summaryrefslogtreecommitdiffstats
path: root/arch/ia64/kernel/entry.S
blob: ac06d44b9b278a83e09c12b3dbc19c912c8d1e10 (plain)
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
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
/* SPDX-License-Identifier: GPL-2.0 */
/*
 * arch/ia64/kernel/entry.S
 *
 * Kernel entry points.
 *
 * Copyright (C) 1998-2003, 2005 Hewlett-Packard Co
 *	David Mosberger-Tang <davidm@hpl.hp.com>
 * Copyright (C) 1999, 2002-2003
 *	Asit Mallick <Asit.K.Mallick@intel.com>
 * 	Don Dugger <Don.Dugger@intel.com>
 *	Suresh Siddha <suresh.b.siddha@intel.com>
 *	Fenghua Yu <fenghua.yu@intel.com>
 * Copyright (C) 1999 VA Linux Systems
 * Copyright (C) 1999 Walt Drummond <drummond@valinux.com>
 */
/*
 * ia64_switch_to now places correct virtual mapping in in TR2 for
 * kernel stack. This allows us to handle interrupts without changing
 * to physical mode.
 *
 * Jonathan Nicklin	<nicklin@missioncriticallinux.com>
 * Patrick O'Rourke	<orourke@missioncriticallinux.com>
 * 11/07/2000
 */
/*
 * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
 *                    VA Linux Systems Japan K.K.
 *                    pv_ops.
 */
/*
 * Global (preserved) predicate usage on syscall entry/exit path:
 *
 *	pKStk:		See entry.h.
 *	pUStk:		See entry.h.
 *	pSys:		See entry.h.
 *	pNonSys:	!pSys
 */

#include <linux/export.h>
#include <linux/pgtable.h>
#include <asm/asmmacro.h>
#include <asm/cache.h>
#include <asm/errno.h>
#include <asm/kregs.h>
#include <asm/asm-offsets.h>
#include <asm/percpu.h>
#include <asm/processor.h>
#include <asm/thread_info.h>
#include <asm/unistd.h>
#include <asm/ftrace.h>

#include "minstate.h"

	/*
	 * execve() is special because in case of success, we need to
	 * setup a null register window frame.
	 */
ENTRY(ia64_execve)
	/*
	 * Allocate 8 input registers since ptrace() may clobber them
	 */
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
	alloc loc1=ar.pfs,8,2,3,0
	mov loc0=rp
	.body
	mov out0=in0			// filename
	;;				// stop bit between alloc and call
	mov out1=in1			// argv
	mov out2=in2			// envp
	br.call.sptk.many rp=sys_execve
.ret0:
	cmp4.ge p6,p7=r8,r0
	mov ar.pfs=loc1			// restore ar.pfs
	sxt4 r8=r8			// return 64-bit result
	;;
	stf.spill [sp]=f0
	mov rp=loc0
(p6)	mov ar.pfs=r0			// clear ar.pfs on success
(p7)	br.ret.sptk.many rp

	/*
	 * In theory, we'd have to zap this state only to prevent leaking of
	 * security sensitive state (e.g., if current->mm->dumpable is zero).  However,
	 * this executes in less than 20 cycles even on Itanium, so it's not worth
	 * optimizing for...).
	 */
	mov ar.unat=0; 		mov ar.lc=0
	mov r4=0;		mov f2=f0;		mov b1=r0
	mov r5=0;		mov f3=f0;		mov b2=r0
	mov r6=0;		mov f4=f0;		mov b3=r0
	mov r7=0;		mov f5=f0;		mov b4=r0
	ldf.fill f12=[sp];	mov f13=f0;		mov b5=r0
	ldf.fill f14=[sp];	ldf.fill f15=[sp];	mov f16=f0
	ldf.fill f17=[sp];	ldf.fill f18=[sp];	mov f19=f0
	ldf.fill f20=[sp];	ldf.fill f21=[sp];	mov f22=f0
	ldf.fill f23=[sp];	ldf.fill f24=[sp];	mov f25=f0
	ldf.fill f26=[sp];	ldf.fill f27=[sp];	mov f28=f0
	ldf.fill f29=[sp];	ldf.fill f30=[sp];	mov f31=f0
	br.ret.sptk.many rp
END(ia64_execve)

/*
 * sys_clone2(u64 flags, u64 ustack_base, u64 ustack_size, u64 parent_tidptr, u64 child_tidptr,
 *	      u64 tls)
 */
GLOBAL_ENTRY(sys_clone2)
	/*
	 * Allocate 8 input registers since ptrace() may clobber them
	 */
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
	alloc r16=ar.pfs,8,2,6,0
	DO_SAVE_SWITCH_STACK
	mov loc0=rp
	mov loc1=r16                             // save ar.pfs across ia64_clone
	.body
	mov out0=in0
	mov out1=in1
	mov out2=in2
	mov out3=in3
	mov out4=in4
	mov out5=in5
	br.call.sptk.many rp=ia64_clone
.ret1:	.restore sp
	adds sp=IA64_SWITCH_STACK_SIZE,sp	// pop the switch stack
	mov ar.pfs=loc1
	mov rp=loc0
	br.ret.sptk.many rp
END(sys_clone2)

/*
 * sys_clone(u64 flags, u64 ustack_base, u64 parent_tidptr, u64 child_tidptr, u64 tls)
 *	Deprecated.  Use sys_clone2() instead.
 */
GLOBAL_ENTRY(sys_clone)
	/*
	 * Allocate 8 input registers since ptrace() may clobber them
	 */
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
	alloc r16=ar.pfs,8,2,6,0
	DO_SAVE_SWITCH_STACK
	mov loc0=rp
	mov loc1=r16                             // save ar.pfs across ia64_clone
	.body
	mov out0=in0
	mov out1=in1
	mov out2=16				// stacksize (compensates for 16-byte scratch area)
	mov out3=in3
	mov out4=in4
	mov out5=in5
	br.call.sptk.many rp=ia64_clone
.ret2:	.restore sp
	adds sp=IA64_SWITCH_STACK_SIZE,sp	// pop the switch stack
	mov ar.pfs=loc1
	mov rp=loc0
	br.ret.sptk.many rp
END(sys_clone)

/*
 * prev_task <- ia64_switch_to(struct task_struct *next)
 *	With Ingo's new scheduler, interrupts are disabled when this routine gets
 *	called.  The code starting at .map relies on this.  The rest of the code
 *	doesn't care about the interrupt masking status.
 */
GLOBAL_ENTRY(ia64_switch_to)
	.prologue
	alloc r16=ar.pfs,1,0,0,0
	DO_SAVE_SWITCH_STACK
	.body

	adds r22=IA64_TASK_THREAD_KSP_OFFSET,r13
	movl r25=init_task
	mov r27=IA64_KR(CURRENT_STACK)
	adds r21=IA64_TASK_THREAD_KSP_OFFSET,in0
	dep r20=0,in0,61,3		// physical address of "next"
	;;
	st8 [r22]=sp			// save kernel stack pointer of old task
	shr.u r26=r20,IA64_GRANULE_SHIFT
	cmp.eq p7,p6=r25,in0
	;;
	/*
	 * If we've already mapped this task's page, we can skip doing it again.
	 */
(p6)	cmp.eq p7,p6=r26,r27
(p6)	br.cond.dpnt .map
	;;
.done:
	ld8 sp=[r21]			// load kernel stack pointer of new task
	MOV_TO_KR(CURRENT, in0, r8, r9)		// update "current" application register
	mov r8=r13			// return pointer to previously running task
	mov r13=in0			// set "current" pointer
	;;
	DO_LOAD_SWITCH_STACK

