|
|
@@ -27,44 +27,82 @@
|
|
|
load subroutines
|
|
|
}
|
|
|
{$ifndef MACOS}
|
|
|
+label
|
|
|
+ _restfpr_14_x,
|
|
|
+ _restfpr_15_x,
|
|
|
+ _restfpr_16_x,
|
|
|
+ _restfpr_17_x,
|
|
|
+ _restfpr_18_x,
|
|
|
+ _restfpr_19_x,
|
|
|
+ _restfpr_20_x,
|
|
|
+ _restfpr_21_x,
|
|
|
+ _restfpr_22_x,
|
|
|
+ _restfpr_23_x,
|
|
|
+ _restfpr_24_x,
|
|
|
+ _restfpr_25_x,
|
|
|
+ _restfpr_26_x,
|
|
|
+ _restfpr_27_x,
|
|
|
+ _restfpr_28_x,
|
|
|
+ _restfpr_29_x,
|
|
|
+ _restfpr_30_x,
|
|
|
+ _restfpr_31_x,
|
|
|
+ _restfpr_14_l,
|
|
|
+ _restfpr_15_l,
|
|
|
+ _restfpr_16_l,
|
|
|
+ _restfpr_17_l,
|
|
|
+ _restfpr_18_l,
|
|
|
+ _restfpr_19_l,
|
|
|
+ _restfpr_20_l,
|
|
|
+ _restfpr_21_l,
|
|
|
+ _restfpr_22_l,
|
|
|
+ _restfpr_23_l,
|
|
|
+ _restfpr_24_l,
|
|
|
+ _restfpr_25_l,
|
|
|
+ _restfpr_26_l,
|
|
|
+ _restfpr_27_l,
|
|
|
+ _restfpr_28_l,
|
|
|
+ _restfpr_29_l,
|
|
|
+ _restfpr_30_l,
|
|
|
+ _restfpr_31_l;
|
|
|
+
|
|
|
procedure saverestorereg;assembler;
|
|
|
asm
|
|
|
{ exit }
|
|
|
-.global _restfpr_14_x
|
|
|
+.globl _restfpr_14_x
|
|
|
_restfpr_14_x: lfd f14, -144(r11)
|
|
|
-.global _restfpr_15_x
|
|
|
+.globl _restfpr_15_x
|
|
|
_restfpr_15_x: lfd f15, -136(r11)
|
|
|
-.global _restfpr_16_x
|
|
|
+.globl _restfpr_16_x
|
|
|
_restfpr_16_x: lfd f16, -128(r11)
|
|
|
-.global _restfpr_17_x
|
|
|
+.globl _restfpr_17_x
|
|
|
_restfpr_17_x: lfd f17, -120(r11)
|
|
|
-.global _restfpr_18_x
|
|
|
+.globl _restfpr_18_x
|
|
|
_restfpr_18_x: lfd f18, -112(r11)
|
|
|
-.global _restfpr_19_x
|
|
|
+.globl _restfpr_19_x
|
|
|
_restfpr_19_x: lfd f19, -104(r11)
|
|
|
-.global _restfpr_20_x
|
|
|
+.globl _restfpr_20_x
|
|
|
_restfpr_20_x: lfd f20, -96(r11)
|
|
|
-.global _restfpr_21_x
|
|
|
+.globl _restfpr_21_x
|
|
|
_restfpr_21_x: lfd f21, -88(r11)
|
|
|
-.global _restfpr_22_x
|
|
|
+.globl _restfpr_22_x
|
|
|
_restfpr_22_x: lfd f22, -80(r11)
|
|
|
-.global _restfpr_23_x
|
|
|
+.globl _restfpr_23_x
|
|
|
_restfpr_23_x: lfd f23, -72(r11)
|
|
|
-.global _restfpr_24_x
|
|
|
+.globl _restfpr_24_x
|
|
|
_restfpr_24_x: lfd f24, -64(r11)
|
|
|
-.global _restfpr_25_x
|
|
|
+.globl _restfpr_25_x
|
|
|
_restfpr_25_x: lfd f25, -56(r11)
|
|
|
-.global _restfpr_26_x
|
|
|
+.globl _restfpr_26_x
|
|
|
_restfpr_26_x: lfd f26, -48(r11)
|
|
|
-.global _restfpr_27_x
|
|
|
+.globl _restfpr_27_x
|
|
|
_restfpr_27_x: lfd f27, -40(r11)
|
|
|
-.global _restfpr_28_x
|
|
|
+.globl _restfpr_28_x
|
|
|
_restfpr_28_x: lfd f28, -32(r11)
|
|
|
-.global _restfpr_29_x
