summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c16
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c38
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_device.c56
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_job.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c10
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c16
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c53
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c26
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h41
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v13_0.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c9
27 files changed, 215 insertions, 127 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 8639a4f9c6e8..2eca58220550 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1293,6 +1293,7 @@ void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
u32 reg, u32 v);
struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
struct dma_fence *gang);
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev);
/* atpx handler */
#if defined(CONFIG_VGA_SWITCHEROO)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 0561812aa0a4..5d9a34601a1a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -706,6 +706,13 @@ err:
void amdgpu_amdkfd_set_compute_idle(struct amdgpu_device *adev, bool idle)
{
+ /* Temporary workaround to fix issues observed in some
+ * compute applications when GFXOFF is enabled on GFX11.
+ */
+ if (IP_VERSION_MAJ(adev->ip_versions[GC_HWIP][0]) == 11) {
+ pr_debug("GFXOFF is %s\n", idle ? "enabled" : "disabled");
+ amdgpu_gfx_off_ctrl(adev, idle);
+ }
amdgpu_dpm_switch_power_profile(adev,
PP_SMC_POWER_PROFILE_COMPUTE,
!idle);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
index c8935d718207..4485bb29bec9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c
@@ -41,5 +41,6 @@ const struct kfd2kgd_calls aldebaran_kfd2kgd = {
.get_atc_vmid_pasid_mapping_info =
kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
+ .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings
};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 84f44f7e4111..1f76e27f1a35 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -171,9 +171,7 @@ int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
(kfd_mem_limit.ttm_mem_used + ttm_mem_needed >
kfd_mem_limit.max_ttm_mem_limit) ||
(adev && adev->kfd.vram_used + vram_needed >
- adev->gmc.real_vram_size -
- atomic64_read(&adev->vram_pin_size) -
- reserved_for_pt)) {
+ adev->gmc.real_vram_size - reserved_for_pt)) {
ret = -ENOMEM;
goto release;
}
@@ -988,6 +986,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
struct amdkfd_process_info *process_info = mem->process_info;
struct amdgpu_bo *bo = mem->bo;
struct ttm_operation_ctx ctx = { true, false };
+ struct hmm_range *range;
int ret = 0;
mutex_lock(&process_info->lock);
@@ -1017,7 +1016,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
return 0;
}
- ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+ ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages, &range);
if (ret) {
pr_err("%s: Failed to get user pages: %d\n", __func__, ret);
goto unregister_out;
@@ -1035,7 +1034,7 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr,
amdgpu_bo_unreserve(bo);
release_out:
- amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+ amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
unregister_out:
if (ret)
amdgpu_mn_unregister(bo);
@@ -2372,6 +2371,8 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
/* Go through userptr_inval_list and update any invalid user_pages */
list_for_each_entry(mem, &process_info->userptr_inval_list,
validate_list.head) {
+ struct hmm_range *range;
+
invalid = atomic_read(&mem->invalid);
if (!invalid)
/* BO hasn't been invalidated since the last
@@ -2382,7 +2383,8 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
bo = mem->bo;
/* Get updated user pages */
- ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+ ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages,
+ &range);
if (ret) {
pr_debug("Failed %d to get user pages\n", ret);
@@ -2401,7 +2403,7 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info,
* FIXME: Cannot ignore the return code, must hold
* notifier_lock
*/
- amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+ amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
}
/* Mark the BO as valid unless it was invalidated
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
index 2168163aad2d..252a876b0725 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
@@ -209,6 +209,7 @@ void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list,
list_add_tail(&e->tv.head, &bucket[priority]);
e->user_pages = NULL;
+ e->range = NULL;
}
/* Connect the sorted buckets in the output list. */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
