#include <linux/types.h>
 #include <linux/highmem.h>
 #include <linux/dma-mapping.h>
-#include <linux/vmalloc.h>
 
 #include <asm/tlbflush.h>
 
+/*
+ * This address range defaults to a value that is safe for all
+ * platforms which currently set CONFIG_NOT_COHERENT_CACHE. It
+ * can be further configured for specific applications under
+ * the "Advanced Setup" menu. -Matt
+ */
+#define CONSISTENT_BASE        (CONFIG_CONSISTENT_START)
+#define CONSISTENT_END (CONFIG_CONSISTENT_START + CONFIG_CONSISTENT_SIZE)
+#define CONSISTENT_OFFSET(x)   (((unsigned long)(x) - CONSISTENT_BASE) >> PAGE_SHIFT)
+
+/*
+ * This is the page table (2MB) covering uncached, DMA consistent allocations
+ */
+static pte_t *consistent_pte;
+static DEFINE_SPINLOCK(consistent_lock);
+
+/*
+ * VM region handling support.
+ *
+ * This should become something generic, handling VM region allocations for
+ * vmalloc and similar (ioremap, module space, etc).
+ *
+ * I envisage vmalloc()'s supporting vm_struct becoming:
+ *
+ *  struct vm_struct {
+ *    struct vm_region region;
+ *    unsigned long    flags;
+ *    struct page      **pages;
+ *    unsigned int     nr_pages;
+ *    unsigned long    phys_addr;
+ *  };
+ *
+ * get_vm_area() would then call vm_region_alloc with an appropriate
+ * struct vm_region head (eg):
+ *
+ *  struct vm_region vmalloc_head = {
+ *     .vm_list        = LIST_HEAD_INIT(vmalloc_head.vm_list),
+ *     .vm_start       = VMALLOC_START,
+ *     .vm_end         = VMALLOC_END,
+ *  };
+ *
+ * However, vmalloc_head.vm_start is variable (typically, it is dependent on
+ * the amount of RAM found at boot time.)  I would imagine that get_vm_area()
+ * would have to initialise this each time prior to calling vm_region_alloc().
+ */
+struct ppc_vm_region {
+       struct list_head        vm_list;
+       unsigned long           vm_start;
+       unsigned long           vm_end;
+};
+
+static struct ppc_vm_region consistent_head = {
+       .vm_list        = LIST_HEAD_INIT(consistent_head.vm_list),
+       .vm_start       = CONSISTENT_BASE,
+       .vm_end         = CONSISTENT_END,
+};
+
+static struct ppc_vm_region *
+ppc_vm_region_alloc(struct ppc_vm_region *head, size_t size, gfp_t gfp)
+{
+       unsigned long addr = head->vm_start, end = head->vm_end - size;
+       unsigned long flags;
+       struct ppc_vm_region *c, *new;
+
+       new = kmalloc(sizeof(struct ppc_vm_region), gfp);
+       if (!new)
+               goto out;
+
+       spin_lock_irqsave(&consistent_lock, flags);
+
+       list_for_each_entry(c, &head->vm_list, vm_list) {
+               if ((addr + size) < addr)
+                       goto nospc;
+               if ((addr + size) <= c->vm_start)
+                       goto found;
+               addr = c->vm_end;
+               if (addr > end)
+                       goto nospc;
+       }
+
+ found:
+       /*
+        * Insert this entry _before_ the one we found.
+        */
+       list_add_tail(&new->vm_list, &c->vm_list);
+       new->vm_start = addr;
+       new->vm_end = addr + size;
+
+       spin_unlock_irqrestore(&consistent_lock, flags);
+       return new;
+
+ nospc:
+       spin_unlock_irqrestore(&consistent_lock, flags);
+       kfree(new);
+ out:
+       return NULL;
+}
+
+static struct ppc_vm_region *ppc_vm_region_find(struct ppc_vm_region *head, unsigned long addr)
+{
+       struct ppc_vm_region *c;
+
+       list_for_each_entry(c, &head->vm_list, vm_list) {
+               if (c->vm_start == addr)
+                       goto out;
+       }
+       c = NULL;
+ out:
+       return c;
+}
+
 /*
  * Allocate DMA-coherent memory space and return both the kernel remapped
  * virtual and bus address for that space.
 __dma_alloc_coherent(size_t size, dma_addr_t *handle, gfp_t gfp)
 {
        struct page *page;
+       struct ppc_vm_region *c;
        unsigned long order;
-       int i;
-       unsigned int nr_pages = PAGE_ALIGN(size)>>PAGE_SHIFT;
-       unsigned int array_size = nr_pages * sizeof(struct page *);
-       struct page **pages;
-       struct page *end;
        u64 mask = 0x00ffffff, limit; /* ISA default */
-       struct vm_struct *area;
 
