1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
|
.globl func_add1, func_add1_end
.proc func_add1
func_add1:
{.mib; add r8 = 1, r32
nop.i 0
br.ret.sptk.many rp
}
func_add1_end:
.endp func_add1
.globl func_add3, func_add3_end
.proc func_add3
func_add3:
{.mmi; alloc loc0 = ar.pfs, 2, 1, 2, 0
mov r2 = sp
add sp = -16, sp
} ;;
{.mii; ld8 r8 = [in1], 8 // load the function pointer
mov r3 = rp
mov rp = loc0 // trash rp
} ;;
{.mmi; ld8 r9 = [r8], 8 // load the entry-point
st8 [r2] = r3
mov out0 = in0
} ;;
{.mii; ld8 gp = [r8] // load the gp
mov b6 = r9
mov out1 = in1
}
{.mib; nop 0
nop 0
br.call.sptk rp = b6
}
{.mmi; add r2 = 16, sp
;;
ld8 r3 = [r2] // r3 = saved rp
mov ar.pfs = loc0
} ;;
{.mii; nop 0
mov rp = r3
adds sp = 16, sp
} ;;
{.mib; st8 [sp] = in0 // trash rp save location
add r8 = 2, r8
br.ret.sptk.many rp
}
func_add3_end:
.endp func_add3
.globl func_vframe, func_vframe_end
.proc func_vframe
func_vframe:
{.mii; alloc r16 = ar.pfs, 1, 2, 0, 0 // 0
mov loc0 = rp
mov loc1 = sp
} ;;
{.mmi; sub sp = sp, in0
st8 [loc1] = r16
mov r2 = -99 // 0
} ;;
{.mii; nop 0
mov rp = r2
mov ar.pfs = r0
}
{.mib; mov r16 = r2
tbit.nz p6, p0 = in0, 4
(p6) br.cond.sptk.many .exit
} ;;
{.mmi; ld8 r16 = [loc1]
;;
mov r3 = loc0 // 8 move saved rp to r3
mov ar.pfs = r16
} ;;
{.mmi; mov sp = loc1 // 10
st8 [loc1] = r0 // trash saved pfs
mov loc0 = r2
} ;;
{.mib; mov r8 = 10
mov rp = r3
br.ret.sptk.many rp
}
.exit:
{.mmi; ld8 r16 = [loc1]
;;
sub sp = 32, sp
mov ar.pfs = r16
} ;;
{.mmi; mov sp = loc1
st8 [loc1] = r0 // trash saved pfs
mov rp = loc0
}
{.mib; nop 0
mov r8 = 4
br.ret.sptk.many rp
}
func_vframe_end:
.endp func_vframe
#ifdef __linux__
/* We do not need executable stack. */
.section .note.GNU-stack,"",@progbits
#endif
|