index 9caea1688fc3..e4d78491bcc7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.h
@@ -26,6 +26,8 @@
#include <drm/ttm/ttm_execbuf_util.h>
#include <drm/amdgpu_drm.h>
+struct hmm_range;
+
struct amdgpu_device;
struct amdgpu_bo;
struct amdgpu_bo_va;
@@ -36,6 +38,7 @@ struct amdgpu_bo_list_entry {
struct amdgpu_bo_va *bo_va;
uint32_t priority;
struct page **user_pages;
+ struct hmm_range *range;
bool user_invalidated;
};
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
index 491d4846fc02..cfb262911bfc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
@@ -328,7 +328,6 @@ static void amdgpu_connector_free_edid(struct drm_connector *connector)
kfree(amdgpu_connector->edid);
amdgpu_connector->edid = NULL;
- drm_connector_update_edid_property(connector, NULL);
}
static int amdgpu_connector_ddc_get_modes(struct drm_connector *connector)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 1bbd39b3b0fc..365e3fb6a9e5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -109,6 +109,7 @@ static int amdgpu_cs_p1_ib(struct amdgpu_cs_parser *p,
return r;
++(num_ibs[r]);
+ p->gang_leader_idx = r;
return 0;
}
@@ -287,8 +288,10 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p,
}
}
- if (!p->gang_size)
- return -EINVAL;
+ if (!p->gang_size) {
+ ret = -EINVAL;
+ goto free_partial_kdata;
+ }
for (i = 0; i < p->gang_size; ++i) {
ret = amdgpu_job_alloc(p->adev, num_ibs[i], &p->jobs[i], vm);
@@ -300,7 +303,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p,
if (ret)
goto free_all_kdata;
}
- p->gang_leader = p->jobs[p->gang_size - 1];
+ p->gang_leader = p->jobs[p->gang_leader_idx];
if (p->ctx->vram_lost_counter != p->gang_leader->vram_lost_counter) {
ret = -ECANCELED;
@@ -910,7 +913,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
goto out_free_user_pages;
}
- r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages);
+ r = amdgpu_ttm_tt_get_user_pages(bo, e->user_pages, &e->range);
if (r) {
kvfree(e->user_pages);
e->user_pages = NULL;
@@ -988,10 +991,12 @@ out_free_user_pages:
if (!e->user_pages)
continue;
- amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+ amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, e->range);
kvfree(e->user_pages);
e->user_pages = NULL;
+ e->range = NULL;
}
+ mutex_unlock(&p->bo_list->bo_list_mutex);
return r;
}
@@ -1194,16 +1199,18 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
return r;
}
- for (i = 0; i < p->gang_size - 1; ++i) {
+ for (i = 0; i < p->gang_size; ++i) {
+ if (p->jobs[i] == leader)
+ continue;
+
r = amdgpu_sync_clone(&leader->sync, &p->jobs[i]->sync);
if (r)
return r;
}
- r = amdgpu_ctx_wait_prev_fence(p->ctx, p->entities[p->gang_size - 1]);
+ r = amdgpu_ctx_wait_prev_fence(p->ctx, p->entities[p->gang_leader_idx]);
if (r && r != -ERESTARTSYS)
DRM_ERROR("amdgpu_ctx_wait_prev_fence failed.\n");
-
return r;
}
@@ -1237,9 +1244,12 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
for (i = 0; i < p->gang_size; ++i)
drm_sched_job_arm(&p->jobs[i]->base);
- for (i = 0; i < (p->gang_size - 1); ++i) {
+ for (i = 0; i < p->gang_size; ++i) {
struct dma_fence *fence;
+ if (p->jobs[i] == leader)
+ continue;
+
fence = &p->jobs[i]->base.s_fence->scheduled;
r = amdgpu_sync_fence(&leader->sync, fence);
if (r)
@@ -1264,7 +1274,8 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
amdgpu_bo_list_for_each_userptr_entry(e, p->bo_list) {
struct amdgpu_bo *bo = ttm_to_amdgpu_bo(e->tv.bo);
- r |= !amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+ r |= !amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, e->range);
+ e->range = NULL;
}
if (r) {
r = -EAGAIN;
@@ -1275,7 +1286,10 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
list_for_each_entry(e, &p->validated, tv.head) {
/* Everybody except for the gang leader uses READ */
- for (i = 0; i < (p->gang_size - 1); ++i) {
+ for (i = 0; i < p->gang_size; ++i) {
+ if (p->jobs[i] == leader)
+ continue;
+
dma_resv_add_fence(e->tv.bo->base.resv,
&p->jobs[i]->base.s_fence->finished,
DMA_RESV_USAGE_READ);
@@ -1285,7 +1299,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
e->tv.num_shared = 0;
}
- seq = amdgpu_ctx_add_fence(p->ctx, p->entities[p->gang_size - 1],
+ seq = amdgpu_ctx_add_fence(p->ctx, p->entities[p->gang_leader_idx],
p->fence);
amdgpu_cs_post_dependencies(p);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h
index cbaa19b2b8a3..f80adf9069ec 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.h
@@ -54,6 +54,7 @@ struct amdgpu_cs_parser {
/* scheduler job objects */
unsigned int gang_size;
+ unsigned int gang_leader_idx;
struct drm_sched_entity *entities[AMDGPU_CS_GANG_SIZE];
struct amdgpu_job *jobs[AMDGPU_CS_GANG_SIZE];
struct amdgpu_job *gang_leader;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index ddaecb2610c9..f1e9663b4051 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4060,15 +4060,18 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
* at suspend time.
