1 files changed, 55 insertions, 15 deletions

diff --git a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
index 6021f8d9efd1..7e82c41a85f1 100644
--- a/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a2xx_gpu.c
@@ -10,6 +10,48 @@ extern bool hang_debug;
 static void a2xx_dump(struct msm_gpu *gpu);
 static bool a2xx_idle(struct msm_gpu *gpu);
 
+static void a2xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit)
+{
+	struct msm_drm_private *priv = gpu->dev->dev_private;
+	struct msm_ringbuffer *ring = submit->ring;
+	unsigned int i;
+
+	for (i = 0; i < submit->nr_cmds; i++) {
+		switch (submit->cmd[i].type) {
+		case MSM_SUBMIT_CMD_IB_TARGET_BUF:
+			/* ignore IB-targets */
+			break;
+		case MSM_SUBMIT_CMD_CTX_RESTORE_BUF:
+			/* ignore if there has not been a ctx switch: */
+			if (priv->lastctx == submit->queue->ctx)
+				break;
+			fallthrough;
+		case MSM_SUBMIT_CMD_BUF:
+			OUT_PKT3(ring, CP_INDIRECT_BUFFER_PFD, 2);
+			OUT_RING(ring, lower_32_bits(submit->cmd[i].iova));
+			OUT_RING(ring, submit->cmd[i].size);
+			OUT_PKT2(ring);
+			break;
+		}
+	}
+
+	OUT_PKT0(ring, REG_AXXX_CP_SCRATCH_REG2, 1);
+	OUT_RING(ring, submit->seqno);
+
+	/* wait for idle before cache flush/interrupt */
+	OUT_PKT3(ring, CP_WAIT_FOR_IDLE, 1);
+	OUT_RING(ring, 0x00000000);
+
+	OUT_PKT3(ring, CP_EVENT_WRITE, 3);
+	OUT_RING(ring, CACHE_FLUSH_TS);
+	OUT_RING(ring, rbmemptr(ring, fence));
+	OUT_RING(ring, submit->seqno);
+	OUT_PKT3(ring, CP_INTERRUPT, 1);
+	OUT_RING(ring, 0x80000000);
+
+	adreno_flush(gpu, ring, REG_AXXX_CP_RB_WPTR);
+}
+
 static bool a2xx_me_init(struct msm_gpu *gpu)
 {
 	struct msm_ringbuffer *ring = gpu->rb[0];
@@ -53,7 +95,7 @@ static bool a2xx_me_init(struct msm_gpu *gpu)
 	OUT_PKT3(ring, CP_SET_PROTECTED_MODE, 1);
 	OUT_RING(ring, 1);
 
-	gpu->funcs->flush(gpu, ring);
+	adreno_flush(gpu, ring, REG_AXXX_CP_RB_WPTR);
 	return a2xx_idle(gpu);
 }
 
@@ -164,6 +206,11 @@ static int a2xx_hw_init(struct msm_gpu *gpu)
 	if (ret)
 		return ret;
 
+	gpu_write(gpu, REG_AXXX_CP_RB_CNTL,
+		MSM_GPU_RB_CNTL_DEFAULT | AXXX_CP_RB_CNTL_NO_UPDATE);
+
+	gpu_write(gpu, REG_AXXX_CP_RB_BASE, lower_32_bits(gpu->rb[0]->iova));
+
 	/* NOTE: PM4/micro-engine firmware registers look to be the same
 	 * for a2xx and a3xx.. we could possibly push that part down to
 	 * adreno_gpu base class.  Or push both PM4 and PFP but
@@ -416,16 +463,11 @@ a2xx_create_address_space(struct msm_gpu *gpu, struct platform_device *pdev)
 	return aspace;
 }
 
-/* Register offset defines for A2XX - copy of A3XX */
-static const unsigned int a2xx_register_offsets[REG_ADRENO_REGISTER_MAX] = {
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_BASE, REG_AXXX_CP_RB_BASE),
-	REG_ADRENO_SKIP(REG_ADRENO_CP_RB_BASE_HI),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR_ADDR, REG_AXXX_CP_RB_RPTR_ADDR),
-	REG_ADRENO_SKIP(REG_ADRENO_CP_RB_RPTR_ADDR_HI),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_RPTR, REG_AXXX_CP_RB_RPTR),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_WPTR, REG_AXXX_CP_RB_WPTR),
-	REG_ADRENO_DEFINE(REG_ADRENO_CP_RB_CNTL, REG_AXXX_CP_RB_CNTL),
-};
+static u32 a2xx_get_rptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring)
+{
+	ring->memptrs->rptr = gpu_read(gpu, REG_AXXX_CP_RB_RPTR);
+	return ring->memptrs->rptr;
+}
 
 static const struct adreno_gpu_funcs funcs = {
 	.base = {
@@ -434,8 +476,7 @@ static const struct adreno_gpu_funcs funcs = {
 		.pm_suspend = msm_gpu_pm_suspend,
 		.pm_resume = msm_gpu_pm_resume,
 		.recover = a2xx_recover,
-		.submit = adreno_submit,
-		.flush = adreno_flush,
+		.submit = a2xx_submit,
 		.active_ring = adreno_active_ring,
 		.irq = a2xx_irq,
 		.destroy = a2xx_destroy,
@@ -445,6 +486,7 @@ static const struct adreno_gpu_funcs funcs = {
 		.gpu_state_get = a2xx_gpu_state_get,
 		.gpu_state_put = adreno_gpu_state_put,
 		.create_address_space = a2xx_create_address_space,
+		.get_rptr = a2xx_get_rptr,
 	},
 };
 
@@ -486,8 +528,6 @@ struct msm_gpu *a2xx_gpu_init(struct drm_device *dev)
 	else
 		adreno_gpu->registers = a220_registers;
 
-	adreno_gpu->reg_offsets = a2xx_register_offsets;
-
 	ret = adreno_gpu_init(dev, pdev, adreno_gpu, &funcs, 1);
 	if (ret)
 		goto fail;