From acd50a9280a85c692874211f27793ec872ac2542 Mon Sep 17 00:00:00 2001 From: Niibe Yutaka Date: Wed, 16 Aug 2000 21:17:40 +0000 Subject: [PATCH] lib1funcs.asm (GLOBAL): Define. * config/sh/lib1funcs.asm (GLOBAL): Define. Use for all references to GLOBAL symbols. Use LOCAL where appropriate. From-SVN: r35752 --- gcc/ChangeLog | 5 + gcc/config/sh/lib1funcs.asm | 282 ++++++++++++++++++------------------ 2 files changed, 148 insertions(+), 139 deletions(-) diff --git a/gcc/ChangeLog b/gcc/ChangeLog index dc038097f4f..5add95cc397 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,8 @@ +2000-08-16 Niibe Yutaka , Kaz Kojima + + * config/sh/lib1funcs.asm (GLOBAL): Define. Use for all + references to GLOBAL symbols. Use LOCAL where appropriate. + 2000-08-16 Kaveh R. Ghazi * configure.in (*-ibm-aix4.[12]*): Delete test for gnu ld. Always diff --git a/gcc/config/sh/lib1funcs.asm b/gcc/config/sh/lib1funcs.asm index c2ba5d344f0..7d7d44c5200 100644 --- a/gcc/config/sh/lib1funcs.asm +++ b/gcc/config/sh/lib1funcs.asm @@ -42,118 +42,122 @@ Boston, MA 02111-1307, USA. */ #define LOCAL(X) L_##X #endif +#ifndef GLOBAL +#define GLOBAL(X) ___##X +#endif + #ifdef L_ashiftrt - .global ___ashiftrt_r4_0 - .global ___ashiftrt_r4_1 - .global ___ashiftrt_r4_2 - .global ___ashiftrt_r4_3 - .global ___ashiftrt_r4_4 - .global ___ashiftrt_r4_5 - .global ___ashiftrt_r4_6 - .global ___ashiftrt_r4_7 - .global ___ashiftrt_r4_8 - .global ___ashiftrt_r4_9 - .global ___ashiftrt_r4_10 - .global ___ashiftrt_r4_11 - .global ___ashiftrt_r4_12 - .global ___ashiftrt_r4_13 - .global ___ashiftrt_r4_14 - .global ___ashiftrt_r4_15 - .global ___ashiftrt_r4_16 - .global ___ashiftrt_r4_17 - .global ___ashiftrt_r4_18 - .global ___ashiftrt_r4_19 - .global ___ashiftrt_r4_20 - .global ___ashiftrt_r4_21 - .global ___ashiftrt_r4_22 - .global ___ashiftrt_r4_23 - .global ___ashiftrt_r4_24 - .global ___ashiftrt_r4_25 - .global ___ashiftrt_r4_26 - .global ___ashiftrt_r4_27 - .global ___ashiftrt_r4_28 - .global ___ashiftrt_r4_29 - .global ___ashiftrt_r4_30 - .global ___ashiftrt_r4_31 - .global ___ashiftrt_r4_32 + .global GLOBAL(ashiftrt_r4_0) + .global GLOBAL(ashiftrt_r4_1) + .global GLOBAL(ashiftrt_r4_2) + .global GLOBAL(ashiftrt_r4_3) + .global GLOBAL(ashiftrt_r4_4) + .global GLOBAL(ashiftrt_r4_5) + .global GLOBAL(ashiftrt_r4_6) + .global GLOBAL(ashiftrt_r4_7) + .global GLOBAL(ashiftrt_r4_8) + .global GLOBAL(ashiftrt_r4_9) + .global GLOBAL(ashiftrt_r4_10) + .global GLOBAL(ashiftrt_r4_11) + .global GLOBAL(ashiftrt_r4_12) + .global GLOBAL(ashiftrt_r4_13) + .global GLOBAL(ashiftrt_r4_14) + .global GLOBAL(ashiftrt_r4_15) + .global GLOBAL(ashiftrt_r4_16) + .global GLOBAL(ashiftrt_r4_17) + .global GLOBAL(ashiftrt_r4_18) + .global GLOBAL(ashiftrt_r4_19) + .global GLOBAL(ashiftrt_r4_20) + .global GLOBAL(ashiftrt_r4_21) + .global GLOBAL(ashiftrt_r4_22) + .global GLOBAL(ashiftrt_r4_23) + .global