[PATCH v1 6/6] drm/msm/a6xx: Support per-instance pagetables

[Jordan Crouse]

Add support for per-instance pagetables for a6xx targets. Add support
to handle split pagetables and create a new instance if the needed
IOMMU support exists and insert the necessary PM4 commands to trigger
a pagetable switch at the beginning of a user command.

Signed-off-by: Jordan Crouse <jcrouse at codeaurora.org>
---

 drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 89 +++++++++++++++++++++++++++++++++++
 1 file changed, 89 insertions(+)

diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
index 9bec603c..e1a257e 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
@@ -12,6 +12,62 @@
 
 #define GPU_PAS_ID 13
 
+static void a6xx_set_pagetable(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
+		struct msm_file_private *ctx)
+{
+	u64 ttbr;
+	u32 asid;
+
+	if (!msm_iommu_get_ptinfo(ctx->aspace->mmu, &ttbr, &asid))
+		return;
+
+	ttbr = ttbr | ((u64) asid) << 48;
+
+	/* Turn off protected mode */
+	OUT_PKT7(ring, CP_SET_PROTECTED_MODE, 1);
+	OUT_RING(ring, 0);
+
+	/* Turn on APIV mode to access critical regions */
+	OUT_PKT4(ring, REG_A6XX_CP_MISC_CNTL, 1);
+	OUT_RING(ring, 1);
+
+	/* Make sure the ME is synchronized before staring the update */
+	OUT_PKT7(ring, CP_WAIT_FOR_ME, 0);
+
+	/* Execute the table update */
+	OUT_PKT7(ring, CP_SMMU_TABLE_UPDATE, 4);
+	OUT_RING(ring, lower_32_bits(ttbr));
+	OUT_RING(ring, upper_32_bits(ttbr));
+	/* CONTEXTIDR is currently unused */
+	OUT_RING(ring, 0);
+	/* CONTEXTBANK is currently unused */
+	OUT_RING(ring, 0);
+
+	/*
+	 * Write the new TTBR0 to the preemption records - this will be used to
+	 * reload the pagetable if the current ring gets preempted out.
+	 */
+	OUT_PKT7(ring, CP_MEM_WRITE, 4);
+	OUT_RING(ring, lower_32_bits(rbmemptr(ring, ttbr0)));
+	OUT_RING(ring, upper_32_bits(rbmemptr(ring, ttbr0)));
+	OUT_RING(ring, lower_32_bits(ttbr));
+	OUT_RING(ring, upper_32_bits(ttbr));
+
+	/* Invalidate the draw state so we start off fresh */
+	OUT_PKT7(ring, CP_SET_DRAW_STATE, 3);
+	OUT_RING(ring, 0x40000);
+	OUT_RING(ring, 1);
+	OUT_RING(ring, 0);
+
+	/* Turn off APRIV */
+	OUT_PKT4(ring, REG_A6XX_CP_MISC_CNTL, 1);
+	OUT_RING(ring, 0);
+
+	/* Turn off protected mode */
+	OUT_PKT7(ring, CP_SET_PROTECTED_MODE, 1);
+	OUT_RING(ring, 1);
+}
+
 static inline bool _a6xx_check_idle(struct msm_gpu *gpu)
 {
 	struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
@@ -89,6 +145,8 @@ static void a6xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit,
 	struct msm_ringbuffer *ring = submit->ring;
 	unsigned int i;
 
+	a6xx_set_pagetable(gpu, ring, ctx);
+
 	get_stats_counter(ring, REG_A6XX_RBBM_PERFCTR_CP_0_LO,
 		rbmemptr_stats(ring, index, cpcycles_start));
 
@@ -878,6 +936,36 @@ static unsigned long a6xx_gpu_busy(struct msm_gpu *gpu)
 	return (unsigned long)busy_time;
 }
 
+static struct msm_gem_address_space*
+a6xx_create_instance_space(struct msm_gpu *gpu)
+{
+	struct msm_gem_address_space *aspace;
+	struct iommu_domain *iommu;
+	struct msm_mmu *mmu;
+
+	if (!iommu_dev_has_feature(&gpu->pdev->dev, IOMMU_DEV_FEAT_AUX))
+		return gpu->aspace;
+
+	iommu = iommu_domain_alloc(&platform_bus_type);
+	if (!iommu)
+		return gpu->aspace;
+
+	mmu = msm_iommu_new_instance(&gpu->pdev->dev, iommu);
+	if (IS_ERR(mmu)) {
+		iommu_domain_free(iommu);
+		return gpu->aspace;
+	}
+
+	aspace = msm_gem_address_space_create(mmu, "gpu",
+		0x100000000ULL, 0x1ffffffffULL);
+	if (IS_ERR(aspace)) {
+		mmu->funcs->destroy(mmu);
+		return gpu->aspace;
+	}
+
+	return aspace;
+}
+
 static struct msm_gem_address_space *
 a6xx_create_address_space(struct msm_gpu *gpu, struct platform_device *pdev)
 {
@@ -951,6 +1039,7 @@ static const struct adreno_gpu_funcs funcs = {
 		.gpu_state_put = a6xx_gpu_state_put,
 #endif
 		.create_address_space = a6xx_create_address_space,
+		.create_instance_space = a6xx_create_instance_space,
 	},
 	.get_timestamp = a6xx_get_timestamp,
 };
-- 
2.7.4