Skip to contentSkip to conversation

iommu/arm-smmu: Split pagetable support

PATCH RESEND Superseded Jordan Crouse 2 participants 7 messages Jul 8, 2019 – Jul 10, 2019

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover iommu/arm-smmu: Split pagetable support

(RESEND because I fat fingered a email address and I didn't want everybody to get a bunch of SMTP errors)

This series implements split pagetable support for arm-smmu-v2 devices. You have seen this code before as part of [1] but I split it apart from the other features to make it easier to review / merge.

This series adds a new format type to io-pgtable-arm to enable TTBR0 and TTBR1. Because of the way that the registers are mixed up it makes the most sense to enable them together instead of trying to create two different pagetables and merge them together later in the arm-smmu drive.

This will be used later by the drm/msm driver to enable split pagetables as part of the effort to implement per-context pagetables [2].

Thanks,
Jordan

[1] https://patchwork.freedesktop.org/series/57441/
[2] https://patchwork.freedesktop.org/patch/307616/?series=57441&rev=3

Jordan Crouse (3):
iommu: Add DOMAIN_ATTR_SPLIT_TABLES iommu/io-pgtable-arm: Add support for AARCH64 split pagetables iommu/arm-smmu: Add support for DOMAIN_ATTR_SPLIT_TABLES

 drivers/iommu/arm-smmu.c       |  16 ++-
 drivers/iommu/io-pgtable-arm.c | 261 +++++++++++++++++++++++++++++++++++++----
 drivers/iommu/io-pgtable.c     |   1 +
 include/linux/io-pgtable.h     |   2 +
 include/linux/iommu.h          |   1 +
 5 files changed, 256 insertions(+), 25 deletions(-)
Signature · 1 line
2.7.4
#1 lore raw

PATCH RESEND v2 1/3 iommu: Add DOMAIN_ATTR_SPLIT_TABLES

Add a new domain attribute to enable split pagetable support for devices devices that support it.

---

 include/linux/iommu.h | 1 +
 1 file changed, 1 insertion(+)
include/linux/iommu.h +1 -0
index fdc355c..b06db6c 100644
@@ -125,6 +125,7 @@ enum iommu_attr {
 	DOMAIN_ATTR_FSL_PAMUV1,
 	DOMAIN_ATTR_NESTING,	/* two stages of translation */
 	DOMAIN_ATTR_DMA_USE_FLUSH_QUEUE,
+	DOMAIN_ATTR_SPLIT_TABLES,
 	DOMAIN_ATTR_MAX,
 };
 
Signature · 1 line
2.7.4
#2 lore raw

PATCH RESEND v2 2/3 iommu/io-pgtable-arm: Add support for AARCH64 split pagetables

In reply to Jordan Crouse

Add a new sub-format ARM_64_LPAE_SPLIT_S1 to create and set up split pagetables (TTBR0 and TTBR1). The initialization function sets up the correct va_size and sign extension bits and programs the TCR registers. Split pagetable formats use their own own map/unmap wrappers to ensure that the correct pagetable is selected based on the incoming iova but most of the heavy lifting is common to the other formats.

---

 drivers/iommu/io-pgtable-arm.c | 261 +++++++++++++++++++++++++++++++++++++----
 drivers/iommu/io-pgtable.c     |   1 +
 include/linux/io-pgtable.h     |   2 +
 3 files changed, 240 insertions(+), 24 deletions(-)
drivers/iommu/io-pgtable-arm.c +237 -24
index 161a7d56..aec35e5 100644
@@ -118,7 +118,12 @@
 #define ARM_LPAE_TCR_TG0_64K		(1 << 14)
 #define ARM_LPAE_TCR_TG0_16K		(2 << 14)
 
+#define ARM_LPAE_TCR_TG1_4K		(0 << 30)
+#define ARM_LPAE_TCR_TG1_64K		(1 << 30)
+#define ARM_LPAE_TCR_TG1_16K		(2 << 30)
+
 #define ARM_LPAE_TCR_SH0_SHIFT		12
+#define ARM_LPAE_TCR_SH1_SHIFT		28
 #define ARM_LPAE_TCR_SH0_MASK		0x3
 #define ARM_LPAE_TCR_SH_NS		0
 #define ARM_LPAE_TCR_SH_OS		2
@@ -126,6 +131,8 @@
 
 #define ARM_LPAE_TCR_ORGN0_SHIFT	10
 #define ARM_LPAE_TCR_IRGN0_SHIFT	8
+#define ARM_LPAE_TCR_ORGN1_SHIFT	26
+#define ARM_LPAE_TCR_IRGN1_SHIFT	24
 #define ARM_LPAE_TCR_RGN_MASK		0x3
 #define ARM_LPAE_TCR_RGN_NC		0
 #define ARM_LPAE_TCR_RGN_WBWA		1
@@ -136,6 +143,7 @@
 #define ARM_LPAE_TCR_SL0_MASK		0x3
 
 #define ARM_LPAE_TCR_T0SZ_SHIFT		0
+#define ARM_LPAE_TCR_T1SZ_SHIFT		16
 #define ARM_LPAE_TCR_SZ_MASK		0xf
 
 #define ARM_LPAE_TCR_PS_SHIFT		16
@@ -152,6 +160,14 @@
 #define ARM_LPAE_TCR_PS_48_BIT		0x5ULL
 #define ARM_LPAE_TCR_PS_52_BIT		0x6ULL
 
+#define ARM_LPAE_TCR_SEP_SHIFT		47
+#define ARM_LPAE_TCR_SEP_31		(0x0ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_35		(0x1ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_39		(0x2ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_41		(0x3ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_43		(0x4ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_UPSTREAM	(0x7ULL << ARM_LPAE_TCR_SEP_SHIFT)
+
 #define ARM_LPAE_MAIR_ATTR_SHIFT(n)	((n) << 3)
 #define ARM_LPAE_MAIR_ATTR_MASK		0xff
 #define ARM_LPAE_MAIR_ATTR_DEVICE	0x04
@@ -179,11 +195,12 @@ struct arm_lpae_io_pgtable {
 	struct io_pgtable	iop;
 
 	int			levels;
+	u32			sep;
 	size_t			pgd_size;
 	unsigned long		pg_shift;
 	unsigned long		bits_per_level;
 
-	void			*pgd;
+	void			*pgd[2];
 };
 
 typedef u64 arm_lpae_iopte;
@@ -426,7 +443,8 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
 	arm_lpae_iopte pte;
 
 	if (data->iop.fmt == ARM_64_LPAE_S1 ||
-	    data->iop.fmt == ARM_32_LPAE_S1) {
+	    data->iop.fmt == ARM_32_LPAE_S1 ||
+	    data->iop.fmt == ARM_64_LPAE_SPLIT_S1) {
 		pte = ARM_LPAE_PTE_nG;
 		if (!(prot & IOMMU_WRITE) && (prot & IOMMU_READ))
 			pte |= ARM_LPAE_PTE_AP_RDONLY;
@@ -470,11 +488,10 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
 	return pte;
 }
 
-static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
-			phys_addr_t paddr, size_t size, int iommu_prot)
+static int _arm_lpae_map(struct arm_lpae_io_pgtable *data, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot,
+			arm_lpae_iopte *ptep)
 {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
 	int ret, lvl = ARM_LPAE_START_LVL(data);
 	arm_lpae_iopte prot;
 
@@ -497,12 +514,39 @@ static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
 	return ret;
 }
 
+static int arm_lpae_split_map(struct io_pgtable_ops *ops, unsigned long iova,
+		phys_addr_t paddr, size_t size, int iommu_prot)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, ptep);
+}
+
+static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, data->pgd[0]);
+}
+
 static void __arm_lpae_free_pgtable(struct arm_lpae_io_pgtable *data, int lvl,
 				    arm_lpae_iopte *ptep)
 {
 	arm_lpae_iopte *start, *end;
 	unsigned long table_size;
 