GLOBAL(ashiftrt_r4_24) + .global GLOBAL(ashiftrt_r4_25) + .global GLOBAL(ashiftrt_r4_26) + .global GLOBAL(ashiftrt_r4_27) + .global GLOBAL(ashiftrt_r4_28) + .global GLOBAL(ashiftrt_r4_29) + .global GLOBAL(ashiftrt_r4_30) + .global GLOBAL(ashiftrt_r4_31) + .global GLOBAL(ashiftrt_r4_32) .align 1 -___ashiftrt_r4_32: -___ashiftrt_r4_31: +GLOBAL(ashiftrt_r4_32): +GLOBAL(ashiftrt_r4_31): rotcl r4 rts subc r4,r4 -___ashiftrt_r4_30: +GLOBAL(ashiftrt_r4_30): shar r4 -___ashiftrt_r4_29: +GLOBAL(ashiftrt_r4_29): shar r4 -___ashiftrt_r4_28: +GLOBAL(ashiftrt_r4_28): shar r4 -___ashiftrt_r4_27: +GLOBAL(ashiftrt_r4_27): shar r4 -___ashiftrt_r4_26: +GLOBAL(ashiftrt_r4_26): shar r4 -___ashiftrt_r4_25: +GLOBAL(ashiftrt_r4_25): shar r4 -___ashiftrt_r4_24: +GLOBAL(ashiftrt_r4_24): shlr16 r4 shlr8 r4 rts exts.b r4,r4 -___ashiftrt_r4_23: +GLOBAL(ashiftrt_r4_23): shar r4 -___ashiftrt_r4_22: +GLOBAL(ashiftrt_r4_22): shar r4 -___ashiftrt_r4_21: +GLOBAL(ashiftrt_r4_21): shar r4 -___ashiftrt_r4_20: +GLOBAL(ashiftrt_r4_20): shar r4 -___ashiftrt_r4_19: +GLOBAL(ashiftrt_r4_19): shar r4 -___ashiftrt_r4_18: +GLOBAL(ashiftrt_r4_18): shar r4 -___ashiftrt_r4_17: +GLOBAL(ashiftrt_r4_17): shar r4 -___ashiftrt_r4_16: +GLOBAL(ashiftrt_r4_16): shlr16 r4 rts exts.w r4,r4 -___ashiftrt_r4_15: +GLOBAL(ashiftrt_r4_15): shar r4 -___ashiftrt_r4_14: +GLOBAL(ashiftrt_r4_14): shar r4 -___ashiftrt_r4_13: +GLOBAL(ashiftrt_r4_13): shar r4 -___ashiftrt_r4_12: +GLOBAL(ashiftrt_r4_12): shar r4 -___ashiftrt_r4_11: +GLOBAL(ashiftrt_r4_11): shar r4 -___ashiftrt_r4_10: +GLOBAL(ashiftrt_r4_10): shar r4 -___ashiftrt_r4_9: +GLOBAL(ashiftrt_r4_9): shar r4 -___ashiftrt_r4_8: +GLOBAL(ashiftrt_r4_8): shar r4 -___ashiftrt_r4_7: +GLOBAL(ashiftrt_r4_7): shar r4 -___ashiftrt_r4_6: +GLOBAL(ashiftrt_r4_6): shar r4 -___ashiftrt_r4_5: +GLOBAL(ashiftrt_r4_5): shar r4 -___ashiftrt_r4_4: +GLOBAL(ashiftrt_r4_4): shar r4 -___ashiftrt_r4_3: +GLOBAL(ashiftrt_r4_3): shar r4 -___ashiftrt_r4_2: +GLOBAL(ashiftrt_r4_2): shar r4 -___ashiftrt_r4_1: +GLOBAL(ashiftrt_r4_1): rts shar r4 -___ashiftrt_r4_0: +GLOBAL(ashiftrt_r4_0): rts nop #endif @@ -161,7 +165,7 @@ ___ashiftrt_r4_0: #ifdef L_ashiftrt_n ! -! ___ashrsi3 +! GLOBAL(ashrsi3) ! ! Entry: ! @@ -177,9 +181,9 @@ ___ashiftrt_r4_0: ! (none) ! - .global ___ashrsi3 + .global GLOBAL(ashrsi3) .align 2 -___ashrsi3: +GLOBAL(ashrsi3): mov #31,r0 and r0,r5 mova LOCAL(ashrsi3_table),r0 @@ -310,7 +314,7 @@ LOCAL(ashrsi3_0): #ifdef L_ashiftlt ! -! ___ashlsi3 +! GLOBAL(ashlsi3) ! ! Entry: ! @@ -325,9 +329,9 @@ LOCAL(ashrsi3_0): ! ! (none) ! - .global ___ashlsi3 + .global GLOBAL(ashlsi3) .align 2 -___ashlsi3: +GLOBAL(ashlsi3): mov #31,r0 and r0,r5 mova LOCAL(ashlsi3_table),r0 @@ -467,7 +471,7 @@ LOCAL(ashlsi3_0): #ifdef L_lshiftrt ! -! ___lshrsi3 +! GLOBAL(lshrsi3) ! ! Entry: ! @@ -482,9 +486,9 @@ LOCAL(ashlsi3_0): ! ! (none) ! - .global ___lshrsi3 + .global GLOBAL(lshrsi3) .align 