forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtlbex.c
2585 lines (2267 loc) · 70.4 KB
/
tlbex.c
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
/*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*
* Synthesize TLB refill handlers at runtime.
*
* Copyright (C) 2004, 2005, 2006, 2008 Thiemo Seufer
* Copyright (C) 2005, 2007, 2008, 2009 Maciej W. Rozycki
* Copyright (C) 2006 Ralf Baechle ([email protected])
* Copyright (C) 2008, 2009 Cavium Networks, Inc.
* Copyright (C) 2011 MIPS Technologies, Inc.
*
* ... and the days got worse and worse and now you see
* I've gone completely out of my mind.
*
* They're coming to take me a away haha
* they're coming to take me a away hoho hihi haha
* to the funny farm where code is beautiful all the time ...
*
* (Condolences to Napoleon XIV)
*/
#include <linux/bug.h>
#include <linux/export.h>
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/smp.h>
#include <linux/string.h>
#include <linux/cache.h>
#include <linux/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/cpu-type.h>
#include <asm/mipsregs.h>
#include <asm/mmu_context.h>
#include <asm/regdef.h>
#include <asm/uasm.h>
#include <asm/setup.h>
#include <asm/tlbex.h>
static int mips_xpa_disabled;
static int __init xpa_disable(char *s)
{
mips_xpa_disabled = 1;
return 1;
}
__setup("noxpa", xpa_disable);
/*
* TLB load/store/modify handlers.
*
* Only the fastpath gets synthesized at runtime, the slowpath for
* do_page_fault remains normal asm.
*/
extern void tlb_do_page_fault_0(void);
extern void tlb_do_page_fault_1(void);
struct work_registers {
int r1;
int r2;
int r3;
};
struct tlb_reg_save {
unsigned long a;
unsigned long b;
} ____cacheline_aligned_in_smp;
static struct tlb_reg_save handler_reg_save[NR_CPUS];
static inline int r45k_bvahwbug(void)
{
/* XXX: We should probe for the presence of this bug, but we don't. */
return 0;
}
static inline int r4k_250MHZhwbug(void)
{
/* XXX: We should probe for the presence of this bug, but we don't. */
return 0;
}
extern int sb1250_m3_workaround_needed(void);
static inline int __maybe_unused bcm1250_m3_war(void)
{
if (IS_ENABLED(CONFIG_SB1_PASS_2_WORKAROUNDS))
return sb1250_m3_workaround_needed();
return 0;
}
static inline int __maybe_unused r10000_llsc_war(void)
{
return IS_ENABLED(CONFIG_WAR_R10000_LLSC);
}
static int use_bbit_insns(void)
{
switch (current_cpu_type()) {
case CPU_CAVIUM_OCTEON:
case CPU_CAVIUM_OCTEON_PLUS:
case CPU_CAVIUM_OCTEON2:
case CPU_CAVIUM_OCTEON3:
return 1;
default:
return 0;
}
}
static int use_lwx_insns(void)
{
switch (current_cpu_type()) {
case CPU_CAVIUM_OCTEON2:
case CPU_CAVIUM_OCTEON3:
return 1;
default:
return 0;
}
}
#if defined(CONFIG_CAVIUM_OCTEON_CVMSEG_SIZE) && \
CONFIG_CAVIUM_OCTEON_CVMSEG_SIZE > 0
static bool scratchpad_available(void)
{
return true;
}
static int scratchpad_offset(int i)
{
/*
* CVMSEG starts at address -32768 and extends for
* CAVIUM_OCTEON_CVMSEG_SIZE 128 byte cache lines.
*/
i += 1; /* Kernel use starts at the top and works down. */
return CONFIG_CAVIUM_OCTEON_CVMSEG_SIZE * 128 - (8 * i) - 32768;
}
#else
static bool scratchpad_available(void)
{
return false;
}
static int scratchpad_offset(int i)
{
BUG();
/* Really unreachable, but evidently some GCC want this. */
return 0;
}
#endif
/*
* Found by experiment: At least some revisions of the 4kc throw under
* some circumstances a machine check exception, triggered by invalid
* values in the index register. Delaying the tlbp instruction until
* after the next branch, plus adding an additional nop in front of
* tlbwi/tlbwr avoids the invalid index register values. Nobody knows
* why; it's not an issue caused by the core RTL.
