1 files changed, 85 insertions, 185 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index fa87a85e1017..d8a4fddab9c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -51,8 +51,6 @@
 #include "athub_v2_0.h"
 #include "athub_v2_1.h"
 
-#include "amdgpu_reset.h"
-
 static int gmc_v10_0_ecc_interrupt_state(struct amdgpu_device *adev,
 					 struct amdgpu_irq_src *src,
 					 unsigned int type,
@@ -145,11 +143,15 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev,
 		 * the new fast GRBM interface.
 		 */
 		if ((entry->vmid_src == AMDGPU_GFXHUB(0)) &&
-		    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0)))
+		    (amdgpu_ip_version(adev, GC_HWIP, 0) <
+		     IP_VERSION(10, 3, 0)))
 			RREG32(hub->vm_l2_pro_fault_status);
 
 		status = RREG32(hub->vm_l2_pro_fault_status);
 		WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
+
+		amdgpu_vm_update_fault_cache(adev, entry->pasid, addr, status,
+					     entry->vmid_src ? AMDGPU_MMHUB0(0) : AMDGPU_GFXHUB(0));
 	}
 
 	if (!printk_ratelimit())
@@ -230,20 +232,47 @@ static bool gmc_v10_0_get_atc_vmid_pasid_mapping_info(
  * by the amdgpu vm/hsa code.
  */
 
-static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
-				   unsigned int vmhub, uint32_t flush_type)
+/**
+ * gmc_v10_0_flush_gpu_tlb - gart tlb flush callback
+ *
+ * @adev: amdgpu_device pointer
+ * @vmid: vm instance to flush
+ * @vmhub: vmhub type
+ * @flush_type: the flush type
+ *
+ * Flush the TLB for the requested page table.
+ */
+static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t vmhub, uint32_t flush_type)
 {
 	bool use_semaphore = gmc_v10_0_use_invalidate_semaphore(adev, vmhub);
 	struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
 	u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
-	u32 tmp;
 	/* Use register 17 for GART */
 	const unsigned int eng = 17;
-	unsigned int i;
 	unsigned char hub_ip = 0;
+	u32 sem, req, ack;
+	unsigned int i;
+	u32 tmp;
+
+	sem = hub->vm_inv_eng0_sem + hub->eng_distance * eng;
+	req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
+	ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
+
+	/* flush hdp cache */
+	adev->hdp.funcs->flush_hdp(adev, NULL);
+
+	/* For SRIOV run time, driver shouldn't access the register through MMIO
+	 * Directly use kiq to do the vm invalidation instead
+	 */
+	if (adev->gfx.kiq[0].ring.sched.ready && !adev->enable_mes &&
+	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
+		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
+				1 << vmid);
+		return;
+	}
 
-	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ?
-		   GC_HWIP : MMHUB_HWIP;
+	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP;
 
 	spin_lock(&adev->gmc.invalidate_lock);
 	/*
@@ -257,9 +286,7 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 	if (use_semaphore) {
 		for (i = 0; i < adev->usec_timeout; i++) {
 			/* a read return value of 1 means semaphore acuqire */
-			tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
-					 hub->eng_distance * eng, hub_ip);
-
+			tmp = RREG32_RLC_NO_KIQ(sem, hub_ip);
 			if (tmp & 0x1)
 				break;
 			udelay(1);
@@ -269,24 +296,19 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 			DRM_ERROR("Timeout waiting for sem acquire in VM flush!\n");
 	}
 
-	WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req +
-			  hub->eng_distance * eng,
-			  inv_req, hub_ip);
+	WREG32_RLC_NO_KIQ(req, inv_req, hub_ip);
 
 	/*
 	 * Issue a dummy read to wait for the ACK register to be cleared
 	 * to avoid a false ACK due to the new fast GRBM interface.
 	 */
 	if ((vmhub == AMDGPU_GFXHUB(0)) &&
-	    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 3, 0)))
-		RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_req +
-				  hub->eng_distance * eng, hub_ip);
+	    (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 3, 0)))
+		RREG32_RLC_NO_KIQ(req, hub_ip);
 
