powerpc/mm: Add support for early ioremap on non-hash 64-bit processors

This adds some code to do early ioremap's using page tables instead of
bolting entries in the hash table. This will be used by the upcoming
64-bits BookE port.

The patch also changes the test for early vs. late ioremap to use
slab_is_available() instead of our old hackish mem_init_done.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

diff --git a/arch/powerpc/mm/pgtable_64.c b/arch/powerpc/mm/pgtable_64.c
index bfa7db6b2fd5446e193669a1afd02d09f7f9aab9..93ed1a3c87295548fb346864c7a51b50b0d70ef4 100644 (file)
--- a/arch/powerpc/mm/pgtable_64.c
+++ b/arch/powerpc/mm/pgtable_64.c
@@ -33,6 +33,8 @@
 #include <linux/stddef.h>
 #include <linux/vmalloc.h>
 #include <linux/init.h>
+#include <linux/bootmem.h>
+#include <linux/lmb.h>
 
 #include <asm/pgalloc.h>
 #include <asm/page.h>
@@ -55,19 +57,36 @@
 
 unsigned long ioremap_bot = IOREMAP_BASE;
 
+
+#ifdef CONFIG_PPC_MMU_NOHASH
+static void *early_alloc_pgtable(unsigned long size)
+{
+       void *pt;
+
+       if (init_bootmem_done)
+               pt = __alloc_bootmem(size, size, __pa(MAX_DMA_ADDRESS));
+       else
+               pt = __va(lmb_alloc_base(size, size,
+                                        __pa(MAX_DMA_ADDRESS)));
+       memset(pt, 0, size);
+
+       return pt;
+}
+#endif /* CONFIG_PPC_MMU_NOHASH */
+
 /*
- * map_io_page currently only called by __ioremap
- * map_io_page adds an entry to the ioremap page table
+ * map_kernel_page currently only called by __ioremap
+ * map_kernel_page adds an entry to the ioremap page table
  * and adds an entry to the HPT, possibly bolting it
  */
-static int map_io_page(unsigned long ea, unsigned long pa, int flags)
+static int map_kernel_page(unsigned long ea, unsigned long pa, int flags)
 {
        pgd_t *pgdp;
        pud_t *pudp;
        pmd_t *pmdp;
        pte_t *ptep;
 
-       if (mem_init_done) {
+       if (slab_is_available()) {
                pgdp = pgd_offset_k(ea);
                pudp = pud_alloc(&init_mm, pgdp, ea);
                if (!pudp)
@@ -81,6 +100,35 @@ static int map_io_page(unsigned long ea, unsigned long pa, int flags)
                set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
                                                          __pgprot(flags)));
        } else {
+#ifdef CONFIG_PPC_MMU_NOHASH
+               /* Warning ! This will blow up if bootmem is not initialized
+                * which our ppc64 code is keen to do that, we'll need to
+                * fix it and/or be more careful
+                */
+               pgdp = pgd_offset_k(ea);
+#ifdef PUD_TABLE_SIZE
+               if (pgd_none(*pgdp)) {
+                       pudp = early_alloc_pgtable(PUD_TABLE_SIZE);
+                       BUG_ON(pudp == NULL);
+                       pgd_populate(&init_mm, pgdp, pudp);
+               }
+#endif /* PUD_TABLE_SIZE */
+               pudp = pud_offset(pgdp, ea);
+               if (pud_none(*pudp)) {
+                       pmdp = early_alloc_pgtable(PMD_TABLE_SIZE);
+                       BUG_ON(pmdp == NULL);
+                       pud_populate(&init_mm, pudp, pmdp);
+               }
+               pmdp = pmd_offset(pudp, ea);
+               if (!pmd_present(*pmdp)) {
+                       ptep = early_alloc_pgtable(PAGE_SIZE);
+                       BUG_ON(ptep == NULL);
+                       pmd_populate_kernel(&init_mm, pmdp, ptep);
+               }
+               ptep = pte_offset_kernel(pmdp, ea);
+               set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
+                                                         __pgprot(flags)));
+#else /* CONFIG_PPC_MMU_NOHASH */
                /*
                 * If the mm subsystem is not fully up, we cannot create a
                 * linux page table entry for this mapping.  Simply bolt an
@@ -93,6 +141,7 @@ static int map_io_page(unsigned long ea, unsigned long pa, int flags)
                               "memory at %016lx !\n", pa);
                        return -ENOMEM;
                }
+#endif /* !CONFIG_PPC_MMU_NOHASH */
        }
        return 0;
 }
@@ -124,7 +173,7 @@ void __iomem * __ioremap_at(phys_addr_t pa, void *ea, unsigned long size,
        WARN_ON(size & ~PAGE_MASK);
 
        for (i = 0; i < size; i += PAGE_SIZE)
-               if (map_io_page((unsigned long)ea+i, pa+i, flags))
+               if (map_kernel_page((unsigned long)ea+i, pa+i, flags))
                        return NULL;
 
        return (void __iomem *)ea;