From: Matt Porter Date: Fri, 5 Aug 2005 23:10:10 +0000 (-0700) Subject: [PATCH] ppc32: fix ppc440 pagetable attributes X-Git-Tag: v2.6.13-rc6~25 X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=534afb90a9cd0b9643f62d660c164e1d924f39cf;p=linux-2.6 [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 for CELF Signed-off-by: Matt Porter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/include/asm-ppc/pgtable.h b/include/asm-ppc/pgtable.h index 4d4b20c9de..c41c7958ea 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 */