+	if (!ptep)
+		return;
+
 	if (lvl == ARM_LPAE_START_LVL(data))
 		table_size = data->pgd_size;
 	else
@@ -532,7 +576,8 @@ static void arm_lpae_free_pgtable(struct io_pgtable *iop)
 {
 	struct arm_lpae_io_pgtable *data = io_pgtable_to_data(iop);
 
-	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[0]);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[1]);
 	kfree(data);
 }
 
@@ -643,11 +688,28 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
 	return __arm_lpae_unmap(data, iova, size, lvl + 1, ptep);
 }
 
+static size_t arm_lpae_split_unmap(struct io_pgtable_ops *ops,
+				   unsigned long iova, size_t size)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+	int lvl = ARM_LPAE_START_LVL(data);
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
+}
+
 static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
 			     size_t size)
 {
 	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
+	arm_lpae_iopte *ptep = data->pgd[0];
 	int lvl = ARM_LPAE_START_LVL(data);
 
 	if (WARN_ON(iova >= (1ULL << data->iop.cfg.ias)))
@@ -656,11 +718,11 @@ static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
 	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
 }
 
-static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
-					 unsigned long iova)
+static phys_addr_t _arm_lpae_iova_to_phys(struct arm_lpae_io_pgtable *data,
+					  unsigned long iova,
+					  arm_lpae_iopte *ptep)
 {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte pte, *ptep = data->pgd;
+	arm_lpae_iopte pte;
 	int lvl = ARM_LPAE_START_LVL(data);
 
 	do {
@@ -692,6 +754,31 @@ static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
 	return iopte_to_paddr(pte, data) | iova;
 }
 
+
+static phys_addr_t arm_lpae_split_iova_to_phys(struct io_pgtable_ops *ops,
+					       unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return _arm_lpae_iova_to_phys(data, iova, ptep);
+}
+
+static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
+					 unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_iova_to_phys(data, iova, data->pgd[0]);
+}
+
 static void arm_lpae_restrict_pgsizes(struct io_pgtable_cfg *cfg)
 {
 	unsigned long granule, page_sizes;
@@ -771,6 +858,9 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
 	pgd_bits = va_bits - (data->bits_per_level * (data->levels - 1));
 	data->pgd_size = 1UL << (pgd_bits + ilog2(sizeof(arm_lpae_iopte)));
 
+	data->pgd[0] = NULL;
+	data->pgd[1] = NULL;
+
 	data->iop.ops = (struct io_pgtable_ops) {
 		.map		= arm_lpae_map,
 		.unmap		= arm_lpae_unmap,
@@ -780,8 +870,8 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
 	return data;
 }
 
-static struct io_pgtable *
-arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+static struct arm_lpae_io_pgtable *
+_arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg)
 {
 	u64 reg;
 	struct arm_lpae_io_pgtable *data;
@@ -845,8 +935,6 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
 
 	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T0SZ_SHIFT;
 
-	/* Disable speculative walks through TTBR1 */
-	reg |= ARM_LPAE_TCR_EPD1;
 	cfg->arm_lpae_s1_cfg.tcr = reg;
 
 	/* MAIRs */
@@ -863,24 +951,143 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
 	cfg->arm_lpae_s1_cfg.mair[1] = 0;
 
 	/* Looking good; allocate a pgd */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
 		goto out_free_data;
 
 	/* Ensure the empty pgd is visible before any actual TTBR write */
 	wmb();
 
 	/* TTBRs */
-	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd[0]);
 	cfg->arm_lpae_s1_cfg.ttbr[1] = 0;
+	return data;
+
+out_free_data:
+	kfree(data);
+	return NULL;
+}
+
+
+/* Allocate split pagetables */
+static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_split_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	u64 reg;
+	struct arm_lpae_io_pgtable *data;
+	u32 sep;
+
+	/* Figure out what the sign extension bit should be */
+	switch (cfg->ias) {
+	case 32:
+	case 36:
+	case 40:
+	case 42:
+	case 44:
+		sep = cfg->ias - 1;
+		/* Adjust the address size to account for the extension bit */
+		cfg->ias--;
+		break;
+	case 48:
+		/* IAS of 48 is a special case, it has a dedicated bit */
+		sep = 48;
+		break;
+	default:
+		return NULL;
+	}
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	/* Add the TTBR1 settings */
+	reg = cfg->arm_lpae_s1_cfg.tcr;
+
+	/* TCR */
+	reg |= (ARM_LPAE_TCR_SH_IS << ARM_LPAE_TCR_SH1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_IRGN1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_ORGN1_SHIFT);
+
+	switch (ARM_LPAE_GRANULE(data)) {
+	case SZ_4K:
+		reg |= ARM_LPAE_TCR_TG1_4K;
+		break;
+	case SZ_16K:
+		reg |= ARM_LPAE_TCR_TG1_16K;
+		break;
+	case SZ_64K:
+		reg |= ARM_LPAE_TCR_TG1_64K;
+		break;
+	}
+
+	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T1SZ_SHIFT;
+
+	switch (sep) {
+	case 31:
+		reg |= ARM_LPAE_TCR_SEP_31;
+		break;
+	case 35:
+		reg |= ARM_LPAE_TCR_SEP_35;
+		break;
+	case 39:
+		reg |= ARM_LPAE_TCR_SEP_39;
+		break;
+	case 41:
+		reg |= ARM_LPAE_TCR_SEP_41;
+		break;
+	case 43:
+		reg |= ARM_LPAE_TCR_SEP_43;
+		break;
+	case 48:
+		reg |= ARM_LPAE_TCR_SEP_UPSTREAM;
+		break;
+	}
+
+	cfg->arm_lpae_s1_cfg.tcr = reg;
+
+	/* Allocate the TTBR1 pagetable */
+	data->pgd[1] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[1])
+		goto out_free_data;
+
+	/* Override the data ops with split table specific ops */
+	data->iop.ops = (struct io_pgtable_ops) {
+		.map		= arm_lpae_split_map,
+		.unmap		= arm_lpae_split_unmap,
+		.iova_to_phys	= arm_lpae_split_iova_to_phys,
+	};
+
+	/*
+	 * remember the sign extension bit, we'll need it later to figure out
+	 * which pagetable to use
+	 */
+	data->sep = sep;
+
+	/* Ensure the empty pgd is visible before any actual TTBR write */
+	wmb();
+
+	cfg->arm_lpae_s1_cfg.ttbr[1] = virt_to_phys(data->pgd[1]);
 	return &data->iop;
 
 out_free_data:
+	__arm_lpae_free_pages(data->pgd[0], data->pgd_size, cfg);
 	kfree(data);
 	return NULL;
 }
 
 static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	struct arm_lpae_io_pgtable *data;
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	return &data->iop;
+}
+
+static struct io_pgtable *
 arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
 {
 	u64 reg, sl;
@@ -960,15 +1167,15 @@ arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
 	cfg->arm_lpae_s2_cfg.vtcr = reg;
 
 	/* Allocate pgd pages */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
 		goto out_free_data;
 
 	/* Ensure the empty pgd is visible before any actual TTBR write */
 	wmb();
 
 	/* VTTBR */
-	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd[0]);
 	return &data->iop;
 
 out_free_data:
@@ -1041,6 +1248,11 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
 	.free	= arm_lpae_free_pgtable,
 };
 
+struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns = {
+	.alloc	= arm_64_lpae_alloc_pgtable_split_s1,
+	.free	= arm_lpae_free_pgtable,
+};
+
 struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
 	.alloc	= arm_64_lpae_alloc_pgtable_s2,
 	.free	= arm_lpae_free_pgtable,
@@ -1095,9 +1307,9 @@ static void __init arm_lpae_dump_ops(struct io_pgtable_ops *ops)
 
 	pr_err("cfg: pgsize_bitmap 0x%lx, ias %u-bit\n",
 		cfg->pgsize_bitmap, cfg->ias);
-	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p\n",
+	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p %p\n",
 		data->levels, data->pgd_size, data->pg_shift,
-		data->bits_per_level, data->pgd);
+		data->bits_per_level, data->pgd[0], data->pgd[1]);
 }
 
 #define __FAIL(ops, i)	({						\
@@ -1112,6 +1324,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg)
 	static const enum io_pgtable_fmt fmts[] = {
 		ARM_64_LPAE_S1,
 		ARM_64_LPAE_S2,
+		ARM_64_LPAE_SPLIT_S1,
 	};
 
 	int i, j;
drivers/iommu/io-pgtable.c +1 -0
index ced53e5..ed1985f 100644
@@ -20,6 +20,7 @@ io_pgtable_init_table[IO_PGTABLE_NUM_FMTS] = {
 	[ARM_64_LPAE_S1] = &io_pgtable_arm_64_lpae_s1_init_fns,
 	[ARM_64_LPAE_S2] = &io_pgtable_arm_64_lpae_s2_init_fns,
 	[ARM_MALI_LPAE] = &io_pgtable_arm_mali_lpae_init_fns,
+	[ARM_64_LPAE_SPLIT_S1] = &io_pgtable_arm_64_lpae_split_s1_init_fns,
 #endif
 #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S
 	[ARM_V7S] = &io_pgtable_arm_v7s_init_fns,
include/linux/io-pgtable.h +2 -0
index b5a450a..2ed0e6a 100644
@@ -10,6 +10,7 @@ enum io_pgtable_fmt {
 	ARM_32_LPAE_S1,
 	ARM_32_LPAE_S2,
 	ARM_64_LPAE_S1,
+	ARM_64_LPAE_SPLIT_S1,
 	ARM_64_LPAE_S2,
 	ARM_V7S,
 	ARM_MALI_LPAE,
@@ -211,6 +212,7 @@ extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns;
+extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_mali_lpae_init_fns;
 
Signature · 1 line
2.7.4
reviewer
#3 lore raw

Re: [RESEND PATCH v2 2/3] iommu/io-pgtable-arm: Add support for AARCH64 split pagetables

Hi Jordan,

Quoted 6 lines · Jordan Crouse

Add a new sub-format ARM_64_LPAE_SPLIT_S1 to create and set up split pagetables (TTBR0 and TTBR1). The initialization function sets up the correct va_size and sign extension bits and programs the TCR registers. Split pagetable formats use their own own map/unmap wrappers to ensure that the correct pagetable is selected based on the incoming iova but most of the heavy lifting is common to the other formats.

I'm somewhat concerned that this implementation is very closely tied to the current Adreno use-case, and won't easily generalise in future to other potential TTBR1 uses which have been tossed around, like SMMUv3-without-substream-ID.

Furthermore, even for the Adreno pretend-PASID case it appears to be a bit too fragile for comfort - given that a DOMAIN_ATTR_SPLIT_TABLES domain doesn't look any different from a regular one from the users' point of view, what's to stop them making "without PASID" mappings in the lower half of the address space, and thus unwittingly pulling the rug out from under their own feet upon attaching an aux domain? In fact allocating a TTBR0 table at all for the main domain seems like little more than a waste of memory.

Quoted 44 lines

---

  drivers/iommu/io-pgtable-arm.c | 261 +++++++++++++++++++++++++++++++++++++----
  drivers/iommu/io-pgtable.c     |   1 +
  include/linux/io-pgtable.h     |   2 +
  3 files changed, 240 insertions(+), 24 deletions(-)
drivers/iommu/io-pgtable-arm.c +8 -0
index 161a7d56..aec35e5 100644
@@ -118,7 +118,12 @@
  #define ARM_LPAE_TCR_TG0_64K		(1 << 14)
  #define ARM_LPAE_TCR_TG0_16K		(2 << 14)
 
+#define ARM_LPAE_TCR_TG1_4K		(0 << 30)
+#define ARM_LPAE_TCR_TG1_64K		(1 << 30)
+#define ARM_LPAE_TCR_TG1_16K		(2 << 30)
+
  #define ARM_LPAE_TCR_SH0_SHIFT		12
+#define ARM_LPAE_TCR_SH1_SHIFT		28
  #define ARM_LPAE_TCR_SH0_MASK		0x3
  #define ARM_LPAE_TCR_SH_NS		0
  #define ARM_LPAE_TCR_SH_OS		2
@@ -126,6 +131,8 @@
 
  #define ARM_LPAE_TCR_ORGN0_SHIFT	10
  #define ARM_LPAE_TCR_IRGN0_SHIFT	8
+#define ARM_LPAE_TCR_ORGN1_SHIFT	26
+#define ARM_LPAE_TCR_IRGN1_SHIFT	24
  #define ARM_LPAE_TCR_RGN_MASK		0x3
  #define ARM_LPAE_TCR_RGN_NC		0
  #define ARM_LPAE_TCR_RGN_WBWA		1
@@ -136,6 +143,7 @@
  #define ARM_LPAE_TCR_SL0_MASK		0x3
 
  #define ARM_LPAE_TCR_T0SZ_SHIFT		0
+#define ARM_LPAE_TCR_T1SZ_SHIFT		16
  #define ARM_LPAE_TCR_SZ_MASK		0xf
 
  #define ARM_LPAE_TCR_PS_SHIFT		16

drivers/iommu/io-pgtable-arm.c line 169

Quoted 10 lines
drivers/iommu/io-pgtable-arm.c +7 -0
@@ -152,6 +160,14 @@
  #define ARM_LPAE_TCR_PS_48_BIT		0x5ULL
  #define ARM_LPAE_TCR_PS_52_BIT		0x6ULL
 
+#define ARM_LPAE_TCR_SEP_SHIFT		47
+#define ARM_LPAE_TCR_SEP_31		(0x0ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_35		(0x1ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_39		(0x2ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_41		(0x3ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_43		(0x4ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_UPSTREAM	(0x7ULL << ARM_LPAE_TCR_SEP_SHIFT)

This is a specific detail of SMMUv2, and nothing to do with the LPAE/AArch64 VMSA formats.

Quoted 49 lines
drivers/iommu/io-pgtable-arm.c +10 -6
+
  #define ARM_LPAE_MAIR_ATTR_SHIFT(n)	((n) << 3)
  #define ARM_LPAE_MAIR_ATTR_MASK		0xff
  #define ARM_LPAE_MAIR_ATTR_DEVICE	0x04
@@ -179,11 +195,12 @@ struct arm_lpae_io_pgtable {
  	struct io_pgtable	iop;
 
  	int			levels;
+	u32			sep;
  	size_t			pgd_size;
  	unsigned long		pg_shift;
  	unsigned long		bits_per_level;
 
-	void			*pgd;
+	void			*pgd[2];
  };
 
  typedef u64 arm_lpae_iopte;
@@ -426,7 +443,8 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
  	arm_lpae_iopte pte;
 
  	if (data->iop.fmt == ARM_64_LPAE_S1 ||
-	    data->iop.fmt == ARM_32_LPAE_S1) {
+	    data->iop.fmt == ARM_32_LPAE_S1 ||
+	    data->iop.fmt == ARM_64_LPAE_SPLIT_S1) {
  		pte = ARM_LPAE_PTE_nG;
  		if (!(prot & IOMMU_WRITE) && (prot & IOMMU_READ))
  			pte |= ARM_LPAE_PTE_AP_RDONLY;
@@ -470,11 +488,10 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
  	return pte;
  }
 
-static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
-			phys_addr_t paddr, size_t size, int iommu_prot)
+static int _arm_lpae_map(struct arm_lpae_io_pgtable *data, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot,
+			arm_lpae_iopte *ptep)
  {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
  	int ret, lvl = ARM_LPAE_START_LVL(data);
  	arm_lpae_iopte prot;
 
@@ -497,12 +514,39 @@ static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
  	return ret;
  }
 
+static int arm_lpae_split_map(struct io_pgtable_ops *ops, unsigned long iova,
+		phys_addr_t paddr, size_t size, int iommu_prot)

drivers/iommu/io-pgtable-arm.c static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova, · line 528

Quoted 10 lines
drivers/iommu/io-pgtable-arm.c +10 -0
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];

AFAICS, this logic effectively aliases the TTBR1 region across the entire upper half of the address hole, so bogus IOVAs may corrupt valid mappings instead of giving errors as expected.

