about summary refs log tree commit diff
path: root/sysdeps/unix/sysv/linux/ia64/__ia64_longjmp.S
blob: 33434338d419ff75d299ba61ea81e9e0ba6e28b9 (plain) (blame)
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
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
/* Copyright (C) 1999-2016 Free Software Foundation, Inc.
   Contributed by David Mosberger-Tang <davidm@hpl.hp.com>.

   The GNU C Library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
   License as published by the Free Software Foundation; either
   version 2.1 of the License, or (at your option) any later version.

   The GNU C Library is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with the GNU C Library; if not, see
   <http://www.gnu.org/licenses/>.  */

#include <sysdep.h>
#include <features.h>

LEAF(__ia64_flush_rbs)
	flushrs
	mov r9 = ar.rsc		// 12 cyc latency
	;;
	mov r8 = ar.bsp		// 12 cyc latency
	;;
	and r16 = ~0x3, r9	// clear ar.rsc.mode
	;;
	mov ar.rsc = r16	// put RSE into enforced-lazy mode
	;;
	mov r10 = ar.rnat	// 5 cyc latency
	ret
END(__ia64_flush_rbs)


#	define	pPos	p6	/* is rotate count positive? */
#	define	pNeg	p7	/* is rotate count negative? */

/* __ia64_longjmp(__jmp_buf buf, int val, long rnat, long rsc)  */


LEAF(__ia64_longjmp)
	alloc r8=ar.pfs,4,0,0,0
	add r2=0x98,in0		// r2 <- &jmpbuf.orig_jmp_buf_addr
	add r3=0x88,in0		// r3 <- &jmpbuf.ar_bsp
	;;
	ld8 r8=[r2]		// r8 <- orig_jmp_buf_addr
	ld8 r23=[r3],8		// r23 <- jmpbuf.ar_bsp
	mov r2=in0
	;;
	//
	// Note: we need to redo the "flushrs" here even though it's
	// already been done by __ia64_flush_rbs.  It is needed to
	// ensure that ar.bspstore == ar.bsp.
	//
	flushrs			// flush dirty regs to backing store
	ld8 r25=[r3]		// r25 <- jmpbuf.ar_unat
	sub r8=r8,in0		// r8 <- &orig_jmpbuf - &jmpbuf
	;;
	add r3=8,in0		// r3 <- &jmpbuf.r1
	extr.u r8=r8,3,6	// r8 <- (&orig_jmpbuf - &jmpbuf)/8 & 0x3f
	;;
	cmp.lt pNeg,pPos=r8,r0
	;;
(pPos)	mov r16=r8
(pNeg)	add r16=64,r8
(pPos)	sub r17=64,r8
(pNeg)	sub r17=r0,r8
	;;
	shr.u r8=r25,r16
	shl r9=r25,r17
	;;
	or r25=r8,r9
	;;
	mov ar.unat=r25 // setup ar.unat (NaT bits for r1, r4-r7, and r12)
	;;
	ld8.fill.nta sp=[r2],16	// r12 (sp)
	ld8.fill.nta gp=[r3],32 // r1 (gp)
	dep r11=-1,r23,3,6	// r11 <- ia64_rse_rnat_addr(jmpbuf.ar_bsp)
	;;
	ld8.nta r16=[r2],16		// caller's unat
//	ld8.nta r17=[r3],16		// fpsr
	;;
	ld8.fill.nta r4=[r2],16		// r4
	ld8.fill.nta r5=[r3],16		// r5 (gp)
	;;
	ld8.fill.nta r6=[r2],16		// r6
	ld8.fill.nta r7=[r3],16		// r7
	;;
	mov ar.unat=r16			// restore caller's unat
//	mov ar.fpsr=r17			// restore fpsr
	;;
	ld8.nta r16=[r2],16		// b0
	ld8.nta r17=[r3],16		// b1
	;;
	mov ar.bspstore=r23	// restore ar.bspstore
	ld8.nta r18=[r2],16		// b2
	;;
	mov ar.rnat=in2		// restore ar.rnat
	ld8.nta r19=[r3],16		// b3
	;;
	ld8.nta r20=[r2],16		// b4
	ld8.nta r21=[r3],16		// b5
#ifdef PTR_DEMANGLE
	PTR_DEMANGLE (r16, r24)
#endif
	;;
	ld8.nta r11=[r2],16		// ar.pfs
	ld8.nta r22=[r3],56		// ar.lc
	;;
	ld8.nta r24=[r2],32		// pr
	mov ar.rsc=in3		// restore ar.rsc
	mov b0=r16
	;;
	ldf.fill.nta f2=[r2],32
	ldf.fill.nta f3=[r3],32
	mov b1=r17
	;;
	ldf.fill.nta f4=[r2],32
	ldf.fill.nta f5=[r3],32
	mov b2=r18
	;;
	ldf.fill.nta f16=[r2],32
	ldf.fill.nta f17=[r3],32
	mov b3=r19
	;;
	ldf.fill.nta f18=[r2],32
	ldf.fill.nta f19=[r3],32
	mov b4=r20
	;;
	ldf.fill.nta f20=[r2],32
	ldf.fill.nta f21=[r3],32
	mov b5=r21
	;;
	ldf.fill.nta f22=[r2],32
	ldf.fill.nta f23=[r3],32
	mov ar.lc=r22
	;;
	ldf.fill.nta f24=[r2],32
	ldf.fill.nta f25=[r3],32
	cmp.eq p8,p9=0,in1
	;;
	ldf.fill.nta f26=[r2],32
	ldf.fill.nta f27=[r3],32
	mov ar.pfs=r11
	;;
	ldf.fill.nta f28=[r2],32
	ldf.fill.nta f29=[r3],32
(p8)	mov r8=1
	;;
	ldf.fill.nta f30=[r2]
	ldf.fill.nta f31=[r3]
(p9)	mov r8=in1

	invala			// virt. -> phys. regnum mapping may change
	mov pr=r24,-1
	ret
END(__ia64_longjmp)