summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu.h4
-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.c38
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c2
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c72
-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_gmc.c19
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_object.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c4
-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/mes_v11_0.c4
-rw-r--r--drivers/gpu/drm/amd/amdgpu/mes_v12_0.c21
-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
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_topology.c9
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c101
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c17
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c19
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c2
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c34
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc.c10
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c20
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c28
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c14
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c15
-rw-r--r--drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c5
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c6
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_dpms.c13
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c54
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c6
-rw-r--r--drivers/gpu/drm/amd/include/kgd_pp_interface.h1
-rw-r--r--drivers/gpu/drm/amd/pm/amdgpu_dpm.c19
-rw-r--r--drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h2
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c36
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h1
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c2
61 files changed, 581 insertions, 241 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 6d83ccfa42ee..c3641331d4de 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -353,7 +353,6 @@ enum amdgpu_kiq_irq {
AMDGPU_CP_KIQ_IRQ_DRIVER0 = 0,
AMDGPU_CP_KIQ_IRQ_LAST
};
-#define SRIOV_USEC_TIMEOUT 1200000 /* wait 12 * 100ms for SRIOV */
#define MAX_KIQ_REG_WAIT 5000 /* in usecs, 5ms */
#define MAX_KIQ_REG_BAILOUT_INTERVAL 5 /* in msecs, 5ms */
#define MAX_KIQ_REG_TRY 1000
@@ -1124,6 +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;
@@ -1614,11 +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 b7f8f2ff143d..707e131f89d2 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 cfdf558b48b6..02138aa55793 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 a30111d2c3ea..f8b3e04d71ed 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]);
@@ -3643,6 +3644,13 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
adev, adev->ip_blocks[i].version->type))
continue;
+ /* Since we skip suspend for S0i3, we need to cancel the delayed
+ * idle work here as the suspend callback never gets called.
+ */
+ if (adev->in_s0ix &&
+ adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX &&
+ amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(10, 0, 0))
+ cancel_delayed_work_sync(&adev->gfx.idle_work);
/* skip suspend of gfx/mes and psp for S0ix
* gfx is in gfxoff state, so on resume it will exit gfxoff just
* like at runtime. PSP is also part of the always on hardware
@@ -4851,7 +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);
}
@@ -4900,28 +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;
}
@@ -4942,15 +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);
@@ -4961,15 +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_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index dc2713ec95a5..9e738fae2b74 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -120,6 +120,8 @@ MODULE_FIRMWARE("amdgpu/vega20_ip_discovery.bin");
MODULE_FIRMWARE("amdgpu/raven_ip_discovery.bin");
MODULE_FIRMWARE("amdgpu/raven2_ip_discovery.bin");
MODULE_FIRMWARE("amdgpu/picasso_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/arcturus_ip_discovery.bin");
+MODULE_FIRMWARE("amdgpu/aldebaran_ip_discovery.bin");
#define mmIP_DISCOVERY_VERSION 0x16A00
#define mmRCC_CONFIG_MEMSIZE 0xde3
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index 9f627caedc3f..44e120f9f764 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);
@@ -75,11 +100,35 @@ static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf,
*/
static int amdgpu_dma_buf_pin(struct dma_buf_attachment *attach)
{
- struct drm_gem_object *obj = attach->dmabuf->priv;
- struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
+ struct dma_buf *dmabuf = attach->dmabuf;
+ struct amdgpu_bo *bo = gem_to_amdgpu_bo(dmabuf->priv);
+ u32 domains = bo->allowed_domains;
+
+ dma_resv_assert_held(dmabuf->resv);
+
+ /* 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.
+ */
+ 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;
+ }
- /* pin buffer into GTT */
- return amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
+ 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);
}
/**
@@ -134,9 +183,6 @@ static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach,
r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);
if (r)
return ERR_PTR(r);
-
- } else if (bo->tbo.resource->mem_type != TTM_PL_TT) {
- return ERR_PTR(-EBUSY);
}
switch (bo->tbo.resource->mem_type) {
@@ -153,6 +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);
@@ -184,7 +235,7 @@ static void amdgpu_dma_buf_unmap(struct dma_buf_attachment *attach,
struct sg_table *sgt,
enum dma_data_direction dir)
{
- if (sgt->sgl->page_link) {
+ if (sg_page(sgt->sgl)) {
dma_unmap_sgtable(attach->dev, sgt, dir, 0);
sg_free_table(sgt);
kfree(sgt);
@@ -459,6 +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 26bf896f1444..72c807f5822e 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 72af5e5a894a..cf2df7790077 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_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 464625282872..ecb74ccf1d90 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -699,12 +699,10 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid,
uint32_t flush_type, bool all_hub,
uint32_t inst)
{
- u32 usec_timeout = amdgpu_sriov_vf(adev) ? SRIOV_USEC_TIMEOUT :
- adev->usec_timeout;
struct amdgpu_ring *ring = &adev->gfx.kiq[inst].ring;
struct amdgpu_kiq *kiq = &adev->gfx.kiq[inst];
unsigned int ndw;
- int r;
+ int r, cnt = 0;
uint32_t seq;
/*
@@ -761,10 +759,21 @@ int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid,
amdgpu_ring_commit(ring);
spin_unlock(&adev->gfx.kiq[inst].ring_lock);
- if (amdgpu_fence_wait_polling(ring, seq, usec_timeout) < 1) {
+
+ r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+
+ might_sleep();
+ while (r < 1 && cnt++ < MAX_KIQ_REG_TRY &&
+ !amdgpu_reset_pending(adev->reset_domain)) {
+ msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
+ r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+ }
+
+ if (cnt > MAX_KIQ_REG_TRY) {
dev_err(adev->dev, "timeout waiting for kiq fence\n");
r = -ETIME;
- }
+ } else
+ r = 0;
}
error_unlock_reset:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 80cd6f5273db..0b9987781f76 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -163,8 +163,8 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
* When GTT is just an alternative to VRAM make sure that we
* only use it as fallback and still try to fill up VRAM first.
*/
- if (domain & abo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM &&
- !(adev->flags & AMD_IS_APU))
+ if (abo->tbo.resource && !(adev->flags & AMD_IS_APU) &&
+ domain & abo->preferred_domains & AMDGPU_GEM_DOMAIN_VRAM)
places[c].flags |= TTM_PL_FLAG_FALLBACK;
c++;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
index cdcdae7f71ce..83adf81defc7 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/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index 6da8994e0469..2d7f82e98df9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -24,6 +24,7 @@
#include <linux/dma-mapping.h>
#include <drm/ttm/ttm_range_manager.h>
+#include <drm/drm_drv.h>
#include "amdgpu.h"
#include "amdgpu_vm.h"
@@ -907,6 +908,9 @@ int amdgpu_vram_mgr_init(struct amdgpu_device *adev)
struct ttm_resource_manager *man = &mgr->manager;
int err;
+ man->cg = drmm_cgroup_register_region(adev_to_drm(adev), "vram", adev->gmc.real_vram_size);
+ if (IS_ERR(man->cg))
+ return PTR_ERR(man->cg);
ttm_resource_manager_init(man, &adev->mman.bdev,
adev->gmc.real_vram_size);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index a63ce747863f..23e6a05359c2 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 d57db42f9536..2a5c2a1ae3c7 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 e7b58e470292..62a257a4a3e9 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 95d894a231fc..809b3a882d0d 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 ad099f136f84..fec9a007533a 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 05c026d0b0d9..c6f290704d47 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 783e0c3b86b4..5effe8327d29 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 f1dc13b3ab38..cbbeadeb53f7 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 43195c079748..086a647308df 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 fcb8dd2876bc..40940b4ab400 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 a88d25a06c29..6ccd31c8bc69 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 49f7eb4fbd11..2c9239a22f39 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/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index e65916ada23b..ef9538fbbf53 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -894,6 +894,10 @@ static void mes_v11_0_get_fw_version(struct amdgpu_device *adev)
{
int pipe;
+ /* return early if we have already fetched these */
+ if (adev->mes.sched_version && adev->mes.kiq_version)
+ return;
+
/* get MES scheduler/KIQ versions */
mutex_lock(&adev->srbm_mutex);
diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
index 183dd3346da5..e6ab617b9a40 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c
@@ -1392,17 +1392,20 @@ static int mes_v12_0_queue_init(struct amdgpu_device *adev,
mes_v12_0_queue_init_register(ring);
}
- /* get MES scheduler/KIQ versions */
- mutex_lock(&adev->srbm_mutex);
- soc21_grbm_select(adev, 3, pipe, 0, 0);
+ if (((pipe == AMDGPU_MES_SCHED_PIPE) && !adev->mes.sched_version) ||
+ ((pipe == AMDGPU_MES_KIQ_PIPE) && !adev->mes.kiq_version)) {
+ /* get MES scheduler/KIQ versions */
+ mutex_lock(&adev->srbm_mutex);
+ soc21_grbm_select(adev, 3, pipe, 0, 0);
- if (pipe == AMDGPU_MES_SCHED_PIPE)
- adev->mes.sched_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
- else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq)
- adev->mes.kiq_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+ if (pipe == AMDGPU_MES_SCHED_PIPE)
+ adev->mes.sched_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
+ else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq)
+ adev->mes.kiq_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO);
- soc21_grbm_select(adev, 0, 0, 0, 0);
- mutex_unlock(&adev->srbm_mutex);
+ soc21_grbm_select(adev, 0, 0, 0, 0);
+ mutex_unlock(&adev->srbm_mutex);
+ }
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c
index 2ece3ae75ec1..bed5ef4d8788 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 bb5dfc410a66..215543575f47 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 cc621064610f..afdf8ce3b4c5 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 7c49c3f3c388..256288c6cd78 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 8e7a36f26e9c..b8d835c9e17e 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 d716510b8dd6..3eec1b8feaee 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 22ae1939476f..0b19f0ab4480 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 c6f6392c1c20..1f777c125b00 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 3e176b4b7c69..012f6ea928ec 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 ba603b2246e2..f11df9c2ec13 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 d99d05f42f1d..b90da3d3e140 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 581d8629b9d9..e0e84ef7f568 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 86d8bc10d90a..9b3510e53112 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 =
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index e477d7509646..9bbee484d57c 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -1983,9 +1983,6 @@ static void kfd_topology_set_capabilities(struct kfd_topology_device *dev)
if (kfd_dbg_has_ttmps_always_setup(dev->gpu))
dev->node_props.debug_prop |= HSA_DBG_DISPATCH_INFO_ALWAYS_VALID;
- if (dev->gpu->adev->sdma.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)
- dev->node_props.capability2 |= HSA_CAP2_PER_SDMA_QUEUE_RESET_SUPPORTED;
-
if (KFD_GC_VERSION(dev->gpu) < IP_VERSION(10, 0, 0)) {
if (KFD_GC_VERSION(dev->gpu) == IP_VERSION(9, 4, 3) ||
KFD_GC_VERSION(dev->gpu) == IP_VERSION(9, 4, 4))
@@ -2001,7 +1998,11 @@ static void kfd_topology_set_capabilities(struct kfd_topology_device *dev)
dev->node_props.capability |=
HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED;
- dev->node_props.capability |= HSA_CAP_PER_QUEUE_RESET_SUPPORTED;
+ if (!amdgpu_sriov_vf(dev->gpu->adev))
+ dev->node_props.capability |= HSA_CAP_PER_QUEUE_RESET_SUPPORTED;
+
+ if (dev->gpu->adev->sdma.supported_reset & AMDGPU_RESET_TYPE_PER_QUEUE)
+ dev->node_props.capability2 |= HSA_CAP2_PER_SDMA_QUEUE_RESET_SUPPORTED;
} else {
dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 |
HSA_DBG_WATCH_ADDR_MASK_HI_BIT;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index d0d8ad5368c3..cc01b9c68b47 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -372,6 +372,8 @@ get_crtc_by_otg_inst(struct amdgpu_device *adev,
static inline bool is_dc_timing_adjust_needed(struct dm_crtc_state *old_state,
struct dm_crtc_state *new_state)
{
+ if (new_state->stream->adjust.timing_adjust_pending)
+ return true;
if (new_state->freesync_config.state == VRR_STATE_ACTIVE_FIXED)
return true;
else if (amdgpu_dm_crtc_vrr_active(old_state) != amdgpu_dm_crtc_vrr_active(new_state))
@@ -673,15 +675,21 @@ static void dm_crtc_high_irq(void *interrupt_params)
spin_lock_irqsave(&adev_to_drm(adev)->event_lock, flags);
if (acrtc->dm_irq_params.stream &&
- acrtc->dm_irq_params.vrr_params.supported &&
- acrtc->dm_irq_params.freesync_config.state ==
- VRR_STATE_ACTIVE_VARIABLE) {
+ acrtc->dm_irq_params.vrr_params.supported) {
+ bool replay_en = acrtc->dm_irq_params.stream->link->replay_settings.replay_feature_enabled;
+ bool psr_en = acrtc->dm_irq_params.stream->link->psr_settings.psr_feature_enabled;
+ bool fs_active_var_en = acrtc->dm_irq_params.freesync_config.state == VRR_STATE_ACTIVE_VARIABLE;
+
mod_freesync_handle_v_update(adev->dm.freesync_module,
acrtc->dm_irq_params.stream,
&acrtc->dm_irq_params.vrr_params);
- dc_stream_adjust_vmin_vmax(adev->dm.dc, acrtc->dm_irq_params.stream,
- &acrtc->dm_irq_params.vrr_params.adjust);
+ /* update vmin_vmax only if freesync is enabled, or only if PSR and REPLAY are disabled */
+ if (fs_active_var_en || (!fs_active_var_en && !replay_en && !psr_en)) {
+ dc_stream_adjust_vmin_vmax(adev->dm.dc,
+ acrtc->dm_irq_params.stream,
+ &acrtc->dm_irq_params.vrr_params.adjust);
+ }
}
/*
@@ -1726,9 +1734,30 @@ static const struct dmi_system_id dmi_quirk_table[] = {
.callback = edp0_on_dp1_callback,
.matches = {
DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 645 14 inch G11 Notebook PC"),
+ },
+ },
+ {
+ .callback = edp0_on_dp1_callback,
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "HP"),
DMI_MATCH(DMI_PRODUCT_NAME, "HP EliteBook 665 16 inch G11 Notebook PC"),
},
},
+ {
+ .callback = edp0_on_dp1_callback,
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 445 14 inch G11 Notebook PC"),
+ },
+ },
+ {
+ .callback = edp0_on_dp1_callback,
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "HP"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "HP ProBook 465 16 inch G11 Notebook PC"),
+ },
+ },
{}
/* TODO: refactor this from a fixed table to a dynamic option */
};
@@ -1899,26 +1928,6 @@ static enum dmub_ips_disable_type dm_get_default_ips_mode(
switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) {
case IP_VERSION(3, 5, 0):
case IP_VERSION(3, 6, 0):
- /*
- * On DCN35 systems with Z8 enabled, it's possible for IPS2 + Z8 to
- * cause a hard hang. A fix exists for newer PMFW.
- *
- * As a workaround, for non-fixed PMFW, force IPS1+RCG as the deepest
- * IPS state in all cases, except for s0ix and all displays off (DPMS),
- * where IPS2 is allowed.
- *
- * When checking pmfw version, use the major and minor only.
- */
- if ((adev->pm.fw_version & 0x00FFFF00) < 0x005D6300)
- ret = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
- else if (amdgpu_ip_version(adev, GC_HWIP, 0) > IP_VERSION(11, 5, 0))
- /*
- * Other ASICs with DCN35 that have residency issues with
- * IPS2 in idle.
- * We want them to use IPS2 only in display off cases.
- */
- ret = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
- break;
case IP_VERSION(3, 5, 1):
ret = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF;
break;
@@ -3334,16 +3343,16 @@ static void dm_gpureset_commit_state(struct dc_state *dc_state,
for (k = 0; k < dc_state->stream_count; k++) {
bundle->stream_update.stream = dc_state->streams[k];
- for (m = 0; m < dc_state->stream_status->plane_count; m++) {
+ for (m = 0; m < dc_state->stream_status[k].plane_count; m++) {
bundle->surface_updates[m].surface =
- dc_state->stream_status->plane_states[m];
+ dc_state->stream_status[k].plane_states[m];
bundle->surface_updates[m].surface->force_full_update =
true;
}
update_planes_and_stream_adapter(dm->dc,
UPDATE_TYPE_FULL,
- dc_state->stream_status->plane_count,
+ dc_state->stream_status[k].plane_count,
dc_state->streams[k],
&bundle->stream_update,
bundle->surface_updates);
@@ -6500,12 +6509,12 @@ decide_crtc_timing_for_drm_display_mode(struct drm_display_mode *drm_mode,
const struct drm_display_mode *native_mode,
bool scale_enabled)
{
- if (scale_enabled) {
- copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
- } else if (native_mode->clock == drm_mode->clock &&
- native_mode->htotal == drm_mode->htotal &&
- native_mode->vtotal == drm_mode->vtotal) {
- copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
+ if (scale_enabled || (
+ native_mode->clock == drm_mode->clock &&
+ native_mode->htotal == drm_mode->htotal &&
+ native_mode->vtotal == drm_mode->vtotal)) {
+ if (native_mode->crtc_clock)
+ copy_crtc_timing_for_drm_display_mode(native_mode, drm_mode);
} else {
/* no scaling nor amdgpu inserted, no need to patch */
}
@@ -11022,6 +11031,9 @@ static bool should_reset_plane(struct drm_atomic_state *state,
state->allow_modeset)
return true;
+ if (amdgpu_in_reset(adev) && state->allow_modeset)
+ return true;
+
/* Exit early if we know that we're adding or removing the plane. */
if (old_plane_state->crtc != new_plane_state->crtc)
return true;
@@ -12739,7 +12751,7 @@ int amdgpu_dm_process_dmub_aux_transfer_sync(
* Transient states before tunneling is enabled could
* lead to this error. We can ignore this for now.
*/
- if (p_notify->result != AUX_RET_ERROR_PROTOCOL_ERROR) {
+ if (p_notify->result == AUX_RET_ERROR_PROTOCOL_ERROR) {
DRM_WARN("DPIA AUX failed on 0x%x(%d), error %d\n",
payload->address, payload->length,
p_notify->result);
@@ -12748,22 +12760,15 @@ int amdgpu_dm_process_dmub_aux_transfer_sync(
goto out;
}
+ payload->reply[0] = adev->dm.dmub_notify->aux_reply.command & 0xF;
+ if (adev->dm.dmub_notify->aux_reply.command & 0xF0)
+ /* The reply is stored in the top nibble of the command. */
+ payload->reply[0] = (adev->dm.dmub_notify->aux_reply.command >> 4) & 0xF;
- payload->reply[0] = adev->dm.dmub_notify->aux_reply.command;
- if (!payload->write && p_notify->aux_reply.length &&
- (payload->reply[0] == AUX_TRANSACTION_REPLY_AUX_ACK)) {
-
- if (payload->length != p_notify->aux_reply.length) {
- DRM_WARN("invalid read length %d from DPIA AUX 0x%x(%d)!\n",
- p_notify->aux_reply.length,
- payload->address, payload->length);
- *operation_result = AUX_RET_ERROR_INVALID_REPLY;
- goto out;
- }
-
+ /*write req may receive a byte indicating partially written number as well*/
+ if (p_notify->aux_reply.length)
memcpy(payload->data, p_notify->aux_reply.data,
p_notify->aux_reply.length);
- }
/* success */
ret = p_notify->aux_reply.length;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c
index 36a830a7440f..e8bdd7f0c460 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c
@@ -113,6 +113,7 @@ bool amdgpu_dm_crtc_vrr_active(const struct dm_crtc_state *dm_state)
*
* Panel Replay and PSR SU
* - Enable when:
+ * - VRR is disabled
* - vblank counter is disabled
* - entry is allowed: usermode demonstrates an adequate number of fast
* commits)
@@ -131,19 +132,20 @@ static void amdgpu_dm_crtc_set_panel_sr_feature(
bool is_sr_active = (link->replay_settings.replay_allow_active ||
link->psr_settings.psr_allow_active);
bool is_crc_window_active = false;
+ bool vrr_active = amdgpu_dm_crtc_vrr_active_irq(vblank_work->acrtc);
#ifdef CONFIG_DRM_AMD_SECURE_DISPLAY
is_crc_window_active =
amdgpu_dm_crc_window_is_activated(&vblank_work->acrtc->base);
#endif
- if (link->replay_settings.replay_feature_enabled &&
+ if (link->replay_settings.replay_feature_enabled && !vrr_active &&
allow_sr_entry && !is_sr_active && !is_crc_window_active) {
amdgpu_dm_replay_enable(vblank_work->stream, true);
} else if (vblank_enabled) {
if (link->psr_settings.psr_version < DC_PSR_VERSION_SU_1 && is_sr_active)
amdgpu_dm_psr_disable(vblank_work->stream, false);
- } else if (link->psr_settings.psr_feature_enabled &&
+ } else if (link->psr_settings.psr_feature_enabled && !vrr_active &&
allow_sr_entry && !is_sr_active && !is_crc_window_active) {
struct amdgpu_dm_connector *aconn =
@@ -244,6 +246,8 @@ static void amdgpu_dm_crtc_vblank_control_worker(struct work_struct *work)
struct vblank_control_work *vblank_work =
container_of(work, struct vblank_control_work, work);
struct amdgpu_display_manager *dm = vblank_work->dm;
+ struct amdgpu_device *adev = drm_to_adev(dm->ddev);
+ int r;
mutex_lock(&dm->dc_lock);
@@ -271,8 +275,15 @@ static void amdgpu_dm_crtc_vblank_control_worker(struct work_struct *work)
vblank_work->acrtc->dm_irq_params.allow_sr_entry);
}
- if (dm->active_vblank_irq_count == 0)
+ if (dm->active_vblank_irq_count == 0) {
+ r = amdgpu_dpm_pause_power_profile(adev, true);
+ if (r)
+ dev_warn(adev->dev, "failed to set default power profile mode\n");
dc_allow_idle_optimizations(dm->dc, true);
+ r = amdgpu_dpm_pause_power_profile(adev, false);
+ if (r)
+ dev_warn(adev->dev, "failed to restore the power profile mode\n");
+ }
mutex_unlock(&dm->dc_lock);
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
index 5198a079b463..8f22ad966543 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_hdcp.c
@@ -173,6 +173,9 @@ void hdcp_update_display(struct hdcp_workqueue *hdcp_work,
unsigned int conn_index = aconnector->base.index;
guard(mutex)(&hdcp_w->mutex);
+ drm_connector_get(&aconnector->base);
+ if (hdcp_w->aconnector[conn_index])
+ drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
hdcp_w->aconnector[conn_index] = aconnector;
memset(&link_adjust, 0, sizeof(link_adjust));
@@ -220,7 +223,6 @@ static void hdcp_remove_display(struct hdcp_workqueue *hdcp_work,
unsigned int conn_index = aconnector->base.index;
guard(mutex)(&hdcp_w->mutex);
- hdcp_w->aconnector[conn_index] = aconnector;
/* the removal of display will invoke auth reset -> hdcp destroy and
* we'd expect the Content Protection (CP) property changed back to
@@ -236,7 +238,10 @@ static void hdcp_remove_display(struct hdcp_workqueue *hdcp_work,
}
mod_hdcp_remove_display(&hdcp_w->hdcp, aconnector->base.index, &hdcp_w->output);
-
+ if (hdcp_w->aconnector[conn_index]) {
+ drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+ hdcp_w->aconnector[conn_index] = NULL;
+ }
process_output(hdcp_w);
}
@@ -254,6 +259,10 @@ void hdcp_reset_display(struct hdcp_workqueue *hdcp_work, unsigned int link_inde
for (conn_index = 0; conn_index < AMDGPU_DM_MAX_DISPLAY_INDEX; conn_index++) {
hdcp_w->encryption_status[conn_index] =
MOD_HDCP_ENCRYPTION_STATUS_HDCP_OFF;
+ if (hdcp_w->aconnector[conn_index]) {
+ drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+ hdcp_w->aconnector[conn_index] = NULL;
+ }
}
process_output(hdcp_w);
@@ -488,6 +497,7 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
struct hdcp_workqueue *hdcp_work = handle;
struct amdgpu_dm_connector *aconnector = config->dm_stream_ctx;
int link_index = aconnector->dc_link->link_index;
+ unsigned int conn_index = aconnector->base.index;
struct mod_hdcp_display *display = &hdcp_work[link_index].display;
struct mod_hdcp_link *link = &hdcp_work[link_index].link;
struct hdcp_workqueue *hdcp_w = &hdcp_work[link_index];
@@ -544,7 +554,10 @@ static void update_config(void *handle, struct cp_psp_stream_config *config)
guard(mutex)(&hdcp_w->mutex);
mod_hdcp_add_display(&hdcp_w->hdcp, link, display, &hdcp_w->output);
-
+ drm_connector_get(&aconnector->base);
+ if (hdcp_w->aconnector[conn_index])
+ drm_connector_put(&hdcp_w->aconnector[conn_index]->base);
+ hdcp_w->aconnector[conn_index] = aconnector;
process_output(hdcp_w);
}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
index 2cd35392e2da..1395a748d726 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c
@@ -918,7 +918,7 @@ dm_helpers_probe_acpi_edid(void *data, u8 *buf, unsigned int block, size_t len)
{
struct drm_connector *connector = data;
struct acpi_device *acpidev = ACPI_COMPANION(connector->dev->dev);
- unsigned char start = block * EDID_LENGTH;
+ unsigned short start = block * EDID_LENGTH;
struct edid *edid;
int r;
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 7ceedf626d23..5cdbc86ef8f5 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -51,6 +51,9 @@
#define PEAK_FACTOR_X1000 1006
+/*
+ * This function handles both native AUX and I2C-Over-AUX transactions.
+ */
static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
struct drm_dp_aux_msg *msg)
{
@@ -59,6 +62,7 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
enum aux_return_code_type operation_result;
struct amdgpu_device *adev;
struct ddc_service *ddc;
+ uint8_t copy[16];
if (WARN_ON(msg->size > 16))
return -E2BIG;
@@ -74,6 +78,11 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
(msg->request & DP_AUX_I2C_WRITE_STATUS_UPDATE) != 0;
payload.defer_delay = 0;
+ if (payload.write) {
+ memcpy(copy, msg->buffer, msg->size);
+ payload.data = copy;
+ }
+
result = dc_link_aux_transfer_raw(TO_DM_AUX(aux)->ddc_service, &payload,
&operation_result);
@@ -87,15 +96,25 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
if (adev->dm.aux_hpd_discon_quirk) {
if (msg->address == DP_SIDEBAND_MSG_DOWN_REQ_BASE &&
operation_result == AUX_RET_ERROR_HPD_DISCON) {
- result = 0;
+ result = msg->size;
operation_result = AUX_RET_SUCCESS;
}
}
- if (payload.write && result >= 0)
- result = msg->size;
+ /*
+ * result equals to 0 includes the cases of AUX_DEFER/I2C_DEFER
+ */
+ if (payload.write && result >= 0) {
+ if (result) {
+ /*one byte indicating partially written bytes*/
+ drm_dbg_dp(adev_to_drm(adev), "amdgpu: AUX partially written\n");
+ result = payload.data[0];
+ } else if (!payload.reply[0])
+ /*I2C_ACK|AUX_ACK*/
+ result = msg->size;
+ }
- if (result < 0)
+ if (result < 0) {
switch (operation_result) {
case AUX_RET_SUCCESS:
break;
@@ -114,6 +133,13 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux,
break;
}
+ drm_dbg_dp(adev_to_drm(adev), "amdgpu: DP AUX transfer fail:%d\n", operation_result);
+ }
+
+ if (payload.reply[0])
+ drm_dbg_dp(adev_to_drm(adev), "amdgpu: AUX reply command not ACK: 0x%02x.",
+ payload.reply[0]);
+
return result;
}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 28d1353f403d..ba4ce8a63158 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -439,9 +439,12 @@ bool dc_stream_adjust_vmin_vmax(struct dc *dc,
* Don't adjust DRR while there's bandwidth optimizations pending to
* avoid conflicting with firmware updates.
*/
- if (dc->ctx->dce_version > DCE_VERSION_MAX)
- if (dc->optimized_required || dc->wm_optimized_required)
+ if (dc->ctx->dce_version > DCE_VERSION_MAX) {
+ if (dc->optimized_required || dc->wm_optimized_required) {
+ stream->adjust.timing_adjust_pending = true;
return false;
+ }
+ }
dc_exit_ips_for_hw_access(dc);
@@ -3168,7 +3171,8 @@ static void copy_stream_update_to_stream(struct dc *dc,
if (update->crtc_timing_adjust) {
if (stream->adjust.v_total_min != update->crtc_timing_adjust->v_total_min ||
- stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max)
+ stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max ||
+ stream->adjust.timing_adjust_pending)
update->crtc_timing_adjust->timing_adjust_pending = true;
stream->adjust = *update->crtc_timing_adjust;
update->crtc_timing_adjust->timing_adjust_pending = false;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
index 0c8ec30ea672..731fbd4bc600 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
@@ -910,7 +910,7 @@ static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dm
}
//TODO : Could be possibly moved to a common helper layer.
-static bool dml21_wrapper_get_plane_id(const struct dc_state *context, const struct dc_plane_state *plane, unsigned int *plane_id)
+static bool dml21_wrapper_get_plane_id(const struct dc_state *context, unsigned int stream_id, const struct dc_plane_state *plane, unsigned int *plane_id)
{
int i, j;
@@ -918,10 +918,12 @@ static bool dml21_wrapper_get_plane_id(const struct dc_state *context, const str
return false;
for (i = 0; i < context->stream_count; i++) {
- for (j = 0; j < context->stream_status[i].plane_count; j++) {
- if (context->stream_status[i].plane_states[j] == plane) {
- *plane_id = (i << 16) | j;
- return true;
+ if (context->streams[i]->stream_id == stream_id) {
+ for (j = 0; j < context->stream_status[i].plane_count; j++) {
+ if (context->stream_status[i].plane_states[j] == plane) {
+ *plane_id = (i << 16) | j;
+ return true;
+ }
}
}
}
@@ -944,14 +946,14 @@ static unsigned int map_stream_to_dml21_display_cfg(const struct dml2_context *d
return location;
}
-static unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx,
+static unsigned int map_plane_to_dml21_display_cfg(const struct dml2_context *dml_ctx, unsigned int stream_id,
const struct dc_plane_state *plane, const struct dc_state *context)
{
unsigned int plane_id;
int i = 0;
int location = -1;
- if (!dml21_wrapper_get_plane_id(context, plane, &plane_id)) {
+ if (!dml21_wrapper_get_plane_id(context, stream_id, plane, &plane_id)) {
ASSERT(false);
return -1;
}
@@ -1037,7 +1039,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s
dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location;
} else {
for (plane_index = 0; plane_index < context->stream_status[stream_index].plane_count; plane_index++) {
- disp_cfg_plane_location = map_plane_to_dml21_display_cfg(dml_ctx, context->stream_status[stream_index].plane_states[plane_index], context);
+ disp_cfg_plane_location = map_plane_to_dml21_display_cfg(dml_ctx, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], context);
if (disp_cfg_plane_location < 0)
disp_cfg_plane_location = dml_dispcfg->num_planes++;
@@ -1048,7 +1050,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s
populate_dml21_plane_config_from_plane_state(dml_ctx, &dml_dispcfg->plane_descriptors[disp_cfg_plane_location], context->stream_status[stream_index].plane_states[plane_index], context, stream_index);
dml_dispcfg->plane_descriptors[disp_cfg_plane_location].stream_index = disp_cfg_stream_location;
- if (dml21_wrapper_get_plane_id(context, context->stream_status[stream_index].plane_states[plane_index], &dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location]))
+ if (dml21_wrapper_get_plane_id(context, context->streams[stream_index]->stream_id, context->stream_status[stream_index].plane_states[plane_index], &dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id[disp_cfg_plane_location]))
dml_ctx->v21.dml_to_dc_pipe_mapping.disp_cfg_to_plane_id_valid[disp_cfg_plane_location] = true;
/* apply forced pstate policy */
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
index be54f0e696ce..ed6584535e89 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
@@ -2,6 +2,7 @@
//
// Copyright 2024 Advanced Micro Devices, Inc.
+#include <linux/vmalloc.h>
#include "dml2_internal_types.h"
#include "dml_top.h"
@@ -13,11 +14,11 @@
static bool dml21_allocate_memory(struct dml2_context **dml_ctx)
{
- *dml_ctx = kzalloc(sizeof(struct dml2_context), GFP_KERNEL);
+ *dml_ctx = vzalloc(sizeof(struct dml2_context));
if (!(*dml_ctx))
return false;
- (*dml_ctx)->v21.dml_init.dml2_instance = kzalloc(sizeof(struct dml2_instance), GFP_KERNEL);
+ (*dml_ctx)->v21.dml_init.dml2_instance = vzalloc(sizeof(struct dml2_instance));
if (!((*dml_ctx)->v21.dml_init.dml2_instance))
return false;
@@ -27,7 +28,7 @@ static bool dml21_allocate_memory(struct dml2_context **dml_ctx)
(*dml_ctx)->v21.mode_support.display_config = &(*dml_ctx)->v21.display_config;
(*dml_ctx)->v21.mode_programming.display_config = (*dml_ctx)->v21.mode_support.display_config;
- (*dml_ctx)->v21.mode_programming.programming = kzalloc(sizeof(struct dml2_display_cfg_programming), GFP_KERNEL);
+ (*dml_ctx)->v21.mode_programming.programming = vzalloc(sizeof(struct dml2_display_cfg_programming));
if (!((*dml_ctx)->v21.mode_programming.programming))
return false;
@@ -86,6 +87,8 @@ static void dml21_init(const struct dc *in_dc, struct dml2_context **dml_ctx, co
/* Store configuration options */
(*dml_ctx)->config = *config;
+ DC_FP_START();
+
/*Initialize SOCBB and DCNIP params */
dml21_initialize_soc_bb_params(&(*dml_ctx)->v21.dml_init, config, in_dc);
dml21_initialize_ip_params(&(*dml_ctx)->v21.dml_init, config, in_dc);
@@ -96,6 +99,8 @@ static void dml21_init(const struct dc *in_dc, struct dml2_context **dml_ctx, co
/*Initialize DML21 instance */
dml2_initialize_instance(&(*dml_ctx)->v21.dml_init);
+
+ DC_FP_END();
}
bool dml21_create(const struct dc *in_dc, struct dml2_context **dml_ctx, const struct dml2_configuration_options *config)
@@ -111,8 +116,8 @@ bool dml21_create(const struct dc *in_dc, struct dml2_context **dml_ctx, const s
void dml21_destroy(struct dml2_context *dml2)
{
- kfree(dml2->v21.dml_init.dml2_instance);
- kfree(dml2->v21.mode_programming.programming);
+ vfree(dml2->v21.dml_init.dml2_instance);
+ vfree(dml2->v21.mode_programming.programming);
}
static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_state *context, struct resource_context *out_new_hw_state,
@@ -229,7 +234,9 @@ static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_s
if (!result)
return false;
+ DC_FP_START();
result = dml2_build_mode_programming(mode_programming);
+ DC_FP_END();
if (!result)
return false;
@@ -272,7 +279,9 @@ static bool dml21_check_mode_support(const struct dc *in_dc, struct dc_state *co
mode_support->dml2_instance = dml_init->dml2_instance;
dml21_map_dc_state_into_dml_display_cfg(in_dc, context, dml_ctx);
dml_ctx->v21.mode_programming.dml2_instance->scratch.build_mode_programming_locals.mode_programming_params.programming = dml_ctx->v21.mode_programming.programming;
+ DC_FP_START();
is_supported = dml2_check_mode_supported(mode_support);
+ DC_FP_END();
if (!is_supported)
return false;
@@ -284,10 +293,11 @@ bool dml21_validate(const struct dc *in_dc, struct dc_state *context, struct dml
bool out = false;
/* Use dml_validate_only for fast_validate path */
- if (fast_validate) {
+ if (fast_validate)
out = dml21_check_mode_support(in_dc, context, dml_ctx);
- } else
+ else
out = dml21_mode_check_and_programming(in_dc, context, dml_ctx);
+
return out;
}
@@ -426,8 +436,12 @@ void dml21_copy(struct dml2_context *dst_dml_ctx,
dst_dml_ctx->v21.mode_programming.programming = dst_dml2_programming;
+ DC_FP_START();
+
/* need to initialize copied instance for internal references to be correct */
dml2_initialize_instance(&dst_dml_ctx->v21.dml_init);
+
+ DC_FP_END();
}
bool dml21_create_copy(struct dml2_context **dst_dml_ctx,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
index 2061d43b92e1..ab6baf269801 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
@@ -973,7 +973,9 @@ static void populate_dml_surface_cfg_from_plane_state(enum dml_project_id dml2_p
}
}
-static void get_scaler_data_for_plane(const struct dc_plane_state *in, struct dc_state *context, struct scaler_data *out)
+static struct scaler_data *get_scaler_data_for_plane(
+ const struct dc_plane_state *in,
+ struct dc_state *context)
{
int i;
struct pipe_ctx *temp_pipe = &context->res_ctx.temp_pipe;
@@ -994,7 +996,7 @@ static void get_scaler_data_for_plane(const struct dc_plane_state *in, struct dc
}
ASSERT(i < MAX_PIPES);
- memcpy(out, &temp_pipe->plane_res.scl_data, sizeof(*out));
+ return &temp_pipe->plane_res.scl_data;
}
static void populate_dummy_dml_plane_cfg(struct dml_plane_cfg_st *out, unsigned int location,
@@ -1057,11 +1059,7 @@ static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out
const struct dc_plane_state *in, struct dc_state *context,
const struct soc_bounding_box_st *soc)
{
- struct scaler_data *scaler_data = kzalloc(sizeof(*scaler_data), GFP_KERNEL);
- if (!scaler_data)
- return;
-
- get_scaler_data_for_plane(in, context, scaler_data);
+ struct scaler_data *scaler_data = get_scaler_data_for_plane(in, context);
out->CursorBPP[location] = dml_cur_32bit;
out->CursorWidth[location] = 256;
@@ -1126,8 +1124,6 @@ static void populate_dml_plane_cfg_from_plane_state(struct dml_plane_cfg_st *out
out->DynamicMetadataTransmittedBytes[location] = 0;
out->NumberOfCursors[location] = 1;
-
- kfree(scaler_data);
}
static unsigned int map_stream_to_dml_display_cfg(const struct dml2_context *dml2,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
index 939ee0708bd2..e89571874185 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
@@ -24,6 +24,8 @@
*
*/
+#include <linux/vmalloc.h>
+
#include "display_mode_core.h"
#include "dml2_internal_types.h"
#include "dml2_utils.h"
@@ -732,17 +734,22 @@ bool dml2_validate(const struct dc *in_dc, struct dc_state *context, struct dml2
return out;
}
+ DC_FP_START();
+
/* Use dml_validate_only for fast_validate path */
if (fast_validate)
out = dml2_validate_only(context);
else
out = dml2_validate_and_build_resource(in_dc, context);
+
+ DC_FP_END();
+
return out;
}
static inline struct dml2_context *dml2_allocate_memory(void)
{
- return (struct dml2_context *) kzalloc(sizeof(struct dml2_context), GFP_KERNEL);
+ return (struct dml2_context *) vzalloc(sizeof(struct dml2_context));
}
static void dml2_init(const struct dc *in_dc, const struct dml2_configuration_options *config, struct dml2_context **dml2)
@@ -779,11 +786,15 @@ static void dml2_init(const struct dc *in_dc, const struct dml2_configuration_op
break;
}
+ DC_FP_START();
+
initialize_dml2_ip_params(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.ip);
initialize_dml2_soc_bbox(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.soc);
initialize_dml2_soc_states(*dml2, in_dc, &(*dml2)->v20.dml_core_ctx.soc, &(*dml2)->v20.dml_core_ctx.states);
+
+ DC_FP_END();
}
bool dml2_create(const struct dc *in_dc, const struct dml2_configuration_options *config, struct dml2_context **dml2)
@@ -812,7 +823,7 @@ void dml2_destroy(struct dml2_context *dml2)
if (dml2->architecture == dml2_architecture_21)
dml21_destroy(dml2);
- kfree(dml2);
+ vfree(dml2);
}
void dml2_extract_dram_and_fclk_change_support(struct dml2_context *dml2,
diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c
index 1236e0f9a256..712aff7e17f7 100644
--- a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c
+++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_cm.c
@@ -120,10 +120,11 @@ void dpp401_set_cursor_attributes(
enum dc_cursor_color_format color_format = cursor_attributes->color_format;
int cur_rom_en = 0;
- // DCN4 should always do Cursor degamma for Cursor Color modes
if (color_format == CURSOR_MODE_COLOR_PRE_MULTIPLIED_ALPHA ||
color_format == CURSOR_MODE_COLOR_UN_PRE_MULTIPLIED_ALPHA) {
- cur_rom_en = 1;
+ if (cursor_attributes->attribute_flags.bits.ENABLE_CURSOR_DEGAMMA) {
+ cur_rom_en = 1;
+ }
}
REG_UPDATE_3(CURSOR0_CONTROL,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
index 5489f3d431f6..3af6a3402b89 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
@@ -1980,9 +1980,9 @@ void dcn401_program_pipe(
dc->res_pool->hubbub, pipe_ctx->plane_res.hubp->inst, pipe_ctx->hubp_regs.det_size);
}
- if (pipe_ctx->update_flags.raw ||
- (pipe_ctx->plane_state && pipe_ctx->plane_state->update_flags.raw) ||
- pipe_ctx->stream->update_flags.raw)
+ if (pipe_ctx->plane_state && (pipe_ctx->update_flags.raw ||
+ pipe_ctx->plane_state->update_flags.raw ||
+ pipe_ctx->stream->update_flags.raw))
dc->hwss.update_dchubp_dpp(dc, pipe_ctx, context);
if (pipe_ctx->plane_state && (pipe_ctx->update_flags.bits.enable ||
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
index 268626e73c54..53c961f86d43 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
@@ -148,6 +148,7 @@ void link_blank_dp_stream(struct dc_link *link, bool hw_init)
void link_set_all_streams_dpms_off_for_link(struct dc_link *link)
{
struct pipe_ctx *pipes[MAX_PIPES];
+ struct dc_stream_state *streams[MAX_PIPES];
struct dc_state *state = link->dc->current_state;
uint8_t count;
int i;
@@ -160,10 +161,18 @@ void link_set_all_streams_dpms_off_for_link(struct dc_link *link)
link_get_master_pipes_with_dpms_on(link, state, &count, pipes);
+ /* The subsequent call to dc_commit_updates_for_stream for a full update
+ * will release the current state and swap to a new state. Releasing the
+ * current state results in the stream pointers in the pipe_ctx structs
+ * to be zero'd. Hence, cache all streams prior to dc_commit_updates_for_stream.
+ */
+ for (i = 0; i < count; i++)
+ streams[i] = pipes[i]->stream;
+
for (i = 0; i < count; i++) {
- stream_update.stream = pipes[i]->stream;
+ stream_update.stream = streams[i];
dc_commit_updates_for_stream(link->ctx->dc, NULL, 0,
- pipes[i]->stream, &stream_update,
+ streams[i], &stream_update,
state);
}
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
index 34d2e097ca2e..5a5d48fadbf2 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
@@ -35,6 +35,17 @@
#define DC_LOGGER \
link->ctx->logger
+static void get_default_8b_10b_lttpr_aux_rd_interval(
+ union training_aux_rd_interval *training_rd_interval)
+{
+ /* LTTPR are required to program DPCD 0000Eh to 0x4 (16ms) upon AUX
+ * read reply to this register. Since old sinks with DPCD rev 1.1
+ * and earlier may not support this register, assume the mandatory
+ * value is programmed by the LTTPR to avoid AUX timeout issues.
+ */
+ training_rd_interval->raw = 0x4;
+}
+
static int32_t get_cr_training_aux_rd_interval(struct dc_link *link,
const struct dc_link_settings *link_settings,
enum lttpr_mode lttpr_mode)
@@ -43,17 +54,22 @@ static int32_t get_cr_training_aux_rd_interval(struct dc_link *link,
uint32_t wait_in_micro_secs = 100;
memset(&training_rd_interval, 0, sizeof(training_rd_interval));
- if (link_dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING &&
- link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
- core_link_read_dpcd(
- link,
- DP_TRAINING_AUX_RD_INTERVAL,
- (uint8_t *)&training_rd_interval,
- sizeof(training_rd_interval));
- if (lttpr_mode != LTTPR_MODE_NON_TRANSPARENT)
- wait_in_micro_secs = 400;
- if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
- wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;
+ if (link_dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING) {
+ if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12)
+ core_link_read_dpcd(
+ link,
+ DP_TRAINING_AUX_RD_INTERVAL,
+ (uint8_t *)&training_rd_interval,
+ sizeof(training_rd_interval));
+ else if (dp_is_lttpr_present(link))
+ get_default_8b_10b_lttpr_aux_rd_interval(&training_rd_interval);
+
+ if (training_rd_interval.raw != 0) {
+ if (lttpr_mode != LTTPR_MODE_NON_TRANSPARENT)
+ wait_in_micro_secs = 400;
+ if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
+ wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;
+ }
}
return wait_in_micro_secs;
}
@@ -71,13 +87,15 @@ static uint32_t get_eq_training_aux_rd_interval(
DP_128B132B_TRAINING_AUX_RD_INTERVAL,
(uint8_t *)&training_rd_interval,
sizeof(training_rd_interval));
- } else if (link_dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING &&
- link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) {
- core_link_read_dpcd(
- link,
- DP_TRAINING_AUX_RD_INTERVAL,
- (uint8_t *)&training_rd_interval,
- sizeof(training_rd_interval));
+ } else if (link_dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING) {
+ if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12)
+ core_link_read_dpcd(
+ link,
+ DP_TRAINING_AUX_RD_INTERVAL,
+ (uint8_t *)&training_rd_interval,
+ sizeof(training_rd_interval));
+ else if (dp_is_lttpr_present(link))
+ get_default_8b_10b_lttpr_aux_rd_interval(&training_rd_interval);
}
switch (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL) {
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
index 2a59cc61ed8c..944650cb13de 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
@@ -2114,8 +2114,6 @@ static bool dcn32_resource_construct(
#define REG_STRUCT dccg_regs
dccg_regs_init();
- DC_FP_START();
-
ctx->dc_bios->regs = &bios_regs;
pool->base.res_cap = &res_cap_dcn32;
@@ -2501,14 +2499,10 @@ static bool dcn32_resource_construct(
if (ASICREV_IS_GC_11_0_3(dc->ctx->asic_id.hw_internal_rev) && (dc->config.sdpif_request_limit_words_per_umc == 0))
dc->config.sdpif_request_limit_words_per_umc = 16;
- DC_FP_END();
-
return true;
create_fail:
- DC_FP_END();
-
dcn32_resource_destruct(pool);
return false;
diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
index 2a9606118d89..21dc956b5f35 100644
--- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h
+++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
@@ -429,6 +429,7 @@ struct amd_pm_funcs {
int (*set_pp_table)(void *handle, const char *buf, size_t size);
void (*debugfs_print_current_performance_level)(void *handle, struct seq_file *m);
int (*switch_power_profile)(void *handle, enum PP_SMC_POWER_PROFILE type, bool en);
+ int (*pause_power_profile)(void *handle, bool pause);
/* export to amdgpu */
struct amd_vce_state *(*get_vce_clock_state)(void *handle, u32 idx);
int (*dispatch_tasks)(void *handle, enum amd_pp_task task_id,
diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
index 81e9b443ca0a..3533d43ed1e7 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
@@ -349,6 +349,25 @@ int amdgpu_dpm_switch_power_profile(struct amdgpu_device *adev,
return ret;
}
+int amdgpu_dpm_pause_power_profile(struct amdgpu_device *adev,
+ bool pause)
+{
+ const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+ int ret = 0;
+
+ if (amdgpu_sriov_vf(adev))
+ return 0;
+
+ if (pp_funcs && pp_funcs->pause_power_profile) {
+ mutex_lock(&adev->pm.mutex);
+ ret = pp_funcs->pause_power_profile(
+ adev->powerplay.pp_handle, pause);
+ mutex_unlock(&adev->pm.mutex);
+ }
+
+ return ret;
+}
+
int amdgpu_dpm_set_xgmi_pstate(struct amdgpu_device *adev,
uint32_t pstate)
{
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
index f93d287dbf13..4c0f7ad14816 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
@@ -410,6 +410,8 @@ int amdgpu_dpm_set_xgmi_pstate(struct amdgpu_device *adev,
int amdgpu_dpm_switch_power_profile(struct amdgpu_device *adev,
enum PP_SMC_POWER_PROFILE type,
bool en);
+int amdgpu_dpm_pause_power_profile(struct amdgpu_device *adev,
+ bool pause);
int amdgpu_dpm_baco_reset(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index 033c3229b555..46cce1d2aaf3 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -2398,7 +2398,11 @@ static int smu_switch_power_profile(void *handle,
smu_power_profile_mode_get(smu, type);
else
smu_power_profile_mode_put(smu, type);
- ret = smu_bump_power_profile_mode(smu, NULL, 0);
+ /* don't switch the active workload when paused */
+ if (smu->pause_workload)
+ ret = 0;
+ else
+ ret = smu_bump_power_profile_mode(smu, NULL, 0);
if (ret) {
if (enable)
smu_power_profile_mode_put(smu, type);
@@ -2411,6 +2415,35 @@ static int smu_switch_power_profile(void *handle,
return 0;
}
+static int smu_pause_power_profile(void *handle,
+ bool pause)
+{
+ struct smu_context *smu = handle;
+ struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
+ u32 workload_mask = 1 << PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;
+ int ret;
+
+ if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
+ return -EOPNOTSUPP;
+
+ if (smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL &&
+ smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM) {
+ smu->pause_workload = pause;
+
+ /* force to bootup default profile */
+ if (smu->pause_workload && smu->ppt_funcs->set_power_profile_mode)
+ ret = smu->ppt_funcs->set_power_profile_mode(smu,
+ workload_mask,
+ NULL,
+ 0);
+ else
+ ret = smu_bump_power_profile_mode(smu, NULL, 0);
+ return ret;
+ }
+
+ return 0;
+}
+
static enum amd_dpm_forced_level smu_get_performance_level(void *handle)
{
struct smu_context *smu = handle;
@@ -3733,6 +3766,7 @@ static const struct amd_pm_funcs swsmu_pm_funcs = {
.get_pp_table = smu_sys_get_pp_table,
.set_pp_table = smu_sys_set_pp_table,
.switch_power_profile = smu_switch_power_profile,
+ .pause_power_profile = smu_pause_power_profile,
/* export to amdgpu */
.dispatch_tasks = smu_handle_dpm_task,
.load_firmware = smu_load_microcode,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
index 3ba169639f54..dd6d0e7aa242 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
@@ -558,6 +558,7 @@ struct smu_context {
/* asic agnostic workload mask */
uint32_t workload_mask;
+ bool pause_workload;
/* default/user workload preference */
uint32_t power_profile_mode;
uint32_t workload_refcount[PP_SMC_POWER_PROFILE_COUNT];
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
index 78391d8f35a9..25fabf336a64 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
@@ -1204,7 +1204,7 @@ int smu_v11_0_set_fan_speed_rpm(struct smu_context *smu,
uint32_t crystal_clock_freq = 2500;
uint32_t tach_period;
- if (speed == 0)
+ if (!speed || speed > UINT_MAX/8)
return -EINVAL;
/*
* To prevent from possible overheat, some ASICs may have requirement