#ifdef CONFIG_SMP
	sync.i				// ensure "fc"s done by this CPU are visible on other CPUs
#endif
	br.ret.sptk.many rp		// boogie on out in new context

.map:
	RSM_PSR_IC(r25)			// interrupts (psr.i) are already disabled here
	movl r25=PAGE_KERNEL
	;;
	srlz.d
	or r23=r25,r20			// construct PA | page properties
	mov r25=IA64_GRANULE_SHIFT<<2
	;;
	MOV_TO_ITIR(p0, r25, r8)
	MOV_TO_IFA(in0, r8)		// VA of next task...
	;;
	mov r25=IA64_TR_CURRENT_STACK
	MOV_TO_KR(CURRENT_STACK, r26, r8, r9)	// remember last page we mapped...
	;;
	itr.d dtr[r25]=r23		// wire in new mapping...
	SSM_PSR_IC_AND_SRLZ_D(r8, r9)	// reenable the psr.ic bit
	br.cond.sptk .done
END(ia64_switch_to)

/*
 * Note that interrupts are enabled during save_switch_stack and load_switch_stack.  This
 * means that we may get an interrupt with "sp" pointing to the new kernel stack while
 * ar.bspstore is still pointing to the old kernel backing store area.  Since ar.rsc,
 * ar.rnat, ar.bsp, and ar.bspstore are all preserved by interrupts, this is not a
 * problem.  Also, we don't need to specify unwind information for preserved registers
 * that are not modified in save_switch_stack as the right unwind information is already
 * specified at the call-site of save_switch_stack.
 */

/*
 * save_switch_stack:
 *	- r16 holds ar.pfs
 *	- b7 holds address to return to
 *	- rp (b0) holds return address to save
 */
GLOBAL_ENTRY(save_switch_stack)
	.prologue
	.altrp b7
	flushrs			// flush dirty regs to backing store (must be first in insn group)
	.save @priunat,r17
	mov r17=ar.unat		// preserve caller's
	.body
#ifdef CONFIG_ITANIUM
	adds r2=16+128,sp
	adds r3=16+64,sp
	adds r14=SW(R4)+16,sp
	;;
	st8.spill [r14]=r4,16		// spill r4
	lfetch.fault.excl.nt1 [r3],128
	;;
	lfetch.fault.excl.nt1 [r2],128
	lfetch.fault.excl.nt1 [r3],128
	;;
	lfetch.fault.excl [r2]
	lfetch.fault.excl [r3]
	adds r15=SW(R5)+16,sp
#else
	add r2=16+3*128,sp
	add r3=16,sp
	add r14=SW(R4)+16,sp
	;;
	st8.spill [r14]=r4,SW(R6)-SW(R4)	// spill r4 and prefetch offset 0x1c0
	lfetch.fault.excl.nt1 [r3],128	//		prefetch offset 0x010
	;;
	lfetch.fault.excl.nt1 [r3],128	//		prefetch offset 0x090
	lfetch.fault.excl.nt1 [r2],128	//		prefetch offset 0x190
	;;
	lfetch.fault.excl.nt1 [r3]	//		prefetch offset 0x110
	lfetch.fault.excl.nt1 [r2]	//		prefetch offset 0x210
	adds r15=SW(R5)+16,sp
#endif
	;;
	st8.spill [r15]=r5,SW(R7)-SW(R5)	// spill r5
	mov.m ar.rsc=0			// put RSE in mode: enforced lazy, little endian, pl 0
	add r2=SW(F2)+16,sp		// r2 = &sw->f2
	;;
	st8.spill [r14]=r6,SW(B0)-SW(R6)	// spill r6
	mov.m r18=ar.fpsr		// preserve fpsr
	add r3=SW(F3)+16,sp		// r3 = &sw->f3
	;;
	stf.spill [r2]=f2,32
	mov.m r19=ar.rnat
	mov r21=b0

	stf.spill [r3]=f3,32
	st8.spill [r15]=r7,SW(B2)-SW(R7)	// spill r7
	mov r22=b1
	;;
	// since we're done with the spills, read and save ar.unat:
	mov.m r29=ar.unat
	mov.m r20=ar.bspstore
	mov r23=b2
	stf.spill [r2]=f4,32
	stf.spill [r3]=f5,32
	mov r24=b3
	;;
	st8 [r14]=r21,SW(B1)-SW(B0)		// save b0
	st8 [r15]=r23,SW(B3)-SW(B2)		// save b2
	mov r25=b4
	mov r26=b5
	;;
	st8 [r14]=r22,SW(B4)-SW(B1)		// save b1
	st8 [r15]=r24,SW(AR_PFS)-SW(B3)		// save b3
	mov r21=ar.lc		// I-unit
	stf.spill [r2]=f12,32
	stf.spill [r3]=f13,32
	;;
	st8 [r14]=r25,SW(B5)-SW(B4)		// save b4
	st8 [r15]=r16,SW(AR_LC)-SW(AR_PFS)	// save ar.pfs
	stf.spill [r2]=f14,32
	stf.spill [r3]=f15,32
	;;
	st8 [r14]=r26				// save b5
	st8 [r15]=r21				// save ar.lc
	stf.spill [r2]=f16,32
	stf.spill [r3]=f17,32
	;;
	stf.spill [r2]=f18,32
	stf.spill [r3]=f19,32
	;;
	stf.spill [r2]=f20,32
	stf.spill [r3]=f21,32
	;;
	stf.spill [r2]=f22,32
	stf.spill [r3]=f23,32
	;;
	stf.spill [r2]=f24,32
	stf.spill [r3]=f25,32
	;;
	stf.spill [r2]=f26,32
	stf.spill [r3]=f27,32
	;;
	stf.spill [r2]=f28,32
	stf.spill [r3]=f29,32
	;;
	stf.spill [r2]=f30,SW(AR_UNAT)-SW(F30)
	stf.spill [r3]=f31,SW(PR)-SW(F31)
	add r14=SW(CALLER_UNAT)+16,sp
	;;
	st8 [r2]=r29,SW(AR_RNAT)-SW(AR_UNAT)	// save ar.unat
	st8 [r14]=r17,SW(AR_FPSR)-SW(CALLER_UNAT) // save caller_unat
	mov r21=pr
	;;
	st8 [r2]=r19,SW(AR_BSPSTORE)-SW(AR_RNAT) // save ar.rnat
	st8 [r3]=r21				// save predicate registers
	;;
	st8 [r2]=r20				// save ar.bspstore
	st8 [r14]=r18				// save fpsr
	mov ar.rsc=3		// put RSE back into eager mode, pl 0
	br.cond.sptk.many b7
END(save_switch_stack)

