ARMv7: support non-LPAE mapping (not xlat_v2)

Support 32bit descriptor MMU table. This is required by ARMv7
architectures that do not support the Large Page Address Extensions.

nonlpae_tables.c source file is dumped from the OP-TEE project:
core_mmu_armv7.c and related header files.

Change-Id: If912d66c374290c49c5a1211ce4c5c27b2d7dc60
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Usama Arif <usama.arif@arm.com>
diff --git a/include/lib/xlat_tables/xlat_tables_defs.h b/include/lib/xlat_tables/xlat_tables_defs.h
index 6d0fb78..f9bbe0f 100644
--- a/include/lib/xlat_tables/xlat_tables_defs.h
+++ b/include/lib/xlat_tables/xlat_tables_defs.h
@@ -73,7 +73,11 @@
 #define PAGE_SIZE_MASK		(PAGE_SIZE - U(1))
 #define IS_PAGE_ALIGNED(addr)	(((addr) & PAGE_SIZE_MASK) == U(0))
 
-#define XLAT_ENTRY_SIZE_SHIFT	U(3) /* Each MMU table entry is 8 bytes (1 << 3) */
+#if (ARM_ARCH_MAJOR == 7) && !ARMV7_SUPPORTS_LARGE_PAGE_ADDRESSING
+#define XLAT_ENTRY_SIZE_SHIFT	U(2)	/* Each MMU table entry is 4 bytes */
+#else
+#define XLAT_ENTRY_SIZE_SHIFT	U(3)	/* Each MMU table entry is 8 bytes */
+#endif
 #define XLAT_ENTRY_SIZE		(U(1) << XLAT_ENTRY_SIZE_SHIFT)
 
 #define XLAT_TABLE_SIZE_SHIFT	PAGE_SIZE_SHIFT /* Size of one complete table */
diff --git a/lib/xlat_tables/aarch32/nonlpae_tables.c b/lib/xlat_tables/aarch32/nonlpae_tables.c
new file mode 100644
index 0000000..eca3be3
--- /dev/null
+++ b/lib/xlat_tables/aarch32/nonlpae_tables.c
@@ -0,0 +1,550 @@
+/*
+ * Copyright (c) 2016-2017, Linaro Limited. All rights reserved.
+ * Copyright (c) 2014-2017, Arm Limited. All rights reserved.
+ * Copyright (c) 2014, STMicroelectronics International N.V.
+ * All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ */
+
+#include <assert.h>
+#include <stdio.h>
+#include <string.h>
+
+#include <platform_def.h>
+
+#include <arch.h>
+#include <arch_helpers.h>
+#include <common/debug.h>
+#include <lib/cassert.h>
+#include <lib/utils.h>
+#include <lib/xlat_tables/xlat_tables.h>
+
+#include "../xlat_tables_private.h"
+
+#ifdef ARMV7_SUPPORTS_LARGE_PAGE_ADDRESSING
+#error "ARMV7_SUPPORTS_LARGE_PAGE_ADDRESSING flag is set. \
+This module is to be used when LPAE is not supported"
+#endif
+
+CASSERT(PLAT_VIRT_ADDR_SPACE_SIZE == (1ULL << 32), invalid_vaddr_space_size);
+CASSERT(PLAT_PHY_ADDR_SPACE_SIZE == (1ULL << 32), invalid_paddr_space_size);
+
+#define MMU32B_UNSET_DESC	~0ul
+#define MMU32B_INVALID_DESC	0ul
+
+#define MT_UNKNOWN	~0U
+
+/*
+ * MMU related values
+ */
+
+/* Sharable */
+#define MMU32B_TTB_S           (1 << 1)
+
+/* Not Outer Sharable */
+#define MMU32B_TTB_NOS         (1 << 5)
+
+/* Normal memory, Inner Non-cacheable */
+#define MMU32B_TTB_IRGN_NC     0
+
+/* Normal memory, Inner Write-Back Write-Allocate Cacheable */
+#define MMU32B_TTB_IRGN_WBWA   (1 << 6)
+
+/* Normal memory, Inner Write-Through Cacheable */
+#define MMU32B_TTB_IRGN_WT     1
+
+/* Normal memory, Inner Write-Back no Write-Allocate Cacheable */
+#define MMU32B_TTB_IRGN_WB     (1 | (1 << 6))
+
+/* Normal memory, Outer Write-Back Write-Allocate Cacheable */
