1 -1
@@ -114,7 +114,7 @@ static void gfx_v9_4_select_se_sh(struct amdgpu_device *adev, u32 se_num,else data =REG_SET_FIELD (data,GRBM_GFX_INDEX ,SH_INDEX , sh_num); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_INDEX, data); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_INDEX, data); } static const struct soc15_ras_field_entry gfx_v9_4_ras_fields[] = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index 1a2f18b..9360d61 100644
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c +4 -4
@@ -71,13 +71,13 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)uint64_t value;/* Program the AGP BAR */ -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_BASE,0 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_BOT, adev->gmc.agp_start >>24 ); -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_AGP_TOP, adev->gmc.agp_end >>24 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_BASE,0 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_BOT, adev->gmc.agp_start >>24 ); +WREG32_SOC15 (GC ,0 , mmMC_VM_AGP_TOP, adev->gmc.agp_end >>24 );if (!amdgpu_sriov_vf (adev) || adev->asic_type <=CHIP_VEGA10 ) {/* Program the system aperture low logical page number. */ -WREG32_SOC15_RLC (GC ,0 , mmMC_VM_SYSTEM_APERTURE_LOW_ADDR, +WREG32_SOC15 (GC ,0 , mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,min (adev->gmc.fb_start, adev->gmc.agp_start) >>18 ); if (adev->asic_type == CHIP_RAVEN && adev->rev_id >= 0x8) @@ -88,12 +88,12 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
* aperture high address (add 1) to get rid of the VM
* fault and hardware hang.
*/
- WREG32_SOC15_RLC(GC, 0,
+ WREG32_SOC15(GC, 0,
mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, max((adev->gmc.fb_end >> 18) + 0x1,
adev->gmc.agp_end >> 18));
else
- WREG32_SOC15_RLC(
+ WREG32_SOC15(
GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR, max(adev->gmc.fb_end, adev->gmc.agp_end) >> 18);
@@ -134,7 +134,7 @@ static void gfxhub_v1_0_init_tlb_regs(struct amdgpu_device *adev)
MTYPE, MTYPE_UC);/* XXX for emulation. */
tmp = REG_SET_FIELD(tmp, MC_VM_MX_L1_TLB_CNTL, ATC_EN, 1);
- WREG32_SOC15_RLC(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
+ WREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
}
static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev) @@ -151,12 +151,12 @@ static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev)
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, PDE_FAULT_CLASSIFICATION, 0);
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, CONTEXT1_IDENTITY_ACCESS_MODE, 1);
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL, IDENTITY_MODE_FRAGMENT_SIZE, 0);
- WREG32_SOC15_RLC(GC, 0, mmVM_L2_CNTL, tmp);
+ WREG32_SOC15(GC, 0, mmVM_L2_CNTL, tmp);
tmp = RREG32_SOC15(GC, 0, mmVM_L2_CNTL2);
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS, 1); tmp = REG_SET_FIELD(tmp, VM_L2_CNTL2, INVALIDATE_L2_CACHE, 1);
- WREG32_SOC15_RLC(GC, 0, mmVM_L2_CNTL2, tmp);
+ WREG32_SOC15(GC, 0, mmVM_L2_CNTL2, tmp);
tmp = mmVM_L2_CNTL3_DEFAULT;
if (adev->gmc.translate_further) {
@@ -168,12 +168,12 @@ static void gfxhub_v1_0_init_cache_regs(struct amdgpu_device *adev)
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL3,
L2_CACHE_BIGK_FRAGMENT_SIZE, 6);
}
- WREG32_SOC15_RLC(GC, 0, mmVM_L2_CNTL3, tmp);
+ WREG32_SOC15(GC, 0, mmVM_L2_CNTL3, tmp);
tmp = mmVM_L2_CNTL4_DEFAULT;
tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PDE_REQUEST_PHYSICAL, 0); tmp = REG_SET_FIELD(tmp, VM_L2_CNTL4, VMC_TAP_PTE_REQUEST_PHYSICAL, 0);
- WREG32_SOC15_RLC(GC, 0, mmVM_L2_CNTL4, tmp);
+ WREG32_SOC15(GC, 0, mmVM_L2_CNTL4, tmp);
}
static void gfxhub_v1_0_enable_system_domain(struct amdgpu_device *adev) @@ -275,9 +275,9 @@ int gfxhub_v1_0_gart_enable(struct amdgpu_device *adev)
* VF copy registers so vbios post doesn't program them, for
* SRIOV driver need to program them
*/
- WREG32_SOC15_RLC(GC, 0, mmMC_VM_FB_LOCATION_BASE,
+ WREG32_SOC15(GC, 0, mmMC_VM_FB_LOCATION_BASE,
adev->gmc.vram_start >> 24);