/*
 * load_switch_stack:
 *	- "invala" MUST be done at call site (normally in DO_LOAD_SWITCH_STACK)
 *	- b7 holds address to return to
 *	- must not touch r8-r11
 */
GLOBAL_ENTRY(load_switch_stack)
	.prologue
	.altrp b7

	.body
	lfetch.fault.nt1 [sp]
	adds r2=SW(AR_BSPSTORE)+16,sp
	adds r3=SW(AR_UNAT)+16,sp
	mov ar.rsc=0						// put RSE into enforced lazy mode
	adds r14=SW(CALLER_UNAT)+16,sp
	adds r15=SW(AR_FPSR)+16,sp
	;;
	ld8 r27=[r2],(SW(B0)-SW(AR_BSPSTORE))	// bspstore
	ld8 r29=[r3],(SW(B1)-SW(AR_UNAT))	// unat
	;;
	ld8 r21=[r2],16		// restore b0
	ld8 r22=[r3],16		// restore b1
	;;
	ld8 r23=[r2],16		// restore b2
	ld8 r24=[r3],16		// restore b3
	;;
	ld8 r25=[r2],16		// restore b4
	ld8 r26=[r3],16		// restore b5
	;;
	ld8 r16=[r2],(SW(PR)-SW(AR_PFS))	// ar.pfs
	ld8 r17=[r3],(SW(AR_RNAT)-SW(AR_LC))	// ar.lc
	;;
	ld8 r28=[r2]		// restore pr
	ld8 r30=[r3]		// restore rnat
	;;
	ld8 r18=[r14],16	// restore caller's unat
	ld8 r19=[r15],24	// restore fpsr
	;;
	ldf.fill f2=[r14],32
	ldf.fill f3=[r15],32
	;;
	ldf.fill f4=[r14],32
	ldf.fill f5=[r15],32
	;;
	ldf.fill f12=[r14],32
	ldf.fill f13=[r15],32
	;;
	ldf.fill f14=[r14],32
	ldf.fill f15=[r15],32
	;;
	ldf.fill f16=[r14],32
	ldf.fill f17=[r15],32
	;;
	ldf.fill f18=[r14],32
	ldf.fill f19=[r15],32
	mov b0=r21
	;;
	ldf.fill f20=[r14],32
	ldf.fill f21=[r15],32
	mov b1=r22
	;;
	ldf.fill f22=[r14],32
	ldf.fill f23=[r15],32
	mov b2=r23
	;;
	mov ar.bspstore=r27
	mov ar.unat=r29		// establish unat holding the NaT bits for r4-r7
	mov b3=r24
	;;
	ldf.fill f24=[r14],32
	ldf.fill f25=[r15],32
	mov b4=r25
	;;
	ldf.fill f26=[r14],32
	ldf.fill f27=[r15],32
	mov b5=r26
	;;
	ldf.fill f28=[r14],32
	ldf.fill f29=[r15],32
	mov ar.pfs=r16
	;;
	ldf.fill f30=[r14],32
	ldf.fill f31=[r15],24
	mov ar.lc=r17
	;;
	ld8.fill r4=[r14],16
	ld8.fill r5=[r15],16
	mov pr=r28,-1
	;;
	ld8.fill r6=[r14],16
	ld8.fill r7=[r15],16

	mov ar.unat=r18				// restore caller's unat
	mov ar.rnat=r30				// must restore after bspstore but before rsc!
	mov ar.fpsr=r19				// restore fpsr
	mov ar.rsc=3				// put RSE back into eager mode, pl 0
	br.cond.sptk.many b7
END(load_switch_stack)

	/*
	 * Invoke a system call, but do some tracing before and after the call.
	 * We MUST preserve the current register frame throughout this routine
	 * because some system calls (such as ia64_execve) directly
	 * manipulate ar.pfs.
	 */
GLOBAL_ENTRY(ia64_trace_syscall)
	PT_REGS_UNWIND_INFO(0)
	/*
	 * We need to preserve the scratch registers f6-f11 in case the system
	 * call is sigreturn.
	 */
	adds r16=PT(F6)+16,sp
	adds r17=PT(F7)+16,sp
	;;
 	stf.spill [r16]=f6,32
 	stf.spill [r17]=f7,32
	;;
 	stf.spill [r16]=f8,32
 	stf.spill [r17]=f9,32
	;;
 	stf.spill [r16]=f10
 	stf.spill [r17]=f11
	br.call.sptk.many rp=syscall_trace_enter // give parent a chance to catch syscall args
	cmp.lt p6,p0=r8,r0			// check tracehook
	adds r2=PT(R8)+16,sp			// r2 = &pt_regs.r8
	adds r3=PT(R10)+16,sp			// r3 = &pt_regs.r10
	mov r10=0
(p6)	br.cond.sptk strace_error		// syscall failed ->
	adds r16=PT(F6)+16,sp
	adds r17=PT(F7)+16,sp
	;;
	ldf.fill f6=[r16],32
	ldf.fill f7=[r17],32
	;;
	ldf.fill f8=[r16],32
	ldf.fill f9=[r17],32
	;;
	ldf.fill f10=[r16]
	ldf.fill f11=[r17]
	// the syscall number may have changed, so re-load it and re-calculate the
	// syscall entry-point:
	adds r15=PT(R15)+16,sp			// r15 = &pt_regs.r15 (syscall #)
	;;
	ld8 r15=[r15]
	mov r3=NR_syscalls - 1
	;;
	adds r15=-1024,r15
	movl r16=sys_call_table
	;;
	shladd r20=r15,3,r16			// r20 = sys_call_table + 8*(syscall-1024)
	cmp.leu p6,p7=r15,r3
	;;
(p6)	ld8 r20=[r20]				// load address of syscall entry point
(p7)	movl r20=sys_ni_syscall
	;;
	mov b6=r20
	br.call.sptk.many rp=b6			// do the syscall
.strace_check_retval:
	cmp.lt p6,p0=r8,r0			// syscall failed?
	adds r2=PT(R8)+16,sp			// r2 = &pt_regs.r8
	adds r3=PT(R10)+16,sp			// r3 = &pt_regs.r10
	mov r10=0
(p6)	br.cond.sptk strace_error		// syscall failed ->
	;;					// avoid RAW on r10
.strace_save_retval:
.mem.offset 0,0; st8.spill [r2]=r8		// store return value in slot for r8
.mem.offset 8,0; st8.spill [r3]=r10		// clear error indication in slot for r10
	br.call.sptk.many rp=syscall_trace_leave // give parent a chance to catch return value
.ret3:
(pUStk)	cmp.eq.unc p6,p0=r0,r0			// p6 <- pUStk
(pUStk)	rsm psr.i				// disable interrupts
	br.cond.sptk ia64_work_pending_syscall_end

strace_error:
	ld8 r3=[r2]				// load pt_regs.r8
	sub r9=0,r8				// negate return value to get errno value
	;;
	cmp.ne p6,p0=r3,r0			// is pt_regs.r8!=0?
	adds r3=16,r2				// r3=&pt_regs.r10
	;;
(p6)	mov r10=-1
(p6)	mov r8=r9
	br.cond.sptk .strace_save_retval
END(ia64_trace_syscall)

