[PATCH v5 1/5] iommu/arm-smmu: add NVIDIA implementation for dual ARM MMU-500 usage

[Krishna Reddy]

NVIDIA's Tegra194 soc uses two ARM MMU-500s together to interleave
IOVA accesses across them.
Add NVIDIA implementation for dual ARM MMU-500s and add new compatible
string for Tegra194 soc.

Signed-off-by: Krishna Reddy <vdumpa@xxxxxxxxxx>
---
 MAINTAINERS                     |   2 +
 drivers/iommu/Makefile          |   2 +-
 drivers/iommu/arm-smmu-impl.c   |   3 +
 drivers/iommu/arm-smmu-nvidia.c | 161 ++++++++++++++++++++++++++++++++
 drivers/iommu/arm-smmu.h        |   1 +
 5 files changed, 168 insertions(+), 1 deletion(-)
 create mode 100644 drivers/iommu/arm-smmu-nvidia.c

diff --git a/MAINTAINERS b/MAINTAINERS
index ecc0749810b0..0d8c966ecf17 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -16560,9 +16560,11 @@ F:	drivers/i2c/busses/i2c-tegra.c
 
 TEGRA IOMMU DRIVERS
 M:	Thierry Reding <thierry.reding@xxxxxxxxx>
+R:	Krishna Reddy <vdumpa@xxxxxxxxxx>
 L:	linux-tegra@xxxxxxxxxxxxxxx
 S:	Supported
 F:	drivers/iommu/tegra*
+F:	drivers/iommu/arm-smmu-nvidia.c
 
