[XTENSA] Add support for cache-aliasing

Add support for processors that have cache-aliasing issues, such as
the Stretch S5000 processor. Cache-aliasing means that the size of
the cache (for one way) is larger than the page size, thus, a page
can end up in several places in cache depending on the virtual to
physical translation. The method used here is to map a user page
temporarily through the auto-refill way 0 and of of the DTLB.
We probably will want to revisit this issue and use a better
approach with kmap/kunmap.

Signed-off-by: Chris Zankel <chris@zankel.net>

1 files changed, 29 insertions, 78 deletions

diff --git a/include/asm-xtensa/pgalloc.h b/include/asm-xtensa/pgalloc.h
index d56ddf2055e1..3e5b56525102 100644
--- a/include/asm-xtensa/pgalloc.h
+++ b/include/asm-xtensa/pgalloc.h
@@ -1,11 +1,11 @@
 /*
- * linux/include/asm-xtensa/pgalloc.h
+ * include/asm-xtensa/pgalloc.h
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
  *
- * Copyright (C) 2001-2005 Tensilica Inc.
+ * Copyright (C) 2001-2007 Tensilica Inc.
  */
 
 #ifndef _XTENSA_PGALLOC_H
@@ -13,103 +13,54 @@
 
 #ifdef __KERNEL__
 
-#include <linux/threads.h>
 #include <linux/highmem.h>
-#include <asm/processor.h>
-#include <asm/cacheflush.h>
-
-
-/* Cache aliasing:
- *
- * If the cache size for one way is greater than the page size, we have to
- * deal with cache aliasing. The cache index is wider than the page size:
- *
- *      |cache |
- * |pgnum |page|        virtual address
- * |xxxxxX|zzzz|
- * |      |    |
- *   \  / |    |
- *  trans.|    |
- *   /  \ |    |
- * |yyyyyY|zzzz|        physical address
- *
- * When the page number is translated to the physical page address, the lowest
- * bit(s) (X) that are also part of the cache index are also translated (Y).
- * If this translation changes this bit (X), the cache index is also afected,
- * thus resulting in a different cache line than before.
- * The kernel does not provide a mechanism to ensure that the page color
- * (represented by this bit) remains the same when allocated or when pages
- * are remapped. When user pages are mapped into kernel space, the color of
- * the page might also change.
- *
- * We use the address space VMALLOC_END ... VMALLOC_END + DCACHE_WAY_SIZE * 2
- * to temporarily map a patch so we can match the color.
- */
-
-#if (DCACHE_WAY_SIZE > PAGE_SIZE)
-# define PAGE_COLOR_MASK        (PAGE_MASK & (DCACHE_WAY_SIZE-1))
-# define PAGE_COLOR(a)          \
-        (((unsigned long)(a)&PAGE_COLOR_MASK) >> PAGE_SHIFT)
-# define PAGE_COLOR_EQ(a,b)     \
-        ((((unsigned long)(a) ^ (unsigned long)(b)) & PAGE_COLOR_MASK) == 0)
-# define PAGE_COLOR_MAP0(v)     \
-        (VMALLOC_END + ((unsigned long)(v) & PAGE_COLOR_MASK))
-# define PAGE_COLOR_MAP1(v)     \
-        (VMALLOC_END + ((unsigned long)(v) & PAGE_COLOR_MASK) + DCACHE_WAY_SIZE)
-#endif
 
 /*
  * Allocating and freeing a pmd is trivial: the 1-entry pmd is
  * inside the pgd, so has no extra memory associated with it.
  */
 
-#define pgd_free(pgd)   free_page((unsigned long)(pgd))
-
-#if (DCACHE_WAY_SIZE > PAGE_SIZE) && XCHAL_DCACHE_IS_WRITEBACK
+#define pmd_populate_kernel(mm, pmdp, ptep)                                  \
+        (pmd_val(*(pmdp)) = ((unsigned long)ptep))
+#define pmd_populate(mm, pmdp, page)                                         \
+        (pmd_val(*(pmdp)) = ((unsigned long)page_to_virt(page)))
 
-static inline void
-pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmdp, pte_t *pte)
+static inline pgd_t*
+pgd_alloc(struct mm_struct *mm)
 {
-        pmd_val(*(pmdp)) = (unsigned long)(pte);
-        __asm__ __volatile__ ("memw; dhwb %0, 0; dsync" :: "a" (pmdp));
+        return (pgd_t*) __get_free_pages(GFP_KERNEL | __GFP_ZERO, PGD_ORDER);
 }
 
-static inline void
-pmd_populate(struct mm_struct *mm, pmd_t *pmdp, struct page *page)
+static inline void pgd_free(pgd_t *pgd)
 {
-        pmd_val(*(pmdp)) = (unsigned long)page_to_virt(page);
-        __asm__ __volatile__ ("memw; dhwb %0, 0; dsync" :: "a" (pmdp));
+        free_page((unsigned long)pgd);
 }
 
+/* Use a slab cache for the pte pages (see also sparc64 implementation) */
 
+extern struct kmem_cache *pgtable_cache;
 
-#else
-
-# define pmd_populate_kernel(mm, pmdp, pte)                                  \
-        (pmd_val(*(pmdp)) = (unsigned long)(pte))
-# define pmd_populate(mm, pmdp, page)                                        \
-        (pmd_val(*(pmdp)) = (unsigned long)page_to_virt(page))
-
-#endif
-
-static inline pgd_t*
-pgd_alloc(struct mm_struct *mm)
+static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm, 
+                                         unsigned long address)
 {
-        pgd_t *pgd;
-
-        pgd = (pgd_t *)__get_free_pages(GFP_KERNEL|__GFP_ZERO, PGD_ORDER);
-
-        if (likely(pgd != NULL))
-                __flush_dcache_page((unsigned long)pgd);
+        return kmem_cache_alloc(pgtable_cache, GFP_KERNEL|__GFP_REPEAT);
+}
 
-        return pgd;
+static inline struct page *pte_alloc_one(struct mm_struct *mm, 
+                                         unsigned long addr)
+{
+        return virt_to_page(pte_alloc_one_kernel(mm, addr));
 }
 
-extern pte_t* pte_alloc_one_kernel(struct mm_struct* mm, unsigned long addr);
-extern struct page* pte_alloc_one(struct mm_struct* mm, unsigned long addr);
+static inline void pte_free_kernel(pte_t *pte)
+{
+        kmem_cache_free(pgtable_cache, pte);
+}
 
-#define pte_free_kernel(pte) free_page((unsigned long)pte)
-#define pte_free(pte) __free_page(pte)
+static inline void pte_free(struct page *page)
+{
+        kmem_cache_free(pgtable_cache, page_address(page));
+}
 
 #endif /* __KERNEL__ */
 #endif /* _XTENSA_PGALLOC_H */