|
|
|
+.globl _restfpr_29_x
|
|
|
_restfpr_29_x: lfd f29, -24(r11)
|
|
|
-.global _restfpr_30_x
|
|
|
+.globl _restfpr_30_x
|
|
|
_restfpr_30_x: lfd f30, -16(r11)
|
|
|
-.global _restfpr_31_x
|
|
|
+.globl _restfpr_31_x
|
|
|
_restfpr_31_x: lwz r0, 4(r11)
|
|
|
lfd f31, -8(r11)
|
|
|
mtlr r0
|
|
|
@@ -72,41 +110,41 @@ _restfpr_31_x: lwz r0, 4(r11)
|
|
|
blr
|
|
|
|
|
|
{ exit with restoring lr }
|
|
|
-.global _restfpr_14_l
|
|
|
+.globl _restfpr_14_l
|
|
|
_restfpr_14_l: lfd f14, -144(r11)
|
|
|
-.global _restfpr_15_l
|
|
|
+.globl _restfpr_15_l
|
|
|
_restfpr_15_l: lfd f15, -136(r11)
|
|
|
-.global _restfpr_16_l
|
|
|
+.globl _restfpr_16_l
|
|
|
_restfpr_16_l: lfd f16, -128(r11)
|
|
|
-.global _restfpr_17_l
|
|
|
+.globl _restfpr_17_l
|
|
|
_restfpr_17_l: lfd f17, -120(r11)
|
|
|
-.global _restfpr_18_l
|
|
|
+.globl _restfpr_18_l
|
|
|
_restfpr_18_l: lfd f18, -112(r11)
|
|
|
-.global _restfpr_19_l
|
|
|
+.globl _restfpr_19_l
|
|
|
_restfpr_19_l: lfd f19, -104(r11)
|
|
|
-.global _restfpr_20_l
|
|
|
+.globl _restfpr_20_l
|
|
|
_restfpr_20_l: lfd f20, -96(r11)
|
|
|
-.global _restfpr_21_l
|
|
|
+.globl _restfpr_21_l
|
|
|
_restfpr_21_l: lfd f21, -88(r11)
|
|
|
-.global _restfpr_22_l
|
|
|
+.globl _restfpr_22_l
|
|
|
_restfpr_22_l: lfd f22, -80(r11)
|
|
|
-.global _restfpr_23_l
|
|
|
+.globl _restfpr_23_l
|
|
|
_restfpr_23_l: lfd f23, -72(r11)
|
|
|
-.global _restfpr_24_l
|
|
|
+.globl _restfpr_24_l
|
|
|
_restfpr_24_l: lfd f24, -64(r11)
|
|
|
-.global _restfpr_25_l
|
|
|
+.globl _restfpr_25_l
|
|
|
_restfpr_25_l: lfd f25, -56(r11)
|
|
|
-.global _restfpr_26_l
|
|
|
+.globl _restfpr_26_l
|
|
|
_restfpr_26_l: lfd f26, -48(r11)
|
|
|
-.global _restfpr_27_l
|
|
|
+.globl _restfpr_27_l
|
|
|
_restfpr_27_l: lfd f27, -40(r11)
|
|
|
-.global _restfpr_28_l
|
|
|
+.globl _restfpr_28_l
|
|
|
_restfpr_28_l: lfd f28, -32(r11)
|
|
|
-.global _restfpr_29_l
|
|
|
+.globl _restfpr_29_l
|
|
|
_restfpr_29_l: lfd f29, -24(r11)
|
|
|
-.global _restfpr_30_l
|
|
|
+.globl _restfpr_30_l
|
|
|
_restfpr_30_l: lfd f30, -16(r11)
|
|
|
-.global _restfpr_31_l
|
|
|
+.globl _restfpr_31_l
|
|
|
_restfpr_31_l: lwz r0, 4(r11)
|
|
|
lfd f31, -8(r11)
|
|
|
mtlr r0
|
|
|
@@ -140,7 +178,7 @@ asm
|
|
|
cmpwi cr7,r5,63
|
|
|
|
|
|
{ if count <= 0, stop }
|
|
|
- ble cr0,LMoveDone
|
|
|
+ ble cr0,.LMoveDone
|
|
|
|
|
|
{ load the begin of the source in the data cache }
|
|
|
dcbt 0,r3
|
|
|
@@ -160,7 +198,7 @@ asm
|
|
|
addi r10,r10,1
|
|
|
|
|
|