- WREG32_SOC15_RLC(GC, 0, mmMC_VM_FB_LOCATION_TOP,
+ WREG32_SOC15(GC, 0, mmMC_VM_FB_LOCATION_TOP,
adev->gmc.vram_end >> 24);
}
@@ -313,7 +313,7 @@ void gfxhub_v1_0_gart_disable(struct amdgpu_device *adev)
MC_VM_MX_L1_TLB_CNTL, ENABLE_ADVANCED_DRIVER_MODEL,
0);
- WREG32_SOC15_RLC(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
+ WREG32_SOC15(GC, 0, mmMC_VM_MX_L1_TLB_CNTL, tmp);
/* Setup L2 cache */
WREG32_FIELD15(GC, 0, VM_L2_CNTL, ENABLE_L2_CACHE, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index a40499d..9800f0c 100644
drivers/gpu/drm/amd/amdgpu/soc15.c +1 -1
@@ -297,7 +297,7 @@ void soc15_grbm_select(struct amdgpu_device *adev, grbm_gfx_cntl =REG_SET_FIELD (grbm_gfx_cntl,GRBM_GFX_CNTL ,VMID , vmid); grbm_gfx_cntl =REG_SET_FIELD (grbm_gfx_cntl,GRBM_GFX_CNTL ,QUEUEID , queue); -WREG32_SOC15_RLC_SHADOW (GC ,0 , mmGRBM_GFX_CNTL, grbm_gfx_cntl); +WREG32_SOC15 (GC ,0 , mmGRBM_GFX_CNTL, grbm_gfx_cntl); } static void soc15_vga_set_state(struct amdgpu_device *adev, bool state) @@ -459,13 +459,7 @@ void soc15_program_register_sequence(struct amdgpu_device *adev,
tmp |= (entry->or_mask & entry->and_mask);
}
- if (reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_BINNER_EVENT_CNTL_3) ||
- reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_ENHANCE) ||
- reg == SOC15_REG_OFFSET(GC, 0, mmPA_SC_ENHANCE_1) ||
- reg == SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG))
- WREG32_RLC(reg, tmp);
- else
- WREG32(reg, tmp);
+ WREG32(reg, tmp);
}
@@ -991,7 +985,6 @@ static uint64_t soc15_get_pcie_replay_count(struct amdgpu_device *adev)
/* Add the total number of NAKs, i.e the number of replays */
return (nak_r + nak_g);
}
-
static const struct amdgpu_asic_funcs soc15_asic_funcs = {
.read_disabled_bios = &soc15_read_disabled_bios, diff --git a/drivers/gpu/drm/amd/amdgpu/soc15_common.h b/drivers/gpu/drm/amd/amdgpu/soc15_common.h
index 19e870c..7f46a0d 100644
drivers/gpu/drm/amd/amdgpu/soc15_common.h +0 -57
61 lines not on this page · raw message
2.7.4
_______________________________________________
amd-gfx mailing list
[email protected]
https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cyintian.tao%40amd.com%7C5078bddfea8d469a8a7208d7c5811eea%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637195030689409966&sdata=%2F9KR2hhLPMZsbqpLlMWJMk4ALglIwkt0YKS%2FMhivp%2FU%3D&reserved=0
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of Monk Liu
Sent: 2020年3月11日 13:58
To: [email protected]
Cc: Liu, Monk <[email protected]>
Subject: [refactor RLCG wreg path 1/2] drm/amdgpu: refactor RLCG access path part 1
what changed:
1)provide new implementation interface for the rlcg access path 2)put SQ_CMD/SQ_IND_INDEX/SQ_IND_DATA to GFX9 RLCG path to align with SRIOV RLCG logic
background:
we what to clear the code path for WREG32_RLC, to make it only covered and handled by amdgpu_mm_wreg() routine, this way we can let RLCG to serve the register access even through UMR (via debugfs interface) the current implementation cannot achieve that goal because it can only hardcode everywhere, but UMR only pass "offset" as varable to driver
--- drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h | 2 + drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 80 ++++++++++++++- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 177 +++++++++++++++++++++++++++++++- drivers/gpu/drm/amd/amdgpu/soc15.h | 7 ++ 4 files changed, 264 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h +3 -0
@@ -127,6 +127,8 @@ struct amdgpu_rlc_funcs {void (*reset)(struct amdgpu_device *adev);void (*start)(struct amdgpu_device *adev);void (*update_spm_vmid)(struct amdgpu_device *adev,unsigned vmid); +void (*rlcg_wreg)(struct amdgpu_device *adev,u32 offset,u32 v); +bool (*is_rlcg_access_range)(struct amdgpu_device *adev,uint32_t +reg); };
struct amdgpu_rlc {
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +76 -0
86 lines not on this page · raw message
.reset = gfx_v10_0_rlc_reset, .start = gfx_v10_0_rlc_start,
- .update_spm_vmid = gfx_v10_0_update_spm_vmid
