| 1 |
/* Unsigned 32 bit division optimized for Epiphany. |
| 2 |
Copyright (C) 2009-2025 Free Software Foundation, Inc. |
| 3 |
Contributed by Embecosm on behalf of Adapteva, Inc. |
| 4 |
|
| 5 |
This file is part of GCC. |
| 6 |
|
| 7 |
This file is free software; you can redistribute it and/or modify it |
| 8 |
under the terms of the GNU General Public License as published by the |
| 9 |
Free Software Foundation; either version 3, or (at your option) any |
| 10 |
later version. |
| 11 |
|
| 12 |
This file is distributed in the hope that it will be useful, but |
| 13 |
WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 15 |
General Public License for more details. |
| 16 |
|
| 17 |
Under Section 7 of GPL version 3, you are granted additional |
| 18 |
permissions described in the GCC Runtime Library Exception, version |
| 19 |
3.1, as published by the Free Software Foundation. |
| 20 |
|
| 21 |
You should have received a copy of the GNU General Public License and |
| 22 |
a copy of the GCC Runtime Library Exception along with this program; |
| 23 |
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see |
| 24 |
<http://www.gnu.org/licenses/>. */ |
| 25 |
|
| 26 |
#include "epiphany-asm.h" |
| 27 |
|
| 28 |
FSTAB (__udivsi3,T_UINT) |
| 29 |
.global SYM(__udivsi3) |
| 30 |
.balign 4 |
| 31 |
HIDDEN_FUNC(__udivsi3) |
| 32 |
SYM(__udivsi3): |
| 33 |
sub TMP0,r0,r1 |
| 34 |
bltu .Lret0 |
| 35 |
float TMP2,r0 |
| 36 |
mov TMP1,%low(0xb0800000) ; ??? this would be faster with small data |
| 37 |
float TMP3,r1 |
| 38 |
movt TMP1,%high(0xb0800000) |
| 39 |
asr TMP0,r0,8 |
| 40 |
sub TMP0,TMP0,TMP1 |
| 41 |
movt TMP1,%high(0x00810000) |
| 42 |
movgteu TMP2,TMP0 |
| 43 |
bblt .Lret1 |
| 44 |
sub TMP2,TMP2,TMP1 |
| 45 |
sub TMP2,TMP2,TMP3 |
| 46 |
mov TMP3,0 |
| 47 |
movltu TMP2,TMP3 |
| 48 |
lsr TMP2,TMP2,23 |
| 49 |
lsl r1,r1,TMP2 |
| 50 |
mov TMP0,1 |
| 51 |
lsl TMP0,TMP0,TMP2 |
| 52 |
sub r0,r0,r1 |
| 53 |
bltu .Ladd_back |
| 54 |
add TMP3,TMP3,TMP0 |
| 55 |
sub r0,r0,r1 |
| 56 |
bltu .Ladd_back |
| 57 |
.Lsub_loop:; More than two iterations are rare, so it makes sense to leave |
| 58 |
; this label here to reduce average branch penalties. |
| 59 |
add TMP3,TMP3,TMP0 |
| 60 |
sub r0,r0,r1 |
| 61 |
bgteu .Lsub_loop |
| 62 |
.Ladd_back: |
| 63 |
add r0,r0,r1 |
| 64 |
sub TMP1,r1,1 |
| 65 |
mov r1,%low(.L0step) |
| 66 |
movt r1,%high(.L0step) |
| 67 |
lsl TMP2,TMP2,3 |
| 68 |
sub r1,r1,TMP2 |
| 69 |
jr r1 |
| 70 |
.rep 30 |
| 71 |
lsl r0,r0,1 |
| 72 |
sub.l r1,r0,TMP1 |
| 73 |
movgteu r0,r1 |
| 74 |
.endr |
| 75 |
.L0step:sub r1,TMP0,1 ; mask result bits from steps ... |
| 76 |
and r0,r0,r1 |
| 77 |
orr r0,r0,TMP3 ; ... and combine with first bits. |
| 78 |
rts |
| 79 |
.Lret0: mov r0,0 |
| 80 |
rts |
| 81 |
.Lret1: mov r0,1 |
| 82 |
rts |
| 83 |
ENDFUNC(__udivsi3) |