{ if count < 15, copy everything byte by byte }
|
|
|
- blt cr1,LMoveBytes
|
|
|
+ blt cr1,.LMoveBytes
|
|
|
|
|
|
{ if no overlap, then source/dest += -1, otherwise they stay }
|
|
|
{ After the next instruction, r3/r4 + r10 = next position to }
|
|
|
@@ -169,14 +207,14 @@ asm
|
|
|
add r4,r4,r6
|
|
|
|
|
|
{ otherwise, guarantee 4 byte alignment for dest for starters }
|
|
|
-LMove4ByteAlignLoop:
|
|
|
+.LMove4ByteAlignLoop:
|
|
|
lbzux r0,r3,r10
|
|
|
stbux r0,r4,r10
|
|
|
{ is dest now 4 aligned? }
|
|
|
andi. r0,r4,3
|
|
|
subi r5,r5,1
|
|
|
{ while not aligned, continue }
|
|
|
- bne cr0,LMove4ByteAlignLoop
|
|
|
+ bne cr0,.LMove4ByteAlignLoop
|
|
|
|
|
|
{$ifndef ppc603}
|
|
|
{ check for 32 byte alignment }
|
|
|
@@ -215,17 +253,17 @@ LMove4ByteAlignLoop:
|
|
|
add r4,r4,r6
|
|
|
|
|
|
{$ifndef ppc603}
|
|
|
- beq cr0,LMove32BytesAligned
|
|
|
-L32BytesAlignMoveLoop:
|
|
|
+ beq cr0,.LMove32BytesAligned
|
|
|
+.L32BytesAlignMoveLoop:
|
|
|
{ count >= 39 -> align to 8 byte boundary and then use the FPU }
|
|
|
{ since we're already at 4 byte alignment, use dword store }
|
|
|
subic. r7,r7,4
|
|
|
lwzux r0,r3,r10
|
|
|
subi r5,r5,4
|
|
|
stwux r0,r4,r10
|
|
|
- bne L32BytesAlignMoveLoop
|
|
|
+ bne .L32BytesAlignMoveLoop
|
|
|
|
|
|
-LMove32BytesAligned:
|
|
|
+.LMove32BytesAligned:
|
|
|
{ count div 32 ( >= 1, since count was >=63 }
|
|
|
srwi r0,r5,5
|
|
|
{ remainder }
|
|
|
@@ -268,7 +306,7 @@ LMove32BytesAligned:
|
|
|
neg r7,r7
|
|
|
subi r7,r7,32
|
|
|
|
|
|
-LMove32ByteDcbz:
|
|
|
+.LMove32ByteDcbz:
|
|
|
lfdux f0,r3,r10
|
|
|
lfdux f1,r3,r10
|
|
|
lfdux f2,r3,r10
|
|
|
@@ -280,10 +318,10 @@ LMove32ByteDcbz:
|
|
|
stfdux f1,r4,r10
|
|
|
stfdux f2,r4,r10
|
|
|
stfdux f3,r4,r10
|
|
|
- bdnz LMove32ByteDcbz
|
|
|
-LMove32ByteLoopDone:
|
|
|
+ bdnz .LMove32ByteDcbz
|
|
|
+.LMove32ByteLoopDone:
|
|
|
{$else not ppc603}
|
|
|
-LMove16ByteLoop:
|
|
|
+.LMove16ByteLoop:
|
|
|
lwzux r11,r3,r10
|
|
|
lwzux r7,r3,r10
|
|
|
lwzux r8,r3,r10
|
|
|
@@ -292,11 +330,11 @@ LMove16ByteLoop:
|
|
|
stwux r7,r4,r10
|
|
|
stwux r8,r4,r10
|
|
|
stwux r9,r4,r10
|
|
|
- bdnz LMove16ByteLoop
|
|
|
+ bdnz .LMove16ByteLoop
|
|
|
{$endif not ppc603}
|
|
|
|
|
|
{ cr0*4+eq is true if "count and 31" = 0 }
|
|
|
- beq cr0,LMoveDone
|
|
|
+ beq cr0,.LMoveDone
|
|
|
|
|
|
{ make r10 again -1 or 1, but first adjust source/dest pointers }
|
|
|
sub r3,r3,r6
|
|
|
@@ -310,9 +348,9 @@ LMove16ByteLoop:
|
|
|
{$endif not ppc603}
|
|
|
|
|
|