Quoted 134 lines
drivers/iommu/io-pgtable-arm.c +69 -8
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, ptep);
+}
+
+static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, data->pgd[0]);
+}
+
  static void __arm_lpae_free_pgtable(struct arm_lpae_io_pgtable *data, int lvl,
  				    arm_lpae_iopte *ptep)
  {
  	arm_lpae_iopte *start, *end;
  	unsigned long table_size;
 
+	if (!ptep)
+		return;
+
  	if (lvl == ARM_LPAE_START_LVL(data))
  		table_size = data->pgd_size;
  	else
@@ -532,7 +576,8 @@ static void arm_lpae_free_pgtable(struct io_pgtable *iop)
  {
  	struct arm_lpae_io_pgtable *data = io_pgtable_to_data(iop);
 
-	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[0]);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[1]);
  	kfree(data);
  }
 
@@ -643,11 +688,28 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
  	return __arm_lpae_unmap(data, iova, size, lvl + 1, ptep);
  }
 
+static size_t arm_lpae_split_unmap(struct io_pgtable_ops *ops,
+				   unsigned long iova, size_t size)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+	int lvl = ARM_LPAE_START_LVL(data);
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
+}
+
  static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
  			     size_t size)
  {
  	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
+	arm_lpae_iopte *ptep = data->pgd[0];
  	int lvl = ARM_LPAE_START_LVL(data);
 
  	if (WARN_ON(iova >= (1ULL << data->iop.cfg.ias)))
@@ -656,11 +718,11 @@ static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
  	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
  }
 
-static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
-					 unsigned long iova)
+static phys_addr_t _arm_lpae_iova_to_phys(struct arm_lpae_io_pgtable *data,
+					  unsigned long iova,
+					  arm_lpae_iopte *ptep)
  {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte pte, *ptep = data->pgd;
+	arm_lpae_iopte pte;
  	int lvl = ARM_LPAE_START_LVL(data);
 
  	do {
@@ -692,6 +754,31 @@ static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
  	return iopte_to_paddr(pte, data) | iova;
  }
 
+
+static phys_addr_t arm_lpae_split_iova_to_phys(struct io_pgtable_ops *ops,
+					       unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return _arm_lpae_iova_to_phys(data, iova, ptep);
+}
+
+static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
+					 unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_iova_to_phys(data, iova, data->pgd[0]);
+}
+
  static void arm_lpae_restrict_pgsizes(struct io_pgtable_cfg *cfg)
  {
  	unsigned long granule, page_sizes;
@@ -771,6 +858,9 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
  	pgd_bits = va_bits - (data->bits_per_level * (data->levels - 1));
  	data->pgd_size = 1UL << (pgd_bits + ilog2(sizeof(arm_lpae_iopte)));
 
+	data->pgd[0] = NULL;
+	data->pgd[1] = NULL;
+
  	data->iop.ops = (struct io_pgtable_ops) {
  		.map		= arm_lpae_map,
  		.unmap		= arm_lpae_unmap,
@@ -780,8 +870,8 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
  	return data;
  }
 
-static struct io_pgtable *
-arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+static struct arm_lpae_io_pgtable *
+_arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg)
  {
  	u64 reg;
  	struct arm_lpae_io_pgtable *data;

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 849

Quoted 5 lines
drivers/iommu/io-pgtable-arm.c +0 -2
@@ -845,8 +935,6 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
 
  	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T0SZ_SHIFT;
 
-	/* Disable speculative walks through TTBR1 */
-	reg |= ARM_LPAE_TCR_EPD1;

Not setting EPD1 at all could potentially lead to problems for systems not using split tables.

Quoted 28 lines
drivers/iommu/io-pgtable-arm.c +11 -3
  	cfg->arm_lpae_s1_cfg.tcr = reg;
 
  	/* MAIRs */
@@ -863,24 +951,143 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
  	cfg->arm_lpae_s1_cfg.mair[1] = 0;
 
  	/* Looking good; allocate a pgd */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
  		goto out_free_data;
 
  	/* Ensure the empty pgd is visible before any actual TTBR write */
  	wmb();
 
  	/* TTBRs */
-	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd[0]);
  	cfg->arm_lpae_s1_cfg.ttbr[1] = 0;
+	return data;
+
+out_free_data:
+	kfree(data);
+	return NULL;
+}
+
+

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 981

Quoted 10 lines
drivers/iommu/io-pgtable-arm.c +10 -0
+/* Allocate split pagetables */
+static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_split_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	u64 reg;
+	struct arm_lpae_io_pgtable *data;
+	u32 sep;
+
+	/* Figure out what the sign extension bit should be */
+	switch (cfg->ias) {

IAS only tells you how wide the SMMU's address input is; what you really care about for sign extension is how wide the relevant masters' address *outputs* are. There's no way we can know that here, so AFAICS the only truly robust answer is for the caller requesting split tables to also provide the appropriate split point.

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 992

Quoted 11 lines
drivers/iommu/io-pgtable-arm.c +11 -0
+	case 32:
+	case 36:
+	case 40:
+	case 42:
+	case 44:
+		sep = cfg->ias - 1;
+		/* Adjust the address size to account for the extension bit */
+		cfg->ias--;
+		break;
+	case 48:
+		/* IAS of 48 is a special case, it has a dedicated bit */

That's not only not VMSA, but not even any architected SMMU behaviour; it just happens to be the way that one user of io-pgtable-arm currently mangles its IAS values.

Robin.

Quoted 185 lines
drivers/iommu/io-pgtable-arm.c +101 -5
+		sep = 48;
+		break;
+	default:
+		return NULL;
+	}
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	/* Add the TTBR1 settings */
+	reg = cfg->arm_lpae_s1_cfg.tcr;
+
+	/* TCR */
+	reg |= (ARM_LPAE_TCR_SH_IS << ARM_LPAE_TCR_SH1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_IRGN1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_ORGN1_SHIFT);
+
+	switch (ARM_LPAE_GRANULE(data)) {
+	case SZ_4K:
+		reg |= ARM_LPAE_TCR_TG1_4K;
+		break;
+	case SZ_16K:
+		reg |= ARM_LPAE_TCR_TG1_16K;
+		break;
+	case SZ_64K:
+		reg |= ARM_LPAE_TCR_TG1_64K;
+		break;
+	}
+
+	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T1SZ_SHIFT;
+
+	switch (sep) {
+	case 31:
+		reg |= ARM_LPAE_TCR_SEP_31;
+		break;
+	case 35:
+		reg |= ARM_LPAE_TCR_SEP_35;
+		break;
+	case 39:
+		reg |= ARM_LPAE_TCR_SEP_39;
+		break;
+	case 41:
+		reg |= ARM_LPAE_TCR_SEP_41;
+		break;
+	case 43:
+		reg |= ARM_LPAE_TCR_SEP_43;
+		break;
+	case 48:
+		reg |= ARM_LPAE_TCR_SEP_UPSTREAM;
+		break;
+	}
+
+	cfg->arm_lpae_s1_cfg.tcr = reg;
+
+	/* Allocate the TTBR1 pagetable */
+	data->pgd[1] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[1])
+		goto out_free_data;
+
+	/* Override the data ops with split table specific ops */
+	data->iop.ops = (struct io_pgtable_ops) {
+		.map		= arm_lpae_split_map,
+		.unmap		= arm_lpae_split_unmap,
+		.iova_to_phys	= arm_lpae_split_iova_to_phys,
+	};
+
+	/*
+	 * remember the sign extension bit, we'll need it later to figure out
+	 * which pagetable to use
+	 */
+	data->sep = sep;
+
+	/* Ensure the empty pgd is visible before any actual TTBR write */
+	wmb();
+
+	cfg->arm_lpae_s1_cfg.ttbr[1] = virt_to_phys(data->pgd[1]);
  	return &data->iop;
 
  out_free_data:
+	__arm_lpae_free_pages(data->pgd[0], data->pgd_size, cfg);
  	kfree(data);
  	return NULL;
  }
 
  static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	struct arm_lpae_io_pgtable *data;
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	return &data->iop;
+}
+
+static struct io_pgtable *
  arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
  {
  	u64 reg, sl;
@@ -960,15 +1167,15 @@ arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
  	cfg->arm_lpae_s2_cfg.vtcr = reg;
 
  	/* Allocate pgd pages */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
  		goto out_free_data;
 
  	/* Ensure the empty pgd is visible before any actual TTBR write */
  	wmb();
 
  	/* VTTBR */
-	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd[0]);
  	return &data->iop;
 
  out_free_data:
@@ -1041,6 +1248,11 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
  	.free	= arm_lpae_free_pgtable,
  };
 
+struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns = {
+	.alloc	= arm_64_lpae_alloc_pgtable_split_s1,
+	.free	= arm_lpae_free_pgtable,
+};
+
  struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
  	.alloc	= arm_64_lpae_alloc_pgtable_s2,
  	.free	= arm_lpae_free_pgtable,
@@ -1095,9 +1307,9 @@ static void __init arm_lpae_dump_ops(struct io_pgtable_ops *ops)
 
  	pr_err("cfg: pgsize_bitmap 0x%lx, ias %u-bit\n",
  		cfg->pgsize_bitmap, cfg->ias);
-	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p\n",
+	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p %p\n",
  		data->levels, data->pgd_size, data->pg_shift,
-		data->bits_per_level, data->pgd);
+		data->bits_per_level, data->pgd[0], data->pgd[1]);
  }
 
  #define __FAIL(ops, i)	({						\
@@ -1112,6 +1324,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg)
  	static const enum io_pgtable_fmt fmts[] = {
  		ARM_64_LPAE_S1,
  		ARM_64_LPAE_S2,
+		ARM_64_LPAE_SPLIT_S1,
  	};
 
  	int i, j;
drivers/iommu/io-pgtable.c +1 -0
index ced53e5..ed1985f 100644
@@ -20,6 +20,7 @@ io_pgtable_init_table[IO_PGTABLE_NUM_FMTS] = {
  	[ARM_64_LPAE_S1] = &io_pgtable_arm_64_lpae_s1_init_fns,
  	[ARM_64_LPAE_S2] = &io_pgtable_arm_64_lpae_s2_init_fns,
  	[ARM_MALI_LPAE] = &io_pgtable_arm_mali_lpae_init_fns,
+	[ARM_64_LPAE_SPLIT_S1] = &io_pgtable_arm_64_lpae_split_s1_init_fns,
  #endif
  #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S
  	[ARM_V7S] = &io_pgtable_arm_v7s_init_fns,
include/linux/io-pgtable.h +2 -0
index b5a450a..2ed0e6a 100644
@@ -10,6 +10,7 @@ enum io_pgtable_fmt {
  	ARM_32_LPAE_S1,
  	ARM_32_LPAE_S2,
  	ARM_64_LPAE_S1,
+	ARM_64_LPAE_SPLIT_S1,
  	ARM_64_LPAE_S2,
  	ARM_V7S,
  	ARM_MALI_LPAE,
@@ -211,6 +212,7 @@ extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
  extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns;
  extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns;
  extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns;
+extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns;
  extern struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns;
  extern struct io_pgtable_init_fns io_pgtable_arm_mali_lpae_init_fns;
 
#4 lore raw

Re: [RESEND PATCH v2 2/3] iommu/io-pgtable-arm: Add support for AARCH64 split pagetables

Quoted 10 lines · Robin Murphy

Hi Jordan,

Quoted 6 lines · Jordan Crouse

Add a new sub-format ARM_64_LPAE_SPLIT_S1 to create and set up split pagetables (TTBR0 and TTBR1). The initialization function sets up the correct va_size and sign extension bits and programs the TCR registers. Split pagetable formats use their own own map/unmap wrappers to ensure that the correct pagetable is selected based on the incoming iova but most of the heavy lifting is common to the other formats.

Quoted 4 lines

I'm somewhat concerned that this implementation is very closely tied to the current Adreno use-case, and won't easily generalise in future to other potential TTBR1 uses which have been tossed around, like SMMUv3-without-substream-ID.

I haven't heard about these use-cases. I figured v3 would be all PASID all the time. If you have details, I can see if I can avoid painting v3 into a corner.

Quoted 7 lines

Furthermore, even for the Adreno pretend-PASID case it appears to be a bit too fragile for comfort - given that a DOMAIN_ATTR_SPLIT_TABLES domain doesn't look any different from a regular one from the users' point of view, what's to stop them making "without PASID" mappings in the lower half of the address space, and thus unwittingly pulling the rug out from under their own feet upon attaching an aux domain? In fact allocating a TTBR0 table at all for the main domain seems like little more than a waste of memory.

That makes sense. Would it work better if we made a type ARM_64_LPAE_TTBR1_S1 that only allocated memory for TTBR1 and set TBR.EDP0 to trigger faults for TTBR0?

Then we could get rid of most of the extra stuff in io-pgtable-arm.c and I think that would meet most of your concerns.

Jordan

Quoted 474 lines
Quoted 40 lines

---

 drivers/iommu/io-pgtable-arm.c | 261 +++++++++++++++++++++++++++++++++++++----
 drivers/iommu/io-pgtable.c     |   1 +
 include/linux/io-pgtable.h     |   2 +
 3 files changed, 240 insertions(+), 24 deletions(-)
drivers/iommu/io-pgtable-arm.c +8 -0
index 161a7d56..aec35e5 100644
@@ -118,7 +118,12 @@
 #define ARM_LPAE_TCR_TG0_64K		(1 << 14)
 #define ARM_LPAE_TCR_TG0_16K		(2 << 14)
+#define ARM_LPAE_TCR_TG1_4K		(0 << 30)
+#define ARM_LPAE_TCR_TG1_64K		(1 << 30)
+#define ARM_LPAE_TCR_TG1_16K		(2 << 30)
+
 #define ARM_LPAE_TCR_SH0_SHIFT		12
+#define ARM_LPAE_TCR_SH1_SHIFT		28
 #define ARM_LPAE_TCR_SH0_MASK		0x3
 #define ARM_LPAE_TCR_SH_NS		0
 #define ARM_LPAE_TCR_SH_OS		2
@@ -126,6 +131,8 @@
 #define ARM_LPAE_TCR_ORGN0_SHIFT	10
 #define ARM_LPAE_TCR_IRGN0_SHIFT	8
+#define ARM_LPAE_TCR_ORGN1_SHIFT	26
+#define ARM_LPAE_TCR_IRGN1_SHIFT	24
 #define ARM_LPAE_TCR_RGN_MASK		0x3
 #define ARM_LPAE_TCR_RGN_NC		0
 #define ARM_LPAE_TCR_RGN_WBWA		1
@@ -136,6 +143,7 @@
 #define ARM_LPAE_TCR_SL0_MASK		0x3
 #define ARM_LPAE_TCR_T0SZ_SHIFT		0
+#define ARM_LPAE_TCR_T1SZ_SHIFT		16
 #define ARM_LPAE_TCR_SZ_MASK		0xf
 #define ARM_LPAE_TCR_PS_SHIFT		16

drivers/iommu/io-pgtable-arm.c line 168

Quoted 9 lines
drivers/iommu/io-pgtable-arm.c +7 -0
@@ -152,6 +160,14 @@
 #define ARM_LPAE_TCR_PS_48_BIT		0x5ULL
 #define ARM_LPAE_TCR_PS_52_BIT		0x6ULL
+#define ARM_LPAE_TCR_SEP_SHIFT		47
+#define ARM_LPAE_TCR_SEP_31		(0x0ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_35		(0x1ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_39		(0x2ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_41		(0x3ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_43		(0x4ULL << ARM_LPAE_TCR_SEP_SHIFT)
+#define ARM_LPAE_TCR_SEP_UPSTREAM	(0x7ULL << ARM_LPAE_TCR_SEP_SHIFT)

This is a specific detail of SMMUv2, and nothing to do with the LPAE/AArch64 VMSA formats.

Quoted 42 lines
drivers/iommu/io-pgtable-arm.c +10 -6
+
 #define ARM_LPAE_MAIR_ATTR_SHIFT(n)	((n) << 3)
 #define ARM_LPAE_MAIR_ATTR_MASK		0xff
 #define ARM_LPAE_MAIR_ATTR_DEVICE	0x04
@@ -179,11 +195,12 @@ struct arm_lpae_io_pgtable {
 	struct io_pgtable	iop;
 	int			levels;
+	u32			sep;
 	size_t			pgd_size;
 	unsigned long		pg_shift;
 	unsigned long		bits_per_level;
-	void			*pgd;
+	void			*pgd[2];
 };
 typedef u64 arm_lpae_iopte;
@@ -426,7 +443,8 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
 	arm_lpae_iopte pte;
 	if (data->iop.fmt == ARM_64_LPAE_S1 ||
-	    data->iop.fmt == ARM_32_LPAE_S1) {
+	    data->iop.fmt == ARM_32_LPAE_S1 ||
+	    data->iop.fmt == ARM_64_LPAE_SPLIT_S1) {
 		pte = ARM_LPAE_PTE_nG;
 		if (!(prot & IOMMU_WRITE) && (prot & IOMMU_READ))
 			pte |= ARM_LPAE_PTE_AP_RDONLY;
@@ -470,11 +488,10 @@ static arm_lpae_iopte arm_lpae_prot_to_pte(struct arm_lpae_io_pgtable *data,
 	return pte;
 }
-static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
-			phys_addr_t paddr, size_t size, int iommu_prot)
+static int _arm_lpae_map(struct arm_lpae_io_pgtable *data, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot,
+			arm_lpae_iopte *ptep)
 {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
 	int ret, lvl = ARM_LPAE_START_LVL(data);
 	arm_lpae_iopte prot;
@@ -497,12 +514,39 @@ static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
 	return ret;
 }
+static int arm_lpae_split_map(struct io_pgtable_ops *ops, unsigned long iova,
+		phys_addr_t paddr, size_t size, int iommu_prot)

drivers/iommu/io-pgtable-arm.c static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova, · line 527

Quoted 10 lines
drivers/iommu/io-pgtable-arm.c +10 -0
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];

AFAICS, this logic effectively aliases the TTBR1 region across the entire upper half of the address hole, so bogus IOVAs may corrupt valid mappings instead of giving errors as expected.

Quoted 124 lines
drivers/iommu/io-pgtable-arm.c +69 -8
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, ptep);
+}
+
+static int arm_lpae_map(struct io_pgtable_ops *ops, unsigned long iova,
+			phys_addr_t paddr, size_t size, int iommu_prot)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_map(data, iova, paddr, size, iommu_prot, data->pgd[0]);
+}
+
 static void __arm_lpae_free_pgtable(struct arm_lpae_io_pgtable *data, int lvl,
 				    arm_lpae_iopte *ptep)
 {
 	arm_lpae_iopte *start, *end;
 	unsigned long table_size;
+	if (!ptep)
+		return;
+
 	if (lvl == ARM_LPAE_START_LVL(data))
 		table_size = data->pgd_size;
 	else
@@ -532,7 +576,8 @@ static void arm_lpae_free_pgtable(struct io_pgtable *iop)
 {
 	struct arm_lpae_io_pgtable *data = io_pgtable_to_data(iop);
-	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[0]);
+	__arm_lpae_free_pgtable(data, ARM_LPAE_START_LVL(data), data->pgd[1]);
 	kfree(data);
 }
@@ -643,11 +688,28 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data,
 	return __arm_lpae_unmap(data, iova, size, lvl + 1, ptep);
 }
+static size_t arm_lpae_split_unmap(struct io_pgtable_ops *ops,
+				   unsigned long iova, size_t size)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+	int lvl = ARM_LPAE_START_LVL(data);
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
+}
+
 static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
 			     size_t size)
 {
 	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte *ptep = data->pgd;
+	arm_lpae_iopte *ptep = data->pgd[0];
 	int lvl = ARM_LPAE_START_LVL(data);
 	if (WARN_ON(iova >= (1ULL << data->iop.cfg.ias)))
@@ -656,11 +718,11 @@ static size_t arm_lpae_unmap(struct io_pgtable_ops *ops, unsigned long iova,
 	return __arm_lpae_unmap(data, iova, size, lvl, ptep);
 }
-static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
-					 unsigned long iova)
+static phys_addr_t _arm_lpae_iova_to_phys(struct arm_lpae_io_pgtable *data,
+					  unsigned long iova,
+					  arm_lpae_iopte *ptep)
 {
-	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
-	arm_lpae_iopte pte, *ptep = data->pgd;
+	arm_lpae_iopte pte;
 	int lvl = ARM_LPAE_START_LVL(data);
 	do {
@@ -692,6 +754,31 @@ static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
 	return iopte_to_paddr(pte, data) | iova;
 }
+
+static phys_addr_t arm_lpae_split_iova_to_phys(struct io_pgtable_ops *ops,
+					       unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+	unsigned long mask = 1UL << data->sep;
+	arm_lpae_iopte *ptep;
+
+	if (iova & mask) {
+		ptep = data->pgd[1];
+		iova &= (mask - 1);
+	} else
+		ptep = data->pgd[0];
+
+	return _arm_lpae_iova_to_phys(data, iova, ptep);
+}
+
+static phys_addr_t arm_lpae_iova_to_phys(struct io_pgtable_ops *ops,
+					 unsigned long iova)
+{
+	struct arm_lpae_io_pgtable *data = io_pgtable_ops_to_data(ops);
+
+	return _arm_lpae_iova_to_phys(data, iova, data->pgd[0]);
+}
+
 static void arm_lpae_restrict_pgsizes(struct io_pgtable_cfg *cfg)
 {
 	unsigned long granule, page_sizes;
@@ -771,6 +858,9 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
 	pgd_bits = va_bits - (data->bits_per_level * (data->levels - 1));
 	data->pgd_size = 1UL << (pgd_bits + ilog2(sizeof(arm_lpae_iopte)));
+	data->pgd[0] = NULL;
+	data->pgd[1] = NULL;
+
 	data->iop.ops = (struct io_pgtable_ops) {
 		.map		= arm_lpae_map,
 		.unmap		= arm_lpae_unmap,
@@ -780,8 +870,8 @@ arm_lpae_alloc_pgtable(struct io_pgtable_cfg *cfg)
 	return data;
 }
-static struct io_pgtable *
-arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+static struct arm_lpae_io_pgtable *
+_arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg)
 {
 	u64 reg;
 	struct arm_lpae_io_pgtable *data;

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 847

Quoted 3 lines
drivers/iommu/io-pgtable-arm.c +0 -2
@@ -845,8 +935,6 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
 	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T0SZ_SHIFT;
-	/* Disable speculative walks through TTBR1 */
-	reg |= ARM_LPAE_TCR_EPD1;

Not setting EPD1 at all could potentially lead to problems for systems not using split tables.

Quoted 24 lines
drivers/iommu/io-pgtable-arm.c +11 -3
 	cfg->arm_lpae_s1_cfg.tcr = reg;
 	/* MAIRs */
@@ -863,24 +951,143 @@ arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
 	cfg->arm_lpae_s1_cfg.mair[1] = 0;
 	/* Looking good; allocate a pgd */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
 		goto out_free_data;
 	/* Ensure the empty pgd is visible before any actual TTBR write */
 	wmb();
 	/* TTBRs */
-	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s1_cfg.ttbr[0] = virt_to_phys(data->pgd[0]);
 	cfg->arm_lpae_s1_cfg.ttbr[1] = 0;
+	return data;
+
+out_free_data:
+	kfree(data);
+	return NULL;
+}
+
+

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 978

Quoted 10 lines
drivers/iommu/io-pgtable-arm.c +10 -0
+/* Allocate split pagetables */
+static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_split_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	u64 reg;
+	struct arm_lpae_io_pgtable *data;
+	u32 sep;
+
+	/* Figure out what the sign extension bit should be */
+	switch (cfg->ias) {

IAS only tells you how wide the SMMU's address input is; what you really care about for sign extension is how wide the relevant masters' address *outputs* are. There's no way we can know that here, so AFAICS the only truly robust answer is for the caller requesting split tables to also provide the appropriate split point.

drivers/iommu/io-pgtable-arm.c arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie) · line 992

Quoted 11 lines
drivers/iommu/io-pgtable-arm.c +11 -0
+	case 32:
+	case 36:
+	case 40:
+	case 42:
+	case 44:
+		sep = cfg->ias - 1;
+		/* Adjust the address size to account for the extension bit */
+		cfg->ias--;
+		break;
+	case 48:
+		/* IAS of 48 is a special case, it has a dedicated bit */

That's not only not VMSA, but not even any architected SMMU behaviour; it just happens to be the way that one user of io-pgtable-arm currently mangles its IAS values.

Robin.

Quoted 174 lines
drivers/iommu/io-pgtable-arm.c +101 -5
+		sep = 48;
+		break;
+	default:
+		return NULL;
+	}
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	/* Add the TTBR1 settings */
+	reg = cfg->arm_lpae_s1_cfg.tcr;
+
+	/* TCR */
+	reg |= (ARM_LPAE_TCR_SH_IS << ARM_LPAE_TCR_SH1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_IRGN1_SHIFT) |
+	      (ARM_LPAE_TCR_RGN_WBWA << ARM_LPAE_TCR_ORGN1_SHIFT);
+
+	switch (ARM_LPAE_GRANULE(data)) {
+	case SZ_4K:
+		reg |= ARM_LPAE_TCR_TG1_4K;
+		break;
+	case SZ_16K:
+		reg |= ARM_LPAE_TCR_TG1_16K;
+		break;
+	case SZ_64K:
+		reg |= ARM_LPAE_TCR_TG1_64K;
+		break;
+	}
+
+	reg |= (64ULL - cfg->ias) << ARM_LPAE_TCR_T1SZ_SHIFT;
+
+	switch (sep) {
+	case 31:
+		reg |= ARM_LPAE_TCR_SEP_31;
+		break;
+	case 35:
+		reg |= ARM_LPAE_TCR_SEP_35;
+		break;
+	case 39:
+		reg |= ARM_LPAE_TCR_SEP_39;
+		break;
+	case 41:
+		reg |= ARM_LPAE_TCR_SEP_41;
+		break;
+	case 43:
+		reg |= ARM_LPAE_TCR_SEP_43;
+		break;
+	case 48:
+		reg |= ARM_LPAE_TCR_SEP_UPSTREAM;
+		break;
+	}
+
+	cfg->arm_lpae_s1_cfg.tcr = reg;
+
+	/* Allocate the TTBR1 pagetable */
+	data->pgd[1] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[1])
+		goto out_free_data;
+
+	/* Override the data ops with split table specific ops */
+	data->iop.ops = (struct io_pgtable_ops) {
+		.map		= arm_lpae_split_map,
+		.unmap		= arm_lpae_split_unmap,
+		.iova_to_phys	= arm_lpae_split_iova_to_phys,
+	};
+
+	/*
+	 * remember the sign extension bit, we'll need it later to figure out
+	 * which pagetable to use
+	 */
+	data->sep = sep;
+
+	/* Ensure the empty pgd is visible before any actual TTBR write */
+	wmb();
+
+	cfg->arm_lpae_s1_cfg.ttbr[1] = virt_to_phys(data->pgd[1]);
 	return &data->iop;
 out_free_data:
+	__arm_lpae_free_pages(data->pgd[0], data->pgd_size, cfg);
 	kfree(data);
 	return NULL;
 }
 static struct io_pgtable *
+arm_64_lpae_alloc_pgtable_s1(struct io_pgtable_cfg *cfg, void *cookie)
+{
+	struct arm_lpae_io_pgtable *data;
+
+	data = _arm_64_lpae_alloc_pgtable_s1(cfg);
+	if (!data)
+		return NULL;
+
+	return &data->iop;
+}
+
+static struct io_pgtable *
 arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
 {
 	u64 reg, sl;
@@ -960,15 +1167,15 @@ arm_64_lpae_alloc_pgtable_s2(struct io_pgtable_cfg *cfg, void *cookie)
 	cfg->arm_lpae_s2_cfg.vtcr = reg;
 	/* Allocate pgd pages */
-	data->pgd = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
-	if (!data->pgd)
+	data->pgd[0] = __arm_lpae_alloc_pages(data->pgd_size, GFP_KERNEL, cfg);
+	if (!data->pgd[0])
 		goto out_free_data;
 	/* Ensure the empty pgd is visible before any actual TTBR write */
 	wmb();
 	/* VTTBR */
-	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd);
+	cfg->arm_lpae_s2_cfg.vttbr = virt_to_phys(data->pgd[0]);
 	return &data->iop;
 out_free_data:
@@ -1041,6 +1248,11 @@ struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns = {
 	.free	= arm_lpae_free_pgtable,
 };
+struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns = {
+	.alloc	= arm_64_lpae_alloc_pgtable_split_s1,
+	.free	= arm_lpae_free_pgtable,
+};
+
 struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns = {
 	.alloc	= arm_64_lpae_alloc_pgtable_s2,
 	.free	= arm_lpae_free_pgtable,
@@ -1095,9 +1307,9 @@ static void __init arm_lpae_dump_ops(struct io_pgtable_ops *ops)
 	pr_err("cfg: pgsize_bitmap 0x%lx, ias %u-bit\n",
 		cfg->pgsize_bitmap, cfg->ias);
-	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p\n",
+	pr_err("data: %d levels, 0x%zx pgd_size, %lu pg_shift, %lu bits_per_level, pgd @ %p %p\n",
 		data->levels, data->pgd_size, data->pg_shift,
-		data->bits_per_level, data->pgd);
+		data->bits_per_level, data->pgd[0], data->pgd[1]);
 }
 #define __FAIL(ops, i)	({						\
@@ -1112,6 +1324,7 @@ static int __init arm_lpae_run_tests(struct io_pgtable_cfg *cfg)
 	static const enum io_pgtable_fmt fmts[] = {
 		ARM_64_LPAE_S1,
 		ARM_64_LPAE_S2,
+		ARM_64_LPAE_SPLIT_S1,
 	};
 	int i, j;
drivers/iommu/io-pgtable.c +1 -0
index ced53e5..ed1985f 100644
@@ -20,6 +20,7 @@ io_pgtable_init_table[IO_PGTABLE_NUM_FMTS] = {
 	[ARM_64_LPAE_S1] = &io_pgtable_arm_64_lpae_s1_init_fns,
 	[ARM_64_LPAE_S2] = &io_pgtable_arm_64_lpae_s2_init_fns,
 	[ARM_MALI_LPAE] = &io_pgtable_arm_mali_lpae_init_fns,
+	[ARM_64_LPAE_SPLIT_S1] = &io_pgtable_arm_64_lpae_split_s1_init_fns,
 #endif
 #ifdef CONFIG_IOMMU_IO_PGTABLE_ARMV7S
 	[ARM_V7S] = &io_pgtable_arm_v7s_init_fns,
include/linux/io-pgtable.h +2 -0
index b5a450a..2ed0e6a 100644
@@ -10,6 +10,7 @@ enum io_pgtable_fmt {
 	ARM_32_LPAE_S1,
 	ARM_32_LPAE_S2,
 	ARM_64_LPAE_S1,
+	ARM_64_LPAE_SPLIT_S1,
 	ARM_64_LPAE_S2,
 	ARM_V7S,
 	ARM_MALI_LPAE,
@@ -211,6 +212,7 @@ extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_32_lpae_s2_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_s2_init_fns;
+extern struct io_pgtable_init_fns io_pgtable_arm_64_lpae_split_s1_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_v7s_init_fns;
 extern struct io_pgtable_init_fns io_pgtable_arm_mali_lpae_init_fns;
 
