[MIPS] Unify memset.S

The 32-bit version and 64-bit version are almost equal.  Unify them.
This makes further improvements (for example, supporting CDEX, etc.)
easier.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

diff --git a/arch/mips/lib-32/Makefile b/arch/mips/lib-32/Makefile
index dcd4d2ed2ac45737e148cfd3a5d74460f5ff1cac..2036cf5e6857fe924e3bd934b9a01535afacdec5 100644 (file)
--- a/arch/mips/lib-32/Makefile
+++ b/arch/mips/lib-32/Makefile
@@ -2,7 +2,7 @@
 # Makefile for MIPS-specific library files..
 #
 
-lib-y  += memset.o watch.o
+lib-y  += watch.o
 
 obj-$(CONFIG_CPU_MIPS32)       += dump_tlb.o
 obj-$(CONFIG_CPU_MIPS64)       += dump_tlb.o

diff --git a/arch/mips/lib-64/Makefile b/arch/mips/lib-64/Makefile
index dcd4d2ed2ac45737e148cfd3a5d74460f5ff1cac..2036cf5e6857fe924e3bd934b9a01535afacdec5 100644 (file)
--- a/arch/mips/lib-64/Makefile
+++ b/arch/mips/lib-64/Makefile
@@ -2,7 +2,7 @@
 # Makefile for MIPS-specific library files..
 #
 
-lib-y  += memset.o watch.o
+lib-y  += watch.o
 
 obj-$(CONFIG_CPU_MIPS32)       += dump_tlb.o
 obj-$(CONFIG_CPU_MIPS64)       += dump_tlb.o

