More SPE changes

AmigaLabs · Jul 14, 2023 · ad97729 · ad97729
1 parent 11d2e48
commit ad97729
Show file tree

Hide file tree

Showing 73 changed files with 1,744 additions and 1,301 deletions.
diff --git a/GNUmakefile.os4 b/GNUmakefile.os4
@@ -117,9 +117,9 @@ AFLAGS     := -Wa,-mregnames
 
 ifdef SPE
     CC := ppc-amigaos-gcc-6.4.0
-    CFLAGS := $(CFLAGS) -DSPE -D_SOFT_FLOAT -D__SPE__ -mspe -mcpu=8540 -mfloat-gprs=double -mabi=spe -fno-inline-functions -fno-partial-inlining \
+    CFLAGS := $(CFLAGS) -D__SPE__ -mspe -mcpu=8540 -mfloat-gprs=double -mabi=spe -fno-inline-functions -fno-partial-inlining \
                 -fno-align-functions -fno-align-jumps -fno-align-loops -fno-align-labels -fno-inline-small-functions -fno-indirect-inlining -Wno-unused-but-set-variable -Wno-uninitialized
-    AFLAGS := $(AFLAGS) -mspe -mcpu=8540 -mfloat-gprs=double
+    AFLAGS := $(AFLAGS) -D__SPE__ -mspe -mcpu=8540 -mfloat-gprs=double
 endif
 
 VERBOSE    ?= @

diff --git a/README.md b/README.md
@@ -1,4 +1,4 @@
-#  A C runtime library for AmigaOS4
+# A C runtime library for AmigaOS4
 
 [![Build Status](https://travis-ci.com/afxgroup/clib2.svg?branch=master)](https://travis-ci.org/afxgroup/clib2)
 [![License](https://img.shields.io/badge/License-BSD%203--Clause-blue.svg)](https://opensource.org/licenses/BSD-3-Clause)

diff --git a/libc.gmk b/libc.gmk
@@ -211,21 +211,14 @@ C_SEARCH := \
 	search/twalk.o
 
 C_SETJMP := \
+    cpu/altivec/longjmp_altivec.o \
+    cpu/altivec/setjmp_altivec.o \
     setjmp/_longjmp.o \
     setjmp/longjmp.o \
     setjmp/setjmp.o \
     setjmp/sigjmp.o \
-    setjmp/siglongjmp.o
-
-ifdef SPE
-    $(info Using SPE sigssetjmp)
-    C_SETJMP := $(C_SETJMP)  \
-        setjmp/sigsetjmp_spe.o
-else
-    $(info Using NON-SPE sigsetjmp)
-    C_SETJMP := $(C_SETJMP)  \
-        setjmp/sigsetjmp.o
-endif
+    setjmp/siglongjmp.o \
+    setjmp/sigsetjmp.o
 
 C_SOCKET := \
     resolv/dns_parse.o \
@@ -614,15 +607,6 @@ C_STRING := \
 	string/strverscmp.o \
 	string/strxfrm.o
 
-ifdef SPE
-    $(info Adding SPE objects)
-    C_STRING := $(C_STRING) \
-		cpu/spe/memcmpE500.o \
-		cpu/spe/strcmpE500.o \
-		cpu/spe/strcpyE500.o \
-		cpu/spe/strlenE500.o
-endif
-
 C_TERMCAP := \
     termcap/termcap.o \
     termcap/tparam.o
@@ -915,6 +899,22 @@ C_WCHAR := \
 	wchar/wctrans.o \
 	wchar/wctype.o
 
+ifdef SPE
+    $(info Adding SPE objects)
+    C_STRING := $(C_STRING) \
+    	cpu/spe/memcmpE500.o \
+    	cpu/spe/strcmpE500.o \
+    	cpu/spe/strcpyE500.o \
+    	cpu/spe/strlenE500.o
+
+    C_SETJMP := $(C_SETJMP) \
+        cpu/spe/_longjmpE500.o \
+        cpu/spe/_setjmpE500.o \
+        cpu/spe/longjmpE500.o \
+        cpu/spe/setjmpE500.o \
+        cpu/spe/sigsetjmpE500.o
+endif
+
 C_LIBRARY := \
 	$(C_BASE) \
 	$(C_ARGZ) \