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
#5 lore raw

PATCH RESEND v2 3/3 iommu/arm-smmu: Add support for DOMAIN_ATTR_SPLIT_TABLES

In reply to Jordan Crouse

When DOMAIN_ATTR_SPLIT_TABLES is specified for pass ARM_64_LPAE_SPLIT_S1 to io_pgtable_ops to allocate and initialize TTBR0 and TTBR1 pagetables.

v3: Moved all the pagetable specific work into io-pgtable-arm in a previous patch.

---

 drivers/iommu/arm-smmu.c | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/iommu/arm-smmu.c +15 -1
index 653b6b3..7a6b4bb 100644
@@ -257,6 +257,7 @@ struct arm_smmu_domain {
 	bool				non_strict;
 	struct mutex			init_mutex; /* Protects smmu pointer */
 	spinlock_t			cb_lock; /* Serialises ATS1* ops and TLB syncs */
+	u32 attributes;
 	struct iommu_domain		domain;
 };
 
@@ -832,7 +833,11 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 		ias = smmu->va_size;
 		oas = smmu->ipa_size;
 		if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH64) {
-			fmt = ARM_64_LPAE_S1;
+			if (smmu_domain->attributes &
+				(1 << DOMAIN_ATTR_SPLIT_TABLES))
+				fmt = ARM_64_LPAE_SPLIT_S1;
+			else
+				fmt = ARM_64_LPAE_S1;
 		} else if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH32_L) {
 			fmt = ARM_32_LPAE_S1;
 			ias = min(ias, 32UL);
@@ -1582,6 +1587,10 @@ static int arm_smmu_domain_get_attr(struct iommu_domain *domain,
 		case DOMAIN_ATTR_NESTING:
 			*(int *)data = (smmu_domain->stage == ARM_SMMU_DOMAIN_NESTED);
 			return 0;
+		case DOMAIN_ATTR_SPLIT_TABLES:
+			*(int *)data = !!(smmu_domain->attributes &
+				(1 << DOMAIN_ATTR_SPLIT_TABLES));
+			return 0;
 		default:
 			return -ENODEV;
 		}
@@ -1622,6 +1631,11 @@ static int arm_smmu_domain_set_attr(struct iommu_domain *domain,
 			else
 				smmu_domain->stage = ARM_SMMU_DOMAIN_S1;
 			break;
+		case DOMAIN_ATTR_SPLIT_TABLES:
+			if (*((int *)data))
+				smmu_domain->attributes |=
+					(1 << DOMAIN_ATTR_SPLIT_TABLES);
+			break;
 		default:
 			ret = -ENODEV;
 		}
Signature · 1 line
2.7.4
reviewer
#6 lore raw

Re: [RESEND PATCH v2 3/3] iommu/arm-smmu: Add support for DOMAIN_ATTR_SPLIT_TABLES

Quoted 38 lines · Jordan Crouse

When DOMAIN_ATTR_SPLIT_TABLES is specified for pass ARM_64_LPAE_SPLIT_S1 to io_pgtable_ops to allocate and initialize TTBR0 and TTBR1 pagetables.

v3: Moved all the pagetable specific work into io-pgtable-arm in a previous patch.

---

  drivers/iommu/arm-smmu.c | 16 +++++++++++++++-
  1 file changed, 15 insertions(+), 1 deletion(-)
drivers/iommu/arm-smmu.c +6 -1
index 653b6b3..7a6b4bb 100644
@@ -257,6 +257,7 @@ struct arm_smmu_domain {
  	bool				non_strict;
  	struct mutex			init_mutex; /* Protects smmu pointer */
  	spinlock_t			cb_lock; /* Serialises ATS1* ops and TLB syncs */
+	u32 attributes;
  	struct iommu_domain		domain;
  };
 
@@ -832,7 +833,11 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
  		ias = smmu->va_size;
  		oas = smmu->ipa_size;
  		if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH64) {
-			fmt = ARM_64_LPAE_S1;
+			if (smmu_domain->attributes &
+				(1 << DOMAIN_ATTR_SPLIT_TABLES))
+				fmt = ARM_64_LPAE_SPLIT_S1;
+			else
+				fmt = ARM_64_LPAE_S1;
  		} else if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH32_L) {
  			fmt = ARM_32_LPAE_S1;
  			ias = min(ias, 32UL);

drivers/iommu/arm-smmu.c static int arm_smmu_domain_get_attr(struct iommu_domain *domain, · line 1592

Quoted 6 lines
drivers/iommu/arm-smmu.c +3 -0
@@ -1582,6 +1587,10 @@ static int arm_smmu_domain_get_attr(struct iommu_domain *domain,
  		case DOMAIN_ATTR_NESTING:
  			*(int *)data = (smmu_domain->stage == ARM_SMMU_DOMAIN_NESTED);
  			return 0;
+		case DOMAIN_ATTR_SPLIT_TABLES:
+			*(int *)data = !!(smmu_domain->attributes &
+				(1 << DOMAIN_ATTR_SPLIT_TABLES));

I'm not really a fan of always claiming to support this but silently ignoring it depending on hardware/kernel configuration - it seems somewhat tricky to make callers properly robust against making false assumptions (e.g. consider if the system is booted with "arm-smmu.force_stage=2").

Robin.

Quoted 17 lines
drivers/iommu/arm-smmu.c +6 -0
+			return 0;
  		default:
  			return -ENODEV;
  		}
@@ -1622,6 +1631,11 @@ static int arm_smmu_domain_set_attr(struct iommu_domain *domain,
  			else
  				smmu_domain->stage = ARM_SMMU_DOMAIN_S1;
  			break;
+		case DOMAIN_ATTR_SPLIT_TABLES:
+			if (*((int *)data))
+				smmu_domain->attributes |=
+					(1 << DOMAIN_ATTR_SPLIT_TABLES);
+			break;
  		default:
  			ret = -ENODEV;
  		}