diff options
author | David S. Miller <davem@davemloft.net> | 2010-02-22 03:59:49 -0800 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2010-02-25 23:58:50 -0800 |
commit | e44f6f89da537816aff71d926a0b774a77545b48 (patch) | |
tree | 9d628525aaa2e83f763e3202ea41a6c11f4ff58b /sysdeps/sparc/sparc64 | |
parent | dbcaf07c326e18b14d19aebe011b9ffbe4a45972 (diff) | |
download | glibc-e44f6f89da537816aff71d926a0b774a77545b48.tar.gz glibc-e44f6f89da537816aff71d926a0b774a77545b48.tar.xz glibc-e44f6f89da537816aff71d926a0b774a77545b48.zip |
sparc: Pull bcopy/memmove out of memcpy.S implementations.
2010-02-25 David S. Miller <davem@davemloft.net> * sysdeps/sparc/sparc32/bcopy.c: Delete. * sysdeps/sparc/sparc32/memmove.c: Delete. * sysdeps/sparc/sparc32/sparcv9/bcopy.c: Delete. * sysdeps/sparc/sparc32/sparcv9/memmove.c: Delete. * sysdeps/sparc/sparc64/bcopy.c: Delete. * sysdeps/sparc/sparc64/memmove.c: Delete. * sysdeps/sparc/sparc64/memcopy.h: New. * sysdeps/sparc/sparc32/memcpy.S (bcopy, memmove): Remove. * sysdeps/sparc/sparc64/memcpy.S (bcopy, memmove): Likewise. * sysdeps/sparc/sparc64/sparcv9b/memcpy.S (bcopy, memmove): Likewise. * sysdeps/sparc/sparc64/sparcv9v/memcpy.S (bcopy, memmove): Likewise. * sysdeps/sparc/sparc64/sparcv9v2/memcpy.S (bcopy, memmove): Likewise.
Diffstat (limited to 'sysdeps/sparc/sparc64')
-rw-r--r-- | sysdeps/sparc/sparc64/bcopy.c | 1 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/memcopy.h | 1 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/memcpy.S | 290 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/memmove.c | 1 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/sparcv9b/memcpy.S | 285 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/sparcv9v/memcpy.S | 249 | ||||
-rw-r--r-- | sysdeps/sparc/sparc64/sparcv9v2/memcpy.S | 249 |
7 files changed, 1 insertions, 1075 deletions
diff --git a/sysdeps/sparc/sparc64/bcopy.c b/sysdeps/sparc/sparc64/bcopy.c deleted file mode 100644 index 9a455f33c4..0000000000 --- a/sysdeps/sparc/sparc64/bcopy.c +++ /dev/null @@ -1 +0,0 @@ -/* bcopy is in memcpy.S */ diff --git a/sysdeps/sparc/sparc64/memcopy.h b/sysdeps/sparc/sparc64/memcopy.h new file mode 100644 index 0000000000..ec978e3c80 --- /dev/null +++ b/sysdeps/sparc/sparc64/memcopy.h @@ -0,0 +1 @@ +#include <sparc32/memcopy.h> diff --git a/sysdeps/sparc/sparc64/memcpy.S b/sysdeps/sparc/sparc64/memcpy.S index 709d366304..71e7100658 100644 --- a/sysdeps/sparc/sparc64/memcpy.S +++ b/sysdeps/sparc/sparc64/memcpy.S @@ -136,79 +136,8 @@ stx %t0, [%dst - offset - 0x10]; \ stx %t1, [%dst - offset - 0x08]; - /* Macros for non-VIS memmove code. */ -#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stw %t0, [%dst - offset - 0x1c]; \ - srlx %t0, 32, %t0; \ - stw %t0, [%dst - offset - 0x20]; \ - stw %t1, [%dst - offset - 0x14]; \ - srlx %t1, 32, %t1; \ - stw %t1, [%dst - offset - 0x18]; \ - stw %t2, [%dst - offset - 0x0c]; \ - srlx %t2, 32, %t2; \ - stw %t2, [%dst - offset - 0x10]; \ - stw %t3, [%dst - offset - 0x04]; \ - srlx %t3, 32, %t3; \ - stw %t3, [%dst - offset - 0x08]; - -#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stx %t0, [%dst - offset - 0x20]; \ - stx %t1, [%dst - offset - 0x18]; \ - stx %t2, [%dst - offset - 0x10]; \ - stx %t3, [%dst - offset - 0x08]; \ - ldx [%src - offset - 0x40], %t0; \ - ldx [%src - offset - 0x38], %t1; \ - ldx [%src - offset - 0x30], %t2; \ - ldx [%src - offset - 0x28], %t3; \ - stx %t0, [%dst - offset - 0x40]; \ - stx %t1, [%dst - offset - 0x38]; \ - stx %t2, [%dst - offset - 0x30]; \ - stx %t3, [%dst - offset - 0x28]; - -#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stw %t0, [%dst + offset + 0x04]; \ - srlx %t0, 32, %t2; \ - stw %t2, [%dst + offset + 0x00]; \ - stw %t1, [%dst + offset + 0x0c]; \ - srlx %t1, 32, %t3; \ - stw %t3, [%dst + offset + 0x08]; - -#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stx %t0, [%dst + offset + 0x00]; \ - stx %t1, [%dst + offset + 0x08]; - .text .align 32 - -ENTRY(bcopy) - sub %o1, %o0, %o4 /* IEU0 Group */ - mov %o0, %g3 /* IEU1 */ - cmp %o4, %o2 /* IEU1 Group */ - mov %o1, %o0 /* IEU0 */ - bgeu,pt %XCC, 210f /* CTI */ - mov %g3, %o1 /* IEU0 Group */ -#ifndef USE_BPR - srl %o2, 0, %o2 /* IEU1 */ -#endif - brnz,pn %o2, 220f /* CTI Group */ - add %o0, %o2, %o0 /* IEU0 */ - retl - nop -END(bcopy) - - .align 32 ENTRY(__memcpy_large) 200: be,pt %xcc, 201f /* CTI */ andcc %o0, 0x38, %g5 /* IEU1 Group */ @@ -640,223 +569,4 @@ ENTRY(memcpy) mov %g4, %o0 END(memcpy) - .align 32 -ENTRY(__memmove_slowpath) -228: andcc %o2, 1, %g0 /* IEU1 Group */ - be,pt %icc, 2f+4 /* CTI */ -1: ldub [%o1 - 1], %o5 /* LOAD Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %o2, 1, %o2 /* IEU1 Group */ - be,pn %xcc, 229f /* CTI */ - stb %o5, [%o0] /* Store */ -2: ldub [%o1 - 1], %o5 /* LOAD Group */ - sub %o0, 2, %o0 /* IEU0 */ - ldub [%o1 - 2], %g5 /* LOAD Group */ - sub %o1, 2, %o1 /* IEU0 */ - subcc %o2, 2, %o2 /* IEU1 Group */ - stb %o5, [%o0 + 1] /* Store */ - bne,pt %xcc, 2b /* CTI */ - stb %g5, [%o0] /* Store */ -229: retl - mov %g4, %o0 -219: retl - nop -END(__memmove_slowpath) - - .align 32 -ENTRY(memmove) -#ifndef USE_BPR - srl %o2, 0, %o2 /* IEU1 Group */ -#endif - brz,pn %o2, 219b /* CTI Group */ - sub %o0, %o1, %o4 /* IEU0 */ - cmp %o4, %o2 /* IEU1 Group */ - bgeu,pt %XCC, 218b /* CTI */ - mov %o0, %g4 /* IEU0 */ - add %o0, %o2, %o0 /* IEU0 Group */ -220: add %o1, %o2, %o1 /* IEU1 */ - cmp %o2, 15 /* IEU1 Group */ - bleu,pn %xcc, 228b /* CTI */ - andcc %o0, 7, %g2 /* IEU1 Group */ - sub %o0, %o1, %g5 /* IEU0 */ - andcc %g5, 3, %o5 /* IEU1 Group */ - bne,pn %xcc, 232f /* CTI */ - andcc %o1, 3, %g0 /* IEU1 Group */ - be,a,pt %xcc, 236f /* CTI */ - andcc %o1, 4, %g0 /* IEU1 Group */ - andcc %o1, 1, %g0 /* IEU1 Group */ - be,pn %xcc, 4f /* CTI */ - andcc %o1, 2, %g0 /* IEU1 Group */ - ldub [%o1 - 1], %g2 /* Load Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - sub %o2, 1, %o2 /* IEU0 Group */ - be,pn %xcc, 5f /* CTI Group */ - stb %g2, [%o0] /* Store */ -4: lduh [%o1 - 2], %g2 /* Load Group */ - sub %o1, 2, %o1 /* IEU0 */ - sub %o0, 2, %o0 /* IEU1 */ - sub %o2, 2, %o2 /* IEU0 */ - sth %g2, [%o0] /* Store Group + bubble */ -5: andcc %o1, 4, %g0 /* IEU1 */ -236: be,a,pn %xcc, 2f /* CTI */ - andcc %o2, -128, %g6 /* IEU1 Group */ - lduw [%o1 - 4], %g5 /* Load Group */ - sub %o1, 4, %o1 /* IEU0 */ - sub %o0, 4, %o0 /* IEU1 */ - sub %o2, 4, %o2 /* IEU0 Group */ - stw %g5, [%o0] /* Store */ - andcc %o2, -128, %g6 /* IEU1 Group */ -2: be,pn %xcc, 235f /* CTI */ - andcc %o0, 4, %g0 /* IEU1 Group */ - be,pn %xcc, 282f + 4 /* CTI Group */ -5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5) - subcc %g6, 128, %g6 /* IEU1 Group */ - sub %o1, 128, %o1 /* IEU0 */ - bne,pt %xcc, 5b /* CTI */ - sub %o0, 128, %o0 /* IEU0 Group */ -235: andcc %o2, 0x70, %g6 /* IEU1 Group */ -41: be,pn %xcc, 280f /* CTI */ - andcc %o2, 8, %g0 /* IEU1 Group */ - /* Clk1 8-( */ - /* Clk2 8-( */ - /* Clk3 8-( */ - /* Clk4 8-( */ -279: rd %pc, %o5 /* PDU Group */ - sll %g6, 1, %g5 /* IEU0 Group */ - sub %o1, %g6, %o1 /* IEU1 */ - sub %o5, %g5, %o5 /* IEU0 Group */ - jmpl %o5 + %lo(280f - 279b), %g0 /* CTI Group brk forced*/ - sub %o0, %g6, %o0 /* IEU0 Group */ - RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5) -280: be,pt %xcc, 281f /* CTI */ - andcc %o2, 4, %g0 /* IEU1 */ - ldx [%o1 - 8], %g2 /* Load Group */ - sub %o0, 8, %o0 /* IEU0 */ - stw %g2, [%o0 + 4] /* Store Group */ - sub %o1, 8, %o1 /* IEU1 */ - srlx %g2, 32, %g2 /* IEU0 Group */ - stw %g2, [%o0] /* Store */ -281: be,pt %xcc, 1f /* CTI */ - andcc %o2, 2, %g0 /* IEU1 Group */ - lduw [%o1 - 4], %g2 /* Load Group */ - sub %o1, 4, %o1 /* IEU0 */ - stw %g2, [%o0 - 4] /* Store Group */ - sub %o0, 4, %o0 /* IEU0 */ -1: be,pt %xcc, 1f /* CTI */ - andcc %o2, 1, %g0 /* IEU1 Group */ - lduh [%o1 - 2], %g2 /* Load Group */ - sub %o1, 2, %o1 /* IEU0 */ - sth %g2, [%o0 - 2] /* Store Group */ - sub %o0, 2, %o0 /* IEU0 */ -1: be,pt %xcc, 211f /* CTI */ - nop /* IEU1 */ - ldub [%o1 - 1], %g2 /* Load Group */ - stb %g2, [%o0 - 1] /* Store Group + bubble */ -211: retl - mov %g4, %o0 - -282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - subcc %g6, 128, %g6 /* IEU1 Group */ - sub %o1, 128, %o1 /* IEU0 */ - bne,pt %xcc, 282b /* CTI */ - sub %o0, 128, %o0 /* IEU0 Group */ - andcc %o2, 0x70, %g6 /* IEU1 */ - be,pn %xcc, 284f /* CTI */ - andcc %o2, 8, %g0 /* IEU1 Group */ - /* Clk1 8-( */ - /* Clk2 8-( */ - /* Clk3 8-( */ - /* Clk4 8-( */ -283: rd %pc, %o5 /* PDU Group */ - sub %o1, %g6, %o1 /* IEU0 Group */ - sub %o5, %g6, %o5 /* IEU1 */ - jmpl %o5 + %lo(284f - 283b), %g0 /* CTI Group brk forced*/ - sub %o0, %g6, %o0 /* IEU0 Group */ - RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3) -284: be,pt %xcc, 285f /* CTI Group */ - andcc %o2, 4, %g0 /* IEU1 */ - ldx [%o1 - 8], %g2 /* Load Group */ - sub %o0, 8, %o0 /* IEU0 */ - sub %o1, 8, %o1 /* IEU0 Group */ - stx %g2, [%o0] /* Store */ -285: be,pt %xcc, 1f /* CTI */ - andcc %o2, 2, %g0 /* IEU1 Group */ - lduw [%o1 - 4], %g2 /* Load Group */ - sub %o0, 4, %o0 /* IEU0 */ - sub %o1, 4, %o1 /* IEU0 Group */ - stw %g2, [%o0] /* Store */ -1: be,pt %xcc, 1f /* CTI */ - andcc %o2, 1, %g0 /* IEU1 Group */ - lduh [%o1 - 2], %g2 /* Load Group */ - sub %o0, 2, %o0 /* IEU0 */ - sub %o1, 2, %o1 /* IEU0 Group */ - sth %g2, [%o0] /* Store */ -1: be,pt %xcc, 1f /* CTI */ - nop /* IEU0 Group */ - ldub [%o1 - 1], %g2 /* Load Group */ - stb %g2, [%o0 - 1] /* Store Group + bubble */ -1: retl - mov %g4, %o0 - -232: brz,pt %g2, 2f /* CTI Group */ - sub %o2, %g2, %o2 /* IEU0 Group */ -1: ldub [%o1 - 1], %g5 /* Load Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %g2, 1, %g2 /* IEU1 Group */ - bne,pt %xcc, 1b /* CTI */ - stb %g5, [%o0] /* Store */ -2: andn %o2, 7, %g5 /* IEU0 Group */ - and %o2, 7, %o2 /* IEU1 */ - fmovd %f0, %f2 /* FPU */ - alignaddr %o1, %g0, %g1 /* GRU Group */ - ldd [%g1], %f4 /* Load Group */ -1: ldd [%g1 - 8], %f6 /* Load Group */ - sub %g1, 8, %g1 /* IEU0 Group */ - subcc %g5, 8, %g5 /* IEU1 */ - faligndata %f6, %f4, %f0 /* GRU Group */ - std %f0, [%o0 - 8] /* Store */ - sub %o1, 8, %o1 /* IEU0 Group */ - be,pn %xcc, 233f /* CTI */ - sub %o0, 8, %o0 /* IEU1 */ - ldd [%g1 - 8], %f4 /* Load Group */ - sub %g1, 8, %g1 /* IEU0 */ - subcc %g5, 8, %g5 /* IEU1 */ - faligndata %f4, %f6, %f0 /* GRU Group */ - std %f0, [%o0 - 8] /* Store */ - sub %o1, 8, %o1 /* IEU0 */ - bne,pn %xcc, 1b /* CTI Group */ - sub %o0, 8, %o0 /* IEU0 */ -233: brz,pn %o2, 234f /* CTI Group */ - nop /* IEU0 */ -237: ldub [%o1 - 1], %g5 /* LOAD */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %o2, 1, %o2 /* IEU1 */ - bne,pt %xcc, 237b /* CTI */ - stb %g5, [%o0] /* Store Group */ -234: wr %g0, FPRS_FEF, %fprs - retl - mov %g4, %o0 -END(memmove) - libc_hidden_builtin_def (memcpy) -libc_hidden_builtin_def (memmove) diff --git a/sysdeps/sparc/sparc64/memmove.c b/sysdeps/sparc/sparc64/memmove.c deleted file mode 100644 index a8d2d49948..0000000000 --- a/sysdeps/sparc/sparc64/memmove.c +++ /dev/null @@ -1 +0,0 @@ -/* memmove is in memcpy.S */ diff --git a/sysdeps/sparc/sparc64/sparcv9b/memcpy.S b/sysdeps/sparc/sparc64/sparcv9b/memcpy.S index 2a4146feeb..389e09d820 100644 --- a/sysdeps/sparc/sparc64/sparcv9b/memcpy.S +++ b/sysdeps/sparc/sparc64/sparcv9b/memcpy.S @@ -36,23 +36,6 @@ .register %g6,#scratch .text - .align 32 - -ENTRY(bcopy) - sub %o1, %o0, %o4 - mov %o0, %g4 - cmp %o4, %o2 - mov %o1, %o0 - bgeu,pt %XCC, 100f - mov %g4, %o1 -#ifndef USE_BPR - srl %o2, 0, %o2 -#endif - brnz,pn %o2, 220f - add %o0, %o2, %o0 - retl - nop -END(bcopy) /* Special/non-trivial issues of this code: * @@ -332,272 +315,4 @@ small_copy_unaligned: END(memcpy) -#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stw %t0, [%dst - offset - 0x1c]; \ - srlx %t0, 32, %t0; \ - stw %t0, [%dst - offset - 0x20]; \ - stw %t1, [%dst - offset - 0x14]; \ - srlx %t1, 32, %t1; \ - stw %t1, [%dst - offset - 0x18]; \ - stw %t2, [%dst - offset - 0x0c]; \ - srlx %t2, 32, %t2; \ - stw %t2, [%dst - offset - 0x10]; \ - stw %t3, [%dst - offset - 0x04]; \ - srlx %t3, 32, %t3; \ - stw %t3, [%dst - offset - 0x08]; - -#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stx %t0, [%dst - offset - 0x20]; \ - stx %t1, [%dst - offset - 0x18]; \ - stx %t2, [%dst - offset - 0x10]; \ - stx %t3, [%dst - offset - 0x08]; \ - ldx [%src - offset - 0x40], %t0; \ - ldx [%src - offset - 0x38], %t1; \ - ldx [%src - offset - 0x30], %t2; \ - ldx [%src - offset - 0x28], %t3; \ - stx %t0, [%dst - offset - 0x40]; \ - stx %t1, [%dst - offset - 0x38]; \ - stx %t2, [%dst - offset - 0x30]; \ - stx %t3, [%dst - offset - 0x28]; - -#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stw %t0, [%dst + offset + 0x04]; \ - srlx %t0, 32, %t2; \ - stw %t2, [%dst + offset + 0x00]; \ - stw %t1, [%dst + offset + 0x0c]; \ - srlx %t1, 32, %t3; \ - stw %t3, [%dst + offset + 0x08]; - -#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stx %t0, [%dst + offset + 0x00]; \ - stx %t1, [%dst + offset + 0x08]; - - .align 32 -228: andcc %o2, 1, %g0 /* IEU1 Group */ - be,pt %icc, 2f+4 /* CTI */ -1: ldub [%o1 - 1], %o5 /* LOAD Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %o2, 1, %o2 /* IEU1 Group */ - be,pn %xcc, 229f /* CTI */ - stb %o5, [%o0] /* Store */ -2: ldub [%o1 - 1], %o5 /* LOAD Group */ - sub %o0, 2, %o0 /* IEU0 */ - ldub [%o1 - 2], %g5 /* LOAD Group */ - sub %o1, 2, %o1 /* IEU0 */ - subcc %o2, 2, %o2 /* IEU1 Group */ - stb %o5, [%o0 + 1] /* Store */ - bne,pt %xcc, 2b /* CTI */ - stb %g5, [%o0] /* Store */ -229: retl - mov %g4, %o0 - - .align 32 -ENTRY(memmove) - mov %o0, %g5 -#ifndef USE_BPR - srl %o2, 0, %o2 /* IEU1 Group */ -#endif - brz,pn %o2, out /* CTI Group */ - sub %o0, %o1, %o4 /* IEU0 */ - cmp %o4, %o2 /* IEU1 Group */ - bgeu,pt %XCC, 218b /* CTI */ - mov %o0, %g4 /* IEU0 */ - add %o0, %o2, %o0 /* IEU0 Group */ -220: add %o1, %o2, %o1 /* IEU1 */ - cmp %o2, 15 /* IEU1 Group */ - bleu,pn %xcc, 228b /* CTI */ - andcc %o0, 7, %g2 /* IEU1 Group */ - sub %o0, %o1, %g5 /* IEU0 */ - andcc %g5, 3, %o5 /* IEU1 Group */ - bne,pn %xcc, 232f /* CTI */ - andcc %o1, 3, %g0 /* IEU1 Group */ - be,a,pt %xcc, 236f /* CTI */ - andcc %o1, 4, %g0 /* IEU1 Group */ - andcc %o1, 1, %g0 /* IEU1 Group */ - be,pn %xcc, 4f /* CTI */ - andcc %o1, 2, %g0 /* IEU1 Group */ - ldub [%o1 - 1], %g2 /* Load Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - sub %o2, 1, %o2 /* IEU0 Group */ - be,pn %xcc, 5f /* CTI Group */ - stb %g2, [%o0] /* Store */ -4: lduh [%o1 - 2], %g2 /* Load Group */ - sub %o1, 2, %o1 /* IEU0 */ - sub %o0, 2, %o0 /* IEU1 */ - sub %o2, 2, %o2 /* IEU0 */ - sth %g2, [%o0] /* Store Group + bubble */ -5: andcc %o1, 4, %g0 /* IEU1 */ -236: be,a,pn %xcc, 2f /* CTI */ - andcc %o2, -128, %g6 /* IEU1 Group */ - lduw [%o1 - 4], %g5 /* Load Group */ - sub %o1, 4, %o1 /* IEU0 */ - sub %o0, 4, %o0 /* IEU1 */ - sub %o2, 4, %o2 /* IEU0 Group */ - stw %g5, [%o0] /* Store */ - andcc %o2, -128, %g6 /* IEU1 Group */ -2: be,pn %xcc, 235f /* CTI */ - andcc %o0, 4, %g0 /* IEU1 Group */ - be,pn %xcc, 282f + 4 /* CTI Group */ -5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5) - subcc %g6, 128, %g6 /* IEU1 Group */ - sub %o1, 128, %o1 /* IEU0 */ - bne,pt %xcc, 5b /* CTI */ - sub %o0, 128, %o0 /* IEU0 Group */ -235: andcc %o2, 0x70, %g6 /* IEU1 Group */ -41: be,pn %xcc, 280f /* CTI */ - andcc %o2, 8, %g0 /* IEU1 Group */ - /* Clk1 8-( */ - /* Clk2 8-( */ - /* Clk3 8-( */ - /* Clk4 8-( */ -279: rd %pc, %o5 /* PDU Group */ - sll %g6, 1, %g5 /* IEU0 Group */ - sub %o1, %g6, %o1 /* IEU1 */ - sub %o5, %g5, %o5 /* IEU0 Group */ - jmpl %o5 + %lo(280f - 279b), %g0 /* CTI Group brk forced*/ - sub %o0, %g6, %o0 /* IEU0 Group */ - RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5) -280: be,pt %xcc, 281f /* CTI */ - andcc %o2, 4, %g0 /* IEU1 */ - ldx [%o1 - 8], %g2 /* Load Group */ - sub %o0, 8, %o0 /* IEU0 */ - stw %g2, [%o0 + 4] /* Store Group */ - sub %o1, 8, %o1 /* IEU1 */ - srlx %g2, 32, %g2 /* IEU0 Group */ - stw %g2, [%o0] /* Store */ -281: be,pt %xcc, 1f /* CTI */ - andcc %o2, 2, %g0 /* IEU1 Group */ - lduw [%o1 - 4], %g2 /* Load Group */ - sub %o1, 4, %o1 /* IEU0 */ - stw %g2, [%o0 - 4] /* Store Group */ - sub %o0, 4, %o0 /* IEU0 */ -1: be,pt %xcc, 1f /* CTI */ - andcc %o2, 1, %g0 /* IEU1 Group */ - lduh [%o1 - 2], %g2 /* Load Group */ - sub %o1, 2, %o1 /* IEU0 */ - sth %g2, [%o0 - 2] /* Store Group */ - sub %o0, 2, %o0 /* IEU0 */ -1: be,pt %xcc, 211f /* CTI */ - nop /* IEU1 */ - ldub [%o1 - 1], %g2 /* Load Group */ - stb %g2, [%o0 - 1] /* Store Group + bubble */ -211: retl - mov %g4, %o0 - -282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - subcc %g6, 128, %g6 /* IEU1 Group */ - sub %o1, 128, %o1 /* IEU0 */ - bne,pt %xcc, 282b /* CTI */ - sub %o0, 128, %o0 /* IEU0 Group */ - andcc %o2, 0x70, %g6 /* IEU1 */ - be,pn %xcc, 284f /* CTI */ - andcc %o2, 8, %g0 /* IEU1 Group */ - /* Clk1 8-( */ - /* Clk2 8-( */ - /* Clk3 8-( */ - /* Clk4 8-( */ -283: rd %pc, %o5 /* PDU Group */ - sub %o1, %g6, %o1 /* IEU0 Group */ - sub %o5, %g6, %o5 /* IEU1 */ - jmpl %o5 + %lo(284f - 283b), %g0 /* CTI Group brk forced*/ - sub %o0, %g6, %o0 /* IEU0 Group */ - RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3) -284: be,pt %xcc, 285f /* CTI Group */ - andcc %o2, 4, %g0 /* IEU1 */ - ldx [%o1 - 8], %g2 /* Load Group */ - sub %o0, 8, %o0 /* IEU0 */ - sub %o1, 8, %o1 /* IEU0 Group */ - stx %g2, [%o0] /* Store */ -285: be,pt %xcc, 1f /* CTI */ - andcc %o2, 2, %g0 /* IEU1 Group */ - lduw [%o1 - 4], %g2 /* Load Group */ - sub %o0, 4, %o0 /* IEU0 */ - sub %o1, 4, %o1 /* IEU0 Group */ - stw %g2, [%o0] /* Store */ -1: be,pt %xcc, 1f /* CTI */ - andcc %o2, 1, %g0 /* IEU1 Group */ - lduh [%o1 - 2], %g2 /* Load Group */ - sub %o0, 2, %o0 /* IEU0 */ - sub %o1, 2, %o1 /* IEU0 Group */ - sth %g2, [%o0] /* Store */ -1: be,pt %xcc, 1f /* CTI */ - nop /* IEU0 Group */ - ldub [%o1 - 1], %g2 /* Load Group */ - stb %g2, [%o0 - 1] /* Store Group + bubble */ -1: retl - mov %g4, %o0 - -232: brz,pt %g2, 2f /* CTI Group */ - sub %o2, %g2, %o2 /* IEU0 Group */ -1: ldub [%o1 - 1], %g5 /* Load Group */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %g2, 1, %g2 /* IEU1 Group */ - bne,pt %xcc, 1b /* CTI */ - stb %g5, [%o0] /* Store */ -2: andn %o2, 7, %g5 /* IEU0 Group */ - and %o2, 7, %o2 /* IEU1 */ - fmovd %f0, %f2 /* FPU */ - alignaddr %o1, %g0, %g1 /* GRU Group */ - ldd [%g1], %f4 /* Load Group */ -1: ldd [%g1 - 8], %f6 /* Load Group */ - sub %g1, 8, %g1 /* IEU0 Group */ - subcc %g5, 8, %g5 /* IEU1 */ - faligndata %f6, %f4, %f0 /* GRU Group */ - std %f0, [%o0 - 8] /* Store */ - sub %o1, 8, %o1 /* IEU0 Group */ - be,pn %xcc, 233f /* CTI */ - sub %o0, 8, %o0 /* IEU1 */ - ldd [%g1 - 8], %f4 /* Load Group */ - sub %g1, 8, %g1 /* IEU0 */ - subcc %g5, 8, %g5 /* IEU1 */ - faligndata %f4, %f6, %f0 /* GRU Group */ - std %f0, [%o0 - 8] /* Store */ - sub %o1, 8, %o1 /* IEU0 */ - bne,pn %xcc, 1b /* CTI Group */ - sub %o0, 8, %o0 /* IEU0 */ -233: brz,pn %o2, 234f /* CTI Group */ - nop /* IEU0 */ -237: ldub [%o1 - 1], %g5 /* LOAD */ - sub %o1, 1, %o1 /* IEU0 */ - sub %o0, 1, %o0 /* IEU1 */ - subcc %o2, 1, %o2 /* IEU1 */ - bne,pt %xcc, 237b /* CTI */ - stb %g5, [%o0] /* Store Group */ -234: wr %g0, FPRS_FEF, %fprs - retl - mov %g4, %o0 -END(memmove) - libc_hidden_builtin_def (memcpy) -libc_hidden_builtin_def (memmove) diff --git a/sysdeps/sparc/sparc64/sparcv9v/memcpy.S b/sysdeps/sparc/sparc64/sparcv9v/memcpy.S index 61db9ff81f..116c7113ff 100644 --- a/sysdeps/sparc/sparc64/sparcv9v/memcpy.S +++ b/sysdeps/sparc/sparc64/sparcv9v/memcpy.S @@ -41,23 +41,6 @@ .register %g6,#scratch .text - .align 32 - -ENTRY(bcopy) - sub %o1, %o0, %o4 - mov %o0, %g4 - cmp %o4, %o2 - mov %o1, %o0 - bgeu,pt %XCC, 100f - mov %g4, %o1 -#ifndef USE_BPR - srl %o2, 0, %o2 -#endif - brnz,pn %o2, 220f - add %o0, %o2, %o0 - retl - nop -END(bcopy) .align 32 ENTRY(memcpy) @@ -354,236 +337,4 @@ ENTRY(memcpy) END(memcpy) -#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stw %t0, [%dst - offset - 0x1c]; \ - srlx %t0, 32, %t0; \ - stw %t0, [%dst - offset - 0x20]; \ - stw %t1, [%dst - offset - 0x14]; \ - srlx %t1, 32, %t1; \ - stw %t1, [%dst - offset - 0x18]; \ - stw %t2, [%dst - offset - 0x0c]; \ - srlx %t2, 32, %t2; \ - stw %t2, [%dst - offset - 0x10]; \ - stw %t3, [%dst - offset - 0x04]; \ - srlx %t3, 32, %t3; \ - stw %t3, [%dst - offset - 0x08]; - -#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stx %t0, [%dst - offset - 0x20]; \ - stx %t1, [%dst - offset - 0x18]; \ - stx %t2, [%dst - offset - 0x10]; \ - stx %t3, [%dst - offset - 0x08]; \ - ldx [%src - offset - 0x40], %t0; \ - ldx [%src - offset - 0x38], %t1; \ - ldx [%src - offset - 0x30], %t2; \ - ldx [%src - offset - 0x28], %t3; \ - stx %t0, [%dst - offset - 0x40]; \ - stx %t1, [%dst - offset - 0x38]; \ - stx %t2, [%dst - offset - 0x30]; \ - stx %t3, [%dst - offset - 0x28]; - -#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stw %t0, [%dst + offset + 0x04]; \ - srlx %t0, 32, %t2; \ - stw %t2, [%dst + offset + 0x00]; \ - stw %t1, [%dst + offset + 0x0c]; \ - srlx %t1, 32, %t3; \ - stw %t3, [%dst + offset + 0x08]; - -#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stx %t0, [%dst + offset + 0x00]; \ - stx %t1, [%dst + offset + 0x08]; - - .align 32 -228: andcc %o2, 1, %g0 - be,pt %icc, 2f+4 -1: ldub [%o1 - 1], %o5 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - subcc %o2, 1, %o2 - be,pn %xcc, 229f - stb %o5, [%o0] -2: ldub [%o1 - 1], %o5 - sub %o0, 2, %o0 - ldub [%o1 - 2], %g5 - sub %o1, 2, %o1 - subcc %o2, 2, %o2 - stb %o5, [%o0 + 1] - bne,pt %xcc, 2b - stb %g5, [%o0] -229: retl - mov %g4, %o0 -out: retl - mov %g5, %o0 - - .align 32 -ENTRY(memmove) - mov %o0, %g5 -#ifndef USE_BPR - srl %o2, 0, %o2 -#endif - brz,pn %o2, out - sub %o0, %o1, %o4 - cmp %o4, %o2 - bgeu,pt %XCC, 218b - mov %o0, %g4 - add %o0, %o2, %o0 -220: add %o1, %o2, %o1 - cmp %o2, 15 - bleu,pn %xcc, 228b - andcc %o0, 7, %g2 - sub %o0, %o1, %g5 - andcc %g5, 3, %o5 - bne,pn %xcc, 232f - andcc %o1, 3, %g0 - be,a,pt %xcc, 236f - andcc %o1, 4, %g0 - andcc %o1, 1, %g0 - be,pn %xcc, 4f - andcc %o1, 2, %g0 - ldub [%o1 - 1], %g2 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - sub %o2, 1, %o2 - be,pn %xcc, 5f - stb %g2, [%o0] -4: lduh [%o1 - 2], %g2 - sub %o1, 2, %o1 - sub %o0, 2, %o0 - sub %o2, 2, %o2 - sth %g2, [%o0] -5: andcc %o1, 4, %g0 -236: be,a,pn %xcc, 2f - andcc %o2, -128, %g6 - lduw [%o1 - 4], %g5 - sub %o1, 4, %o1 - sub %o0, 4, %o0 - sub %o2, 4, %o2 - stw %g5, [%o0] - andcc %o2, -128, %g6 -2: be,pn %xcc, 235f - andcc %o0, 4, %g0 - be,pn %xcc, 282f + 4 -5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5) - subcc %g6, 128, %g6 - sub %o1, 128, %o1 - bne,pt %xcc, 5b - sub %o0, 128, %o0 -235: andcc %o2, 0x70, %g6 -41: be,pn %xcc, 280f - andcc %o2, 8, %g0 - -279: rd %pc, %o5 - sll %g6, 1, %g5 - sub %o1, %g6, %o1 - sub %o5, %g5, %o5 - jmpl %o5 + %lo(280f - 279b), %g0 - sub %o0, %g6, %o0 - RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5) -280: be,pt %xcc, 281f - andcc %o2, 4, %g0 - ldx [%o1 - 8], %g2 - sub %o0, 8, %o0 - stw %g2, [%o0 + 4] - sub %o1, 8, %o1 - srlx %g2, 32, %g2 - stw %g2, [%o0] -281: be,pt %xcc, 1f - andcc %o2, 2, %g0 - lduw [%o1 - 4], %g2 - sub %o1, 4, %o1 - stw %g2, [%o0 - 4] - sub %o0, 4, %o0 -1: be,pt %xcc, 1f - andcc %o2, 1, %g0 - lduh [%o1 - 2], %g2 - sub %o1, 2, %o1 - sth %g2, [%o0 - 2] - sub %o0, 2, %o0 -1: be,pt %xcc, 211f - nop - ldub [%o1 - 1], %g2 - stb %g2, [%o0 - 1] -211: retl - mov %g4, %o0 - -282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - subcc %g6, 128, %g6 - sub %o1, 128, %o1 - bne,pt %xcc, 282b - sub %o0, 128, %o0 - andcc %o2, 0x70, %g6 - be,pn %xcc, 284f - andcc %o2, 8, %g0 - -283: rd %pc, %o5 - sub %o1, %g6, %o1 - sub %o5, %g6, %o5 - jmpl %o5 + %lo(284f - 283b), %g0 - sub %o0, %g6, %o0 - RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3) -284: be,pt %xcc, 285f - andcc %o2, 4, %g0 - ldx [%o1 - 8], %g2 - sub %o0, 8, %o0 - sub %o1, 8, %o1 - stx %g2, [%o0] -285: be,pt %xcc, 1f - andcc %o2, 2, %g0 - lduw [%o1 - 4], %g2 - sub %o0, 4, %o0 - sub %o1, 4, %o1 - stw %g2, [%o0] -1: be,pt %xcc, 1f - andcc %o2, 1, %g0 - lduh [%o1 - 2], %g2 - sub %o0, 2, %o0 - sub %o1, 2, %o1 - sth %g2, [%o0] -1: be,pt %xcc, 1f - nop - ldub [%o1 - 1], %g2 - stb %g2, [%o0 - 1] -1: retl - mov %g4, %o0 - -232: ldub [%o1 - 1], %g5 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - subcc %o2, 1, %o2 - bne,pt %xcc, 232b - stb %g5, [%o0] -234: retl - mov %g4, %o0 -END(memmove) - libc_hidden_builtin_def (memcpy) -libc_hidden_builtin_def (memmove) diff --git a/sysdeps/sparc/sparc64/sparcv9v2/memcpy.S b/sysdeps/sparc/sparc64/sparcv9v2/memcpy.S index 2b1444ec7f..300b12f427 100644 --- a/sysdeps/sparc/sparc64/sparcv9v2/memcpy.S +++ b/sysdeps/sparc/sparc64/sparcv9v2/memcpy.S @@ -143,23 +143,6 @@ .register %g6,#scratch .text - .align 32 - -ENTRY(bcopy) - sub %o1, %o0, %o4 - mov %o0, %g4 - cmp %o4, %o2 - mov %o1, %o0 - bgeu,pt %XCC, 100f - mov %g4, %o1 -#ifndef USE_BPR - srl %o2, 0, %o2 -#endif - brnz,pn %o2, 220f - add %o0, %o2, %o0 - retl - nop -END(bcopy) .align 32 ENTRY(memcpy) @@ -504,236 +487,4 @@ ENTRY(memcpy) END(memcpy) -#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stw %t0, [%dst - offset - 0x1c]; \ - srlx %t0, 32, %t0; \ - stw %t0, [%dst - offset - 0x20]; \ - stw %t1, [%dst - offset - 0x14]; \ - srlx %t1, 32, %t1; \ - stw %t1, [%dst - offset - 0x18]; \ - stw %t2, [%dst - offset - 0x0c]; \ - srlx %t2, 32, %t2; \ - stw %t2, [%dst - offset - 0x10]; \ - stw %t3, [%dst - offset - 0x04]; \ - srlx %t3, 32, %t3; \ - stw %t3, [%dst - offset - 0x08]; - -#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src - offset - 0x20], %t0; \ - ldx [%src - offset - 0x18], %t1; \ - ldx [%src - offset - 0x10], %t2; \ - ldx [%src - offset - 0x08], %t3; \ - stx %t0, [%dst - offset - 0x20]; \ - stx %t1, [%dst - offset - 0x18]; \ - stx %t2, [%dst - offset - 0x10]; \ - stx %t3, [%dst - offset - 0x08]; \ - ldx [%src - offset - 0x40], %t0; \ - ldx [%src - offset - 0x38], %t1; \ - ldx [%src - offset - 0x30], %t2; \ - ldx [%src - offset - 0x28], %t3; \ - stx %t0, [%dst - offset - 0x40]; \ - stx %t1, [%dst - offset - 0x38]; \ - stx %t2, [%dst - offset - 0x30]; \ - stx %t3, [%dst - offset - 0x28]; - -#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stw %t0, [%dst + offset + 0x04]; \ - srlx %t0, 32, %t2; \ - stw %t2, [%dst + offset + 0x00]; \ - stw %t1, [%dst + offset + 0x0c]; \ - srlx %t1, 32, %t3; \ - stw %t3, [%dst + offset + 0x08]; - -#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \ - ldx [%src + offset + 0x00], %t0; \ - ldx [%src + offset + 0x08], %t1; \ - stx %t0, [%dst + offset + 0x00]; \ - stx %t1, [%dst + offset + 0x08]; - - .align 32 -228: andcc %o2, 1, %g0 - be,pt %icc, 2f+4 -1: ldub [%o1 - 1], %o5 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - subcc %o2, 1, %o2 - be,pn %XCC, 229f - stb %o5, [%o0] -2: ldub [%o1 - 1], %o5 - sub %o0, 2, %o0 - ldub [%o1 - 2], %g5 - sub %o1, 2, %o1 - subcc %o2, 2, %o2 - stb %o5, [%o0 + 1] - bne,pt %XCC, 2b - stb %g5, [%o0] -229: retl - mov %g4, %o0 -out: retl - mov %g5, %o0 - - .align 32 -ENTRY(memmove) - mov %o0, %g5 -#ifndef USE_BPR - srl %o2, 0, %o2 -#endif - brz,pn %o2, out - sub %o0, %o1, %o4 - cmp %o4, %o2 - bgeu,pt %XCC, 218b - mov %o0, %g4 - add %o0, %o2, %o0 -220: add %o1, %o2, %o1 - cmp %o2, 15 - bleu,pn %XCC, 228b - andcc %o0, 7, %g2 - sub %o0, %o1, %g5 - andcc %g5, 3, %o5 - bne,pn %XCC, 232f - andcc %o1, 3, %g0 - be,a,pt %XCC, 236f - andcc %o1, 4, %g0 - andcc %o1, 1, %g0 - be,pn %XCC, 4f - andcc %o1, 2, %g0 - ldub [%o1 - 1], %g2 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - sub %o2, 1, %o2 - be,pn %XCC, 5f - stb %g2, [%o0] -4: lduh [%o1 - 2], %g2 - sub %o1, 2, %o1 - sub %o0, 2, %o0 - sub %o2, 2, %o2 - sth %g2, [%o0] -5: andcc %o1, 4, %g0 -236: be,a,pn %XCC, 2f - andcc %o2, -128, %g6 - lduw [%o1 - 4], %g5 - sub %o1, 4, %o1 - sub %o0, 4, %o0 - sub %o2, 4, %o2 - stw %g5, [%o0] - andcc %o2, -128, %g6 -2: be,pn %XCC, 235f - andcc %o0, 4, %g0 - be,pn %XCC, 282f + 4 -5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5) - subcc %g6, 128, %g6 - sub %o1, 128, %o1 - bne,pt %XCC, 5b - sub %o0, 128, %o0 -235: andcc %o2, 0x70, %g6 -41: be,pn %XCC, 280f - andcc %o2, 8, %g0 - -279: rd %pc, %o5 - sll %g6, 1, %g5 - sub %o1, %g6, %o1 - sub %o5, %g5, %o5 - jmpl %o5 + %lo(280f - 279b), %g0 - sub %o0, %g6, %o0 - RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5) - RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5) -280: be,pt %XCC, 281f - andcc %o2, 4, %g0 - ldx [%o1 - 8], %g2 - sub %o0, 8, %o0 - stw %g2, [%o0 + 4] - sub %o1, 8, %o1 - srlx %g2, 32, %g2 - stw %g2, [%o0] -281: be,pt %XCC, 1f - andcc %o2, 2, %g0 - lduw [%o1 - 4], %g2 - sub %o1, 4, %o1 - stw %g2, [%o0 - 4] - sub %o0, 4, %o0 -1: be,pt %XCC, 1f - andcc %o2, 1, %g0 - lduh [%o1 - 2], %g2 - sub %o1, 2, %o1 - sth %g2, [%o0 - 2] - sub %o0, 2, %o0 -1: be,pt %XCC, 211f - nop - ldub [%o1 - 1], %g2 - stb %g2, [%o0 - 1] -211: retl - mov %g4, %o0 - -282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5) - RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5) - subcc %g6, 128, %g6 - sub %o1, 128, %o1 - bne,pt %XCC, 282b - sub %o0, 128, %o0 - andcc %o2, 0x70, %g6 - be,pn %XCC, 284f - andcc %o2, 8, %g0 - -283: rd %pc, %o5 - sub %o1, %g6, %o1 - sub %o5, %g6, %o5 - jmpl %o5 + %lo(284f - 283b), %g0 - sub %o0, %g6, %o0 - RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3) - RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3) -284: be,pt %XCC, 285f - andcc %o2, 4, %g0 - ldx [%o1 - 8], %g2 - sub %o0, 8, %o0 - sub %o1, 8, %o1 - stx %g2, [%o0] -285: be,pt %XCC, 1f - andcc %o2, 2, %g0 - lduw [%o1 - 4], %g2 - sub %o0, 4, %o0 - sub %o1, 4, %o1 - stw %g2, [%o0] -1: be,pt %XCC, 1f - andcc %o2, 1, %g0 - lduh [%o1 - 2], %g2 - sub %o0, 2, %o0 - sub %o1, 2, %o1 - sth %g2, [%o0] -1: be,pt %XCC, 1f - nop - ldub [%o1 - 1], %g2 - stb %g2, [%o0 - 1] -1: retl - mov %g4, %o0 - -232: ldub [%o1 - 1], %g5 - sub %o1, 1, %o1 - sub %o0, 1, %o0 - subcc %o2, 1, %o2 - bne,pt %XCC, 232b - stb %g5, [%o0] -234: retl - mov %g4, %o0 -END(memmove) - libc_hidden_builtin_def (memcpy) -libc_hidden_builtin_def (memmove) |