+#define MMU32B_TTB_RNG_WBWA    (1 << 3)
+
+#define MMU32B_DEFAULT_ATTRS \
+		(MMU32B_TTB_S | MMU32B_TTB_NOS | \
+		 MMU32B_TTB_IRGN_WBWA | MMU32B_TTB_RNG_WBWA)
+
+/* armv7 memory mapping attributes: section mapping */
+#define SECTION_SECURE			(0 << 19)
+#define SECTION_NOTSECURE		(1 << 19)
+#define SECTION_SHARED			(1 << 16)
+#define SECTION_NOTGLOBAL		(1 << 17)
+#define SECTION_ACCESS_FLAG		(1 << 10)
+#define SECTION_UNPRIV			(1 << 11)
+#define SECTION_RO			(1 << 15)
+#define SECTION_TEX(tex)		((((tex) >> 2) << 12) | \
+					((((tex) >> 1) & 0x1) << 3) | \
+					(((tex) & 0x1) << 2))
+#define SECTION_DEVICE			SECTION_TEX(MMU32B_ATTR_DEVICE_INDEX)
+#define SECTION_NORMAL			SECTION_TEX(MMU32B_ATTR_DEVICE_INDEX)
+#define SECTION_NORMAL_CACHED		\
+				SECTION_TEX(MMU32B_ATTR_IWBWA_OWBWA_INDEX)
+
+#define SECTION_XN			(1 << 4)
+#define SECTION_PXN			(1 << 0)
+#define SECTION_SECTION			(2 << 0)
+
+#define SECTION_PT_NOTSECURE		(1 << 3)
+#define SECTION_PT_PT			(1 << 0)
+
+#define SMALL_PAGE_SMALL_PAGE		(1 << 1)
+#define SMALL_PAGE_SHARED		(1 << 10)
+#define SMALL_PAGE_NOTGLOBAL		(1 << 11)
+#define SMALL_PAGE_TEX(tex)		((((tex) >> 2) << 6) | \
+					((((tex) >> 1) & 0x1) << 3) | \
+					(((tex) & 0x1) << 2))
+#define SMALL_PAGE_DEVICE		\
+				SMALL_PAGE_TEX(MMU32B_ATTR_DEVICE_INDEX)
+#define SMALL_PAGE_NORMAL		\
+				SMALL_PAGE_TEX(MMU32B_ATTR_DEVICE_INDEX)
+#define SMALL_PAGE_NORMAL_CACHED	\
+				SMALL_PAGE_TEX(MMU32B_ATTR_IWBWA_OWBWA_INDEX)
+#define SMALL_PAGE_ACCESS_FLAG		(1 << 4)
+#define SMALL_PAGE_UNPRIV		(1 << 5)
+#define SMALL_PAGE_RO			(1 << 9)
+#define SMALL_PAGE_XN			(1 << 0)
+
+/* The TEX, C and B bits concatenated */
+#define MMU32B_ATTR_DEVICE_INDEX		0x0
+#define MMU32B_ATTR_IWBWA_OWBWA_INDEX		0x1
+
+#define MMU32B_PRRR_IDX(idx, tr, nos)	(((tr) << (2 * (idx))) | \
+					 ((uint32_t)(nos) << ((idx) + 24)))
+#define MMU32B_NMRR_IDX(idx, ir, or)	(((ir) << (2 * (idx))) | \
+					 ((uint32_t)(or) << (2 * (idx) + 16)))
+#define MMU32B_PRRR_DS0			(1 << 16)
+#define MMU32B_PRRR_DS1			(1 << 17)
+#define MMU32B_PRRR_NS0			(1 << 18)
+#define MMU32B_PRRR_NS1			(1 << 19)
+
+#define DACR_DOMAIN(num, perm)		((perm) << ((num) * 2))
+#define DACR_DOMAIN_PERM_NO_ACCESS	0x0
+#define DACR_DOMAIN_PERM_CLIENT		0x1
+#define DACR_DOMAIN_PERM_MANAGER	0x3
+
+#define NUM_1MB_IN_4GB		(1 << 12)
+#define NUM_4K_IN_1MB		(1 << 8)
+
+#define ONE_MB_SHIFT		20
+
+/* mmu 32b integration */
+#define MMU32B_L1_TABLE_SIZE		(NUM_1MB_IN_4GB * 4)
+#define MMU32B_L2_TABLE_SIZE		(NUM_4K_IN_1MB * 4)
+#define MMU32B_L1_TABLE_ALIGN		(1 << 14)
+#define MMU32B_L2_TABLE_ALIGN		(1 << 10)
+
+static unsigned int next_xlat;
+static unsigned long long xlat_max_pa;
+static uintptr_t xlat_max_va;
+
+static uint32_t mmu_l1_base[NUM_1MB_IN_4GB]
+	__aligned(MMU32B_L1_TABLE_ALIGN) __attribute__((section("xlat_table")));
+
+static uint32_t mmu_l2_base[MAX_XLAT_TABLES][NUM_4K_IN_1MB]
+	__aligned(MMU32B_L2_TABLE_ALIGN) __attribute__((section("xlat_table")));
+
+/*
+ * Array of all memory regions stored in order of ascending base address.