diff --git a/libm.gmk b/libm.gmk
@@ -220,7 +220,6 @@ C_MATH := \
 	math/logf.o \
 	math/logl.o \
 	math/lrint.o \
-	math/lrintf.o \
 	math/lrintl.o \
 	math/lround.o \
 	math/lroundf.o \
@@ -293,11 +292,16 @@ C_MATH := \
 	math/truncl.o
 
 ifdef SPE
-	C_MATH := $(C_MATH)\
-		math/fabs_spe.o
+    $(info Using SPE math objects)
+	C_MATH := $(C_MATH) \
+		math/fabs_spe.o \
+		math/fabsf_spe.o \
+		math/lrintf_spe.o
 else
 	C_MATH := $(C_MATH)\
-		math/fabs.o
+		math/fabs.o \
+		math/fabsf.o \
+		math/lrintf.o
 endif
 
 MATH_LIB := \

diff --git a/library/cpu/altivec/longjmp_altivec.sx b/library/cpu/altivec/longjmp_altivec.sx
@@ -0,0 +1,94 @@
+/*
+ * $Id: stdlib_longjmp.c,v 1.6 2010-10-20 13:50:17 clib2devs Exp $
+ */
+
+/* This is a simple version of setjmp and longjmp for the PowerPC.
+   Ian Lance Taylor, Cygnus Support, 9 Feb 1994.
+   Modified by Jeff Johnston, Red Hat Inc. 2 Oct 2001. */
+
+#include "ppc-asm.h"
+
+FUNC_START(longjmp_altivec)
+    addi	3,3,15		# align Altivec to 16 byte boundary
+    rlwinm	3,3,0,0,27
+
+    lwz		1,0(3)		# offset 0
+    lwzu	2,4(3)		# offset 4
+    lwzu	13,4(3)		# offset 8
+    lwzu	14,4(3)		# offset 12
+    lwzu	15,4(3)		# offset 16
+    lwzu	16,4(3)		# offset 20
+    lwzu	17,4(3)		# offset 24
+    lwzu	18,4(3)		# offset 28
+    lwzu	19,4(3)		# offset 32
+    lwzu	20,4(3)		# offset 36
+    lwzu	21,4(3)		# offset 40
+    lwzu	22,4(3)		# offset 44
+    lwzu	23,4(3)		# offset 48
+    lwzu	24,4(3)		# offset 52
+    lwzu	25,4(3)		# offset 56
+    lwzu	26,4(3)		# offset 60
+    lwzu	27,4(3)		# offset 64
+    lwzu	28,4(3)		# offset 68
+    lwzu	29,4(3)		# offset 72
+    lwzu	30,4(3)		# offset 76
+    lwzu	31,4(3)		# offset 80
+    lwzu	5,4(3)		# offset 84
+    mtlr	5
+    lwzu	5,4(3)		# offset 88
+    mtcrf	255,5
+//						# one word pad to get floating point aligned on 8 byte boundary
+#ifndef _SOFT_FLOAT
+    lfdu	14,8(3)         # offset 96
+    lfdu	15,8(3)         # offset 104
+    lfdu	16,8(3)         # offset 112
+    lfdu	17,8(3)         # offset 120
+    lfdu	18,8(3)         # offset 128
+    lfdu	19,8(3)         # offset 136
+    lfdu	20,8(3)         # offset 144
+    lfdu	21,8(3)         # offset 152
+    lfdu	22,8(3)         # offset 160
+    lfdu	23,8(3)         # offset 168
+    lfdu	24,8(3)         # offset 176
+    lfdu	25,8(3)         # offset 184
+    lfdu	26,8(3)         # offset 192
+    lfdu	27,8(3)         # offset 200
+    lfdu	28,8(3)         # offset 208
+    lfdu	29,8(3)         # offset 216
+    lfdu	30,8(3)         # offset 224
+    lfdu	31,8(3)         # offset 232
+#endif
+
+/* restore Altivec vrsave and v20-v31 registers */
+    lwzu	5,16(3)		# offset 248
+    mtspr	256,5		# vrsave