*
*/
-static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
+static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
{
+ int ret;
+
/* No need to evict vram on APUs for suspend to ram or s2idle */
if ((adev->in_s3 || adev->in_s0ix) && (adev->flags & AMD_IS_APU))
- return;
+ return 0;
- if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM))
+ ret = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
+ if (ret)
DRM_WARN("evicting device resources failed\n");
-
+ return ret;
}
/*
@@ -4118,7 +4121,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
if (!adev->in_s0ix)
amdgpu_amdkfd_suspend(adev, adev->in_runpm);
- amdgpu_device_evict_resources(adev);
+ r = amdgpu_device_evict_resources(adev);
+ if (r)
+ return r;
amdgpu_fence_driver_hw_fini(adev);
@@ -6039,3 +6044,44 @@ struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
dma_fence_put(old);
return NULL;
}
+
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
+{
+ switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+ case CHIP_HAINAN:
+#endif
+ case CHIP_TOPAZ:
+ /* chips with no display hardware */
+ return false;
+#ifdef CONFIG_DRM_AMDGPU_SI
+ case CHIP_TAHITI:
+ case CHIP_PITCAIRN:
+ case CHIP_VERDE:
+ case CHIP_OLAND:
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+ case CHIP_BONAIRE:
+ case CHIP_HAWAII:
+ case CHIP_KAVERI:
+ case CHIP_KABINI:
+ case CHIP_MULLINS:
+#endif
+ case CHIP_TONGA:
+ case CHIP_FIJI:
+ case CHIP_POLARIS10:
+ case CHIP_POLARIS11:
+ case CHIP_POLARIS12:
+ case CHIP_VEGAM:
+ case CHIP_CARRIZO:
+ case CHIP_STONEY:
+ /* chips with display hardware */
+ return true;
+ default:
+ /* IP discovery */
+ if (!adev->ip_versions[DCE_HWIP][0] ||
+ (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
+ return false;
+ return true;
+ }
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 3c9fecdd6b2f..bf2d50c8c92a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2201,7 +2201,8 @@ amdgpu_pci_remove(struct pci_dev *pdev)
pm_runtime_forbid(dev->dev);
}
- if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 2)) {
+ if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 2) &&
+ !amdgpu_sriov_vf(adev)) {
bool need_to_reset_gpu = false;
if (adev->gmc.xgmi.num_physical_nodes > 1) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 8ef31d687ef3..91571b1324f2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -378,6 +378,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
struct amdgpu_device *adev = drm_to_adev(dev);
struct drm_amdgpu_gem_userptr *args = data;
struct drm_gem_object *gobj;
+ struct hmm_range *range;
struct amdgpu_bo *bo;
uint32_t handle;
int r;
@@ -413,14 +414,13 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
if (r)
goto release_object;
- if (args->flags & AMDGPU_GEM_USERPTR_REGISTER) {
- r = amdgpu_mn_register(bo, args->addr);
- if (r)
- goto release_object;
- }
+ r = amdgpu_mn_register(bo, args->addr);
+ if (r)
+ goto release_object;
if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE) {
- r = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages);
+ r = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages,
+ &range);
if (r)
goto release_object;
@@ -443,7 +443,7 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data,
user_pages_done:
if (args->flags & AMDGPU_GEM_USERPTR_VALIDATE)
- amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm);
+ amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range);
release_object:
drm_gem_object_put(gobj);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 34233a74248c..28612e56d0d4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -479,6 +479,12 @@ int amdgpu_gmc_allocate_vm_inv_eng(struct amdgpu_device *adev)
unsigned i;
unsigned vmhub, inv_eng;
+ if (adev->enable_mes) {
+ /* reserve engine 5 for firmware */
+ for (vmhub = 0; vmhub < AMDGPU_MAX_VMHUBS; vmhub++)
+ vm_inv_engs[vmhub] &= ~(1 << 5);
+ }
+
for (i = 0; i < adev->num_rings; ++i) {
ring = adev->rings[i];
vmhub = ring->funcs->vmhub;
@@ -656,7 +662,7 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
}
if (amdgpu_sriov_vf(adev) ||
- !amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_DCE)) {
+ !amdgpu_device_has_display_hardware(adev)) {
size = 0;
} else {
size = amdgpu_gmc_get_vbios_fb_size(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index cd968e781077..adac650cf544 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -169,7 +169,11 @@ static void amdgpu_job_free_cb(struct drm_sched_job *s_job)
amdgpu_sync_free(&job->sync);
amdgpu_sync_free(&job->sched_sync);
- dma_fence_put(&job->hw_fence);
+ /* only put the hw fence if has embedded fence */
+ if (!job->hw_fence.ops)
+ kfree(job);
+ else
+ dma_fence_put(&job->hw_fence);
}
void amdgpu_job_set_gang_leader(struct amdgpu_job *job,
@@ -254,6 +258,9 @@ static struct dma_fence *amdgpu_job_dependency(struct drm_sched_job *sched_job,
DRM_ERROR("Error adding fence (%d)\n", r);
}
+ if (!fence && job->gang_submit)
+ fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit);
+
while (fence == NULL && vm && !job->vmid) {
r = amdgpu_vmid_grab(vm, ring, &job->sync,
&job->base.s_fence->finished,
@@ -264,9 +271,6 @@ static struct dma_fence *amdgpu_job_dependency(struct drm_sched_job *sched_job,
fence = amdgpu_sync_get_fence(&job->sync);
}
- if (!fence && job->gang_submit)
- fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit);
-
return fence;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index bf1ff8f0e712..4e42dcb1950f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -337,12 +337,14 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
fw_info->feature = adev->psp.cap_feature_version;
break;
case AMDGPU_INFO_FW_MES_KIQ:
- fw_info->ver = adev->mes.ucode_fw_version[0];
- fw_info->feature = 0;
+ fw_info->ver = adev->mes.kiq_version & AMDGPU_MES_VERSION_MASK;
+ fw_info->feature = (adev->mes.kiq_version & AMDGPU_MES_FEAT_VERSION_MASK)
+ >> AMDGPU_MES_FEAT_VERSION_SHIFT;
break;
case AMDGPU_INFO_FW_MES:
- fw_info->ver = adev->mes.ucode_fw_version[1];
- fw_info->feature = 0;
+ fw_info->ver = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK;
+ fw_info->feature = (adev->mes.sched_version & AMDGPU_MES_FEAT_VERSION_MASK)
+ >> AMDGPU_MES_FEAT_VERSION_SHIFT;
break;
case AMDGPU_INFO_FW_IMU:
fw_info->ver = adev->gfx.imu_fw_version;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index effa7df3ddbf..7978307e1d6d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -172,6 +172,7 @@ void psp_ta_free_shared_buf(struct ta_mem_context *mem_ctx)
{
amdgpu_bo_free_kernel(&mem_ctx->shared_bo, &mem_ctx->shared_mc_addr,
&mem_ctx->shared_buf);
+ mem_ctx->shared_bo = NULL;
}
static void psp_free_shared_bufs(struct psp_context *psp)
@@ -182,6 +183,7 @@ static void psp_free_shared_bufs(struct psp_context *psp)
/* free TMR memory buffer */
pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
amdgpu_bo_free_kernel(&psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+ psp->tmr_bo = NULL;