-       BUG_ON(!mem_init_done);
+       if (!consistent_pte) {
+               printk(KERN_ERR "%s: not initialised\n", __func__);
+               dump_stack();
+               return NULL;
+       }
+
        size = PAGE_ALIGN(size);
        limit = (mask + 1) & ~mask;
-       if (limit && size >= limit) {
-               printk(KERN_WARNING "coherent allocation too big (requested "
-                               "%#x mask %#Lx)\n", size, mask);
+       if ((limit && size >= limit) || size >= (CONSISTENT_END - CONSISTENT_BASE)) {
+               printk(KERN_WARNING "coherent allocation too big (requested %#x mask %#Lx)\n",
+                      size, mask);
                return NULL;
        }
 
        if (!page)
                goto no_page;
 
-       end = page + (1 << order);
-
        /*
         * Invalidate any data that might be lurking in the
         * kernel direct-mapped region for device DMA.
                flush_dcache_range(kaddr, kaddr + size);
        }
 
-       split_page(page, order);
-
        /*
-        * Set the "dma handle"
+        * Allocate a virtual address in the consistent mapping region.
         */
-       *handle = page_to_phys(page);
-
-       area = get_vm_area_caller(size, VM_IOREMAP,
-                       __builtin_return_address(1));
-       if (!area)
-               goto out_free_pages;
-
-       if (array_size > PAGE_SIZE) {
-               pages = vmalloc(array_size);
-               area->flags |= VM_VPAGES;
-       } else {
-               pages = kmalloc(array_size, GFP_KERNEL);
-       }
-       if (!pages)
-               goto out_free_area;
+       c = ppc_vm_region_alloc(&consistent_head, size,
+                           gfp & ~(__GFP_DMA | __GFP_HIGHMEM));
+       if (c) {
+               unsigned long vaddr = c->vm_start;
+               pte_t *pte = consistent_pte + CONSISTENT_OFFSET(vaddr);
+               struct page *end = page + (1 << order);
 
-       area->pages = pages;
-       area->nr_pages = nr_pages;
+               split_page(page, order);
 
-       for (i = 0; i < nr_pages; i++)
-               pages[i] = page + i;
+               /*
+                * Set the "dma handle"
+                */
+               *handle = page_to_phys(page);
 
-       if (map_vm_area(area, pgprot_noncached(PAGE_KERNEL), &pages))
-               goto out_unmap;
+               do {
+                       BUG_ON(!pte_none(*pte));
 
-       /*
-        * Free the otherwise unused pages.
-        */
-       page += nr_pages;
-       while (page < end) {
-               __free_page(page);
-               page++;
+                       SetPageReserved(page);
+                       set_pte_at(&init_mm, vaddr,
+                                  pte, mk_pte(page, pgprot_noncached(PAGE_KERNEL)));
+                       page++;
+                       pte++;
+                       vaddr += PAGE_SIZE;
+               } while (size -= PAGE_SIZE);
+
+               /*
+                * Free the otherwise unused pages.
+                */
+               while (page < end) {
+                       __free_page(page);
+                       page++;
+               }
+
+               return (void *)c->vm_start;
        }
 