{ cr1 contains whether count <= 11 }
|
|
|
- ble cr1,LMoveBytes
|
|
|
+ ble cr1,.LMoveBytes
|
|
|
|
|
|
-LMoveDWords:
|
|
|
+.LMoveDWords:
|
|
|
mtctr r0
|
|
|
andi. r5,r5,3
|
|
|
{ r10 * 4 }
|
|
|
@@ -321,26 +359,26 @@ LMoveDWords:
|
|
|
add r3,r3,r6
|
|
|
add r4,r4,r6
|
|
|
|
|
|
-LMoveDWordsLoop:
|
|
|
+.LMoveDWordsLoop:
|
|
|
lwzux r0,r3,r10
|
|
|
stwux r0,r4,r10
|
|
|
- bdnz LMoveDWordsLoop
|
|
|
+ bdnz .LMoveDWordsLoop
|
|
|
|
|
|
- beq cr0,LMoveDone
|
|
|
+ beq cr0,.LMoveDone
|
|
|
{ make r10 again -1 or 1 }
|
|
|
sub r3,r3,r6
|
|
|
sub r4,r4,r6
|
|
|
srawi r10,r10,2
|
|
|
srawi r6,r6,2
|
|
|
-LMoveBytes:
|
|
|
+.LMoveBytes:
|
|
|
add r3,r3,r6
|
|
|
add r4,r4,r6
|
|
|
mtctr r5
|
|
|
-LMoveBytesLoop:
|
|
|
+.LMoveBytesLoop:
|
|
|
lbzux r0,r3,r10
|
|
|
stbux r0,r4,r10
|
|
|
- bdnz LMoveBytesLoop
|
|
|
-LMoveDone:
|
|
|
+ bdnz .LMoveBytesLoop
|
|
|
+.LMoveDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -371,12 +409,12 @@ asm
|
|
|
{ setup for aligning x to multiple of 4}
|
|
|
rlwinm r10,r3,0,31-2+1,31
|
|
|
rlwimi r5,r5,16,0,15
|
|
|
- ble cr6,LFillCharDone
|
|
|
+ ble cr6,.LFillCharDone
|
|
|
{ get the start of the data in the cache (and mark it as "will be }
|
|
|
{ modified") }
|
|
|
dcbtst 0,r3
|
|
|
subfic r10,r10,4
|
|
|
- blt cr7,LFillCharVerySmall
|
|
|
+ blt cr7,.LFillCharVerySmall
|
|
|
{ just store 4 bytes instead of using a loop to align (there are }
|
|
|
{ plenty of other instructions now to keep the processor busy }
|
|
|
{ while it handles the (possibly unaligned) store) }
|
|
|
@@ -385,7 +423,7 @@ asm
|
|
|
add r3,r3,r10
|
|
|
{ decrease count with number of bytes already stored }
|
|
|
sub r4,r4,r10
|
|
|
- blt cr1,LFillCharSmall
|
|
|
+ blt cr1,.LFillCharSmall
|
|
|
{ if we have to fill with 0 (which happens a lot), we can simply use }
|
|
|
{ dcbz for the most part, which is very fast, so make a special case }
|
|
|
{ for that }
|
|
|
@@ -398,16 +436,16 @@ asm
|
|
|
{ get number of bytes to store }
|
|
|
subfic r10,r10,32
|
|
|
{ if already 32byte aligned, skip align loop }
|
|
|
- beq L32ByteAlignLoopDone
|
|
|
+ beq .L32ByteAlignLoopDone
|
|
|
{ substract from the total count }
|
|
|
sub r4,r4,r10
|
|
|
-L32ByteAlignLoop:
|
|
|
+.L32ByteAlignLoop:
|
|
|
{ we were already aligned to 4 byres, so this will count down to }
|
|
|
{ exactly 0 }
|
|
|
subic. r10,r10,4
|
|
|
stwu r5,4(r3)
|
|
|
- bne L32ByteAlignLoop
|
|
|
-L32ByteAlignLoopDone:
|
|
|
+ bne .L32ByteAlignLoop
|
|
|
+.L32ByteAlignLoopDone:
|
|