/* free xgmi shared memory */
psp_ta_free_shared_buf(&psp->xgmi_context.context.mem_context);
@@ -743,7 +745,7 @@ static int psp_load_toc(struct psp_context *psp,
/* Set up Trusted Memory Region */
static int psp_tmr_init(struct psp_context *psp)
{
- int ret;
+ int ret = 0;
int tmr_size;
void *tmr_buf;
void **pptr;
@@ -770,10 +772,12 @@ static int psp_tmr_init(struct psp_context *psp)
}
}
- pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
- ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_ALIGNMENT,
- AMDGPU_GEM_DOMAIN_VRAM,
- &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+ if (!psp->tmr_bo) {
+ pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
+ ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_ALIGNMENT,
+ AMDGPU_GEM_DOMAIN_VRAM,
+ &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
+ }
return ret;
}
@@ -2732,8 +2736,6 @@ static int psp_suspend(void *handle)
}
out:
- psp_free_shared_bufs(psp);
-
return ret;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 57277b1cf183..b64938ed8cb6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -643,9 +643,6 @@ struct amdgpu_ttm_tt {
struct task_struct *usertask;
uint32_t userflags;
bool bound;
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
- struct hmm_range *range;
-#endif
};
#define ttm_to_amdgpu_ttm_tt(ptr) container_of(ptr, struct amdgpu_ttm_tt, ttm)
@@ -658,7 +655,8 @@ struct amdgpu_ttm_tt {
* Calling function must call amdgpu_ttm_tt_userptr_range_done() once and only
* once afterwards to stop HMM tracking
*/
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages,
+ struct hmm_range **range)
{
struct ttm_tt *ttm = bo->tbo.ttm;
struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
@@ -668,16 +666,15 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
bool readonly;
int r = 0;
+ /* Make sure get_user_pages_done() can cleanup gracefully */
+ *range = NULL;
+
mm = bo->notifier.mm;
if (unlikely(!mm)) {
DRM_DEBUG_DRIVER("BO is not registered?\n");
return -EFAULT;
}
- /* Another get_user_pages is running at the same time?? */
- if (WARN_ON(gtt->range))
- return -EFAULT;
-
if (!mmget_not_zero(mm)) /* Happens during process shutdown */
return -ESRCH;
@@ -695,7 +692,7 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
readonly = amdgpu_ttm_tt_is_readonly(ttm);
r = amdgpu_hmm_range_get_pages(&bo->notifier, mm, pages, start,
- ttm->num_pages, &gtt->range, readonly,
+ ttm->num_pages, range, readonly,
true, NULL);
out_unlock:
mmap_read_unlock(mm);
@@ -713,30 +710,24 @@ out_unlock:
*
* Returns: true if pages are still valid
*/
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
+bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+ struct hmm_range *range)
{
struct amdgpu_ttm_tt *gtt = ttm_to_amdgpu_ttm_tt(ttm);
- bool r = false;
- if (!gtt || !gtt->userptr)
+ if (!gtt || !gtt->userptr || !range)
return false;
DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%x\n",
gtt->userptr, ttm->num_pages);
- WARN_ONCE(!gtt->range || !gtt->range->hmm_pfns,
- "No user pages to check\n");
+ WARN_ONCE(!range->hmm_pfns, "No user pages to check\n");
- if (gtt->range) {
- /*
- * FIXME: Must always hold notifier_lock for this, and must
- * not ignore the return code.
- */
- r = amdgpu_hmm_range_get_pages_done(gtt->range);
- gtt->range = NULL;
- }
-
- return !r;
+ /*
+ * FIXME: Must always hold notifier_lock for this, and must
+ * not ignore the return code.