+    addi	3,3,8
+    lvx		20,0,3		# offset 256
+    addi	3,3,16
+    lvx		21,0,3		# offset 272
+    addi	3,3,16
+    lvx		22,0,3		# offset 288
+    addi	3,3,16
+    lvx		23,0,3		# offset 304
+    addi	3,3,16
+    lvx		24,0,3		# offset 320
+    addi	3,3,16
+    lvx		25,0,3		# offset 336
+    addi	3,3,16
+    lvx		26,0,3		# offset 352
+    addi	3,3,16
+    lvx		27,0,3		# offset 368
+    addi	3,3,16
+    lvx		28,0,3		# offset 384
+    addi	3,3,16
+    lvx		29,0,3		# offset 400
+    addi	3,3,16
+    lvx		30,0,3		# offset 416
+    addi	3,3,16
+    lvx		31,0,3		# offset 432
+
+    mr.		3,4
+    bclr+	4,2
+    li		3,1
+    blr
+FUNC_END(longjmp_altivec)
diff --git a/library/cpu/altivec/setjmp_altivec.sx b/library/cpu/altivec/setjmp_altivec.sx
@@ -0,0 +1,98 @@
+/*
+ * $Id: stdlib_setjmp.c,v 2.0 2010-10-20 13:50:17 clib2devs Exp $
+ */
+/* This is a simple version of setjmp and longjmp for the PowerPC.
+   Ian Lance Taylor, Cygnus Support, 9 Feb 1994.
+   Modified by Jeff Johnston, Red Hat Inc. 2 Oct 2001. */
+
+#include "ppc-asm.h"
+
+FUNC_START(setjmp_altivec)
+    addi	3,3,15		# align Altivec to 16 byte boundary
+    rlwinm	3,3,0,0,27
+
+    stw		1,0(3)		# offset 0
+    stwu	2,4(3)		# offset 4
+    stwu	13,4(3)		# offset 8
+    stwu	14,4(3)		# offset 12
+    stwu	15,4(3)		# offset 16
+    stwu	16,4(3)		# offset 20
+    stwu	17,4(3)		# offset 24
+    stwu	18,4(3)		# offset 28
+    stwu	19,4(3)		# offset 32
+    stwu	20,4(3)		# offset 36
+    stwu	21,4(3)		# offset 40
+    stwu	22,4(3)		# offset 44
+    stwu	23,4(3)		# offset 48
+    stwu	24,4(3)		# offset 52
+    stwu	25,4(3)		# offset 56
+    stwu	26,4(3)		# offset 60
+    stwu	27,4(3)		# offset 64
+    stwu	28,4(3)		# offset 68
+    stwu	29,4(3)		# offset 72
+    stwu	30,4(3)		# offset 76
+    stwu	31,4(3)		# offset 80
+    mflr	4
+    stwu	4,4(3)		# offset 84
+    mfcr	4
+    stwu	4,4(3)		# offset 88
+//						# one word pad to get floating point aligned on 8 byte boundary
+#ifndef _SOFT_FLOAT
+    stfdu	14,8(3)		# offset 96
+    stfdu	15,8(3)		# offset 104
+    stfdu	16,8(3)		# offset 112
+    stfdu	17,8(3)		# offset 120
+    stfdu	18,8(3)		# offset 128
+    stfdu	19,8(3)		# offset 136
+    stfdu	20,8(3)		# offset 144
+    stfdu	21,8(3)		# offset 152
+    stfdu	22,8(3)		# offset 160
+    stfdu	23,8(3)		# offset 168
+    stfdu	24,8(3)		# offset 176
+    stfdu	25,8(3)		# offset 184
+    stfdu	26,8(3)		# offset 192
+    stfdu	27,8(3)		# offset 200
+    stfdu	28,8(3)		# offset 208
+    stfdu	29,8(3)		# offset 216
+    stfdu	30,8(3)		# offset 224
+    stfdu	31,8(3)		# offset 232
+#endif
+
+/* This requires a total of 21 * 4 + 18 * 8 + 4 + 4 + 4
+   bytes == 60 * 4 bytes == 240 bytes.  */