diff --git a/arch/mips/lib-64/memset.S b/arch/mips/lib-64/memset.S
deleted file mode 100644 (file)
index e2c42c8..0000000
--- a/arch/mips/lib-64/memset.S
+++ /dev/null
@@ -1,142 +0,0 @@
-/*
- * This file is subject to the terms and conditions of the GNU General Public
- * License.  See the file "COPYING" in the main directory of this archive
- * for more details.
- *
- * Copyright (C) 1998, 1999, 2000 by Ralf Baechle
- * Copyright (C) 1999, 2000 Silicon Graphics, Inc.
- */
-#include <asm/asm.h>
-#include <asm/asm-offsets.h>
-#include <asm/regdef.h>
-
-#define EX(insn,reg,addr,handler)                      \
-9:     insn    reg, addr;                              \
-       .section __ex_table,"a";                        \
-       PTR     9b, handler;                            \
-       .previous
-
-       .macro  f_fill64 dst, offset, val, fixup
-       EX(LONG_S, \val, (\offset +  0 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  1 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  2 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  3 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  4 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  5 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  6 * LONGSIZE)(\dst), \fixup)
-       EX(LONG_S, \val, (\offset +  7 * LONGSIZE)(\dst), \fixup)
-       .endm
-
-/*
- * memset(void *s, int c, size_t n)
- *
- * a0: start of area to clear
- * a1: char to fill with
- * a2: size of area to clear
- */
-       .set    noreorder
-       .align  5
-LEAF(memset)
-       beqz            a1, 1f
-        move           v0, a0                  /* result */
-
-       andi            a1, 0xff                /* spread fillword */
-       dsll            t1, a1, 8
-       or              a1, t1
-       dsll            t1, a1, 16
-       or              a1, t1
-       dsll            t1, a1, 32
-       or              a1, t1
-1:
-
-FEXPORT(__bzero)
-       sltiu           t0, a2, LONGSIZE        /* very small region? */
-       bnez            t0, small_memset
-        andi           t0, a0, LONGMASK        /* aligned? */
-
-       beqz            t0, 1f
-        PTR_SUBU       t0, LONGSIZE            /* alignment in bytes */
-
-#ifdef __MIPSEB__
-       EX(sdl, a1, (a0), first_fixup)          /* make dword aligned */
-#endif
-#ifdef __MIPSEL__
-       EX(sdr, a1, (a0), first_fixup)          /* make dword aligned */
-#endif
-       PTR_SUBU        a0, t0                  /* long align ptr */
-       PTR_ADDU        a2, t0                  /* correct size */
-
-1:     ori             t1, a2, 0x3f            /* # of full blocks */
-       xori            t1, 0x3f
-       beqz            t1, memset_partial      /* no block to fill */
-        andi           t0, a2, 0x38
-
-       PTR_ADDU        t1, a0                  /* end address */
-       .set            reorder
-1:     PTR_ADDIU       a0, 64
-       f_fill64 a0, -64, a1, fwd_fixup
-       bne             t1, a0, 1b
-       .set            noreorder
-
-memset_partial:
-       PTR_LA          t1, 2f                  /* where to start */
-       .set            noat
-       dsrl            AT, t0, 1
-       PTR_SUBU        t1, AT
-       .set            noat
-       jr              t1
-        PTR_ADDU       a0, t0                  /* dest ptr */
-
-       .set            push
-       .set            noreorder
-       .set            nomacro
-       f_fill64 a0, -64, a1, partial_fixup     /* ... but first do longs ... */
-2:     .set            pop
-       andi            a2, LONGMASK            /* At most one long to go */
-
-       beqz            a2, 1f
-        PTR_ADDU       a0, a2                  /* What's left */
-#ifdef __MIPSEB__
-       EX(sdr, a1, -1(a0), last_fixup)
-#endif
-#ifdef __MIPSEL__
-       EX(sdl, a1, -1(a0), last_fixup)
-#endif
-1:     jr              ra
-        move           a2, zero
-
-small_memset:
-       beqz            a2, 2f
-        PTR_ADDU       t1, a0, a2
-
-1:     PTR_ADDIU       a0, 1                   /* fill bytewise */
-       bne             t1, a0, 1b
-        sb             a1, -1(a0)
-
-2:     jr              ra                      /* done */
-        move           a2, zero
-       END(memset)
-
-first_fixup:
-       jr      ra
-        nop
-
-fwd_fixup:
-       PTR_L           t0, TI_TASK($28)
-       LONG_L          t0, THREAD_BUADDR(t0)
-       andi            a2, 0x3f
-       LONG_ADDU       a2, t1
-       jr              ra
-        LONG_SUBU      a2, t0
-
-partial_fixup:
-       PTR_L           t0, TI_TASK($28)
-       LONG_L          t0, THREAD_BUADDR(t0)
-       andi            a2, LONGMASK
-       LONG_ADDU       a2, t1
-       jr              ra
-        LONG_SUBU      a2, t0
-
-last_fixup:
-       jr              ra
-        andi           v1, a2, LONGMASK

diff --git a/arch/mips/lib/Makefile b/arch/mips/lib/Makefile
index 989c900b8b14a5f4b3b1422552044fcd9ac24f52..5ad501b30b430f387f219c575d4f7fbe9c8c517e 100644 (file)
--- a/arch/mips/lib/Makefile
+++ b/arch/mips/lib/Makefile
@@ -2,7 +2,7 @@
 # Makefile for MIPS-specific library files..
 #
 
-lib-y  += csum_partial.o memcpy.o promlib.o \
+lib-y  += csum_partial.o memcpy.o memset.o promlib.o \
           strlen_user.o strncpy_user.o strnlen_user.o uncached.o
 
 obj-y  += iomap.o

diff --git a/arch/mips/lib-32/memset.S b/arch/mips/lib/memset.S
similarity index 85%
rename from arch/mips/lib-32/memset.S
rename to arch/mips/lib/memset.S
index 1981485bd48b7f34120d0a70aa1fb4dcc74e5d3c..3f8b8b3d0b23575139ba263fb22a19eb301544f1 100644 (file)
--- a/arch/mips/lib-32/memset.S
+++ b/arch/mips/lib/memset.S
@@ -10,6 +10,14 @@
 #include <asm/asm-offsets.h>
 #include <asm/regdef.h>
 
