Skip to contentSkip to conversation

drm/amdgpu: refactor RLCG access path part 1

PATCH New Monk Liu 1 participant 2 messages Mar 10, 2020

Subsystems (3)
lore
Share & tools
lore

Monk Liu

author
#0 lore raw

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
index 52509c2..60bb3e8 100644
@@ -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
index d0fb7a6..b03f950 100644
@@ -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
author
#1 lore raw

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
index df841c2..a21f005 100644
@@ -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
index 572eb6e..9029407 100644
@@ -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
index 3222cd3..78291b5 100644
@@ -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
index 17f1e7b..6b9127b 100644
@@ -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
index 1a2f18b..9360d61 100644
@@ -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
index a40499d..9800f0c 100644
@@ -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