	/*
	 * When traced and returning from sigreturn, we invoke syscall_trace but then
	 * go straight to ia64_leave_kernel rather than ia64_leave_syscall.
	 */
GLOBAL_ENTRY(ia64_strace_leave_kernel)
	PT_REGS_UNWIND_INFO(0)
{	/*
	 * Some versions of gas generate bad unwind info if the first instruction of a
	 * procedure doesn't go into the first slot of a bundle.  This is a workaround.
	 */
	nop.m 0
	nop.i 0
	br.call.sptk.many rp=syscall_trace_leave // give parent a chance to catch return value
}
.ret4:	br.cond.sptk ia64_leave_kernel
END(ia64_strace_leave_kernel)

ENTRY(call_payload)
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(0)
	/* call the kernel_thread payload; fn is in r4, arg - in r5 */
	alloc loc1=ar.pfs,0,3,1,0
	mov loc0=rp
	mov loc2=gp
	mov out0=r5		// arg
	ld8 r14 = [r4], 8	// fn.address
	;;
	mov b6 = r14
	ld8 gp = [r4]		// fn.gp
	;;
	br.call.sptk.many rp=b6	// fn(arg)
.ret12:	mov gp=loc2
	mov rp=loc0
	mov ar.pfs=loc1
	/* ... and if it has returned, we are going to userland */
	cmp.ne pKStk,pUStk=r0,r0
	br.ret.sptk.many rp
END(call_payload)

GLOBAL_ENTRY(ia64_ret_from_clone)
	PT_REGS_UNWIND_INFO(0)
{	/*
	 * Some versions of gas generate bad unwind info if the first instruction of a
	 * procedure doesn't go into the first slot of a bundle.  This is a workaround.
	 */
	nop.m 0
	nop.i 0
	/*
	 * We need to call schedule_tail() to complete the scheduling process.
	 * Called by ia64_switch_to() after ia64_clone()->copy_thread().  r8 contains the
	 * address of the previously executing task.
	 */
	br.call.sptk.many rp=ia64_invoke_schedule_tail
}
.ret8:
(pKStk)	br.call.sptk.many rp=call_payload
	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
	;;
	ld4 r2=[r2]
	;;
	mov r8=0
	and r2=_TIF_SYSCALL_TRACEAUDIT,r2
	;;
	cmp.ne p6,p0=r2,r0
(p6)	br.cond.spnt .strace_check_retval
	;;					// added stop bits to prevent r8 dependency
END(ia64_ret_from_clone)
	// fall through
GLOBAL_ENTRY(ia64_ret_from_syscall)
	PT_REGS_UNWIND_INFO(0)
	cmp.ge p6,p7=r8,r0			// syscall executed successfully?
	adds r2=PT(R8)+16,sp			// r2 = &pt_regs.r8
	mov r10=r0				// clear error indication in r10
(p7)	br.cond.spnt handle_syscall_error	// handle potential syscall failure
END(ia64_ret_from_syscall)
	// fall through

/*
 * ia64_leave_syscall(): Same as ia64_leave_kernel, except that it doesn't
 *	need to switch to bank 0 and doesn't restore the scratch registers.
 *	To avoid leaking kernel bits, the scratch registers are set to
 *	the following known-to-be-safe values:
 *
 *		  r1: restored (global pointer)
 *		  r2: cleared
 *		  r3: 1 (when returning to user-level)
 *	      r8-r11: restored (syscall return value(s))
 *		 r12: restored (user-level stack pointer)
 *		 r13: restored (user-level thread pointer)
 *		 r14: set to __kernel_syscall_via_epc
 *		 r15: restored (syscall #)
 *	     r16-r17: cleared
 *		 r18: user-level b6
 *		 r19: cleared
 *		 r20: user-level ar.fpsr
 *		 r21: user-level b0
 *		 r22: cleared
 *		 r23: user-level ar.bspstore
 *		 r24: user-level ar.rnat
 *		 r25: user-level ar.unat
 *		 r26: user-level ar.pfs
 *		 r27: user-level ar.rsc
 *		 r28: user-level ip
 *		 r29: user-level psr
 *		 r30: user-level cfm
 *		 r31: user-level pr
 *	      f6-f11: cleared
 *		  pr: restored (user-level pr)
 *		  b0: restored (user-level rp)
 *	          b6: restored
 *		  b7: set to __kernel_syscall_via_epc
 *	     ar.unat: restored (user-level ar.unat)
 *	      ar.pfs: restored (user-level ar.pfs)
 *	      ar.rsc: restored (user-level ar.rsc)
 *	     ar.rnat: restored (user-level ar.rnat)
 *	 ar.bspstore: restored (user-level ar.bspstore)
 *	     ar.fpsr: restored (user-level ar.fpsr)
 *	      ar.ccv: cleared
 *	      ar.csd: cleared
 *	      ar.ssd: cleared
 */
GLOBAL_ENTRY(ia64_leave_syscall)
	PT_REGS_UNWIND_INFO(0)
	/*
	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
	 * user- or fsys-mode, hence we disable interrupts early on.
	 *
	 * p6 controls whether current_thread_info()->flags needs to be check for
	 * extra work.  We always check for extra work when returning to user-level.
	 * With CONFIG_PREEMPTION, we also check for extra work when the preempt_count
	 * is 0.  After extra work processing has been completed, execution
	 * resumes at ia64_work_processed_syscall with p6 set to 1 if the extra-work-check
	 * needs to be redone.
	 */
#ifdef CONFIG_PREEMPTION
	RSM_PSR_I(p0, r2, r18)			// disable interrupts
	cmp.eq pLvSys,p0=r0,r0			// pLvSys=1: leave from syscall
(pKStk) adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
	;;
	.pred.rel.mutex pUStk,pKStk
(pKStk) ld4 r21=[r20]			// r21 <- preempt_count
(pUStk)	mov r21=0			// r21 <- 0
	;;
	cmp.eq p6,p0=r21,r0		// p6 <- pUStk || (preempt_count == 0)
#else /* !CONFIG_PREEMPTION */
	RSM_PSR_I(pUStk, r2, r18)
	cmp.eq pLvSys,p0=r0,r0		// pLvSys=1: leave from syscall
(pUStk)	cmp.eq.unc p6,p0=r0,r0		// p6 <- pUStk
#endif
.global ia64_work_processed_syscall;
ia64_work_processed_syscall:
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
	adds r2=PT(LOADRS)+16,r12
	MOV_FROM_ITC(pUStk, p9, r22, r19)	// fetch time at leave
	adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
	;;
(p6)	ld4 r31=[r18]				// load current_thread_info()->flags
	ld8 r19=[r2],PT(B6)-PT(LOADRS)		// load ar.rsc value for "loadrs"
	adds r3=PT(AR_BSPSTORE)+16,r12		// deferred
	;;
#else
	adds r2=PT(LOADRS)+16,r12
	adds r3=PT(AR_BSPSTORE)+16,r12
	adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
	;;
(p6)	ld4 r31=[r18]				// load current_thread_info()->flags
	ld8 r19=[r2],PT(B6)-PT(LOADRS)		// load ar.rsc value for "loadrs"
	nop.i 0
	;;
#endif
	mov r16=ar.bsp				// M2  get existing backing store pointer
	ld8 r18=[r2],PT(R9)-PT(B6)		// load b6
(p6)	and r15=TIF_WORK_MASK,r31		// any work other than TIF_SYSCALL_TRACE?
	;;
	ld8 r23=[r3],PT(R11)-PT(AR_BSPSTORE)	// load ar.bspstore (may be garbage)
(p6)	cmp4.ne.unc p6,p0=r15, r0		// any special work pending?
(p6)	br.cond.spnt .work_pending_syscall
	;;
	// start restoring the state saved on the kernel stack (struct pt_regs):
	ld8 r9=[r2],PT(CR_IPSR)-PT(R9)
	ld8 r11=[r3],PT(CR_IIP)-PT(R11)
(pNonSys) break 0		//      bug check: we shouldn't be here if pNonSys is TRUE!
	;;
	invala			// M0|1 invalidate ALAT
	RSM_PSR_I_IC(r28, r29, r30)	// M2   turn off interrupts and interruption collection
	cmp.eq p9,p0=r0,r0	// A    set p9 to indicate that we should restore cr.ifs