+ */
+ return !amdgpu_hmm_range_get_pages_done(range);
}
#endif
@@ -813,20 +804,6 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev,
/* unmap the pages mapped to the device */
dma_unmap_sgtable(adev->dev, ttm->sg, direction, 0);
sg_free_table(ttm->sg);
-
-#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
- if (gtt->range) {
- unsigned long i;
-
- for (i = 0; i < ttm->num_pages; i++) {
- if (ttm->pages[i] !=
- hmm_pfn_to_page(gtt->range->hmm_pfns[i]))
- break;
- }
-
- WARN((i == ttm->num_pages), "Missing get_user_page_done\n");
- }
-#endif
}
static void amdgpu_ttm_gart_bind(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 6a70818039dd..a37207011a69 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -39,6 +39,8 @@
#define AMDGPU_POISON 0xd0bed0be
+struct hmm_range;
+
struct amdgpu_gtt_mgr {
struct ttm_resource_manager manager;
struct drm_mm mm;
@@ -149,15 +151,19 @@ void amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo);
uint64_t amdgpu_ttm_domain_start(struct amdgpu_device *adev, uint32_t type);
#if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR)
-int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages);
-bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm);
+int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages,
+ struct hmm_range **range);
+bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+ struct hmm_range *range);
#else
static inline int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo,
- struct page **pages)
+ struct page **pages,
+ struct hmm_range **range)
{
return -EPERM;
}
-static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
+static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm,
+ struct hmm_range *range)
{
return false;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index 0b52af415b28..ce64ca1c6e66 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -156,6 +156,9 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
break;
case IP_VERSION(3, 0, 2):
fw_name = FIRMWARE_VANGOGH;
+ if ((adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) &&
+ (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG))
+ adev->vcn.indirect_sram = true;
break;
case IP_VERSION(3, 0, 16):
fw_name = FIRMWARE_DIMGREY_CAVEFISH;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
index f4b5301ea2a0..500a1dc4fe02 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
@@ -500,6 +500,8 @@ static int amdgpu_vkms_sw_init(void *handle)
adev_to_drm(adev)->mode_config.fb_base = adev->gmc.aper_base;
+ adev_to_drm(adev)->mode_config.fb_modifiers_not_supported = true;
+
r = amdgpu_display_modeset_create_props(adev);
if (r)
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 2291aa14d888..003aa9e47085 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -143,32 +143,6 @@ int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm,
return 0;
}
-/*
- * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
- * happens while holding this lock anywhere to prevent deadlocks when
- * an MMU notifier runs in reclaim-FS context.
- */
-static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
-{
- mutex_lock(&vm->eviction_lock);
- vm->saved_flags = memalloc_noreclaim_save();
-}
-
-static inline int amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
-{
- if (mutex_trylock(&vm->eviction_lock)) {
- vm->saved_flags = memalloc_noreclaim_save();
- return 1;
- }
- return 0;
-}
-
-static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
-{
- memalloc_noreclaim_restore(vm->saved_flags);
- mutex_unlock(&vm->eviction_lock);
-}
-
/**
* amdgpu_vm_bo_evicted - vm_bo is evicted
*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 83acb7bd80fe..6546e786bf00 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -492,7 +492,48 @@ void amdgpu_debugfs_vm_bo_info(struct amdgpu_vm *vm, struct seq_file *m);
*/
static inline uint64_t amdgpu_vm_tlb_seq(struct amdgpu_vm *vm)
{
+ unsigned long flags;
+ spinlock_t *lock;
+
+ /*
+ * Workaround to stop racing between the fence signaling and handling
+ * the cb. The lock is static after initially setting it up, just make
+ * sure that the dma_fence structure isn't freed up.
+ */
+ rcu_read_lock();
+ lock = vm->last_tlb_flush->lock;
+ rcu_read_unlock();
+
+ spin_lock_irqsave(lock, flags);
+ spin_unlock_irqrestore(lock, flags);
+
return atomic64_read(&vm->tlb_seq);
}
+/*
+ * vm eviction_lock can be taken in MMU notifiers. Make sure no reclaim-FS
+ * happens while holding this lock anywhere to prevent deadlocks when
+ * an MMU notifier runs in reclaim-FS context.