|
{ get the amount of 32 byte blocks }
|
|
|
srwi r10,r4,5
|
|
|
{ and keep the rest in r4 (recording whether there is any rest) }
|
|
|
@@ -421,14 +459,14 @@ L32ByteAlignLoopDone:
|
|
|
bne cr1, LFillCharNoZero
|
|
|
{ make r3 point again to the actual store position }
|
|
|
addi r3,r3,4
|
|
|
-LFillCharDCBZLoop:
|
|
|
+.LFillCharDCBZLoop:
|
|
|
dcbz 0,r3
|
|
|
addi r3,r3,32
|
|
|
- bdnz LFillCharDCBZLoop
|
|
|
+ bdnz .LFillCharDCBZLoop
|
|
|
{ if there was no rest, we're finished }
|
|
|
- beq LFillCharDone
|
|
|
- b LFillCharVerySmall
|
|
|
-LFillCharNoZero:
|
|
|
+ beq .LFillCharDone
|
|
|
+ b .LFillCharVerySmall
|
|
|
+.LFillCharNoZero:
|
|
|
{$ifdef FPC_ABI_AIX}
|
|
|
stw r5,0(r1)
|
|
|
stw r5,4(r1)
|
|
|
@@ -443,19 +481,19 @@ LFillCharNoZero:
|
|
|
subi r3,r3,4
|
|
|
{ load r10 with 8, so that dcbz uses the correct address }
|
|
|
li r10, 8
|
|
|
-LFillChar32ByteLoop:
|
|
|
+.LFillChar32ByteLoop:
|
|
|
dcbz r3,r10
|
|
|
stfdu f0,8(r3)
|
|
|
stfdu f0,8(r3)
|
|
|
stfdu f0,8(r3)
|
|
|
stfdu f0,8(r3)
|
|
|
- bdnz LFillChar32ByteLoop
|
|
|
+ bdnz .LFillChar32ByteLoop
|
|
|
{ if there was no rest, we're finished }
|
|
|
- beq LFillCharDone
|
|
|
+ beq .LFillCharDone
|
|
|
{ make r3 point again to the actual next byte that must be written }
|
|
|
addi r3,r3,8
|
|
|
- b LFillCharVerySmall
|
|
|
-LFillCharSmall:
|
|
|
+ b .LFillCharVerySmall
|
|
|
+.LFillCharSmall:
|
|
|
{ when we arrive here, we're already 4 byte aligned }
|
|
|
{ get count div 4 to store dwords }
|
|
|
srwi r10,r4,2
|
|
|
@@ -463,20 +501,20 @@ LFillCharSmall:
|
|
|
subi r3,r3,4
|
|
|
mtctr r10
|
|
|
rlwinm. r4,r4,0,31-2+1,31
|
|
|
-LFillCharSmallLoop:
|
|
|
+.LFillCharSmallLoop:
|
|
|
stwu r5,4(r3)
|
|
|
- bdnz LFillCharSmallLoop
|
|
|
+ bdnz .LFillCharSmallLoop
|
|
|
{ if nothing left, stop }
|
|
|
- beq LFillCharDone
|
|
|
+ beq .LFillCharDone
|
|
|
{ get ready to store bytes }
|
|
|
addi r3,r3,4
|
|
|
-LFillCharVerySmall:
|
|
|
+.LFillCharVerySmall:
|
|
|
mtctr r4
|
|
|
subi r3,r3,1
|
|
|
-LFillCharVerySmallLoop:
|
|
|
+.LFillCharVerySmallLoop:
|
|
|
stbu r5,1(r3)
|
|
|
- bdnz LFillCharVerySmallLoop
|
|
|
-LFillCharDone:
|
|
|
+ bdnz .LFillCharVerySmallLoop
|
|
|
+.LFillCharDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -492,11 +530,11 @@ asm
|
|
|
cmpwi cr0,r4,0
|
|
|
mtctr r4
|
|
|
subi r3,r3,4
|
|
|
- ble LFillDWordEnd //if count<=0 Then Exit
|
|
|
-LFillDWordLoop:
|
|
|
+ ble .LFillDWordEnd //if count<=0 Then Exit
|
|
|
+.LFillDWordLoop:
|
|
|
stwu r5,4(r3)
|
|
|
- bdnz LFillDWordLoop
|
|
|
-LFillDWordEnd:
|
|
|
+ bdnz .LFillDWordLoop
|
|
|
+.LFillDWordEnd:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -513,15 +551,15 @@ asm
|
|
|
mr r0,r3
|
|
|
{ assume not found }
|
|
|
li r3,-1
|
|
|
- ble LIndexByteDone
|
|
|
-LIndexByteLoop:
|
|
|
+ ble .LIndexByteDone
|
|
|
+.LIndexByteLoop:
|
|
|
lbzu r9,1(r10)
|
|
|
cmplw r9,r5
|
|
|
- bdnzf cr0*4+eq,LIndexByteLoop
|
|
|
+ bdnzf cr0*4+eq,.LIndexByteLoop
|
|
|
{ r3 still contains -1 here }
|
|
|
- bne LIndexByteDone
|
|
|
+ bne .LIndexByteDone
|
|
|
sub r3,r10,r0
|
|
|
-LIndexByteDone:
|
|
|
+.LIndexByteDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -538,15 +576,15 @@ asm
|
|
|
mr r0,r3
|
|
|
{ assume not found }
|
|
|
li r3,-1
|
|
|
- ble LIndexWordDone
|
|
|
-LIndexWordLoop:
|
|
|
+ ble .LIndexWordDone
|
|
|
+.LIndexWordLoop:
|
|
|
lhzu r9,2(r10)
|
|
|
cmplw r9,r5
|
|
|
- bdnzf cr0*4+eq,LIndexWordLoop
|
|
|
+ bdnzf cr0*4+eq,.LIndexWordLoop
|
|
|
{ r3 still contains -1 here }
|
|
|
- bne LIndexWordDone
|
|
|
+ bne .LIndexWordDone
|
|
|
sub r3,r10,r0
|
|
|
-LIndexWordDone:
|
|
|
+.LIndexWordDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -563,15 +601,15 @@ asm
|
|
|
mr r0,r3
|
|
|
{ assume not found }
|
|
|
li r3,-1
|
|
|
- ble LIndexDWordDone
|
|
|
-LIndexDWordLoop:
|
|
|
+ ble .LIndexDWordDone
|
|
|
+.LIndexDWordLoop:
|
|
|
lwzu r9,4(r10)
|
|
|
cmplw r9,r5
|
|
|
- bdnzf cr0*4+eq, LIndexDWordLoop
|
|
|
+ bdnzf cr0*4+eq, .LIndexDWordLoop
|
|
|
{ r3 still contains -1 here }
|
|
|
- bne LIndexDWordDone
|
|
|
+ bne .LIndexDWordDone
|
|
|
sub r3,r10,r0
|
|
|
-LIndexDWordDone:
|
|
|
+.LIndexDWordDone:
|
|
|
end;
|
|
|
|
|
|
{$define FPC_SYSTEM_HAS_COMPAREBYTE}
|
|
|
@@ -588,16 +626,16 @@ asm
|
|
|
subi r11,r3,1
|
|
|
subi r4,r4,1
|
|
|
li r3,0
|
|
|
- ble LCompByteDone
|
|
|
-LCompByteLoop:
|
|
|
+ ble .LCompByteDone
|
|
|
+.LCompByteLoop:
|
|
|
{ load next chars }
|
|
|
lbzu r9,1(r11)
|
|
|
lbzu r10,1(r4)
|
|
|
{ calculate difference }
|
|
|
sub. r3,r9,r10
|
|
|
{ if chars not equal or at the end, we're ready }
|
|
|
- bdnzt cr0*4+eq, LCompByteLoop
|
|
|
-LCompByteDone:
|
|
|
+ bdnzt cr0*4+eq, .LCompByteLoop
|
|
|
+.LCompByteDone:
|
|
|
end;
|
|
|
|
|
|
{$define FPC_SYSTEM_HAS_COMPAREWORD}
|
|
|
@@ -614,16 +652,16 @@ asm
|
|
|
subi r11,r3,2
|
|
|
subi r4,r4,2
|
|
|
li r3,0
|
|
|
- ble LCompWordDone
|
|
|
-LCompWordLoop:
|
|
|
+ ble .LCompWordDone
|
|
|
+.LCompWordLoop:
|
|
|
{ load next chars }
|
|