	ld8 r29=[r2],16		// M0|1 load cr.ipsr
	ld8 r28=[r3],16		// M0|1 load cr.iip
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
(pUStk) add r14=TI_AC_LEAVE+IA64_TASK_SIZE,r13
	;;
	ld8 r30=[r2],16		// M0|1 load cr.ifs
	ld8 r25=[r3],16		// M0|1 load ar.unat
(pUStk) add r15=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
	;;
#else
	mov r22=r0		// A    clear r22
	;;
	ld8 r30=[r2],16		// M0|1 load cr.ifs
	ld8 r25=[r3],16		// M0|1 load ar.unat
(pUStk) add r14=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
	;;
#endif
	ld8 r26=[r2],PT(B0)-PT(AR_PFS)	// M0|1 load ar.pfs
	MOV_FROM_PSR(pKStk, r22, r21)	// M2   read PSR now that interrupts are disabled
	nop 0
	;;
	ld8 r21=[r2],PT(AR_RNAT)-PT(B0) // M0|1 load b0
	ld8 r27=[r3],PT(PR)-PT(AR_RSC)	// M0|1 load ar.rsc
	mov f6=f0			// F    clear f6
	;;
	ld8 r24=[r2],PT(AR_FPSR)-PT(AR_RNAT)	// M0|1 load ar.rnat (may be garbage)
	ld8 r31=[r3],PT(R1)-PT(PR)		// M0|1 load predicates
	mov f7=f0				// F    clear f7
	;;
	ld8 r20=[r2],PT(R12)-PT(AR_FPSR)	// M0|1 load ar.fpsr
	ld8.fill r1=[r3],16			// M0|1 load r1
(pUStk) mov r17=1				// A
	;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
(pUStk) st1 [r15]=r17				// M2|3
#else
(pUStk) st1 [r14]=r17				// M2|3
#endif
	ld8.fill r13=[r3],16			// M0|1
	mov f8=f0				// F    clear f8
	;;
	ld8.fill r12=[r2]			// M0|1 restore r12 (sp)
	ld8.fill r15=[r3]			// M0|1 restore r15
	mov b6=r18				// I0   restore b6

	LOAD_PHYS_STACK_REG_SIZE(r17)
	mov f9=f0					// F    clear f9
(pKStk) br.cond.dpnt.many skip_rbs_switch		// B

	srlz.d				// M0   ensure interruption collection is off (for cover)
	shr.u r18=r19,16		// I0|1 get byte size of existing "dirty" partition
	COVER				// B    add current frame into dirty partition & set cr.ifs
	;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
	mov r19=ar.bsp			// M2   get new backing store pointer
	st8 [r14]=r22			// M	save time at leave
	mov f10=f0			// F    clear f10

	mov r22=r0			// A	clear r22
	movl r14=__kernel_syscall_via_epc // X
	;;
#else
	mov r19=ar.bsp			// M2   get new backing store pointer
	mov f10=f0			// F    clear f10

	nop.m 0
	movl r14=__kernel_syscall_via_epc // X
	;;
#endif
	mov.m ar.csd=r0			// M2   clear ar.csd
	mov.m ar.ccv=r0			// M2   clear ar.ccv
	mov b7=r14			// I0   clear b7 (hint with __kernel_syscall_via_epc)

	mov.m ar.ssd=r0			// M2   clear ar.ssd
	mov f11=f0			// F    clear f11
	br.cond.sptk.many rbs_switch	// B
END(ia64_leave_syscall)

GLOBAL_ENTRY(ia64_leave_kernel)
	PT_REGS_UNWIND_INFO(0)
	/*
	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
	 * user- or fsys-mode, hence we disable interrupts early on.
	 *
	 * p6 controls whether current_thread_info()->flags needs to be check for
	 * extra work.  We always check for extra work when returning to user-level.
	 * With CONFIG_PREEMPTION, we also check for extra work when the preempt_count
	 * is 0.  After extra work processing has been completed, execution
	 * resumes at .work_processed_syscall with p6 set to 1 if the extra-work-check
	 * needs to be redone.
	 */
#ifdef CONFIG_PREEMPTION
	RSM_PSR_I(p0, r17, r31)			// disable interrupts
	cmp.eq p0,pLvSys=r0,r0			// pLvSys=0: leave from kernel
(pKStk)	adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
	;;
	.pred.rel.mutex pUStk,pKStk
(pKStk)	ld4 r21=[r20]			// r21 <- preempt_count
(pUStk)	mov r21=0			// r21 <- 0
	;;
	cmp.eq p6,p0=r21,r0		// p6 <- pUStk || (preempt_count == 0)
#else
	RSM_PSR_I(pUStk, r17, r31)
	cmp.eq p0,pLvSys=r0,r0		// pLvSys=0: leave from kernel
(pUStk)	cmp.eq.unc p6,p0=r0,r0		// p6 <- pUStk
#endif
.work_processed_kernel:
	adds r17=TI_FLAGS+IA64_TASK_SIZE,r13
	;;
(p6)	ld4 r31=[r17]				// load current_thread_info()->flags
	adds r21=PT(PR)+16,r12
	;;

	lfetch [r21],PT(CR_IPSR)-PT(PR)
	adds r2=PT(B6)+16,r12
	adds r3=PT(R16)+16,r12
	;;
	lfetch [r21]
	ld8 r28=[r2],8		// load b6
	adds r29=PT(R24)+16,r12

