Add ring reset callbacks for gfx and compute.
v2: fix gfx handling
v3: wait for KIQ to complete
Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 91 ++++++++++++++++++++++++++
1 file changed, 91 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 32c0cc52861c..3d0446337751 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -9454,6 +9454,95 @@ static void gfx_v10_0_emit_mem_sync(struct amdgpu_ring *ring)
amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */
}
+static int gfx_v10_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid)
+{
+ struct amdgpu_device *adev = ring->adev;
+ struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+ struct amdgpu_ring *kiq_ring = &kiq->ring;
+ unsigned long flags;
+ u32 tmp;
+ u64 addr;
+ int r;
+
+ if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+ return -EINVAL;
+
+ spin_lock_irqsave(&kiq->ring_lock, flags);
+
+ if (amdgpu_ring_alloc(kiq_ring, 5 + 7 + 7 + kiq->pmf->map_queues_size)) {
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+ return -ENOMEM;
+ }
+
+ addr = amdgpu_bo_gpu_offset(ring->mqd_obj) +
+ offsetof(struct v10_gfx_mqd, cp_gfx_hqd_active);
+ tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+ if (ring->pipe == 0)
+ tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE0_QUEUES, 1 << ring->queue);
+ else
+ tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE1_QUEUES, 1 << ring->queue);
+
+ gfx_v10_0_ring_emit_wreg(kiq_ring,
+ SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), tmp);
+ gfx_v10_0_wait_reg_mem(kiq_ring, 0, 1, 0,
+ lower_32_bits(addr), upper_32_bits(addr),
+ 0, 1, 0x20);
+ gfx_v10_0_ring_emit_reg_wait(kiq_ring,
+ SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), 0, 0xffffffff);
+ kiq->pmf->kiq_map_queues(kiq_ring, ring);
+ amdgpu_ring_commit(kiq_ring);
+
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+ r = amdgpu_ring_test_ring(kiq_ring);
+ if (r)
+ return r;
+
+ /* reset the ring */
+ ring->wptr = 0;
+ *ring->wptr_cpu_addr = 0;
+ amdgpu_ring_clear_ring(ring);
+
+ return amdgpu_ring_test_ring(ring);
+}
+
+static int gfx_v10_0_reset_kcq(struct amdgpu_ring *ring,
+ unsigned int vmid)
+{
+ struct amdgpu_device *adev = ring->adev;
+ struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+ struct amdgpu_ring *kiq_ring = &kiq->ring;
+ unsigned long flags;
+ int r;
+
+ if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+ return -EINVAL;
+
+ spin_lock_irqsave(&kiq->ring_lock, flags);
+
+ if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) {
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+ return -ENOMEM;
+ }
+
+ kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES,
+ 0, 0);
+ amdgpu_ring_commit(kiq_ring);
+
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+ r = amdgpu_ring_test_ring(kiq_ring);
+ if (r)
+ return r;
+
+ /* reset the ring */
+ ring->wptr = 0;
+ *ring->wptr_cpu_addr = 0;
+ amdgpu_ring_clear_ring(ring);
+
+ return amdgpu_ring_test_ring(ring);
+}
+
static void gfx_v10_ip_print(void *handle, struct drm_printer *p)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@@ -9657,6 +9746,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = {
.emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait,
.soft_recovery = gfx_v10_0_ring_soft_recovery,
.emit_mem_sync = gfx_v10_0_emit_mem_sync,
+ .reset = gfx_v10_0_reset_kgq,
};
static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
@@ -9693,6 +9783,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = {
.emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait,
.soft_recovery = gfx_v10_0_ring_soft_recovery,
.emit_mem_sync = gfx_v10_0_emit_mem_sync,
+ .reset = gfx_v10_0_reset_kcq,
};
static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = {
--
2.45.2