 	/* Wait for ACK with a delay.*/
 	for (i = 0; i < adev->usec_timeout; i++) {
-		tmp = RREG32_RLC_NO_KIQ(hub->vm_inv_eng0_ack +
-				  hub->eng_distance * eng, hub_ip);
-
+		tmp = RREG32_RLC_NO_KIQ(ack, hub_ip);
 		tmp &= 1 << vmid;
 		if (tmp)
 			break;
@@ -296,109 +318,13 @@ static void gmc_v10_0_flush_vm_hub(struct amdgpu_device *adev, uint32_t vmid,
 
 	/* TODO: It needs to continue working on debugging with semaphore for GFXHUB as well. */
 	if (use_semaphore)
-		/*
-		 * add semaphore release after invalidation,
-		 * write with 0 means semaphore release
-		 */
-		WREG32_RLC_NO_KIQ(hub->vm_inv_eng0_sem +
-				  hub->eng_distance * eng, 0, hub_ip);
+		WREG32_RLC_NO_KIQ(sem, 0, hub_ip);
 
 	spin_unlock(&adev->gmc.invalidate_lock);
 
-	if (i < adev->usec_timeout)
-		return;
-
-	DRM_ERROR("Timeout waiting for VM flush hub: %d!\n", vmhub);
-}
-
-/**
- * gmc_v10_0_flush_gpu_tlb - gart tlb flush callback
- *
- * @adev: amdgpu_device pointer
- * @vmid: vm instance to flush
- * @vmhub: vmhub type
- * @flush_type: the flush type
- *
- * Flush the TLB for the requested page table.
- */
-static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
-					uint32_t vmhub, uint32_t flush_type)
-{
-	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	struct dma_fence *fence;
-	struct amdgpu_job *job;
-
-	int r;
-
-	/* flush hdp cache */
-	adev->hdp.funcs->flush_hdp(adev, NULL);
-
-	/* For SRIOV run time, driver shouldn't access the register through MMIO
-	 * Directly use kiq to do the vm invalidation instead
-	 */
-	if (adev->gfx.kiq[0].ring.sched.ready && !adev->enable_mes &&
-	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev)) &&
-	    down_read_trylock(&adev->reset_domain->sem)) {
-		struct amdgpu_vmhub *hub = &adev->vmhub[vmhub];
-		const unsigned int eng = 17;
-		u32 inv_req = hub->vmhub_funcs->get_invalidate_req(vmid, flush_type);
-		u32 req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
-		u32 ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
-
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-				1 << vmid);
-
-		up_read(&adev->reset_domain->sem);
-		return;
-	}
-
-	mutex_lock(&adev->mman.gtt_window_lock);
-
-	if (vmhub == AMDGPU_MMHUB0(0)) {
-		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_MMHUB0(0), 0);
-		mutex_unlock(&adev->mman.gtt_window_lock);
-		return;
-	}
-
-	BUG_ON(vmhub != AMDGPU_GFXHUB(0));
-
-	if (!adev->mman.buffer_funcs_enabled ||
-	    !adev->ib_pool_ready ||
-	    amdgpu_in_reset(adev) ||
-	    ring->sched.ready == false) {
-		gmc_v10_0_flush_vm_hub(adev, vmid, AMDGPU_GFXHUB(0), 0);
-		mutex_unlock(&adev->mman.gtt_window_lock);
-		return;
-	}
-
-	/* The SDMA on Navi has a bug which can theoretically result in memory
-	 * corruption if an invalidation happens at the same time as an VA
-	 * translation. Avoid this by doing the invalidation from the SDMA
-	 * itself.
-	 */
-	r = amdgpu_job_alloc_with_ib(ring->adev, &adev->mman.high_pr,
-				     AMDGPU_FENCE_OWNER_UNDEFINED,
-				     16 * 4, AMDGPU_IB_POOL_IMMEDIATE,
-				     &job);
-	if (r)
-		goto error_alloc;
-
-	job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gart.bo);
-	job->vm_needs_flush = true;
-	job->ibs->ptr[job->ibs->length_dw++] = ring->funcs->nop;
-	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
-	fence = amdgpu_job_submit(job);
-
-	mutex_unlock(&adev->mman.gtt_window_lock);
-
-	dma_fence_wait(fence, false);
-	dma_fence_put(fence);
-
-	return;
-
-error_alloc:
-	mutex_unlock(&adev->mman.gtt_window_lock);
-	DRM_ERROR("Error flushing GPU TLB using the SDMA (%d)!\n", r);
+	if (i >= adev->usec_timeout)
+		dev_err(adev->dev, "Timeout waiting for VM flush hub: %d!\n",
+			vmhub);
 }
 