+ * The list is terminated by the first entry with size == 0.
+ */
+static mmap_region_t mmap[MAX_MMAP_REGIONS + 1];
+
+void print_mmap(void)
+{
+#if LOG_LEVEL >= LOG_LEVEL_VERBOSE
+	mmap_region_t *mm = mmap;
+
+	printf("init xlat - l1:%p  l2:%p (%d)\n",
+		    (void *)mmu_l1_base, (void *)mmu_l2_base, MAX_XLAT_TABLES);
+	printf("mmap:\n");
+	while (mm->size) {
+		printf(" VA:%p  PA:0x%llx  size:0x%zx  attr:0x%x\n",
+				(void *)mm->base_va, mm->base_pa,
+				mm->size, mm->attr);
+		++mm;
+	};
+	printf("\n");
+#endif
+}
+
+void mmap_add(const mmap_region_t *mm)
+{
+	const mmap_region_t *mm_cursor = mm;
+
+	while ((mm_cursor->size != 0U) || (mm_cursor->attr != 0U)) {
+		mmap_add_region(mm_cursor->base_pa, mm_cursor->base_va,
+				mm_cursor->size, mm_cursor->attr);
+		mm_cursor++;
+	}
+}
+
+void mmap_add_region(unsigned long long base_pa, uintptr_t base_va,
+		     size_t size, unsigned int attr)
+{
+	mmap_region_t *mm = mmap;
+	const mmap_region_t *mm_last = mm + ARRAY_SIZE(mmap) - 1U;
+	unsigned long long end_pa = base_pa + size - 1U;
+	uintptr_t end_va = base_va + size - 1U;
+
+	assert(IS_PAGE_ALIGNED(base_pa));
+	assert(IS_PAGE_ALIGNED(base_va));
+	assert(IS_PAGE_ALIGNED(size));
+
+	if (size == 0U)
+		return;
+
+	assert(base_pa < end_pa); /* Check for overflows */
+	assert(base_va < end_va);
+
+	assert((base_va + (uintptr_t)size - (uintptr_t)1) <=
+					(PLAT_VIRT_ADDR_SPACE_SIZE - 1U));
+	assert((base_pa + (unsigned long long)size - 1ULL) <=
+					(PLAT_PHY_ADDR_SPACE_SIZE - 1U));
+
+#if ENABLE_ASSERTIONS
+
+	/* Check for PAs and VAs overlaps with all other regions */
+	for (mm = mmap; mm->size; ++mm) {
+
+		uintptr_t mm_end_va = mm->base_va + mm->size - 1U;
+
+		/*
+		 * Check if one of the regions is completely inside the other
+		 * one.
+		 */
+		bool fully_overlapped_va =
+			((base_va >= mm->base_va) && (end_va <= mm_end_va)) ||
+			((mm->base_va >= base_va) && (mm_end_va <= end_va));
+
+		/*
+		 * Full VA overlaps are only allowed if both regions are
+		 * identity mapped (zero offset) or have the same VA to PA
+		 * offset. Also, make sure that it's not the exact same area.