|
lhzu r9,2(r11)
|
|
|
lhzu r10,2(r4)
|
|
|
{ calculate difference }
|
|
|
sub. r3,r9,r10
|
|
|
{ if chars not equal or at the end, we're ready }
|
|
|
- bdnzt cr0*4+eq, LCompWordLoop
|
|
|
-LCompWordDone:
|
|
|
+ bdnzt cr0*4+eq, .LCompWordLoop
|
|
|
+.LCompWordDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -641,16 +679,16 @@ asm
|
|
|
subi r11,r3,4
|
|
|
subi r4,r4,4
|
|
|
li r3,0
|
|
|
- ble LCompDWordDone
|
|
|
-LCompDWordLoop:
|
|
|
+ ble .LCompDWordDone
|
|
|
+.LCompDWordLoop:
|
|
|
{ load next chars }
|
|
|
lwzu r9,4(r11)
|
|
|
lwzu r10,4(r4)
|
|
|
{ calculate difference }
|
|
|
sub. r3,r9,r10
|
|
|
{ if chars not equal or at the end, we're ready }
|
|
|
- bdnzt cr0*4+eq, LCompDWordLoop
|
|
|
-LCompDWordDone:
|
|
|
+ bdnzt cr0*4+eq, .LCompDWordLoop
|
|
|
+.LCompDWordDone:
|
|
|
end;
|
|
|
|
|
|
{$define FPC_SYSTEM_HAS_INDEXCHAR0}
|
|
|
@@ -668,16 +706,16 @@ asm
|
|
|
{ assume not found }
|
|
|
li r3,-1
|
|
|
{ if yes, do nothing }
|
|
|
- ble LIndexChar0Done
|
|
|
-LIndexChar0Loop:
|
|
|
+ ble .LIndexChar0Done
|
|
|
+.LIndexChar0Loop:
|
|
|
lbzu r10,1(r9)
|
|
|
cmplwi cr1,r10,0
|
|
|
cmplw r10,r5
|
|
|
- beq cr1,LIndexChar0Done
|
|
|
- bdnzf cr0*4+eq, LIndexChar0Loop
|
|
|
- bne LIndexChar0Done
|
|
|
+ beq cr1,.LIndexChar0Done
|
|
|
+ bdnzf cr0*4+eq, .LIndexChar0Loop
|
|
|
+ bne .LIndexChar0Done
|
|
|
sub r3,r9,r0
|
|
|
-LIndexChar0Done:
|
|
|
+.LIndexChar0Done:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -705,12 +743,12 @@ asm
|
|
|
{ put length in ctr }
|
|
|
mtctr r4
|
|
|
stb r4,0(r3)
|
|
|
- beq LShortStrCopyDone
|
|
|
-LShortStrCopyLoop:
|
|
|
+ beq .LShortStrCopyDone
|
|
|
+.LShortStrCopyLoop:
|
|
|
lbzu r0,1(r5)
|
|
|
stbu r0,1(r3)
|
|
|
- bdnz LShortStrCopyLoop
|
|
|
-LShortStrCopyDone:
|
|
|
+ bdnz .LShortStrCopyLoop
|
|
|
+.LShortStrCopyDone:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -738,12 +776,12 @@ asm
|
|
|
{ put length in ctr }
|
|
|
mtctr r3
|
|
|
stb r3,0(r5)
|
|
|
- beq LShortStrCopyDone2
|
|
|
-LShortStrCopyLoop2:
|
|
|
+ beq .LShortStrCopyDone2
|
|
|
+.LShortStrCopyLoop2:
|
|
|
lbzu r0,1(r4)
|
|
|
stbu r0,1(r5)
|
|
|
- bdnz LShortStrCopyLoop2
|
|
|
-LShortStrCopyDone2:
|
|
|
+ bdnz .LShortStrCopyLoop2
|
|
|
+.LShortStrCopyDone2:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -776,15 +814,15 @@ asm
|
|
|
stb r10,0(r3)
|
|
|
{ go to last current character of result }
|
|
|
add r3,r6,r3
|
|
|
-
|
|
|
+
|
|
|
{ if nothing to do, exit }
|
|
|
- beq LShortStrAppendDone