	ld8.fill r16=[r3],PT(AR_CSD)-PT(R16)
	adds r30=PT(AR_CCV)+16,r12
(p6)	and r19=TIF_WORK_MASK,r31		// any work other than TIF_SYSCALL_TRACE?
	;;
	ld8.fill r24=[r29]
	ld8 r15=[r30]		// load ar.ccv
(p6)	cmp4.ne.unc p6,p0=r19, r0		// any special work pending?
	;;
	ld8 r29=[r2],16		// load b7
	ld8 r30=[r3],16		// load ar.csd
(p6)	br.cond.spnt .work_pending
	;;
	ld8 r31=[r2],16		// load ar.ssd
	ld8.fill r8=[r3],16
	;;
	ld8.fill r9=[r2],16
	ld8.fill r10=[r3],PT(R17)-PT(R10)
	;;
	ld8.fill r11=[r2],PT(R18)-PT(R11)
	ld8.fill r17=[r3],16
	;;
	ld8.fill r18=[r2],16
	ld8.fill r19=[r3],16
	;;
	ld8.fill r20=[r2],16
	ld8.fill r21=[r3],16
	mov ar.csd=r30
	mov ar.ssd=r31
	;;
	RSM_PSR_I_IC(r23, r22, r25)	// initiate turning off of interrupt and interruption collection
	invala			// invalidate ALAT
	;;
	ld8.fill r22=[r2],24
	ld8.fill r23=[r3],24
	mov b6=r28
	;;
	ld8.fill r25=[r2],16
	ld8.fill r26=[r3],16
	mov b7=r29
	;;
	ld8.fill r27=[r2],16
	ld8.fill r28=[r3],16
	;;
	ld8.fill r29=[r2],16
	ld8.fill r30=[r3],24
	;;
	ld8.fill r31=[r2],PT(F9)-PT(R31)
	adds r3=PT(F10)-PT(F6),r3
	;;
	ldf.fill f9=[r2],PT(F6)-PT(F9)
	ldf.fill f10=[r3],PT(F8)-PT(F10)
	;;
	ldf.fill f6=[r2],PT(F7)-PT(F6)
	;;
	ldf.fill f7=[r2],PT(F11)-PT(F7)
	ldf.fill f8=[r3],32
	;;
	srlz.d	// ensure that inter. collection is off (VHPT is don't care, since text is pinned)
	mov ar.ccv=r15
	;;
	ldf.fill f11=[r2]
	BSW_0(r2, r3, r15)	// switch back to bank 0 (no stop bit required beforehand...)
	;;
(pUStk)	mov r18=IA64_KR(CURRENT)// M2 (12 cycle read latency)
	adds r16=PT(CR_IPSR)+16,r12
	adds r17=PT(CR_IIP)+16,r12

#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
	.pred.rel.mutex pUStk,pKStk
	MOV_FROM_PSR(pKStk, r22, r29)	// M2 read PSR now that interrupts are disabled
	MOV_FROM_ITC(pUStk, p9, r22, r29)	// M  fetch time at leave
	nop.i 0
	;;
#else
	MOV_FROM_PSR(pKStk, r22, r29)	// M2 read PSR now that interrupts are disabled
	nop.i 0
	nop.i 0
	;;
#endif
	ld8 r29=[r16],16	// load cr.ipsr
	ld8 r28=[r17],16	// load cr.iip
	;;
	ld8 r30=[r16],16	// load cr.ifs
	ld8 r25=[r17],16	// load ar.unat
	;;
	ld8 r26=[r16],16	// load ar.pfs
	ld8 r27=[r17],16	// load ar.rsc
	cmp.eq p9,p0=r0,r0	// set p9 to indicate that we should restore cr.ifs
	;;
	ld8 r24=[r16],16	// load ar.rnat (may be garbage)
	ld8 r23=[r17],16	// load ar.bspstore (may be garbage)
	;;
	ld8 r31=[r16],16	// load predicates
	ld8 r21=[r17],16	// load b0
	;;
	ld8 r19=[r16],16	// load ar.rsc value for "loadrs"
	ld8.fill r1=[r17],16	// load r1
	;;
	ld8.fill r12=[r16],16
	ld8.fill r13=[r17],16
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
(pUStk)	adds r3=TI_AC_LEAVE+IA64_TASK_SIZE,r18
#else
(pUStk)	adds r18=IA64_TASK_THREAD_ON_USTACK_OFFSET,r18
#endif
	;;
	ld8 r20=[r16],16	// ar.fpsr
	ld8.fill r15=[r17],16
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
(pUStk)	adds r18=IA64_TASK_THREAD_ON_USTACK_OFFSET,r18	// deferred
#endif
	;;
	ld8.fill r14=[r16],16
	ld8.fill r2=[r17]
(pUStk)	mov r17=1
	;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE
	//  mmi_ :  ld8 st1 shr;;         mmi_ : st8 st1 shr;;
	//  mib  :  mov add br        ->  mib  : ld8 add br
	//  bbb_ :  br  nop cover;;       mbb_ : mov br  cover;;
	//
	//  no one require bsp in r16 if (pKStk) branch is selected.
(pUStk)	st8 [r3]=r22		// save time at leave
(pUStk)	st1 [r18]=r17		// restore current->thread.on_ustack
	shr.u r18=r19,16	// get byte size of existing "dirty" partition
	;;
	ld8.fill r3=[r16]	// deferred
	LOAD_PHYS_STACK_REG_SIZE(r17)
(pKStk)	br.cond.dpnt skip_rbs_switch
	mov r16=ar.bsp		// get existing backing store pointer
#else
	ld8.fill r3=[r16]
(pUStk)	st1 [r18]=r17		// restore current->thread.on_ustack
	shr.u r18=r19,16	// get byte size of existing "dirty" partition
	;;
	mov r16=ar.bsp		// get existing backing store pointer
	LOAD_PHYS_STACK_REG_SIZE(r17)
(pKStk)	br.cond.dpnt skip_rbs_switch
#endif

	/*
	 * Restore user backing store.
	 *
	 * NOTE: alloc, loadrs, and cover can't be predicated.
	 */
(pNonSys) br.cond.dpnt dont_preserve_current_frame
	COVER				// add current frame into dirty partition and set cr.ifs
	;;
	mov r19=ar.bsp			// get new backing store pointer
rbs_switch:
	sub r16=r16,r18			// krbs = old bsp - size of dirty partition
	cmp.ne p9,p0=r0,r0		// clear p9 to skip restore of cr.ifs
	;;
	sub r19=r19,r16			// calculate total byte size of dirty partition
	add r18=64,r18			// don't force in0-in7 into memory...
	;;
	shl r19=r19,16			// shift size of dirty partition into loadrs position
	;;
dont_preserve_current_frame:
	/*
	 * To prevent leaking bits between the kernel and user-space,
	 * we must clear the stacked registers in the "invalid" partition here.
	 * Not pretty, but at least it's fast (3.34 registers/cycle on Itanium,
	 * 5 registers/cycle on McKinley).
	 */
#	define pRecurse	p6
#	define pReturn	p7
#ifdef CONFIG_ITANIUM
#	define Nregs	10
#else
#	define Nregs	14
#endif
	alloc loc0=ar.pfs,2,Nregs-2,2,0
	shr.u loc1=r18,9		// RNaTslots <= floor(dirtySize / (64*8))
	sub r17=r17,r18			// r17 = (physStackedSize + 8) - dirtySize
	;;
	mov ar.rsc=r19			// load ar.rsc to be used for "loadrs"
	shladd in0=loc1,3,r17
	mov in1=0
	;;
	TEXT_ALIGN(32)
rse_clear_invalid:
#ifdef CONFIG_ITANIUM
	// cycle 0
 { .mii
	alloc loc0=ar.pfs,2,Nregs-2,2,0
	cmp.lt pRecurse,p0=Nregs*8,in0	// if more than Nregs regs left to clear, (re)curse
	add out0=-Nregs*8,in0
}{ .mfb
	add out1=1,in1			// increment recursion count
	nop.f 0
	nop.b 0				// can't do br.call here because of alloc (WAW on CFM)
	;;
}{ .mfi	// cycle 1
	mov loc1=0
	nop.f 0
	mov loc2=0
}{ .mib
	mov loc3=0
	mov loc4=0
(pRecurse) br.call.sptk.many b0=rse_clear_invalid