+		 */
+		if (fully_overlapped_va) {
+			assert((mm->base_va - mm->base_pa) ==
+			       (base_va - base_pa));
+			assert((base_va != mm->base_va) || (size != mm->size));
+		} else {
+			/*
+			 * If the regions do not have fully overlapping VAs,
+			 * then they must have fully separated VAs and PAs.
+			 * Partial overlaps are not allowed
+			 */
+
+			unsigned long long mm_end_pa =
+						     mm->base_pa + mm->size - 1;
+
+			bool separated_pa = (end_pa < mm->base_pa) ||
+				(base_pa > mm_end_pa);
+			bool separated_va = (end_va < mm->base_va) ||
+				(base_va > mm_end_va);
+
+			assert(separated_va && separated_pa);
+		}
+	}
+
+	mm = mmap; /* Restore pointer to the start of the array */
+
+#endif /* ENABLE_ASSERTIONS */
+
+	/* Find correct place in mmap to insert new region */
+	while ((mm->base_va < base_va) && (mm->size != 0U))
+		++mm;
+
+	/*
+	 * If a section is contained inside another one with the same base
+	 * address, it must be placed after the one it is contained in:
+	 *
+	 * 1st |-----------------------|
+	 * 2nd |------------|
+	 * 3rd |------|
+	 *
+	 * This is required for mmap_region_attr() to get the attributes of the
+	 * small region correctly.
+	 */
+	while ((mm->base_va == base_va) && (mm->size > size))
+		++mm;
+
+	/* Make room for new region by moving other regions up by one place */
+	(void)memmove(mm + 1, mm, (uintptr_t)mm_last - (uintptr_t)mm);
+
+	/* Check we haven't lost the empty sentinal from the end of the array */
+	assert(mm_last->size == 0U);
+
+	mm->base_pa = base_pa;
+	mm->base_va = base_va;
+	mm->size = size;
+	mm->attr = attr;
+
+	if (end_pa > xlat_max_pa)
+		xlat_max_pa = end_pa;
+	if (end_va > xlat_max_va)
+		xlat_max_va = end_va;
+}
+
+/* map all memory as shared/global/domain0/no-usr access */
+static unsigned long mmap_desc(unsigned attr, unsigned long addr_pa,
+					unsigned int level)
+{
+	unsigned long desc;
+
+	switch (level) {
+	case 1:
+		assert(!(addr_pa & (MMU32B_L1_TABLE_ALIGN - 1)));
+
+		desc = SECTION_SECTION | SECTION_SHARED;
+
+		desc |= attr & MT_NS ? SECTION_NOTSECURE : 0;
+
+		desc |= SECTION_ACCESS_FLAG;
+		desc |= attr & MT_RW ? 0 : SECTION_RO;
+
+		desc |= attr & MT_MEMORY ?
+			SECTION_NORMAL_CACHED : SECTION_DEVICE;
+
+		if ((attr & MT_RW) || !(attr & MT_MEMORY))
+			desc |= SECTION_XN;
+		break;
+	case 2:
+		assert(!(addr_pa & (MMU32B_L2_TABLE_ALIGN - 1)));
+
+		desc = SMALL_PAGE_SMALL_PAGE | SMALL_PAGE_SHARED;
+
+		desc |= SMALL_PAGE_ACCESS_FLAG;
+		desc |= attr & MT_RW ? 0 : SMALL_PAGE_RO;
+
+		desc |= attr & MT_MEMORY ?
+			SMALL_PAGE_NORMAL_CACHED : SMALL_PAGE_DEVICE;
+
+		if ((attr & MT_RW) || !(attr & MT_MEMORY))
+			desc |= SMALL_PAGE_XN;
+		break;
+	default:
+		panic();
+	}
+#if LOG_LEVEL >= LOG_LEVEL_VERBOSE
+	/* dump only the non-lpae level 2 tables */
+	if (level == 2) {
+		printf(attr & MT_MEMORY ? "MEM" : "dev");
+		printf(attr & MT_RW ? "-rw" : "-RO");
+		printf(attr & MT_NS ? "-NS" : "-S");
+	}
+#endif
+	return desc | addr_pa;
+}
+
+static unsigned int mmap_region_attr(const mmap_region_t *mm, uintptr_t base_va,
+				     size_t size, unsigned int *attr)
+{
+	/* Don't assume that the area is contained in the first region */
+	unsigned int ret = MT_UNKNOWN;
+
+	/*
+	 * Get attributes from last (innermost) region that contains the
+	 * requested area. Don't stop as soon as one region doesn't contain it
+	 * because there may be other internal regions that contain this area:
+	 *
+	 * |-----------------------------1-----------------------------|
+	 * |----2----|     |-------3-------|    |----5----|
+	 *                   |--4--|
+	 *
+	 *                   |---| <- Area we want the attributes of.
