X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=arch%2Farm%2Fmm%2Fioremap.c;h=0ac615c0f7987f81ccc7aa3a16c8ace113f65192;hb=refs%2Fheads%2Fvserver;hp=da9b35974118d34d2b6fac17bb246306a6e19228;hpb=76828883507a47dae78837ab5dec5a5b4513c667;p=linux-2.6.git

diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
index da9b35974..0ac615c0f 100644
--- a/arch/arm/mm/ioremap.c
+++ b/arch/arm/mm/ioremap.c
@@ -26,97 +26,245 @@
 #include <linux/vmalloc.h>
 
 #include <asm/cacheflush.h>
-#include <asm/hardware.h>
 #include <asm/io.h>
+#include <asm/mmu_context.h>
+#include <asm/pgalloc.h>
 #include <asm/tlbflush.h>
+#include <asm/sizes.h>
 
-static inline void
-remap_area_pte(pte_t * pte, unsigned long address, unsigned long size,
-	       unsigned long phys_addr, pgprot_t pgprot)
+/*
+ * Used by ioremap() and iounmap() code to mark (super)section-mapped
+ * I/O regions in vm_struct->flags field.
+ */
+#define VM_ARM_SECTION_MAPPING	0x80000000
+
+static int remap_area_pte(pmd_t *pmd, unsigned long addr, unsigned long end,
+			  unsigned long phys_addr, pgprot_t prot)
 {
-	unsigned long end;
+	pte_t *pte;
+
+	pte = pte_alloc_kernel(pmd, addr);
+	if (!pte)
+		return -ENOMEM;
 
-	address &= ~PMD_MASK;
-	end = address + size;
-	if (end > PMD_SIZE)
-		end = PMD_SIZE;
-	BUG_ON(address >= end);
 	do {
 		if (!pte_none(*pte))
 			goto bad;
 
-		set_pte(pte, pfn_pte(phys_addr >> PAGE_SHIFT, pgprot));
-		address += PAGE_SIZE;
+		set_pte_ext(pte, pfn_pte(phys_addr >> PAGE_SHIFT, prot), 0);
 		phys_addr += PAGE_SIZE;
-		pte++;
-	} while (address && (address < end));
-	return;
+	} while (pte++, addr += PAGE_SIZE, addr != end);
+	return 0;
 
  bad:
-	printk("remap_area_pte: page already exists\n");
+	printk(KERN_CRIT "remap_area_pte: page already exists\n");
 	BUG();
 }
 
-static inline int
-remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size,
-	       unsigned long phys_addr, unsigned long flags)
+static inline int remap_area_pmd(pgd_t *pgd, unsigned long addr,
+				 unsigned long end, unsigned long phys_addr,
+				 pgprot_t prot)
 {
-	unsigned long end;
-	pgprot_t pgprot;
-
-	address &= ~PGDIR_MASK;
-	end = address + size;
-
-	if (end > PGDIR_SIZE)
-		end = PGDIR_SIZE;
+	unsigned long next;
+	pmd_t *pmd;
+	int ret = 0;
 
-	phys_addr -= address;
-	BUG_ON(address >= end);
+	pmd = pmd_alloc(&init_mm, pgd, addr);
+	if (!pmd)
+		return -ENOMEM;
 
-	pgprot = __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_DIRTY | L_PTE_WRITE | flags);
 	do {
-		pte_t * pte = pte_alloc_kernel(pmd, address);
-		if (!pte)
-			return -ENOMEM;
-		remap_area_pte(pte, address, end - address, address + phys_addr, pgprot);
-		address = (address + PMD_SIZE) & PMD_MASK;
-		pmd++;
-	} while (address && (address < end));
-	return 0;
+		next = pmd_addr_end(addr, end);
+		ret = remap_area_pte(pmd, addr, next, phys_addr, prot);
+		if (ret)
+			return ret;
+		phys_addr += next - addr;
+	} while (pmd++, addr = next, addr != end);
+	return ret;
 }
 