+ */
+static inline void amdgpu_vm_eviction_lock(struct amdgpu_vm *vm)
+{
+ mutex_lock(&vm->eviction_lock);
+ vm->saved_flags = memalloc_noreclaim_save();
+}
+
+static inline bool amdgpu_vm_eviction_trylock(struct amdgpu_vm *vm)
+{
+ if (mutex_trylock(&vm->eviction_lock)) {
+ vm->saved_flags = memalloc_noreclaim_save();
+ return true;
+ }
+ return false;
+}
+
+static inline void amdgpu_vm_eviction_unlock(struct amdgpu_vm *vm)
+{
+ memalloc_noreclaim_restore(vm->saved_flags);
+ mutex_unlock(&vm->eviction_lock);
+}
+
#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
index 358b91243e37..b5f3bba851db 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c
@@ -597,7 +597,9 @@ static int amdgpu_vm_pt_alloc(struct amdgpu_device *adev,
if (entry->bo)
return 0;
+ amdgpu_vm_eviction_unlock(vm);
r = amdgpu_vm_pt_create(adev, vm, cursor->level, immediate, &pt);
+ amdgpu_vm_eviction_lock(vm);
if (r)
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index 73a517bcf5c1..80dd1343594c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -435,7 +435,7 @@ static int amdgpu_vram_mgr_new(struct ttm_resource_manager *man,
if (place->flags & TTM_PL_FLAG_TOPDOWN)
vres->flags |= DRM_BUDDY_TOPDOWN_ALLOCATION;
- if (fpfn || lpfn != man->size)
+ if (fpfn || lpfn != mgr->mm.size)
/* Allocate blocks in desired range */
vres->flags |= DRM_BUDDY_RANGE_ALLOCATION;
diff --git a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
index f772bb499f3e..0312c71c3af9 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/mmsch_v4_0.h
@@ -32,7 +32,6 @@
#define RB_ENABLED (1 << 0)
#define RB4_ENABLED (1 << 1)
-#define MMSCH_DOORBELL_OFFSET 0x8
#define MMSCH_VF_ENGINE_STATUS__PASS 0x1
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index 21d822b1d589..88f9b327183a 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -45,6 +45,7 @@ MODULE_FIRMWARE("amdgpu/psp_13_0_0_ta.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_7_sos.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_7_ta.bin");
MODULE_FIRMWARE("amdgpu/psp_13_0_10_sos.bin");
+MODULE_FIRMWARE("amdgpu/psp_13_0_10_ta.bin");
/* For large FW files the time to complete can be very long */
#define USBC_PD_POLLING_LIMIT_S 240
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
index 897a5ce9c9da..dcc49b01bd59 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -100,7 +100,6 @@ static int vcn_v4_0_sw_init(void *handle)
struct amdgpu_ring *ring;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
int i, r;
- int vcn_doorbell_index = 0;
r = amdgpu_vcn_sw_init(adev);
if (r)
@@ -112,12 +111,6 @@ static int vcn_v4_0_sw_init(void *handle)
if (r)
return r;
- if (amdgpu_sriov_vf(adev)) {
- vcn_doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 - MMSCH_DOORBELL_OFFSET;
- /* get DWORD offset */
- vcn_doorbell_index = vcn_doorbell_index << 1;
- }
-
for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
volatile struct amdgpu_vcn4_fw_shared *fw_shared;
@@ -135,7 +128,7 @@ static int vcn_v4_0_sw_init(void *handle)
ring = &adev->vcn.inst[i].ring_enc[0];
ring->use_doorbell = true;
if (amdgpu_sriov_vf(adev))
- ring->doorbell_index = vcn_doorbell_index + i * (adev->vcn.num_enc_rings + 1) + 1;
+ ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + i * (adev->vcn.num_enc_rings + 1) + 1;
else
ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + 8 * i;