[PATCH] ppc32: fix ppc440 pagetable attributes

This patch fixes a bug in the PPC440 pagetable attributes that breaks swap
support.  It also adds some notes on the PPC440 attribute fields.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com> for CELF
Signed-off-by: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

1 files changed, 49 insertions, 3 deletions

diff --git a/include/asm-ppc/pgtable.h b/include/asm-ppc/pgtable.h
index 4d4b20c9de78..c41c7958ea1f 100644
--- a/include/asm-ppc/pgtable.h
+++ b/include/asm-ppc/pgtable.h
@@ -202,18 +202,64 @@ extern unsigned long ioremap_bot, ioremap_base;
  *
  * Note that these bits preclude future use of a page size
  * less than 4KB.
+ *
+ *
+ * PPC 440 core has following TLB attribute fields;
+ *
+ *   TLB1:
+ *   0  1  2  3  4  ... 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
+ *   RPN.................................  -  -  -  -  -  - ERPN.......
+ *
+ *   TLB2:
+ *   0  1  2  3  4  ... 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
+ *   -  -  -  -  -    - U0 U1 U2 U3 W  I  M  G  E   - UX UW UR SX SW SR
+ *
+ * There are some constrains and options, to decide mapping software bits
+ * into TLB entry.
+ *
+ *   - PRESENT *must* be in the bottom three bits because swap cache
+ *     entries use the top 29 bits for TLB2.
+ *
+ *   - FILE *must* be in the bottom three bits because swap cache
+ *     entries use the top 29 bits for TLB2.
+ *
+ *   - CACHE COHERENT bit (M) has no effect on PPC440 core, because it
+ *     doesn't support SMP. So we can use this as software bit, like
+ *     DIRTY.
+ *
+ * PPC Book-E Linux implementation uses PPC HW PTE bit field definition,
+ * even it doesn't have HW PTE. 0-11th LSB of PTE stand for memory
+ * protection-related function. (See PTE structure in include/asm-ppc/mmu.h)
+ * Definition of _PAGE_XXX in "include/asm-ppc/pagetable.h" stands for
+ * above bits. Note that those bits values are CPU dependent, not
+ * architecture.
+ *
+ * Kernel PTE entry holds arch-dependent swp_entry structure under certain
+ * situation. In other words, in such situation, some portion of PTE bits
+ * are used as swp_entry. In PPC implementation, 3-24th LSB are shared with
+ * swp_entry, however 0-2nd three LSB still hold protection values.
+ * That means three protection bits are reserved for both PTE and SWAP
+ * entry at the most three LSBs.
+ *
+ * There are three protection bits available for SWAP entry;
+ *      _PAGE_PRESENT
+ *      _PAGE_FILE
+ *      _PAGE_HASHPTE (if HW has)
+ *
+ * So those three bits have to be inside of 0-2nd LSB of PTE.
+ *
  */
+
 #define _PAGE_PRESENT   0x00000001              /* S: PTE valid */
 #define _PAGE_RW        0x00000002              /* S: Write permission */
-#define _PAGE_DIRTY     0x00000004              /* S: Page dirty */
+#define _PAGE_FILE      0x00000004              /* S: nonlinear file mapping */
 #define _PAGE_ACCESSED  0x00000008              /* S: Page referenced */
 #define _PAGE_HWWRITE   0x00000010              /* H: Dirty & RW */
 #define _PAGE_HWEXEC    0x00000020              /* H: Execute permission */
 #define _PAGE_USER      0x00000040              /* S: User page */
 #define _PAGE_ENDIAN    0x00000080              /* H: E bit */
 #define _PAGE_GUARDED   0x00000100              /* H: G bit */
-#define _PAGE_COHERENT  0x00000200              /* H: M bit */
-#define _PAGE_FILE      0x00000400              /* S: nonlinear file mapping */
+#define _PAGE_DIRTY     0x00000200              /* S: Page dirty */
 #define _PAGE_NO_CACHE  0x00000400              /* H: I bit */
 #define _PAGE_WRITETHRU 0x00000800              /* H: W bit */