-static int
-remap_area_pages(unsigned long start, unsigned long pfn,
-		 unsigned long size, unsigned long flags)
+static int remap_area_pages(unsigned long start, unsigned long pfn,
+			    unsigned long size, unsigned long flags)
 {
-	unsigned long address = start;
-	unsigned long end = start + size;
+	unsigned long addr = start;
+	unsigned long next, end = start + size;
 	unsigned long phys_addr = __pfn_to_phys(pfn);
+	pgprot_t prot = __pgprot(L_PTE_PRESENT | L_PTE_YOUNG |
+				 L_PTE_DIRTY | L_PTE_WRITE | flags);
+	pgd_t *pgd;
 	int err = 0;
-	pgd_t * dir;
 
-	phys_addr -= address;
-	dir = pgd_offset(&init_mm, address);
-	BUG_ON(address >= end);
+	BUG_ON(addr >= end);
+	pgd = pgd_offset_k(addr);
 	do {
-		pmd_t *pmd = pmd_alloc(&init_mm, dir, address);
-		if (!pmd) {
-			err = -ENOMEM;
+		next = pgd_addr_end(addr, end);
+		err = remap_area_pmd(pgd, addr, next, phys_addr, prot);
+		if (err)
 			break;
+		phys_addr += next - addr;
+	} while (pgd++, addr = next, addr != end);
+
+	return err;
+}
+
+
+void __check_kvm_seq(struct mm_struct *mm)
+{
+	unsigned int seq;
+
+	do {
+		seq = init_mm.context.kvm_seq;
+		memcpy(pgd_offset(mm, VMALLOC_START),
+		       pgd_offset_k(VMALLOC_START),
+		       sizeof(pgd_t) * (pgd_index(VMALLOC_END) -
+					pgd_index(VMALLOC_START)));
+		mm->context.kvm_seq = seq;
+	} while (seq != init_mm.context.kvm_seq);
+}
+
+#ifndef CONFIG_SMP
+/*
+ * Section support is unsafe on SMP - If you iounmap and ioremap a region,
+ * the other CPUs will not see this change until their next context switch.
+ * Meanwhile, (eg) if an interrupt comes in on one of those other CPUs
+ * which requires the new ioremap'd region to be referenced, the CPU will
+ * reference the _old_ region.
+ *
+ * Note that get_vm_area() allocates a guard 4K page, so we need to mask
+ * the size back to 1MB aligned or we will overflow in the loop below.
+ */
+static void unmap_area_sections(unsigned long virt, unsigned long size)
+{
+	unsigned long addr = virt, end = virt + (size & ~SZ_1M);
+	pgd_t *pgd;
+
+	flush_cache_vunmap(addr, end);
+	pgd = pgd_offset_k(addr);
+	do {
+		pmd_t pmd, *pmdp = pmd_offset(pgd, addr);
+
+		pmd = *pmdp;
+		if (!pmd_none(pmd)) {
+			/*
+			 * Clear the PMD from the page table, and
+			 * increment the kvm sequence so others
+			 * notice this change.
+			 *
+			 * Note: this is still racy on SMP machines.
+			 */
+			pmd_clear(pmdp);
+			init_mm.context.kvm_seq++;
+
+			/*
+			 * Free the page table, if there was one.
+			 */
