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.h3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c18
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_device.c31
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c57
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c24
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c14
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c8
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c16
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v11_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v13_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/psp_v14_0.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c9
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c3
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c54
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vi.c7
33 files changed, 235 insertions, 96 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 2c04ae1338484..c3641331d4de7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1123,6 +1123,7 @@ struct amdgpu_device {
bool in_s3;
bool in_s4;
bool in_s0ix;
+ suspend_state_t last_suspend_state;
enum pp_mp1_state mp1_state;
struct amdgpu_doorbell_index doorbell_index;
@@ -1613,11 +1614,9 @@ static inline void amdgpu_acpi_get_backlight_caps(struct amdgpu_dm_backlight_cap
#if defined(CONFIG_ACPI) && defined(CONFIG_SUSPEND)
bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev);
bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev);
-void amdgpu_choose_low_power_state(struct amdgpu_device *adev);
#else
static inline bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) { return false; }
static inline bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev) { return false; }
-static inline void amdgpu_choose_low_power_state(struct amdgpu_device *adev) { }
#endif
void amdgpu_register_gpu_instance(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index b7f8f2ff143dd..707e131f89d23 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -1533,22 +1533,4 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev)
#endif /* CONFIG_AMD_PMC */
}
-/**
- * amdgpu_choose_low_power_state
- *
- * @adev: amdgpu_device_pointer
- *
- * Choose the target low power state for the GPU
- */
-void amdgpu_choose_low_power_state(struct amdgpu_device *adev)
-{
- if (adev->in_runpm)
- return;
-
- if (amdgpu_acpi_is_s0ix_active(adev))
- adev->in_s0ix = true;
- else if (amdgpu_acpi_is_s3_active(adev))
- adev->in_s3 = true;
-}
-
#endif /* CONFIG_SUSPEND */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
index cfdf558b48b64..02138aa557935 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
@@ -109,7 +109,7 @@ int amdgpu_unmap_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm,
struct drm_exec exec;
int r;
- drm_exec_init(&exec, DRM_EXEC_INTERRUPTIBLE_WAIT, 0);
+ drm_exec_init(&exec, 0, 0);
drm_exec_until_all_locked(&exec) {
r = amdgpu_vm_lock_pd(vm, &exec, 0);
if (likely(!r))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b34b915203f25..f8b3e04d71eda 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -3510,6 +3510,7 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
amdgpu_device_mem_scratch_fini(adev);
amdgpu_ib_pool_fini(adev);
amdgpu_seq64_fini(adev);
+ amdgpu_doorbell_fini(adev);
}
if (adev->ip_blocks[i].version->funcs->sw_fini) {
r = adev->ip_blocks[i].version->funcs->sw_fini(&adev->ip_blocks[i]);
@@ -4858,7 +4859,6 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
iounmap(adev->rmmio);
adev->rmmio = NULL;
- amdgpu_doorbell_fini(adev);
drm_dev_exit(idx);
}
@@ -4907,28 +4907,20 @@ static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
* @data: data
*
* This function is called when the system is about to suspend or hibernate.
- * It is used to evict resources from the device before the system goes to
- * sleep while there is still access to swap.
+ * It is used to set the appropriate flags so that eviction can be optimized
+ * in the pm prepare callback.
*/
static int amdgpu_device_pm_notifier(struct notifier_block *nb, unsigned long mode,
void *data)
{
struct amdgpu_device *adev = container_of(nb, struct amdgpu_device, pm_nb);
- int r;
switch (mode) {
case PM_HIBERNATION_PREPARE:
adev->in_s4 = true;
- fallthrough;
- case PM_SUSPEND_PREPARE:
- r = amdgpu_device_evict_resources(adev);
- /*
- * This is considered non-fatal at this time because
- * amdgpu_device_prepare() will also fatally evict resources.