|
|
|
+ beq .LShortStrAppendDone
|
|
|
{ and concatenate }
|
|
|
-LShortStrAppendLoop:
|
|
|
+.LShortStrAppendLoop:
|
|
|
lbzu r10,1(r5)
|
|
|
stbu r10,1(r3)
|
|
|
- bdnz LShortStrAppendLoop
|
|
|
-LShortStrAppendDone:
|
|
|
+ bdnz .LShortStrAppendLoop
|
|
|
+.LShortStrAppendDone:
|
|
|
end;
|
|
|
|
|
|
(*
|
|
|
@@ -870,16 +908,16 @@ end;
|
|
|
function get_caller_addr(framebp:pointer):pointer;assembler;
|
|
|
asm
|
|
|
cmplwi r3,0
|
|
|
- beq Lcaller_addr_frame_null
|
|
|
+ beq .Lcaller_addr_frame_null
|
|
|
lwz r3,0(r3)
|
|
|
cmplwi r3,0
|
|
|
- beq Lcaller_addr_frame_null
|
|
|
+ beq .Lcaller_addr_frame_null
|
|
|
{$ifdef FPC_ABI_AIX}
|
|
|
lwz r3,8(r3)
|
|
|
{$else FPC_ABI_AIX}
|
|
|
lwz r3,4(r3)
|
|
|
{$endif FPC_ABI_AIX}
|
|
|
-Lcaller_addr_frame_null:
|
|
|
+.Lcaller_addr_frame_null:
|
|
|
end;
|
|
|
|
|
|
|
|
|
@@ -887,9 +925,9 @@ end;
|
|
|
function get_caller_frame(framebp:pointer):pointer;assembler;
|
|
|
asm
|
|
|
cmplwi r3,0
|
|
|
- beq Lcaller_frame_null
|
|
|
+ beq .Lcaller_frame_null
|
|
|
lwz r3,0(r3)
|
|
|
-Lcaller_frame_null:
|
|
|
+.Lcaller_frame_null:
|
|
|
end;
|
|
|
|
|
|
{$define FPC_SYSTEM_HAS_ABS_LONGINT}
|
|
|
@@ -920,7 +958,7 @@ end;
|
|
|
|
|
|
|
|
|
{$define FPC_SYSTEM_HAS_SPTR}
|
|
|
-Function Sptr : Longint;assembler;
|
|
|
+Function Sptr : Pointer;assembler;
|
|
|
asm
|
|
|
mr r3,r1
|
|
|
end;
|
|
|
@@ -944,11 +982,11 @@ function declocked(var l : longint) : boolean;assembler;
|
|
|
{ input: address of l in r3 }
|
|
|
{ output: boolean indicating whether l is zero after decrementing }
|
|
|
asm
|
|
|
-LDecLockedLoop:
|
|
|
+.LDecLockedLoop:
|
|
|
lwarx r10,0,r3
|
|
|
subi r10,r10,1
|
|
|
stwcx. r10,0,r3
|
|
|
- bne- LDecLockedLoop
|
|
|
+ bne- .LDecLockedLoop
|
|
|
cntlzw r3,r10
|
|
|
srwi r3,r3,5
|
|
|
end;
|
|
|
@@ -956,17 +994,20 @@ end;
|
|
|
{$define FPC_SYSTEM_HAS_INCLOCKED}
|
|
|
procedure inclocked(var l : longint);assembler;
|
|
|
asm
|
|
|
-LIncLockedLoop:
|
|
|
+.LIncLockedLoop:
|
|
|
lwarx r10,0,r3
|
|
|
addi r10,r10,1
|
|
|
stwcx. r10,0,r3
|
|
|
- bne- LIncLockedLoop
|
|
|
+ bne- .LIncLockedLoop
|
|
|
end;
|
|
|
|
|
|
|
|
|
{
|
|
|
$Log$
|
|
|
- Revision 1.54 2003-09-14 20:33:28 jonas
|
|
|
+ Revision 1.55 2003-11-15 19:01:27 florian
|
|
|
+ * fixed rtl to work with the integrated fpc ppc assembler reader
|
|
|
+
|
|
|
+ Revision 1.54 2003/09/14 20:33:28 jonas
|
|
|
* renamed sp to r1, gnu as doesn't understand sp
|
|
|
|
|
|
Revision 1.53 2003/09/06 10:44:41 olle
|
florian