+			if ((pmd_val(pmd) & PMD_TYPE_MASK) == PMD_TYPE_TABLE)
+				pte_free_kernel(pmd_page_vaddr(pmd));
 		}
-		if (remap_area_pmd(pmd, address, end - address,
-					 phys_addr + address, flags)) {
-			err = -ENOMEM;
-			break;
+
+		addr += PGDIR_SIZE;
+		pgd++;
+	} while (addr < end);
+
+	/*
+	 * Ensure that the active_mm is up to date - we want to
+	 * catch any use-after-iounmap cases.
+	 */
+	if (current->active_mm->context.kvm_seq != init_mm.context.kvm_seq)
+		__check_kvm_seq(current->active_mm);
+
+	flush_tlb_kernel_range(virt, end);
+}
+
+static int
+remap_area_sections(unsigned long virt, unsigned long pfn,
+		    unsigned long size, unsigned long flags)
+{
+	unsigned long prot, addr = virt, end = virt + size;
+	pgd_t *pgd;
+
+	/*
+	 * Remove and free any PTE-based mapping, and
+	 * sync the current kernel mapping.
+	 */
+	unmap_area_sections(virt, size);
+
+	prot = PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_DOMAIN(DOMAIN_IO) |
+	       (flags & (L_PTE_CACHEABLE | L_PTE_BUFFERABLE));
+
+	/*
+	 * ARMv6 and above need XN set to prevent speculative prefetches
+	 * hitting IO.
+	 */
+	if (cpu_architecture() >= CPU_ARCH_ARMv6)
+		prot |= PMD_SECT_XN;
+
+	pgd = pgd_offset_k(addr);
+	do {
+		pmd_t *pmd = pmd_offset(pgd, addr);
+
+		pmd[0] = __pmd(__pfn_to_phys(pfn) | prot);
+		pfn += SZ_1M >> PAGE_SHIFT;
+		pmd[1] = __pmd(__pfn_to_phys(pfn) | prot);
+		pfn += SZ_1M >> PAGE_SHIFT;
+		flush_pmd_entry(pmd);
+
+		addr += PGDIR_SIZE;
+		pgd++;
+	} while (addr < end);
+
+	return 0;
+}
+
+static int
+remap_area_supersections(unsigned long virt, unsigned long pfn,
+			 unsigned long size, unsigned long flags)
+{
+	unsigned long prot, addr = virt, end = virt + size;
+	pgd_t *pgd;
+
+	/*
+	 * Remove and free any PTE-based mapping, and
+	 * sync the current kernel mapping.
+	 */
+	unmap_area_sections(virt, size);
+
+	prot = PMD_TYPE_SECT | PMD_SECT_SUPER | PMD_SECT_AP_WRITE |
+			PMD_DOMAIN(DOMAIN_IO) |
+			(flags & (L_PTE_CACHEABLE | L_PTE_BUFFERABLE));
+
+	/*
+	 * ARMv6 and above need XN set to prevent speculative prefetches
+	 * hitting IO.
+	 */
+	if (cpu_architecture() >= CPU_ARCH_ARMv6)
+		prot |= PMD_SECT_XN;
+
+	pgd = pgd_offset_k(virt);
+	do {
+		unsigned long super_pmd_val, i;
+
+		super_pmd_val = __pfn_to_phys(pfn) | prot;
+		super_pmd_val |= ((pfn >> (32 - PAGE_SHIFT)) & 0xf) << 20;
+
+		for (i = 0; i < 8; i++) {
+			pmd_t *pmd = pmd_offset(pgd, addr);
+
+			pmd[0] = __pmd(super_pmd_val);
+			pmd[1] = __pmd(super_pmd_val);
+			flush_pmd_entry(pmd);
+
+			addr += PGDIR_SIZE;
+			pgd++;
 		}
 