 /**
@@ -412,62 +338,31 @@ error_alloc:
  *
  * Flush the TLB for the requested pasid.
  */
-static int gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
-					uint16_t pasid, uint32_t flush_type,
-					bool all_hub, uint32_t inst)
+static void gmc_v10_0_flush_gpu_tlb_pasid(struct amdgpu_device *adev,
+					  uint16_t pasid, uint32_t flush_type,
+					  bool all_hub, uint32_t inst)
 {
+	uint16_t queried;
 	int vmid, i;
-	signed long r;
-	uint32_t seq;
-	uint16_t queried_pasid;
-	bool ret;
-	u32 usec_timeout = amdgpu_sriov_vf(adev) ? SRIOV_USEC_TIMEOUT : adev->usec_timeout;
-	struct amdgpu_ring *ring = &adev->gfx.kiq[0].ring;
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
-
-	if (amdgpu_emu_mode == 0 && ring->sched.ready) {
-		spin_lock(&adev->gfx.kiq[0].ring_lock);
-		/* 2 dwords flush + 8 dwords fence */
-		amdgpu_ring_alloc(ring, kiq->pmf->invalidate_tlbs_size + 8);
-		kiq->pmf->kiq_invalidate_tlbs(ring,
-					pasid, flush_type, all_hub);
-		r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
-		if (r) {
-			amdgpu_ring_undo(ring);
-			spin_unlock(&adev->gfx.kiq[0].ring_lock);
-			return -ETIME;
-		}
-
-		amdgpu_ring_commit(ring);
-		spin_unlock(&adev->gfx.kiq[0].ring_lock);
-		r = amdgpu_fence_wait_polling(ring, seq, usec_timeout);
-		if (r < 1) {
-			dev_err(adev->dev, "wait for kiq fence error: %ld.\n", r);
-			return -ETIME;
-		}
-
-		return 0;
-	}
 
 	for (vmid = 1; vmid < AMDGPU_NUM_VMID; vmid++) {
-
-		ret = gmc_v10_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
-				&queried_pasid);
-		if (ret	&& queried_pasid == pasid) {
-			if (all_hub) {
-				for_each_set_bit(i, adev->vmhubs_mask, AMDGPU_MAX_VMHUBS)
-					gmc_v10_0_flush_gpu_tlb(adev, vmid,
-							i, flush_type);
-			} else {
-				gmc_v10_0_flush_gpu_tlb(adev, vmid,
-						AMDGPU_GFXHUB(0), flush_type);
-			}
-			if (!adev->enable_mes)
-				break;
+		bool valid;
+
+		valid = gmc_v10_0_get_atc_vmid_pasid_mapping_info(adev, vmid,
+								  &queried);
+		if (!valid || queried != pasid)
+			continue;
+
+		if (all_hub) {
+			for_each_set_bit(i, adev->vmhubs_mask,
+					 AMDGPU_MAX_VMHUBS)
+				gmc_v10_0_flush_gpu_tlb(adev, vmid, i,
+							flush_type);
+		} else {
+			gmc_v10_0_flush_gpu_tlb(adev, vmid, AMDGPU_GFXHUB(0),
+						flush_type);
 		}
 	}
-
-	return 0;
 }
 
 static uint64_t gmc_v10_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
@@ -634,6 +529,7 @@ static void gmc_v10_0_get_vm_pte(struct amdgpu_device *adev,
 	}
 
 	if (bo && bo->flags & (AMDGPU_GEM_CREATE_COHERENT |
+			       AMDGPU_GEM_CREATE_EXT_COHERENT |
 			       AMDGPU_GEM_CREATE_UNCACHED))
 		*flags = (*flags & ~AMDGPU_PTE_MTYPE_NV10_MASK) |
 			 AMDGPU_PTE_MTYPE_NV10(MTYPE_UC);
@@ -680,7 +576,7 @@ static void gmc_v10_0_set_gmc_funcs(struct amdgpu_device *adev)
 
 static void gmc_v10_0_set_umc_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[UMC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, UMC_HWIP, 0)) {
 	case IP_VERSION(8, 7, 0):
 		adev->umc.max_ras_err_cnt_per_query = UMC_V8_7_TOTAL_CHANNEL_NUM;
 		adev->umc.channel_inst_num = UMC_V8_7_CHANNEL_INSTANCE_NUM;
@@ -697,7 +593,7 @@ static void gmc_v10_0_set_umc_funcs(struct amdgpu_device *adev)
 