- * See https://gitlab.freedesktop.org/drm/amd/-/issues/3781
- */
- if (r)
- drm_warn(adev_to_drm(adev), "Failed to evict resources, freeze active processes if problems occur: %d\n", r);
+ break;
+ case PM_POST_HIBERNATION:
+ adev->in_s4 = false;
break;
}
@@ -4949,15 +4941,13 @@ int amdgpu_device_prepare(struct drm_device *dev)
struct amdgpu_device *adev = drm_to_adev(dev);
int i, r;
- amdgpu_choose_low_power_state(adev);
-
if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
return 0;
/* Evict the majority of BOs before starting suspend sequence */
r = amdgpu_device_evict_resources(adev);
if (r)
- goto unprepare;
+ return r;
flush_delayed_work(&adev->gfx.gfx_off_delay_work);
@@ -4968,15 +4958,10 @@ int amdgpu_device_prepare(struct drm_device *dev)
continue;
r = adev->ip_blocks[i].version->funcs->prepare_suspend(&adev->ip_blocks[i]);
if (r)
- goto unprepare;
+ return r;
}
return 0;
-
-unprepare:
- adev->in_s0ix = adev->in_s3 = adev->in_s4 = false;
-
- return r;
}
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 667080cc9ae1c..44e120f9f7649 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -43,6 +43,29 @@
#include <linux/dma-fence-array.h>
#include <linux/pci-p2pdma.h>
+static const struct dma_buf_attach_ops amdgpu_dma_buf_attach_ops;
+
+/**
+ * dma_buf_attach_adev - Helper to get adev of an attachment
+ *
+ * @attach: attachment
+ *
+ * Returns:
+ * A struct amdgpu_device * if the attaching device is an amdgpu device or
+ * partition, NULL otherwise.
+ */
+static struct amdgpu_device *dma_buf_attach_adev(struct dma_buf_attachment *attach)
+{
+ if (attach->importer_ops == &amdgpu_dma_buf_attach_ops) {
+ struct drm_gem_object *obj = attach->importer_priv;
+ struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+
+ return amdgpu_ttm_adev(bo->tbo.bdev);
+ }
+
+ return NULL;
+}
+
/**
* amdgpu_dma_buf_attach - &dma_buf_ops.attach implementation
*
@@ -54,11 +77,13 @@
static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
struct dma_buf_attachment *attach)
{
+ struct amdgpu_device *attach_adev = dma_buf_attach_adev(attach);
struct drm_gem_object *obj = dmabuf->priv;
struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
- if (pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
+ if (!amdgpu_dmabuf_is_xgmi_accessible(attach_adev, bo) &&
+ pci_p2pdma_distance(adev->pdev, attach->dev, false) < 0)
attach->peer2peer = false;
amdgpu_vm_bo_update_shared(bo);
@@ -77,22 +102,32 @@ static int amdgpu_dma_buf_pin(struct dma_buf_attachment *attach)
{
struct dma_buf *dmabuf = attach->dmabuf;
struct amdgpu_bo *bo = gem_to_amdgpu_bo(dmabuf->priv);
- u32 domains = bo->preferred_domains;
+ u32 domains = bo->allowed_domains;
dma_resv_assert_held(dmabuf->resv);
- /*
- * Try pinning into VRAM to allow P2P with RDMA NICs without ODP
+ /* Try pinning into VRAM to allow P2P with RDMA NICs without ODP
* support if all attachments can do P2P. If any attachment can't do
* P2P just pin into GTT instead.
+ *
+ * To avoid with conflicting pinnings between GPUs and RDMA when move
+ * notifiers are disabled, only allow pinning in VRAM when move
+ * notiers are enabled.