+#if LONGSIZE == 4
+#define LONG_S_L swl
+#define LONG_S_R swr
+#else
+#define LONG_S_L sdl
+#define LONG_S_R sdr
+#endif
+
 #define EX(insn,reg,addr,handler)                      \
 9:     insn    reg, addr;                              \
        .section __ex_table,"a";                        \
@@ -25,6 +33,7 @@
        EX(LONG_S, \val, (\offset +  5 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset +  6 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset +  7 * LONGSIZE)(\dst), \fixup)
+#if LONGSIZE == 4
        EX(LONG_S, \val, (\offset +  8 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset +  9 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset + 10 * LONGSIZE)(\dst), \fixup)
@@ -33,6 +42,7 @@
        EX(LONG_S, \val, (\offset + 13 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset + 14 * LONGSIZE)(\dst), \fixup)
        EX(LONG_S, \val, (\offset + 15 * LONGSIZE)(\dst), \fixup)
+#endif
        .endm
 
 /*
@@ -49,9 +59,13 @@ LEAF(memset)
         move           v0, a0                  /* result */
 
        andi            a1, 0xff                /* spread fillword */
-       sll             t1, a1, 8
+       LONG_SLL                t1, a1, 8
        or              a1, t1
-       sll             t1, a1, 16
+       LONG_SLL                t1, a1, 16
+#if LONGSIZE == 8
+       or              a1, t1
+       LONG_SLL                t1, a1, 32
+#endif
        or              a1, t1
 1:
 
@@ -64,10 +78,10 @@ FEXPORT(__bzero)
         PTR_SUBU       t0, LONGSIZE            /* alignment in bytes */
 
 #ifdef __MIPSEB__
-       EX(swl, a1, (a0), first_fixup)          /* make word aligned */
+       EX(LONG_S_L, a1, (a0), first_fixup)     /* make word/dword aligned */
 #endif
 #ifdef __MIPSEL__
-       EX(swr, a1, (a0), first_fixup)          /* make word aligned */
+       EX(LONG_S_R, a1, (a0), first_fixup)     /* make word/dword aligned */
 #endif
        PTR_SUBU        a0, t0                  /* long align ptr */
        PTR_ADDU        a2, t0                  /* correct size */
@@ -75,7 +89,7 @@ FEXPORT(__bzero)
 1:     ori             t1, a2, 0x3f            /* # of full blocks */
        xori            t1, 0x3f
        beqz            t1, memset_partial      /* no block to fill */
-        andi           t0, a2, 0x3c
+        andi           t0, a2, 0x40-LONGSIZE
 
        PTR_ADDU        t1, a0                  /* end address */
        .set            reorder
@@ -86,7 +100,14 @@ FEXPORT(__bzero)
 
 memset_partial:
        PTR_LA          t1, 2f                  /* where to start */
+#if LONGSIZE == 4
        PTR_SUBU        t1, t0
+#else
+       .set            noat
+       LONG_SRL                AT, t0, 1
+       PTR_SUBU        t1, AT
+       .set            noat
+#endif
        jr              t1
         PTR_ADDU       a0, t0                  /* dest ptr */
 
@@ -100,10 +121,10 @@ memset_partial:
        beqz            a2, 1f
         PTR_ADDU       a0, a2                  /* What's left */
 #ifdef __MIPSEB__
-       EX(swr, a1, -1(a0), last_fixup)
+       EX(LONG_S_R, a1, -1(a0), last_fixup)
 #endif
 #ifdef __MIPSEL__
-       EX(swl, a1, -1(a0), last_fixup)
+       EX(LONG_S_L, a1, -1(a0), last_fixup)
 #endif
 1:     jr              ra
         move           a2, zero