2 -___lshrsi3: +GLOBAL(lshrsi3): mov #31,r0 and r0,r5 mova LOCAL(lshrsi3_table),r0 @@ -628,84 +632,84 @@ LOCAL(lshrsi3_0): ! jump to movstr+ done: add #64,r5 - mova ___movstrSI0,r0 + mova GLOBAL(movstrSI0),r0 shll2 r6 add r6,r0 jmp @r0 add #64,r4 .align 4 - .global ___movstrSI64 -___movstrSI64: + .global GLOBAL(movstrSI64) +GLOBAL(movstrSI64): mov.l @(60,r5),r0 mov.l r0,@(60,r4) - .global ___movstrSI60 -___movstrSI60: + .global GLOBAL(movstrSI60) +GLOBAL(movstrSI60): mov.l @(56,r5),r0 mov.l r0,@(56,r4) - .global ___movstrSI56 -___movstrSI56: + .global GLOBAL(movstrSI56) +GLOBAL(movstrSI56): mov.l @(52,r5),r0 mov.l r0,@(52,r4) - .global ___movstrSI52 -___movstrSI52: + .global GLOBAL(movstrSI52) +GLOBAL(movstrSI52): mov.l @(48,r5),r0 mov.l r0,@(48,r4) - .global ___movstrSI48 -___movstrSI48: + .global GLOBAL(movstrSI48) +GLOBAL(movstrSI48): mov.l @(44,r5),r0 mov.l r0,@(44,r4) - .global ___movstrSI44 -___movstrSI44: + .global GLOBAL(movstrSI44) +GLOBAL(movstrSI44): mov.l @(40,r5),r0 mov.l r0,@(40,r4) - .global ___movstrSI40 -___movstrSI40: + .global GLOBAL(movstrSI40) +GLOBAL(movstrSI40): mov.l @(36,r5),r0 mov.l r0,@(36,r4) - .global ___movstrSI36 -___movstrSI36: + .global GLOBAL(movstrSI36) +GLOBAL(movstrSI36): mov.l @(32,r5),r0 mov.l r0,@(32,r4) - .global ___movstrSI32 -___movstrSI32: + .global GLOBAL(movstrSI32) +GLOBAL(movstrSI32): mov.l @(28,r5),r0 mov.l r0,@(28,r4) - .global ___movstrSI28 -___movstrSI28: + .global GLOBAL(movstrSI28) +GLOBAL(movstrSI28): mov.l @(24,r5),r0 mov.l r0,@(24,r4) - .global ___movstrSI24 -___movstrSI24: + .global GLOBAL(movstrSI24) +GLOBAL(movstrSI24): mov.l @(20,r5),r0 mov.l r0,@(20,r4) - .global ___movstrSI20 -___movstrSI20: + .global GLOBAL(movstrSI20) +GLOBAL(movstrSI20): mov.l @(16,r5),r0 mov.l r0,@(16,r4) - .global ___movstrSI16 -___movstrSI16: + .global GLOBAL(movstrSI16) +GLOBAL(movstrSI16): mov.l @(12,r5),r0 mov.l r0,@(12,r4) - .global ___movstrSI12 -___movstrSI12: + .global GLOBAL(movstrSI12) +GLOBAL(movstrSI12): mov.l @(8,r5),r0 mov.l r0,@(8,r4) - .global ___movstrSI8 -___movstrSI8: + .global GLOBAL(movstrSI8) +GLOBAL(movstrSI8): mov.l @(4,r5),r0 mov.l r0,@(4,r4) - .global ___movstrSI4 -___movstrSI4: + .global GLOBAL(movstrSI4) +GLOBAL(movstrSI4): mov.l @(0,r5),r0 mov.l r0,@(0,r4) -___movstrSI0: +GLOBAL(movstrSI0): rts nop .align 4 - .global ___movstr -___movstr: + .global GLOBAL(movstr) +GLOBAL(movstr): mov.l @(60,r5),r0 mov.l r0,@(60,r4) @@ -759,16 +763,16 @@ ___movstr: bf done add #64,r5 - bra ___movstr + bra GLOBAL(movstr) add #64,r4 #endif #ifdef L_movstr_i4 #if defined(__SH4__) || defined(__SH4_SINGLE__) || defined(__SH4_SINGLE_ONLY__) .text - .global ___movstr_i4_even - .global ___movstr_i4_odd - .global ___movstrSI12_i4 + .global GLOBAL(movstr_i4_even) + .global GLOBAL(movstr_i4_odd) + .global GLOBAL(movstrSI12_i4) .p2align 5 L_movstr_2mod4_end: @@ -778,7 +782,7 @@ L_movstr_2mod4_end: .p2align 2 -___movstr_i4_odd: +GLOBAL(movstr_i4_odd): mov.l @r5+,r1 add #-4,r4 mov.l @r5+,r2 @@ -804,13 +808,13 @@ L_movstr_start_even: rts mov.l r3,@(12,r4) -___movstr_i4_even: +GLOBAL(movstr_i4_even): mov.l @r5+,r0 bra L_movstr_start_even mov.l @r5+,r1 .p2align 4 -___movstrSI12_i4: +GLOBAL(movstrSI12_i4): mov.l @r5,r0 mov.l @(4,r5),r1 mov.l @(8,r5),r2 @@ -824,7 +828,7 @@ ___movstrSI12_i4: #ifdef L_mulsi3 - .global ___mulsi3 + .global GLOBAL(mulsi3) ! r4 = aabb ! r5 = ccdd @@ -837,7 +841,7 @@ ___movstrSI12_i4: ! aa = bb*dd + (aa*dd*65536) + (cc*bb*65536) ! -___mulsi3: +GLOBAL(mulsi3): mulu r4,r5 ! multiply the lsws macl=bb*dd mov r5,r3 ! r3 = ccdd swap.w r4,r2 ! r2 = bbaa @@ -865,8 +869,8 @@ hiset: sts macl,r0 ! r0 = bb*dd #ifdef __SH4__ !! args in r4 and r5, result in fpul, clobber dr0, dr2 - .global ___sdivsi3_i4 -___sdivsi3_i4: + .global GLOBAL(sdivsi3_i4) +GLOBAL(sdivsi3_i4): lds r4,fpul float fpul,dr0 lds r5,fpul @@ -878,8 +882,8 @@ ___sdivsi3_i4: #elif defined(__SH4_SINGLE__) || defined(__SH4_SINGLE_ONLY__) !! args in r4 and r5, result in fpul, clobber r2, dr0, dr2 - .global ___sdivsi3_i4 -___sdivsi3_i4: + .global GLOBAL(sdivsi3_i4) +GLOBAL(sdivsi3_i4): sts.l fpscr,@-r15 mov #8,r2 swap.w r2,r2 @@ -908,8 +912,8 @@ ___sdivsi3_i4: !! args in r4 and r5, result in r0 clobber r1,r2,r3 - .global ___sdivsi3 -___sdivsi3: + .global GLOBAL(sdivsi3) +GLOBAL(sdivsi3): mov r4,r1 mov r5,r0 @@ -1002,8 +1006,8 @@ div0: rts #ifdef __SH4__ !! args in r4 and r5, result in fpul, clobber r0, r1, r4, r5, dr0, dr2, dr4 - .global ___udivsi3_i4 -___udivsi3_i4: + .global GLOBAL(udivsi3_i4) +GLOBAL(udivsi3_i4): mov #1,r1 cmp/hi r1,r5 bf trivial @@ -1046,8 +1050,8 @@ L1: #elif defined(__SH4_SINGLE__) || defined(__SH4_SINGLE_ONLY__) !! args in r4 and r5, result in fpul, clobber r0, r1, r4, r5, dr0, dr2, dr4 - .global ___udivsi3_i4 -___udivsi3_i4: + .global GLOBAL(udivsi3_i4) +GLOBAL(udivsi3_i4): mov #1,r1 cmp/hi r1,r5 bf trivial @@ -1109,9 +1113,9 @@ L1: !! !! args in r4 and r5, result in r0, clobbers r4, pr, and t bit - .global ___udivsi3 + .global GLOBAL(udivsi3) -___udivsi3: +GLOBAL(udivsi3): longway: mov #0,r0 div0u @@ -1163,10 +1167,10 @@ ret: rts #endif #ifdef L_set_fpscr #if defined (__SH3E__) || defined(__SH4_SINGLE__) || defined(__SH4__) || defined(__SH4_SINGLE_ONLY__) - .global ___set_fpscr -___set_fpscr: + .global GLOBAL(set_fpscr) +GLOBAL(set_fpscr): lds r4,fpscr - mov.l ___set_fpscr_L1,r1 + mov.l LOCAL(set_fpscr_L1),r1 swap.w r4,r0 or #24,r0 #ifndef FMOVD_WORKS @@ -1194,12 +1198,12 @@ ___set_fpscr: mov.l r3,@(4,r1) #endif .align 2 -___set_fpscr_L1: - .long ___fpscr_values +LOCAL(set_fpscr_L1): + .long GLOBAL(fpscr_values) #ifdef __ELF__ - .comm ___fpscr_values,8,4 + .comm GLOBAL(fpscr_values),8,4 #else - .comm ___fpscr_values,8 + .comm GLOBAL(fpscr_values),8 #endif /* ELF */ #endif /* SH3E / SH4 */ #endif /* L_set_fpscr */