*/
- list_for_each_entry(attach, &dmabuf->attachments, node)
- if (!attach->peer2peer)
- domains &= ~AMDGPU_GEM_DOMAIN_VRAM;
+ if (!IS_ENABLED(CONFIG_DMABUF_MOVE_NOTIFY)) {
+ domains &= ~AMDGPU_GEM_DOMAIN_VRAM;
+ } else {
+ list_for_each_entry(attach, &dmabuf->attachments, node)
+ if (!attach->peer2peer)
+ domains &= ~AMDGPU_GEM_DOMAIN_VRAM;
+ }
if (domains & AMDGPU_GEM_DOMAIN_VRAM)
bo->flags |= AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
+ if (WARN_ON(!domains))
+ return -EINVAL;
+
return amdgpu_bo_pin(bo, domains);
}
@@ -164,6 +199,11 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
break;
case TTM_PL_VRAM:
+ /* XGMI-accessible memory should never be DMA-mapped */
+ if (WARN_ON(amdgpu_dmabuf_is_xgmi_accessible(
+ dma_buf_attach_adev(attach), bo)))
+ return ERR_PTR(-EINVAL);
+
r = amdgpu_vram_mgr_alloc_sgt(adev, bo->tbo.resource, 0,
bo->tbo.base.size, attach->dev,
dir, &sgt);
@@ -470,6 +510,9 @@ bool amdgpu_dmabuf_is_xgmi_accessible(struct amdgpu_device *adev,
struct drm_gem_object *obj = &bo->tbo.base;
struct drm_gem_object *gobj;
+ if (!adev)
+ return false;
+
if (obj->import_attach) {
struct dma_buf *dma_buf = obj->import_attach->dmabuf;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 26bf896f1444e..72c807f5822e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -2548,8 +2548,20 @@ static int amdgpu_pmops_suspend(struct device *dev)
adev->in_s0ix = true;
else if (amdgpu_acpi_is_s3_active(adev))
adev->in_s3 = true;
- if (!adev->in_s0ix && !adev->in_s3)
+ if (!adev->in_s0ix && !adev->in_s3) {
+ /* don't allow going deep first time followed by s2idle the next time */
+ if (adev->last_suspend_state != PM_SUSPEND_ON &&
+ adev->last_suspend_state != pm_suspend_target_state) {
+ drm_err_once(drm_dev, "Unsupported suspend state %d\n",
+ pm_suspend_target_state);
+ return -EINVAL;
+ }
return 0;
+ }
+
+ /* cache the state last used for suspend */
+ adev->last_suspend_state = pm_suspend_target_state;
+
return amdgpu_device_suspend(drm_dev, true);
}
@@ -2603,13 +2615,8 @@ static int amdgpu_pmops_freeze(struct device *dev)
static int amdgpu_pmops_thaw(struct device *dev)
{
struct drm_device *drm_dev = dev_get_drvdata(dev);
- struct amdgpu_device *adev = drm_to_adev(drm_dev);
- int r;
-
- r = amdgpu_device_resume(drm_dev, true);
- adev->in_s4 = false;
- return r;
+ return amdgpu_device_resume(drm_dev, true);
}
static int amdgpu_pmops_poweroff(struct device *dev)
@@ -2622,9 +2629,6 @@ static int amdgpu_pmops_poweroff(struct device *dev)
static int amdgpu_pmops_restore(struct device *dev)
{
struct drm_device *drm_dev = dev_get_drvdata(dev);
- struct amdgpu_device *adev = drm_to_adev(drm_dev);
-
- adev->in_s4 = false;
return amdgpu_device_resume(drm_dev, true);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 72af5e5a894a2..cf2df7790077d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -1438,9 +1438,11 @@ static int amdgpu_gfx_run_cleaner_shader_job(struct amdgpu_ring *ring)
struct amdgpu_device *adev = ring->adev;
struct drm_gpu_scheduler *sched = &ring->sched;
struct drm_sched_entity entity;
+ static atomic_t counter;
struct dma_fence *f;
struct amdgpu_job *job;
struct amdgpu_ib *ib;
+ void *owner;
int i, r;
/* Initialize the scheduler entity */
@@ -1451,9 +1453,15 @@ static int amdgpu_gfx_run_cleaner_shader_job(struct amdgpu_ring *ring)
goto err;
}
- r = amdgpu_job_alloc_with_ib(ring->adev, &entity, NULL,
- 64, 0,
- &job);
+ /*
+ * Use some unique dummy value as the owner to make sure we execute
+ * the cleaner shader on each submission. The value just need to change
+ * for each submission and is otherwise meaningless.
+ */
+ owner = (void *)(unsigned long)atomic_inc_return(&counter);
+
+ r = amdgpu_job_alloc_with_ib(ring->adev, &entity, owner,
+ 64, 0, &job);
if (r)
goto err;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index cdcdae7f71ce9..83adf81defc71 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
@@ -66,7 +66,6 @@
#define VCN_ENC_CMD_REG_WAIT 0x0000000c
#define VCN_AON_SOC_ADDRESS_2_0 0x1f800
-#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
#define VCN_VID_IP_ADDRESS_2_0 0x0
#define VCN_AON_IP_ADDRESS_2_0 0x30000
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index a63ce747863f1..23e6a05359c24 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -6114,7 +6114,7 @@ static int gfx_v10_0_cp_gfx_load_pfp_microcode(struct amdgpu_device *adev)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, mmCP_PFP_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
@@ -6192,7 +6192,7 @@ static int gfx_v10_0_cp_gfx_load_ce_microcode(struct amdgpu_device *adev)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, mmCP_CE_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_CE_IC_BASE_CNTL, VMID, 0);
@@ -6269,7 +6269,7 @@ static int gfx_v10_0_cp_gfx_load_me_microcode(struct amdgpu_device *adev)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, mmCP_ME_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
@@ -6644,7 +6644,7 @@ static int gfx_v10_0_cp_compute_load_microcode(struct amdgpu_device *adev)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, mmCP_CPC_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index d57db42f95360..2a5c2a1ae3c74 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -2428,7 +2428,7 @@ static int gfx_v11_0_config_me_cache(struct amdgpu_device *adev, uint64_t addr)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, regCP_ME_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_ME_IC_BASE_CNTL, VMID, 0);
@@ -2472,7 +2472,7 @@ static int gfx_v11_0_config_pfp_cache(struct amdgpu_device *adev, uint64_t addr)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_PFP_IC_BASE_CNTL, VMID, 0);
@@ -2517,7 +2517,7 @@ static int gfx_v11_0_config_mec_cache(struct amdgpu_device *adev, uint64_t addr)
}
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
tmp = RREG32_SOC15(GC, 0, regCP_CPC_IC_BASE_CNTL);
tmp = REG_SET_FIELD(tmp, CP_CPC_IC_BASE_CNTL, CACHE_POLICY, 0);
@@ -3153,7 +3153,7 @@ static int gfx_v11_0_cp_gfx_load_pfp_microcode_rs64(struct amdgpu_device *adev)
amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_data_obj);
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
lower_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
@@ -3371,7 +3371,7 @@ static int gfx_v11_0_cp_gfx_load_me_microcode_rs64(struct amdgpu_device *adev)
amdgpu_bo_unreserve(adev->gfx.me.me_fw_data_obj);
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
lower_32_bits(adev->gfx.me.me_fw_gpu_addr));
@@ -4541,7 +4541,7 @@ static int gfx_v11_0_gfxhub_enable(struct amdgpu_device *adev)
if (r)
return r;
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index e7b58e4702927..62a257a4a3e9b 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -2324,7 +2324,7 @@ static int gfx_v12_0_cp_gfx_load_pfp_microcode_rs64(struct amdgpu_device *adev)
amdgpu_bo_unreserve(adev->gfx.pfp.pfp_fw_data_obj);
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
WREG32_SOC15(GC, 0, regCP_PFP_IC_BASE_LO,
lower_32_bits(adev->gfx.pfp.pfp_fw_gpu_addr));
@@ -2468,7 +2468,7 @@ static int gfx_v12_0_cp_gfx_load_me_microcode_rs64(struct amdgpu_device *adev)
amdgpu_bo_unreserve(adev->gfx.me.me_fw_data_obj);
if (amdgpu_emu_mode == 1)
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
WREG32_SOC15(GC, 0, regCP_ME_IC_BASE_LO,
lower_32_bits(adev->gfx.me.me_fw_gpu_addr));
@@ -3426,7 +3426,7 @@ static int gfx_v12_0_gfxhub_enable(struct amdgpu_device *adev)
if (r)
return r;
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 95d894a231fcf..809b3a882d0d7 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -268,7 +268,7 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
/* flush hdp cache */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
/* This is necessary for SRIOV as well as for GFXOFF to function
* properly under bare metal
@@ -969,7 +969,7 @@ static int gmc_v10_0_gart_enable(struct amdgpu_device *adev)
adev->hdp.funcs->init_registers(adev);
/* Flush HDP after it is initialized */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index ad099f136f84e..fec9a007533ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -229,7 +229,7 @@ static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
/* flush hdp cache */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
/* This is necessary for SRIOV as well as for GFXOFF to function
* properly under bare metal
@@ -752,6 +752,18 @@ static int gmc_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
adev->gmc.vram_type = vram_type;
adev->gmc.vram_vendor = vram_vendor;
+ /* The mall_size is already calculated as mall_size_per_umc * num_umc.
+ * However, for gfx1151, which features a 2-to-1 UMC mapping,
+ * the result must be multiplied by 2 to determine the actual mall size.
+ */
+ switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+ case IP_VERSION(11, 5, 1):
+ adev->gmc.mall_size *= 2;
+ break;
+ default:
+ break;
+ }
+
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(11, 0, 0):
case IP_VERSION(11, 0, 1):
@@ -899,7 +911,7 @@ static int gmc_v11_0_gart_enable(struct amdgpu_device *adev)
return r;
/* Flush HDP after it is initialized */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
index 05c026d0b0d96..c6f290704d473 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_0.c
@@ -297,7 +297,7 @@ static void gmc_v12_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
return;
/* flush hdp cache */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
/* This is necessary for SRIOV as well as for GFXOFF to function
* properly under bare metal
@@ -881,7 +881,7 @@ static int gmc_v12_0_gart_enable(struct amdgpu_device *adev)
return r;
/* Flush HDP after it is initialized */
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
value = (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS) ?
false : true;
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 783e0c3b86b4c..5effe8327d29f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -2435,7 +2435,7 @@ static int gmc_v9_0_hw_init(struct amdgpu_ip_block *ip_block)
adev->hdp.funcs->init_registers(adev);
/* After HDP is initialized, flush HDP.*/
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
if (amdgpu_vm_fault_stop == AMDGPU_VM_FAULT_STOP_ALWAYS)
value = false;
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
index f1dc13b3ab38e..cbbeadeb53f72 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
@@ -41,7 +41,12 @@ static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev,
{
if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ /* We just need to read back a register to post the write.
+ * Reading back the remapped register causes problems on
+ * some platforms so just read back the memory size register.
+ */
+ if (adev->nbio.funcs->get_memsize)
+ adev->nbio.funcs->get_memsize(adev);
} else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
index 43195c0797480..086a647308df0 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
@@ -32,7 +32,12 @@ static void hdp_v5_0_flush_hdp(struct amdgpu_device *adev,
{
if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ /* We just need to read back a register to post the write.
+ * Reading back the remapped register causes problems on
+ * some platforms so just read back the memory size register.
+ */
+ if (adev->nbio.funcs->get_memsize)
+ adev->nbio.funcs->get_memsize(adev);
} else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
index fcb8dd2876bcc..40940b4ab4007 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
@@ -33,7 +33,17 @@ static void hdp_v5_2_flush_hdp(struct amdgpu_device *adev,
if (!ring || !ring->funcs->emit_wreg) {
WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
0);
- RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ if (amdgpu_sriov_vf(adev)) {
+ /* this is fine because SR_IOV doesn't remap the register */
+ RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
+ /* We just need to read back a register to post the write.
+ * Reading back the remapped register causes problems on
+ * some platforms so just read back the memory size register.
+ */
+ if (adev->nbio.funcs->get_memsize)
+ adev->nbio.funcs->get_memsize(adev);
+ }
} else {
amdgpu_ring_emit_wreg(ring,
(adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
index a88d25a06c29b..6ccd31c8bc692 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
@@ -35,7 +35,12 @@ static void hdp_v6_0_flush_hdp(struct amdgpu_device *adev,
{
if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ /* We just need to read back a register to post the write.
+ * Reading back the remapped register causes problems on
+ * some platforms so just read back the memory size register.
+ */
+ if (adev->nbio.funcs->get_memsize)
+ adev->nbio.funcs->get_memsize(adev);
} else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
index 49f7eb4fbd117..2c9239a22f398 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
@@ -32,7 +32,12 @@ static void hdp_v7_0_flush_hdp(struct amdgpu_device *adev,
{
if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ /* We just need to read back a register to post the write.
+ * Reading back the remapped register causes problems on
+ * some platforms so just read back the memory size register.
+ */
+ if (adev->nbio.funcs->get_memsize)
+ adev->nbio.funcs->get_memsize(adev);
} else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
}
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
index 2ece3ae75ec12..bed5ef4d87889 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
@@ -360,7 +360,7 @@ static void nbio_v7_11_get_clockgating_state(struct amdgpu_device *adev,
*flags |= AMD_CG_SUPPORT_BIF_LS;
}
-#define MMIO_REG_HOLE_OFFSET (0x80000 - PAGE_SIZE)
+#define MMIO_REG_HOLE_OFFSET 0x44000
static void nbio_v7_11_set_reg_remap(struct amdgpu_device *adev)
{
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
index bb5dfc410a667..215543575f477 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
@@ -533,7 +533,7 @@ static int psp_v11_0_memory_training(struct psp_context *psp, uint32_t ops)
}
memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
vfree(buf);
drm_dev_exit(idx);
} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index cc621064610f1..afdf8ce3b4c59 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -610,7 +610,7 @@ static int psp_v13_0_memory_training(struct psp_context *psp, uint32_t ops)
}
memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
vfree(buf);
drm_dev_exit(idx);
} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c
index 7c49c3f3c3881..256288c6cd78e 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v14_0.c
@@ -498,7 +498,7 @@ static int psp_v14_0_memory_training(struct psp_context *psp, uint32_t ops)
}
memcpy_toio(adev->mman.aper_base_kaddr, buf, sz);
- adev->hdp.funcs->flush_hdp(adev, NULL);
+ amdgpu_device_flush_hdp(adev, NULL);
vfree(buf);
drm_dev_exit(idx);
} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
index 8e7a36f26e9cb..b8d835c9e17ed 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
@@ -39,6 +39,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fa00
#define VCN1_VID_SOC_ADDRESS_3_0 0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
#define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x1fd
#define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x503
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
index d716510b8dd68..3eec1b8feaeea 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
@@ -39,6 +39,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fa00
#define VCN1_VID_SOC_ADDRESS_3_0 0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
#define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x27
#define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x0f
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
index 22ae1939476f0..0b19f0ab4480d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
@@ -40,6 +40,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fa00
#define VCN1_VID_SOC_ADDRESS_3_0 0x48200
+#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
#define mmUVD_CONTEXT_ID_INTERNAL_OFFSET 0x27
#define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET 0x0f
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
index c6f6392c1c20b..1f777c125b00d 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
@@ -46,6 +46,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fb00
#define VCN1_VID_SOC_ADDRESS_3_0 0x48300
+#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
#define VCN_HARVEST_MMSCH 0
@@ -614,7 +615,8 @@ static void vcn_v4_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
/* VCN global tiling registers */
WREG32_SOC15_DPG_MODE(inst_idx, SOC15_DPG_MODE_OFFSET(
- VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+ VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG),
+ adev->gfx.config.gb_addr_config, 0, indirect);
}
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
index 3e176b4b7c69d..012f6ea928ec6 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
@@ -45,6 +45,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fb00
#define VCN1_VID_SOC_ADDRESS_3_0 0x48300
+#define VCN1_AON_SOC_ADDRESS_3_0 0x48000
static const struct amdgpu_hwip_reg_entry vcn_reg_list_4_0_3[] = {
SOC15_REG_ENTRY_STR(VCN, 0, regUVD_POWER_STATUS),
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
index ba603b2246e2e..f11df9c2ec131 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
@@ -46,6 +46,7 @@
#define VCN_VID_SOC_ADDRESS_2_0 0x1fb00
#define VCN1_VID_SOC_ADDRESS_3_0 (0x48300 + 0x38000)
+#define VCN1_AON_SOC_ADDRESS_3_0 (0x48000 + 0x38000)
#define VCN_HARVEST_MMSCH 0
@@ -1022,6 +1023,10 @@ static int vcn_v4_0_5_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
ring->doorbell_index << VCN_RB1_DB_CTRL__OFFSET__SHIFT |
VCN_RB1_DB_CTRL__EN_MASK);
+ /* Keeping one read-back to ensure all register writes are done, otherwise
+ * it may introduce race conditions */
+ RREG32_SOC15(VCN, inst_idx, regVCN_RB1_DB_CTRL);
+
return 0;
}
@@ -1204,6 +1209,10 @@ static int vcn_v4_0_5_start(struct amdgpu_vcn_inst *vinst)
WREG32_SOC15(VCN, i, regVCN_RB_ENABLE, tmp);
fw_shared->sq.queue_mode &= ~(FW_QUEUE_RING_RESET | FW_QUEUE_DPG_HOLD_OFF);
+ /* Keeping one read-back to ensure all register writes are done, otherwise
+ * it may introduce race conditions */
+ RREG32_SOC15(VCN, i, regVCN_RB_ENABLE);
+
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
index d99d05f42f1d9..b90da3d3e1406 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
@@ -533,7 +533,8 @@ static void vcn_v5_0_0_mc_resume_dpg_mode(struct amdgpu_vcn_inst *vinst,
/* VCN global tiling registers */
WREG32_SOC24_DPG_MODE(inst_idx, SOC24_DPG_MODE_OFFSET(
- VCN, 0, regUVD_GFX10_ADDR_CONFIG), adev->gfx.config.gb_addr_config, 0, indirect);
+ VCN, inst_idx, regUVD_GFX10_ADDR_CONFIG),
+ adev->gfx.config.gb_addr_config, 0, indirect);
return;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
index 581d8629b9d95..e0e84ef7f5686 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
@@ -503,6 +503,52 @@ static void vcn_v5_0_1_enable_clock_gating(struct amdgpu_vcn_inst *vinst)
}
/**
+ * vcn_v5_0_1_pause_dpg_mode - VCN pause with dpg mode
+ *
+ * @vinst: VCN instance
+ * @new_state: pause state
+ *
+ * Pause dpg mode for VCN block
+ */
+static int vcn_v5_0_1_pause_dpg_mode(struct amdgpu_vcn_inst *vinst,
+ struct dpg_pause_state *new_state)
+{
+ struct amdgpu_device *adev = vinst->adev;
+ uint32_t reg_data = 0;
+ int vcn_inst;
+
+ vcn_inst = GET_INST(VCN, vinst->inst);
+
+ /* pause/unpause if state is changed */
+ if (vinst->pause_state.fw_based != new_state->fw_based) {
+ DRM_DEV_DEBUG(adev->dev, "dpg pause state changed %d -> %d %s\n",
+ vinst->pause_state.fw_based, new_state->fw_based,
+ new_state->fw_based ? "VCN_DPG_STATE__PAUSE" : "VCN_DPG_STATE__UNPAUSE");
+ reg_data = RREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE) &
+ (~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+
+ if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
+ /* pause DPG */
+ reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE, reg_data);
+
+ /* wait for ACK */
+ SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_DPG_PAUSE,
+ UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
+ UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
+ } else {
+ /* unpause DPG, no need to wait */
+ reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
+ WREG32_SOC15(VCN, vcn_inst, regUVD_DPG_PAUSE, reg_data);
+ }
+ vinst->pause_state.fw_based = new_state->fw_based;
+ }
+
+ return 0;
+}
+
+
+/**
* vcn_v5_0_1_start_dpg_mode - VCN start with dpg mode
*
* @vinst: VCN instance
@@ -518,6 +564,7 @@ static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
volatile struct amdgpu_vcn5_fw_shared *fw_shared =
adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
struct amdgpu_ring *ring;
+ struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__PAUSE};
int vcn_inst;
uint32_t tmp;
@@ -582,6 +629,9 @@ static int vcn_v5_0_1_start_dpg_mode(struct amdgpu_vcn_inst *vinst,
if (indirect)
amdgpu_vcn_psp_update_sram(adev, inst_idx, AMDGPU_UCODE_ID_VCN0_RAM);
+ /* Pause dpg */
+ vcn_v5_0_1_pause_dpg_mode(vinst, &state);
+
ring = &adev->vcn.inst[inst_idx].ring_enc[0];
WREG32_SOC15(VCN, vcn_inst, regUVD_RB_BASE_LO, lower_32_bits(ring->gpu_addr));
@@ -775,9 +825,13 @@ static void vcn_v5_0_1_stop_dpg_mode(struct amdgpu_vcn_inst *vinst)
int inst_idx = vinst->inst;
uint32_t tmp;
int vcn_inst;
+ struct dpg_pause_state state = {.fw_based = VCN_DPG_STATE__UNPAUSE};
vcn_inst = GET_INST(VCN, inst_idx);
+ /* Unpause dpg */
+ vcn_v5_0_1_pause_dpg_mode(vinst, &state);
+
/* Wait for power status to be 1 */
SOC15_WAIT_ON_RREG(VCN, vcn_inst, regUVD_POWER_STATUS, 1,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c
index 86d8bc10d90ab..9b3510e531127 100644
--- a/drivers/gpu/drm/amd/amdgpu/vi.c
+++ b/drivers/gpu/drm/amd/amdgpu/vi.c
@@ -239,6 +239,13 @@ static const struct amdgpu_video_codec_info cz_video_codecs_decode_array[] =
.max_pixels_per_frame = 4096 * 4096,
.max_level = 186,
},
+ {
+ .codec_type = AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG,
+ .max_width = 4096,
+ .max_height = 4096,
+ .max_pixels_per_frame = 4096 * 4096,
+ .max_level = 0,
+ },
};
static const struct amdgpu_video_codecs cz_video_codecs_decode =