+	 *
+	 * In this example, the area is contained in regions 1, 3 and 4 but not
+	 * in region 2. The loop shouldn't stop at region 2 as inner regions
+	 * have priority over outer regions, it should stop at region 5.
+	 */
+	for ( ; ; ++mm) {
+
+		if (mm->size == 0U)
+			return ret; /* Reached end of list */
+
+		if (mm->base_va > (base_va + size - 1U))
+			return ret; /* Next region is after area so end */
+
+		if ((mm->base_va + mm->size - 1U) < base_va)
+			continue; /* Next region has already been overtaken */
+
+		if ((ret == 0U) && (mm->attr == *attr))
+			continue; /* Region doesn't override attribs so skip */
+
+		if ((mm->base_va > base_va) ||
+			((mm->base_va + mm->size - 1U) < (base_va + size - 1U)))
+			return MT_UNKNOWN; /* Region doesn't fully cover area */
+
+		*attr = mm->attr;
+		ret = 0U;
+	}
+	return ret;
+}
+
+static mmap_region_t *init_xlation_table_inner(mmap_region_t *mm,
+						unsigned long base_va,
+						unsigned long *table,
+						unsigned int level)
+{
+	unsigned int level_size_shift = (level == 1) ?
+					ONE_MB_SHIFT : FOUR_KB_SHIFT;
+	unsigned int level_size = 1 << level_size_shift;
+	unsigned long level_index_mask = (level == 1) ?
+					(NUM_1MB_IN_4GB - 1) << ONE_MB_SHIFT :
+					(NUM_4K_IN_1MB - 1) << FOUR_KB_SHIFT;
+
+	assert(level == 1 || level == 2);
+
+	VERBOSE("init xlat table at %p (level%1d)\n", (void *)table, level);
+
+	do  {
+		unsigned long desc = MMU32B_UNSET_DESC;
+
+		if (mm->base_va + mm->size <= base_va) {
+			/* Area now after the region so skip it */
+			++mm;
+			continue;
+		}
+#if LOG_LEVEL >= LOG_LEVEL_VERBOSE
+		/* dump only non-lpae level 2 tables content */
+		if (level == 2)
+			printf("      0x%lx %x " + 6 - 2 * level,
+						base_va, level_size);
+#endif
+		if (mm->base_va >= base_va + level_size) {
+			/* Next region is after area so nothing to map yet */
+			desc = MMU32B_INVALID_DESC;
+		} else if (mm->base_va <= base_va && mm->base_va + mm->size >=
+				base_va + level_size) {
+			/* Next region covers all of area */
+			unsigned int attr = mm->attr;
+			unsigned int r = mmap_region_attr(mm, base_va,
+							  level_size, &attr);
+
+			if (r == 0U) {
+				desc = mmap_desc(attr,
+					base_va - mm->base_va + mm->base_pa,
+					level);
+			}
+		}
+
+		if (desc == MMU32B_UNSET_DESC) {
+			unsigned long xlat_table;
+
+			/*
+			 * Area not covered by a region so need finer table
+			 * Reuse next level table if any (assert attrib matching).
+			 * Otherwise allocate a xlat table.
+			 */
+			if (*table) {
+				assert((*table & 3) == SECTION_PT_PT);
+				assert(!(*table & SECTION_PT_NOTSECURE) ==
+							!(mm->attr & MT_NS));
+
+				xlat_table = (*table) &
+						~(MMU32B_L1_TABLE_ALIGN - 1);
+				desc = *table;
+			} else {
+				xlat_table = (unsigned long)mmu_l2_base +
+					next_xlat * MMU32B_L2_TABLE_SIZE;
+				assert(++next_xlat <= MAX_XLAT_TABLES);
+				memset((char *)xlat_table, 0,
+					MMU32B_L2_TABLE_SIZE);
+
+				desc = xlat_table | SECTION_PT_PT;
+				desc |= mm->attr & MT_NS ?