+ .update_spm_vmid = gfx_v10_0_update_spm_vmid,
+ .rlcg_wreg = gfx_v10_rlcg_wreg,
+ .is_rlcg_access_range = gfx_v10_0_is_rlcg_access_range,
};
static int gfx_v10_0_set_powergating_state(void *handle, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 87747b3..1fc430d 100644
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +90 -0
96 lines not on this page · raw message
+{
+ static void *scratch_reg0;
+ static void *scratch_reg1;
+ static void *scratch_reg2;
+ static void *scratch_reg3;
+ static void *spare_int;
+ static uint32_t grbm_cntl;
+ static uint32_t grbm_idx;
+ bool shadow;
+
+ scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;
+ scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;
+ scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;
+ scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;
+ spare_int = adev->rmmio +
+(adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] +
+mmRLC_SPARE_INT)*4;
+
+ grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;
+ grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] +
+mmGRBM_GFX_INDEX;
+
+ if (amdgpu_sriov_runtime(adev)) {
+ pr_err("shoudn't call rlcg write register during runtime\n");
+ return;
+ }
+
+ if (offset == grbm_cntl || offset == grbm_idx)
+ shadow = true;
+
+ if (shadow) {
+ if (offset == grbm_cntl)
+ writel(v, scratch_reg2);
+ else if (offset == grbm_idx)
+ writel(v, scratch_reg3);
+
+ writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));
+ } else {
+ uint32_t i = 0;
+ uint32_t retries = 50000;
+
+ writel(v, scratch_reg0);
+ writel(offset | 0x80000000, scratch_reg1);
+ writel(1, spare_int);
+ for (i = 0; i < retries; i++) {
+ u32 tmp;
+
+ tmp = readl(scratch_reg1);
+ if (!(tmp & 0x80000000))
+ break;
+
+ udelay(10);
+ }
+ if (i >= retries)
+ pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);
+ }
+
+}
+
#define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042 #define VEGA12_GB_ADDR_CONFIG_GOLDEN 0x24104041 #define RAVEN_GB_ADDR_CONFIG_GOLDEN 0x24000042 @@ -4783,6 +4930,32 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data); }
+static bool gfx_v9_0_check_rlcg_range(struct amdgpu_device *adev,
+ uint32_t offset,
+ struct soc15_reg_rlcg *entries, int arr_size) {
+ int i;
+ uint32_t reg;
+
+ for (i = 0; i < arr_size; i++) {
+ const struct soc15_reg_rlcg *entry;
+
+ entry = &entries[i];
+ reg = adev->reg_offset[entry->hwip][entry->instance][entry->segment] + entry->reg;
+ if (offset == reg)
+ return true;
+ }
+
+ return false;
+}
+
+static bool gfx_v9_0_is_rlcg_access_range(struct amdgpu_device *adev,
+u32 offset) {
+ return gfx_v9_0_check_rlcg_range(adev, offset,
+ (void *)rlcg_access_gc_9_0,
+ ARRAY_SIZE(rlcg_access_gc_9_0));
+}
+
static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {
.is_rlc_enabled = gfx_v9_0_is_rlc_enabled,
.set_safe_mode = gfx_v9_0_set_safe_mode, @@ -4795,7 +4968,9 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = { .stop = gfx_v9_0_rlc_stop,
.reset = gfx_v9_0_rlc_reset,
.start = gfx_v9_0_rlc_start,
- .update_spm_vmid = gfx_v9_0_update_spm_vmid
+ .update_spm_vmid = gfx_v9_0_update_spm_vmid,
+ .rlcg_wreg = gfx_v9_0_rlcg_wreg,
+ .is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range,
};
static int gfx_v9_0_set_powergating_state(void *handle, diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h
index d0fb7a6..b03f950 100644
drivers/gpu/drm/amd/amdgpu/soc15.h +7 -0
@@ -42,6 +42,13 @@ struct soc15_reg_golden {u32 or_mask; }; +struct soc15_reg_rlcg { +u32 hwip; +u32 instance; +u32 segment; +u32 reg; +}; +struct soc15_reg_entry {uint32_t hwip;uint32_t inst;
--
2.7.4
_______________________________________________
amd-gfx mailing list
[email protected]
https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cyintian.tao%40amd.com%7Cfbc7d7e64f324f9e357d08d7c5811dec%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637195030680452029&sdata=VIGkDYltuTDMU%2FvfIl3WBa%2B54F945ZMFTeWIEqNKDaY%3D&reserved=0
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx