[PATCH 2/2] drm/amdgpu: schedule fault recover work to another CPU

[Philip Yang]

Flooding GPU vm fault interrupt may hold CPU and delay the scheduled
recover work to handle vm fault, schedule recover work to another CPU of
same NUMA node.

Signed-off-by: Philip Yang <Philip.Yang at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/vega10_ih.c | 13 +++++++++++--
 drivers/gpu/drm/amd/amdgpu/vega20_ih.c | 13 +++++++++++--
 2 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
index cdd599a08125..3a47107737a3 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c
@@ -435,12 +435,21 @@ static int vega10_ih_self_irq(struct amdgpu_device *adev,
 			      struct amdgpu_irq_src *source,
 			      struct amdgpu_iv_entry *entry)
 {
+	int cpu, new_cpu;
+
+	cpu = new_cpu = smp_processor_id();
+	do {
+		new_cpu = cpumask_next(new_cpu, cpu_online_mask) % nr_cpu_ids;
+		if (cpu_to_node(new_cpu) == numa_node_id())
+			break;
+	} while (cpu != new_cpu);
+
 	switch (entry->ring_id) {
 	case 1:
-		schedule_work(&adev->irq.ih1_work);
+		schedule_work_on(new_cpu, &adev->irq.ih1_work);
 		break;
 	case 2:
-		schedule_work(&adev->irq.ih2_work);
+		schedule_work_on(new_cpu, &adev->irq.ih2_work);
 		break;
 	default: break;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
index 3b4eb8285943..a9465f0d8fbe 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
@@ -487,12 +487,21 @@ static int vega20_ih_self_irq(struct amdgpu_device *adev,
 			      struct amdgpu_irq_src *source,
 			      struct amdgpu_iv_entry *entry)
 {
+	int cpu, new_cpu;
+
+	cpu = new_cpu = smp_processor_id();
+	do {
+		new_cpu = cpumask_next(new_cpu, cpu_online_mask) % nr_cpu_ids;
+		if (cpu_to_node(new_cpu) == numa_node_id())
+			break;
+	} while (cpu != new_cpu);
+
 	switch (entry->ring_id) {
 	case 1:
-		schedule_work(&adev->irq.ih1_work);
+		schedule_work_on(new_cpu, &adev->irq.ih1_work);
 		break;
 	case 2:
-		schedule_work(&adev->irq.ih2_work);
+		schedule_work_on(new_cpu, &adev->irq.ih2_work);
 		break;
 	default: break;
 	}
-- 
2.35.1