 TEGRA KBC DRIVER
 M:	Laxman Dewangan <ldewangan@xxxxxxxxxx>
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index 57cf4ba5e27c..35542df00da7 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -15,7 +15,7 @@ obj-$(CONFIG_AMD_IOMMU) += amd_iommu.o amd_iommu_init.o amd_iommu_quirks.o
 obj-$(CONFIG_AMD_IOMMU_DEBUGFS) += amd_iommu_debugfs.o
 obj-$(CONFIG_AMD_IOMMU_V2) += amd_iommu_v2.o
 obj-$(CONFIG_ARM_SMMU) += arm_smmu.o
-arm_smmu-objs += arm-smmu.o arm-smmu-impl.o arm-smmu-qcom.o
+arm_smmu-objs += arm-smmu.o arm-smmu-impl.o arm-smmu-qcom.o arm-smmu-nvidia.o
 obj-$(CONFIG_ARM_SMMU_V3) += arm-smmu-v3.o
 obj-$(CONFIG_DMAR_TABLE) += dmar.o
 obj-$(CONFIG_INTEL_IOMMU) += intel-iommu.o intel-pasid.o
diff --git a/drivers/iommu/arm-smmu-impl.c b/drivers/iommu/arm-smmu-impl.c
index 74d97a886e93..dcdd513323aa 100644
--- a/drivers/iommu/arm-smmu-impl.c
+++ b/drivers/iommu/arm-smmu-impl.c
@@ -158,6 +158,9 @@ struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu)
 	 */
 	switch (smmu->model) {
 	case ARM_MMU500:
+		if (of_device_is_compatible(smmu->dev->of_node,
+					    "nvidia,tegra194-smmu-500"))
+			return nvidia_smmu_impl_init(smmu);
 		smmu->impl = &arm_mmu500_impl;
 		break;
 	case CAVIUM_SMMUV2:
diff --git a/drivers/iommu/arm-smmu-nvidia.c b/drivers/iommu/arm-smmu-nvidia.c
new file mode 100644
index 000000000000..dafc293a4521
--- /dev/null
+++ b/drivers/iommu/arm-smmu-nvidia.c
@@ -0,0 +1,161 @@
+// SPDX-License-Identifier: GPL-2.0-only
+// Nvidia ARM SMMU v2 implementation quirks
+// Copyright (C) 2019 NVIDIA CORPORATION.  All rights reserved.
+
+#define pr_fmt(fmt) "nvidia-smmu: " fmt
+
+#include <linux/bitfield.h>
+#include <linux/delay.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include "arm-smmu.h"
+
+/* Tegra194 has three ARM MMU-500 Instances.
+ * Two of them are used together for Interleaved IOVA accesses and
+ * used by Non-Isochronous Hw devices for SMMU translations.
+ * Third one is used for SMMU translations from Isochronous HW devices.
+ * It is possible to use this Implementation to program either
+ * all three or two of the instances identically as desired through
+ * DT node.
+ *
+ * Programming all the three instances identically comes with redundant tlb
+ * invalidations as all three never need to be tlb invalidated for a HW device.
+ *
+ * When Linux Kernel supports multiple SMMU devices, The SMMU device used for
+ * Isochornous HW devices should be added as a separate ARM MMU-500 device
+ * in DT and be programmed independently for efficient tlb invalidates.
+ *
+ */
+#define MAX_SMMU_INSTANCES 3
+
+#define TLB_LOOP_TIMEOUT		1000000	/* 1s! */
+#define TLB_SPIN_COUNT			10
+
+struct nvidia_smmu {
+	struct arm_smmu_device	smmu;
+	unsigned int		num_inst;
+	void __iomem		*bases[MAX_SMMU_INSTANCES];
+};
+
+#define to_nvidia_smmu(s) container_of(s, struct nvidia_smmu, smmu)
+
+#define nsmmu_page(smmu, inst, page) \
+	(((inst) ? to_nvidia_smmu(smmu)->bases[(inst)] : smmu->base) + \
+	((page) << smmu->pgshift))
+
+static u32 nsmmu_read_reg(struct arm_smmu_device *smmu,
+			      int page, int offset)
+{
+	return readl_relaxed(nsmmu_page(smmu, 0, page) + offset);
+}
+
+static void nsmmu_write_reg(struct arm_smmu_device *smmu,
+			    int page, int offset, u32 val)
+{
+	unsigned int i;
+
+	for (i = 0; i < to_nvidia_smmu(smmu)->num_inst; i++)
+		writel_relaxed(val, nsmmu_page(smmu, i, page) + offset);
+}
+
+static u64 nsmmu_read_reg64(struct arm_smmu_device *smmu,
+				int page, int offset)
+{
+	return readq_relaxed(nsmmu_page(smmu, 0, page) + offset);
+}
+
+static void nsmmu_write_reg64(struct arm_smmu_device *smmu,
+				  int page, int offset, u64 val)
+{
+	unsigned int i;
+
+	for (i = 0; i < to_nvidia_smmu(smmu)->num_inst; i++)
+		writeq_relaxed(val, nsmmu_page(smmu, i, page) + offset);
+}
+
+static void nsmmu_tlb_sync(struct arm_smmu_device *smmu, int page,
+			   int sync, int status)
+{
+	u32 reg;
+	unsigned int i;
+	unsigned int spin_cnt, delay;
+
+	arm_smmu_writel(smmu, page, sync, 0);
+
+	for (delay = 1; delay < TLB_LOOP_TIMEOUT; delay *= 2) {
+		for (spin_cnt = TLB_SPIN_COUNT; spin_cnt > 0; spin_cnt--) {
+			reg = 0;
+			for (i = 0; i < to_nvidia_smmu(smmu)->num_inst; i++) {
+				reg |= readl_relaxed(
+					nsmmu_page(smmu, i, page) + status);
+			}
+			if (!(reg & ARM_SMMU_sTLBGSTATUS_GSACTIVE))
+				return;
+			cpu_relax();
+		}
+		udelay(delay);
+	}
+	dev_err_ratelimited(smmu->dev,
+			    "TLB sync timed out -- SMMU may be deadlocked\n");
+}
+
+static int nsmmu_reset(struct arm_smmu_device *smmu)
+{
+	u32 reg;
+	unsigned int i;
+
+	for (i = 0; i < to_nvidia_smmu(smmu)->num_inst; i++) {
+		/* clear global FSR */
+		reg = readl_relaxed(nsmmu_page(smmu, i, ARM_SMMU_GR0) +
+				    ARM_SMMU_GR0_sGFSR);
+		writel_relaxed(reg, nsmmu_page(smmu, i, ARM_SMMU_GR0) +
+				    ARM_SMMU_GR0_sGFSR);
+	}
+
+	return 0;
+}
+
+static const struct arm_smmu_impl nvidia_smmu_impl = {
+	.read_reg = nsmmu_read_reg,
+	.write_reg = nsmmu_write_reg,
+	.read_reg64 = nsmmu_read_reg64,
+	.write_reg64 = nsmmu_write_reg64,
+	.reset = nsmmu_reset,
+	.tlb_sync = nsmmu_tlb_sync,
+};
+
+struct arm_smmu_device *nvidia_smmu_impl_init(struct arm_smmu_device *smmu)
+{
+	unsigned int i;
+	struct nvidia_smmu *nsmmu;
+	struct resource *res;
+	struct device *dev = smmu->dev;
+	struct platform_device *pdev = to_platform_device(smmu->dev);
+
+	nsmmu = devm_kzalloc(smmu->dev, sizeof(*nsmmu), GFP_KERNEL);
+	if (!nsmmu)
+		return ERR_PTR(-ENOMEM);
+
+	nsmmu->smmu = *smmu;
+	/* Instance 0 is ioremapped by arm-smmu.c */
+	nsmmu->num_inst = 1;
+
+	for (i = 1; i < MAX_SMMU_INSTANCES; i++) {
+		res = platform_get_resource(pdev, IORESOURCE_MEM, i);
+		if (!res)
+			break;
+		nsmmu->bases[i] = devm_ioremap_resource(dev, res);
+		if (IS_ERR(nsmmu->bases[i]))
+			return (struct arm_smmu_device *)nsmmu->bases[i];
+		nsmmu->num_inst++;
+	}
+
+	nsmmu->smmu.impl = &nvidia_smmu_impl;
+	devm_kfree(smmu->dev, smmu);
+	pr_info("NVIDIA ARM SMMU Implementation, Instances=%d\n",
+		nsmmu->num_inst);
+
+	return &nsmmu->smmu;
+}
diff --git a/drivers/iommu/arm-smmu.h b/drivers/iommu/arm-smmu.h
index 8d1cd54d82a6..67c3c6f5c49e 100644
--- a/drivers/iommu/arm-smmu.h
+++ b/drivers/iommu/arm-smmu.h
@@ -450,6 +450,7 @@ static inline void arm_smmu_writeq(struct arm_smmu_device *smmu, int page,
 
 struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu);
 struct arm_smmu_device *qcom_smmu_impl_init(struct arm_smmu_device *smmu);
+struct arm_smmu_device *nvidia_smmu_impl_init(struct arm_smmu_device *smmu);
 
 int arm_mmu500_reset(struct arm_smmu_device *smmu);
 
-- 
2.26.2