-       return area->addr;
-out_unmap:
-       vunmap(area->addr);
-       if (array_size > PAGE_SIZE)
-               vfree(pages);
-       else
-               kfree(pages);
-       goto out_free_pages;
-out_free_area:
-       free_vm_area(area);
-out_free_pages:
        if (page)
                __free_pages(page, order);
-no_page:
+ no_page:
        return NULL;
 }
 EXPORT_SYMBOL(__dma_alloc_coherent);
  */
 void __dma_free_coherent(size_t size, void *vaddr)
 {
-       vfree(vaddr);
+       struct ppc_vm_region *c;
+       unsigned long flags, addr;
+       pte_t *ptep;
+
+       size = PAGE_ALIGN(size);
+
+       spin_lock_irqsave(&consistent_lock, flags);
+
+       c = ppc_vm_region_find(&consistent_head, (unsigned long)vaddr);
+       if (!c)
+               goto no_area;
+
+       if ((c->vm_end - c->vm_start) != size) {
+               printk(KERN_ERR "%s: freeing wrong coherent size (%ld != %d)\n",
+                      __func__, c->vm_end - c->vm_start, size);
+               dump_stack();
+               size = c->vm_end - c->vm_start;
+       }
+
+       ptep = consistent_pte + CONSISTENT_OFFSET(c->vm_start);
+       addr = c->vm_start;
+       do {
+               pte_t pte = ptep_get_and_clear(&init_mm, addr, ptep);
+               unsigned long pfn;
+
+               ptep++;
+               addr += PAGE_SIZE;
 
+               if (!pte_none(pte) && pte_present(pte)) {
+                       pfn = pte_pfn(pte);
+
+                       if (pfn_valid(pfn)) {
+                               struct page *page = pfn_to_page(pfn);
+                               ClearPageReserved(page);
+
+                               __free_page(page);
+                               continue;
+                       }
+               }
+
+               printk(KERN_CRIT "%s: bad page in kernel page table\n",
+                      __func__);
+       } while (size -= PAGE_SIZE);
+
+       flush_tlb_kernel_range(c->vm_start, c->vm_end);
+
+       list_del(&c->vm_list);
+
+       spin_unlock_irqrestore(&consistent_lock, flags);
+
+       kfree(c);
+       return;
+
+ no_area:
+       spin_unlock_irqrestore(&consistent_lock, flags);
+       printk(KERN_ERR "%s: trying to free invalid coherent area: %p\n",
+              __func__, vaddr);
+       dump_stack();
 }
 EXPORT_SYMBOL(__dma_free_coherent);
 
+/*
+ * Initialise the consistent memory allocation.
+ */
+static int __init dma_alloc_init(void)
+{
+       pgd_t *pgd;
+       pud_t *pud;
+       pmd_t *pmd;
+       pte_t *pte;
+       int ret = 0;
+
+       do {
+               pgd = pgd_offset(&init_mm, CONSISTENT_BASE);
+               pud = pud_alloc(&init_mm, pgd, CONSISTENT_BASE);
+               pmd = pmd_alloc(&init_mm, pud, CONSISTENT_BASE);
+               if (!pmd) {
+                       printk(KERN_ERR "%s: no pmd tables\n", __func__);
+                       ret = -ENOMEM;
+                       break;
+               }
+
+               pte = pte_alloc_kernel(pmd, CONSISTENT_BASE);
+               if (!pte) {
+                       printk(KERN_ERR "%s: no pte tables\n", __func__);
+                       ret = -ENOMEM;
+                       break;
+               }
+
+               consistent_pte = pte;
+       } while (0);
+
+       return ret;
+}
+
+core_initcall(dma_alloc_init);
+
 /*
  * make an area consistent.
  */