+					SECTION_PT_NOTSECURE : 0;
+			}
+			/* Recurse to fill in new table */
+			mm = init_xlation_table_inner(mm, base_va,
+						(unsigned long *)xlat_table,
+						level + 1);
+		}
+#if LOG_LEVEL >= LOG_LEVEL_VERBOSE
+		/* dump only non-lpae level 2 tables content */
+		if (level == 2)
+			printf("\n");
+#endif
+		*table++ = desc;
+		base_va += level_size;
+	} while (mm->size && (base_va & level_index_mask));
+
+	return mm;
+}
+
+void init_xlat_tables(void)
+{
+	print_mmap();
+
+	assert(!((unsigned int)mmu_l1_base & (MMU32B_L1_TABLE_ALIGN - 1)));
+	assert(!((unsigned int)mmu_l2_base & (MMU32B_L2_TABLE_ALIGN - 1)));
+
+	memset(mmu_l1_base, 0, MMU32B_L1_TABLE_SIZE);
+
+	init_xlation_table_inner(mmap, 0, (unsigned long *)mmu_l1_base, 1);
+
+	VERBOSE("init xlat - max_va=%p, max_pa=%llx\n",
+			(void *)xlat_max_va, xlat_max_pa);
+	assert(xlat_max_va <= PLAT_VIRT_ADDR_SPACE_SIZE - 1);
+	assert(xlat_max_pa <= PLAT_VIRT_ADDR_SPACE_SIZE - 1);
+}
+
+/*******************************************************************************
+ * Function for enabling the MMU in Secure PL1, assuming that the
+ * page-tables have already been created.
+ ******************************************************************************/
+void enable_mmu_svc_mon(unsigned int flags)
+{
+	unsigned int prrr;
+	unsigned int nmrr;
+	unsigned int sctlr;
+
+	assert(IS_IN_SECURE());
+	assert((read_sctlr() & SCTLR_M_BIT) == 0);
+
+	/* Enable Access flag (simplified access permissions) and TEX remap */
+	write_sctlr(read_sctlr() | SCTLR_AFE_BIT | SCTLR_TRE_BIT);
+
+	prrr = MMU32B_PRRR_IDX(MMU32B_ATTR_DEVICE_INDEX, 1, 0) \
+			| MMU32B_PRRR_IDX(MMU32B_ATTR_IWBWA_OWBWA_INDEX, 2, 1);
+	nmrr = MMU32B_NMRR_IDX(MMU32B_ATTR_DEVICE_INDEX, 0, 0) \
+			| MMU32B_NMRR_IDX(MMU32B_ATTR_IWBWA_OWBWA_INDEX, 1, 1);
+
+	prrr |= MMU32B_PRRR_NS1 | MMU32B_PRRR_DS1;
+
+	write_prrr(prrr);
+	write_nmrr(nmrr);
+
+	/* Program Domain access control register: domain 0 only */
+	write_dacr(DACR_DOMAIN(0, DACR_DOMAIN_PERM_CLIENT));
+
+	/* Invalidate TLBs at the current exception level */
+	tlbiall();
+
+	/* set MMU base xlat table entry (use only TTBR0) */
+	write_ttbr0((uint32_t)mmu_l1_base | MMU32B_DEFAULT_ATTRS);
+	write_ttbr1(0);
+
+	/*
+	 * Ensure all translation table writes have drained
+	 * into memory, the TLB invalidation is complete,
+	 * and translation register writes are committed
+	 * before enabling the MMU
+	 */
+	dsb();
+	isb();
+
+	sctlr = read_sctlr();
+	sctlr |= SCTLR_M_BIT;
+#if ARMV7_SUPPORTS_VIRTUALIZATION
+	sctlr |= SCTLR_WXN_BIT;
+#endif
+
+	if (flags & DISABLE_DCACHE)
+		sctlr &= ~SCTLR_C_BIT;
+	else
+		sctlr |= SCTLR_C_BIT;
+
+	write_sctlr(sctlr);
+
+	/* Ensure the MMU enable takes effect immediately */
+	isb();
+}