+
+/* save Altivec vrsave and vr20-vr31 registers */
+    mfspr	4,256		# vrsave register
+    stwu	4,16(3)		# offset 248
+    addi	3,3,8
+    stvx	20,0,3		# offset 256
+    addi	3,3,16
+    stvx	21,0,3		# offset 272
+    addi	3,3,16
+    stvx	22,0,3		# offset 288
+    addi	3,3,16
+    stvx	23,0,3		# offset 304
+    addi	3,3,16
+    stvx	24,0,3		# offset 320
+    addi	3,3,16
+    stvx	25,0,3		# offset 336
+    addi	3,3,16
+    stvx	26,0,3		# offset 352
+    addi	3,3,16
+    stvx	27,0,3		# offset 368
+    addi	3,3,16
+    stvx	28,0,3		# offset 384
+    addi	3,3,16
+    stvx	29,0,3		# offset 400
+    addi	3,3,16
+    stvx	30,0,3		# offset 416
+    addi	3,3,16
+    stvx	31,0,3		# offset 432
+
+/* This requires a total of 240 + 8 + 8 + 12 * 16 == 448 bytes. */
+
+    li		3,0
+    blr
+FUNC_END(setjmp_altivec)
+
+
diff --git a/library/cpu/spe/_longjmpE500.S b/library/cpu/spe/_longjmpE500.S
@@ -0,0 +1,40 @@
+/*
+ * $Id: _longjmpE500.c,v 1.0 2023-07-14 13:50:17 clib2devs Exp $
+ */
+
+#include "ppc-asm.h"
+
+FUNC_START(_longjmp_spe)
+    evldd	r1,24+0*8(r3)
+    evldd	r2,24+1*8(r3)
+    evldd	r11,24+2*8(r3)
+    evldd	r12,24+3*8(r3)
+    evldd	r13,24+4*8(r3)
+    evldd	r14,24+5*8(r3)
+    evldd	r15,24+6*8(r3)
+    evldd	r16,24+7*8(r3)
+    evldd	r17,24+8*8(r3)
+    evldd	r18,24+9*8(r3)
+    evldd	r19,24+10*8(r3)
+    evldd	r20,24+11*8(r3)
+    evldd	r21,24+12*8(r3)
+    evldd	r22,24+13*8(r3)
+    evldd	r23,24+14*8(r3)
+    evldd	r24,24+15*8(r3)
+    evldd	r25,24+16*8(r3)
+    evldd	r26,24+17*8(r3)
+    evldd	r27,24+18*8(r3)
+    evldd	r28,24+19*8(r3)
+    evldd	r29,24+20*8(r3)
+    evldd	r30,24+21*8(r3)
+    evldd	r31,24+22*8(r3)
+
+    mtlr	r11
+    mtcr	r12
+    or.	r3,r4,r4
+    bnelr
+    li	r3,1
+    blr
+FUNC_END(_longjmp_spe)
+
+.section .note.GNU-stack,"",%progbits
diff --git a/library/cpu/spe/_setjmpE500.S b/library/cpu/spe/_setjmpE500.S
@@ -0,0 +1,38 @@
+/*
+ * $Id: _setjmpE500.c,v 1.0 2023-07-14 13:50:17 clib2devs Exp $
+ */
+
+#include "ppc-asm.h"
+
+FUNC_START(_setjmp_spe)
+	mflr	r11
+	mfcr	r12
+	evstdd	r1,24+0*8(r3)
+	evstdd	r2,24+1*8(r3)
+	evstdd	r11,24+2*8(r3)
+	evstdd	r12,24+3*8(r3)
+	evstdd	r13,24+4*8(r3)
+	evstdd	r14,24+5*8(r3)
+	evstdd	r15,24+6*8(r3)
+	evstdd	r16,24+7*8(r3)
+	evstdd	r17,24+8*8(r3)
+	evstdd	r18,24+9*8(r3)
+	evstdd	r19,24+10*8(r3)
+	evstdd	r20,24+11*8(r3)
+	evstdd	r21,24+12*8(r3)
+	evstdd	r22,24+13*8(r3)
+	evstdd	r23,24+14*8(r3)
+	evstdd	r24,24+15*8(r3)
+	evstdd	r25,24+16*8(r3)
+	evstdd	r26,24+17*8(r3)
+	evstdd	r27,24+18*8(r3)
+	evstdd	r28,24+19*8(r3)
+	evstdd	r29,24+20*8(r3)
+	evstdd	r30,24+21*8(r3)
+	evstdd	r31,24+22*8(r3)
+
+	li	r3,0
+	blr
+FUNC_END(_setjmp_spe)
+
+.section .note.GNU-stack,"",%progbits