}{ .mfi	// cycle 2
	mov loc5=0
	nop.f 0
	cmp.ne pReturn,p0=r0,in1	// if recursion count != 0, we need to do a br.ret
}{ .mib
	mov loc6=0
	mov loc7=0
(pReturn) br.ret.sptk.many b0
}
#else /* !CONFIG_ITANIUM */
	alloc loc0=ar.pfs,2,Nregs-2,2,0
	cmp.lt pRecurse,p0=Nregs*8,in0	// if more than Nregs regs left to clear, (re)curse
	add out0=-Nregs*8,in0
	add out1=1,in1			// increment recursion count
	mov loc1=0
	mov loc2=0
	;;
	mov loc3=0
	mov loc4=0
	mov loc5=0
	mov loc6=0
	mov loc7=0
(pRecurse) br.call.dptk.few b0=rse_clear_invalid
	;;
	mov loc8=0
	mov loc9=0
	cmp.ne pReturn,p0=r0,in1	// if recursion count != 0, we need to do a br.ret
	mov loc10=0
	mov loc11=0
(pReturn) br.ret.dptk.many b0
#endif /* !CONFIG_ITANIUM */
#	undef pRecurse
#	undef pReturn
	;;
	alloc r17=ar.pfs,0,0,0,0	// drop current register frame
	;;
	loadrs
	;;
skip_rbs_switch:
	mov ar.unat=r25		// M2
(pKStk)	extr.u r22=r22,21,1	// I0 extract current value of psr.pp from r22
(pLvSys)mov r19=r0		// A  clear r19 for leave_syscall, no-op otherwise
	;;
(pUStk)	mov ar.bspstore=r23	// M2
(pKStk)	dep r29=r22,r29,21,1	// I0 update ipsr.pp with psr.pp
(pLvSys)mov r16=r0		// A  clear r16 for leave_syscall, no-op otherwise
	;;
	MOV_TO_IPSR(p0, r29, r25)	// M2
	mov ar.pfs=r26		// I0
(pLvSys)mov r17=r0		// A  clear r17 for leave_syscall, no-op otherwise

	MOV_TO_IFS(p9, r30, r25)// M2
	mov b0=r21		// I0
(pLvSys)mov r18=r0		// A  clear r18 for leave_syscall, no-op otherwise

	mov ar.fpsr=r20		// M2
	MOV_TO_IIP(r28, r25)	// M2
	nop 0
	;;
(pUStk)	mov ar.rnat=r24		// M2 must happen with RSE in lazy mode
	nop 0
(pLvSys)mov r2=r0

	mov ar.rsc=r27		// M2
	mov pr=r31,-1		// I0
	RFI			// B

	/*
	 * On entry:
	 *	r20 = &current->thread_info->pre_count (if CONFIG_PREEMPTION)
	 *	r31 = current->thread_info->flags
	 * On exit:
	 *	p6 = TRUE if work-pending-check needs to be redone
	 *
	 * Interrupts are disabled on entry, reenabled depend on work, and
	 * disabled on exit.
	 */
.work_pending_syscall:
	add r2=-8,r2
	add r3=-8,r3
	;;
	st8 [r2]=r8
	st8 [r3]=r10
.work_pending:
	tbit.z p6,p0=r31,TIF_NEED_RESCHED	// is resched not needed?
(p6)	br.cond.sptk.few .notify
	br.call.spnt.many rp=preempt_schedule_irq
.ret9:	cmp.eq p6,p0=r0,r0	// p6 <- 1 (re-check)
(pLvSys)br.cond.sptk.few  ia64_work_pending_syscall_end
	br.cond.sptk.many .work_processed_kernel

.notify:
(pUStk)	br.call.spnt.many rp=notify_resume_user
.ret10:	cmp.ne p6,p0=r0,r0	// p6 <- 0 (don't re-check)
(pLvSys)br.cond.sptk.few  ia64_work_pending_syscall_end
	br.cond.sptk.many .work_processed_kernel

.global ia64_work_pending_syscall_end;
ia64_work_pending_syscall_end:
	adds r2=PT(R8)+16,r12
	adds r3=PT(R10)+16,r12
	;;
	ld8 r8=[r2]
	ld8 r10=[r3]
	br.cond.sptk.many ia64_work_processed_syscall
END(ia64_leave_kernel)

ENTRY(handle_syscall_error)
	/*
	 * Some system calls (e.g., ptrace, mmap) can return arbitrary values which could
	 * lead us to mistake a negative return value as a failed syscall.  Those syscall
	 * must deposit a non-zero value in pt_regs.r8 to indicate an error.  If
	 * pt_regs.r8 is zero, we assume that the call completed successfully.
	 */
	PT_REGS_UNWIND_INFO(0)
	ld8 r3=[r2]		// load pt_regs.r8
	;;
	cmp.eq p6,p7=r3,r0	// is pt_regs.r8==0?
	;;
(p7)	mov r10=-1
(p7)	sub r8=0,r8		// negate return value to get errno
	br.cond.sptk ia64_leave_syscall
END(handle_syscall_error)

	/*
	 * Invoke schedule_tail(task) while preserving in0-in7, which may be needed
	 * in case a system call gets restarted.
	 */
GLOBAL_ENTRY(ia64_invoke_schedule_tail)
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
	alloc loc1=ar.pfs,8,2,1,0
	mov loc0=rp
	mov out0=r8				// Address of previous task
	;;
	br.call.sptk.many rp=schedule_tail
.ret11:	mov ar.pfs=loc1
	mov rp=loc0
	br.ret.sptk.many rp
END(ia64_invoke_schedule_tail)

	/*
	 * Setup stack and call do_notify_resume_user(), keeping interrupts
	 * disabled.
	 *
	 * Note that pSys and pNonSys need to be set up by the caller.
	 * We declare 8 input registers so the system call args get preserved,
	 * in case we need to restart a system call.
	 */
GLOBAL_ENTRY(notify_resume_user)
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
	alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs in case of syscall restart!
	mov r9=ar.unat
	mov loc0=rp				// save return address
	mov out0=0				// there is no "oldset"
	adds out1=8,sp				// out1=&sigscratch->ar_pfs
(pSys)	mov out2=1				// out2==1 => we're in a syscall
	;;
(pNonSys) mov out2=0				// out2==0 => not a syscall
	.fframe 16
	.spillsp ar.unat, 16
	st8 [sp]=r9,-16				// allocate space for ar.unat and save it
	st8 [out1]=loc1,-8			// save ar.pfs, out1=&sigscratch
	.body
	br.call.sptk.many rp=do_notify_resume_user
.ret15:	.restore sp
	adds sp=16,sp				// pop scratch stack space
	;;
	ld8 r9=[sp]				// load new unat from sigscratch->scratch_unat
	mov rp=loc0
	;;
	mov ar.unat=r9
	mov ar.pfs=loc1
	br.ret.sptk.many rp
END(notify_resume_user)