 static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[MMHUB_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, MMHUB_HWIP, 0)) {
 	case IP_VERSION(2, 3, 0):
 	case IP_VERSION(2, 4, 0):
 	case IP_VERSION(2, 4, 1):
@@ -711,7 +607,7 @@ static void gmc_v10_0_set_mmhub_funcs(struct amdgpu_device *adev)
 
 static void gmc_v10_0_set_gfxhub_funcs(struct amdgpu_device *adev)
 {
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 1):
@@ -777,8 +673,9 @@ static void gmc_v10_0_vram_gtt_location(struct amdgpu_device *adev,
 	base += adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
 
 	amdgpu_gmc_vram_location(adev, &adev->gmc, base);
-	amdgpu_gmc_gart_location(adev, mc);
-	amdgpu_gmc_agp_location(adev, mc);
+	amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT);
+	if (!amdgpu_sriov_vf(adev))
+		amdgpu_gmc_agp_location(adev, mc);
 
 	/* base offset of vram pages */
 	adev->vm_manager.vram_base_offset = adev->gfxhub.funcs->get_mc_fb_offset(adev);
@@ -825,7 +722,7 @@ static int gmc_v10_0_mc_init(struct amdgpu_device *adev)
 
 	/* set the gart size */
 	if (amdgpu_gart_size == -1) {
-		switch (adev->ip_versions[GC_HWIP][0]) {
+		switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 		default:
 			adev->gmc.gart_size = 512ULL << 20;
 			break;
@@ -892,7 +789,7 @@ static int gmc_v10_0_sw_init(void *handle)
 		adev->gmc.vram_vendor = vram_vendor;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 3, 0):
 		adev->gmc.mall_size = 128 * 1024 * 1024;
 		break;
@@ -910,7 +807,7 @@ static int gmc_v10_0_sw_init(void *handle)
 		break;
 	}
 
-	switch (adev->ip_versions[GC_HWIP][0]) {
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
 	case IP_VERSION(10, 1, 10):
 	case IP_VERSION(10, 1, 1):
 	case IP_VERSION(10, 1, 2):
@@ -1084,8 +981,10 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
 
 static int gmc_v10_0_hw_init(void *handle)
 {
-	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int r;
+
+	adev->gmc.flush_pasid_uses_kiq = !amdgpu_emu_mode;
 
 	/* The sequence of these two function calls matters.*/
 	gmc_v10_0_init_golden_registers(adev);
@@ -1195,7 +1094,8 @@ static int gmc_v10_0_set_clockgating_state(void *handle,
 	 * is a new problem observed at DF 3.0.3, however with the same suspend sequence not
 	 * seen any issue on the DF 3.0.2 series platform.
 	 */
-	if (adev->in_s0ix && adev->ip_versions[DF_HWIP][0] > IP_VERSION(3, 0, 2)) {
+	if (adev->in_s0ix &&
+	    amdgpu_ip_version(adev, DF_HWIP, 0) > IP_VERSION(3, 0, 2)) {
 		dev_dbg(adev->dev, "keep mmhub clock gating being enabled for s0ix\n");
 		return 0;
 	}
@@ -1204,7 +1104,7 @@ static int gmc_v10_0_set_clockgating_state(void *handle,
 	if (r)
 		return r;
 
-	if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))
+	if (amdgpu_ip_version(adev, ATHUB_HWIP, 0) >= IP_VERSION(2, 1, 0))
 		return athub_v2_1_set_clockgating(adev, state);
 	else
 		return athub_v2_0_set_clockgating(adev, state);
@@ -1214,13 +1114,13 @@ static void gmc_v10_0_get_clockgating_state(void *handle, u64 *flags)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 3) ||
-	    adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 1, 4))
+	if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 3) ||
+	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 1, 4))
 		return;
 
 	adev->mmhub.funcs->get_clockgating(adev, flags);
 
-	if (adev->ip_versions[ATHUB_HWIP][0] >= IP_VERSION(2, 1, 0))
+	if (amdgpu_ip_version(adev, ATHUB_HWIP, 0) >= IP_VERSION(2, 1, 0))
 		athub_v2_1_get_clockgating(adev, flags);
 	else
 		athub_v2_0_get_clockgating(adev, flags);