-		address = (address + PGDIR_SIZE) & PGDIR_MASK;
-		dir++;
-	} while (address && (address < end));
+		pfn += SUPERSECTION_SIZE >> PAGE_SHIFT;
+	} while (addr < end);
 
-	flush_cache_vmap(start, end);
-	return err;
+	return 0;
 }
+#endif
+
 
 /*
  * Remap an arbitrary physical address space into the kernel virtual
@@ -134,18 +282,44 @@ void __iomem *
 __ioremap_pfn(unsigned long pfn, unsigned long offset, size_t size,
 	      unsigned long flags)
 {
+	int err;
 	unsigned long addr;
  	struct vm_struct * area;
 
+	/*
+	 * High mappings must be supersection aligned
+	 */
+	if (pfn >= 0x100000 && (__pfn_to_phys(pfn) & ~SUPERSECTION_MASK))
+		return NULL;
+
+	size = PAGE_ALIGN(size);
+
  	area = get_vm_area(size, VM_IOREMAP);
  	if (!area)
  		return NULL;
  	addr = (unsigned long)area->addr;
- 	if (remap_area_pages(addr, pfn, size, flags)) {
- 		vfree((void *)addr);
+
+#ifndef CONFIG_SMP
+	if (DOMAIN_IO == 0 &&
+	    (((cpu_architecture() >= CPU_ARCH_ARMv6) && (get_cr() & CR_XP)) ||
+	       cpu_is_xsc3()) &&
+	       !((__pfn_to_phys(pfn) | size | addr) & ~SUPERSECTION_MASK)) {
+		area->flags |= VM_ARM_SECTION_MAPPING;
+		err = remap_area_supersections(addr, pfn, size, flags);
+	} else if (!((__pfn_to_phys(pfn) | size | addr) & ~PMD_MASK)) {
+		area->flags |= VM_ARM_SECTION_MAPPING;
+		err = remap_area_sections(addr, pfn, size, flags);
+	} else
+#endif
+		err = remap_area_pages(addr, pfn, size, flags);
+
+	if (err) {
+ 		vunmap((void *)addr);
  		return NULL;
  	}
- 	return (void __iomem *) (offset + (char *)addr);
+
+	flush_cache_vmap(addr, addr + size);
+	return (void __iomem *) (offset + addr);
 }
 EXPORT_SYMBOL(__ioremap_pfn);
 
@@ -172,55 +346,40 @@ __ioremap(unsigned long phys_addr, size_t size, unsigned long flags)
 }
 EXPORT_SYMBOL(__ioremap);
 
-void __iounmap(void __iomem *addr)
+void __iounmap(volatile void __iomem *addr)
 {
-	vfree((void *) (PAGE_MASK & (unsigned long) addr));
-}
-EXPORT_SYMBOL(__iounmap);
-
-#ifdef __io
-void __iomem *ioport_map(unsigned long port, unsigned int nr)
-{
-	return __io(port);
-}
-EXPORT_SYMBOL(ioport_map);
-
-void ioport_unmap(void __iomem *addr)
-{
-}
-EXPORT_SYMBOL(ioport_unmap);
+#ifndef CONFIG_SMP
+	struct vm_struct **p, *tmp;
 #endif
+	unsigned int section_mapping = 0;
 
-#ifdef CONFIG_PCI
-#include <linux/pci.h>
-#include <linux/ioport.h>
+	addr = (volatile void __iomem *)(PAGE_MASK & (unsigned long)addr);
 
-void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long maxlen)
-{
-	unsigned long start = pci_resource_start(dev, bar);
-	unsigned long len   = pci_resource_len(dev, bar);
-	unsigned long flags = pci_resource_flags(dev, bar);
-
-	if (!len || !start)
-		return NULL;
-	if (maxlen && len > maxlen)
-		len = maxlen;
-	if (flags & IORESOURCE_IO)
-		return ioport_map(start, len);
-	if (flags & IORESOURCE_MEM) {
-		if (flags & IORESOURCE_CACHEABLE)
-			return ioremap(start, len);
-		return ioremap_nocache(start, len);
+#ifndef CONFIG_SMP
+	/*
+	 * If this is a section based mapping we need to handle it
+	 * specially as the VM subysystem does not know how to handle
+	 * such a beast. We need the lock here b/c we need to clear
+	 * all the mappings before the area can be reclaimed
+	 * by someone else.
+	 */
+	write_lock(&vmlist_lock);
+	for (p = &vmlist ; (tmp = *p) ; p = &tmp->next) {
+		if((tmp->flags & VM_IOREMAP) && (tmp->addr == addr)) {
+			if (tmp->flags & VM_ARM_SECTION_MAPPING) {
+				*p = tmp->next;
+				unmap_area_sections((unsigned long)tmp->addr,
+						    tmp->size);
+				kfree(tmp);
+				section_mapping = 1;
+			}
+			break;
+		}
 	}
-	return NULL;
-}
-EXPORT_SYMBOL(pci_iomap);
+	write_unlock(&vmlist_lock);
+#endif
 
-void pci_iounmap(struct pci_dev *dev, void __iomem *addr)
-{
-	if ((unsigned long)addr >= VMALLOC_START &&
-	    (unsigned long)addr < VMALLOC_END)
-		iounmap(addr);
+	if (!section_mapping)
+		vunmap((void __force *)addr);
 }
-EXPORT_SYMBOL(pci_iounmap);
-#endif
+EXPORT_SYMBOL(__iounmap);