ENTRY(sys_rt_sigreturn)
	PT_REGS_UNWIND_INFO(0)
	/*
	 * Allocate 8 input registers since ptrace() may clobber them
	 */
	alloc r2=ar.pfs,8,0,1,0
	.prologue
	PT_REGS_SAVES(16)
	adds sp=-16,sp
	.body
	cmp.eq pNonSys,pSys=r0,r0		// sigreturn isn't a normal syscall...
	;;
	/*
	 * leave_kernel() restores f6-f11 from pt_regs, but since the streamlined
	 * syscall-entry path does not save them we save them here instead.  Note: we
	 * don't need to save any other registers that are not saved by the stream-lined
	 * syscall path, because restore_sigcontext() restores them.
	 */
	adds r16=PT(F6)+32,sp
	adds r17=PT(F7)+32,sp
	;;
 	stf.spill [r16]=f6,32
 	stf.spill [r17]=f7,32
	;;
 	stf.spill [r16]=f8,32
 	stf.spill [r17]=f9,32
	;;
 	stf.spill [r16]=f10
 	stf.spill [r17]=f11
	adds out0=16,sp				// out0 = &sigscratch
	br.call.sptk.many rp=ia64_rt_sigreturn
.ret19:	.restore sp,0
	adds sp=16,sp
	;;
	ld8 r9=[sp]				// load new ar.unat
	mov.sptk b7=r8,ia64_leave_kernel
	;;
	mov ar.unat=r9
	br.many b7
END(sys_rt_sigreturn)

GLOBAL_ENTRY(ia64_prepare_handle_unaligned)
	.prologue
	/*
	 * r16 = fake ar.pfs, we simply need to make sure privilege is still 0
	 */
	mov r16=r0
	DO_SAVE_SWITCH_STACK
	br.call.sptk.many rp=ia64_handle_unaligned	// stack frame setup in ivt
.ret21:	.body
	DO_LOAD_SWITCH_STACK
	br.cond.sptk.many rp				// goes to ia64_leave_kernel
END(ia64_prepare_handle_unaligned)

	//
	// unw_init_running(void (*callback)(info, arg), void *arg)
	//
#	define EXTRA_FRAME_SIZE	((UNW_FRAME_INFO_SIZE+15)&~15)

GLOBAL_ENTRY(unw_init_running)
	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(2)
	alloc loc1=ar.pfs,2,3,3,0
	;;
	ld8 loc2=[in0],8
	mov loc0=rp
	mov r16=loc1
	DO_SAVE_SWITCH_STACK
	.body

	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(2)
	.fframe IA64_SWITCH_STACK_SIZE+EXTRA_FRAME_SIZE
	SWITCH_STACK_SAVES(EXTRA_FRAME_SIZE)
	adds sp=-EXTRA_FRAME_SIZE,sp
	.body
	;;
	adds out0=16,sp				// &info
	mov out1=r13				// current
	adds out2=16+EXTRA_FRAME_SIZE,sp	// &switch_stack
	br.call.sptk.many rp=unw_init_frame_info
1:	adds out0=16,sp				// &info
	mov b6=loc2
	mov loc2=gp				// save gp across indirect function call
	;;
	ld8 gp=[in0]
	mov out1=in1				// arg
	br.call.sptk.many rp=b6			// invoke the callback function
1:	mov gp=loc2				// restore gp

	// For now, we don't allow changing registers from within
	// unw_init_running; if we ever want to allow that, we'd
	// have to do a load_switch_stack here:
	.restore sp
	adds sp=IA64_SWITCH_STACK_SIZE+EXTRA_FRAME_SIZE,sp

	mov ar.pfs=loc1
	mov rp=loc0
	br.ret.sptk.many rp
END(unw_init_running)
EXPORT_SYMBOL(unw_init_running)

#ifdef CONFIG_FUNCTION_TRACER
#ifdef CONFIG_DYNAMIC_FTRACE
GLOBAL_ENTRY(_mcount)
	br ftrace_stub
END(_mcount)
EXPORT_SYMBOL(_mcount)

.here:
	br.ret.sptk.many b0

GLOBAL_ENTRY(ftrace_caller)
	alloc out0 = ar.pfs, 8, 0, 4, 0
	mov out3 = r0
	;;
	mov out2 = b0
	add r3 = 0x20, r3
	mov out1 = r1;
	br.call.sptk.many b0 = ftrace_patch_gp
	//this might be called from module, so we must patch gp
ftrace_patch_gp:
	movl gp=__gp
	mov b0 = r3
	;;
.global ftrace_call;
ftrace_call:
{
	.mlx
	nop.m 0x0
	movl r3 = .here;;
}
	alloc loc0 = ar.pfs, 4, 4, 2, 0
	;;
	mov loc1 = b0
	mov out0 = b0
	mov loc2 = r8
	mov loc3 = r15
	;;
	adds out0 = -MCOUNT_INSN_SIZE, out0
	mov out1 = in2
	mov b6 = r3

	br.call.sptk.many b0 = b6
	;;
	mov ar.pfs = loc0
	mov b0 = loc1
	mov r8 = loc2
	mov r15 = loc3
	br ftrace_stub
	;;
END(ftrace_caller)

#else
GLOBAL_ENTRY(_mcount)
	movl r2 = ftrace_stub
	movl r3 = ftrace_trace_function;;
	ld8 r3 = [r3];;
	ld8 r3 = [r3];;
	cmp.eq p7,p0 = r2, r3
(p7)	br.sptk.many ftrace_stub
	;;

	alloc loc0 = ar.pfs, 4, 4, 2, 0
	;;
	mov loc1 = b0
	mov out0 = b0
	mov loc2 = r8
	mov loc3 = r15
	;;
	adds out0 = -MCOUNT_INSN_SIZE, out0
	mov out1 = in2
	mov b6 = r3

	br.call.sptk.many b0 = b6
	;;
	mov ar.pfs = loc0
	mov b0 = loc1
	mov r8 = loc2
	mov r15 = loc3
	br ftrace_stub
	;;
END(_mcount)
#endif

GLOBAL_ENTRY(ftrace_stub)
	mov r3 = b0
	movl r2 = _mcount_ret_helper
	;;
	mov b6 = r2
	mov b7 = r3
	br.ret.sptk.many b6

_mcount_ret_helper:
	mov b0 = r42
	mov r1 = r41
	mov ar.pfs = r40
	br b7
END(ftrace_stub)

#endif /* CONFIG_FUNCTION_TRACER */

#define __SYSCALL(nr, entry) data8 entry
	.rodata
	.align 8
	.globl sys_call_table
sys_call_table:
#include <asm/syscall_table.h>