*
*/
static int m4kc_tlbp_war(void)
{
return current_cpu_type() == CPU_4KC;
}
/* Handle labels (which must be positive integers). */
enum label_id {
label_second_part = 1,
label_leave,
label_vmalloc,
label_vmalloc_done,
label_tlbw_hazard_0,
label_split = label_tlbw_hazard_0 + 8,
label_tlbl_goaround1,
label_tlbl_goaround2,
label_nopage_tlbl,
label_nopage_tlbs,
label_nopage_tlbm,
label_smp_pgtable_change,
label_r3000_write_probe_fail,
label_large_segbits_fault,
#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT
label_tlb_huge_update,
#endif
};
UASM_L_LA(_second_part)
UASM_L_LA(_leave)
UASM_L_LA(_vmalloc)
UASM_L_LA(_vmalloc_done)
/* _tlbw_hazard_x is handled differently. */
UASM_L_LA(_split)
UASM_L_LA(_tlbl_goaround1)
UASM_L_LA(_tlbl_goaround2)
UASM_L_LA(_nopage_tlbl)
UASM_L_LA(_nopage_tlbs)
UASM_L_LA(_nopage_tlbm)
UASM_L_LA(_smp_pgtable_change)
UASM_L_LA(_r3000_write_probe_fail)
UASM_L_LA(_large_segbits_fault)
#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT
UASM_L_LA(_tlb_huge_update)
#endif
static int hazard_instance;
static void uasm_bgezl_hazard(u32 **p, struct uasm_reloc **r, int instance)
{
switch (instance) {
case 0 ... 7:
uasm_il_bgezl(p, r, 0, label_tlbw_hazard_0 + instance);
return;
default:
BUG();
}
}
static void uasm_bgezl_label(struct uasm_label **l, u32 **p, int instance)
{
switch (instance) {
case 0 ... 7:
uasm_build_label(l, *p, label_tlbw_hazard_0 + instance);
break;
default:
BUG();
}
}
/*
* pgtable bits are assigned dynamically depending on processor feature
* and statically based on kernel configuration. This spits out the actual
* values the kernel is using. Required to make sense from disassembled
* TLB exception handlers.
*/
static void output_pgtable_bits_defines(void)
{
#define pr_define(fmt, ...) \
pr_debug("#define " fmt, ##__VA_ARGS__)
pr_debug("#include <asm/asm.h>\n");
pr_debug("#include <asm/regdef.h>\n");
pr_debug("\n");
pr_define("_PAGE_PRESENT_SHIFT %d\n", _PAGE_PRESENT_SHIFT);
pr_define("_PAGE_NO_READ_SHIFT %d\n", _PAGE_NO_READ_SHIFT);
pr_define("_PAGE_WRITE_SHIFT %d\n", _PAGE_WRITE_SHIFT);
pr_define("_PAGE_ACCESSED_SHIFT %d\n", _PAGE_ACCESSED_SHIFT);
pr_define("_PAGE_MODIFIED_SHIFT %d\n", _PAGE_MODIFIED_SHIFT);
#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT
pr_define("_PAGE_HUGE_SHIFT %d\n", _PAGE_HUGE_SHIFT);
#endif
#ifdef _PAGE_NO_EXEC_SHIFT
if (cpu_has_rixi)
pr_define("_PAGE_NO_EXEC_SHIFT %d\n", _PAGE_NO_EXEC_SHIFT);
#endif
pr_define("_PAGE_GLOBAL_SHIFT %d\n", _PAGE_GLOBAL_SHIFT);
pr_define("_PAGE_VALID_SHIFT %d\n", _PAGE_VALID_SHIFT);
pr_define("_PAGE_DIRTY_SHIFT %d\n", _PAGE_DIRTY_SHIFT);
pr_define("PFN_PTE_SHIFT %d\n", PFN_PTE_SHIFT);
pr_debug("\n");
}
static inline void dump_handler(const char *symbol, const void *start, const void *end)
{
unsigned int count = (end - start) / sizeof(u32);
const u32 *handler = start;
int i;
pr_debug("LEAF(%s)\n", symbol);
pr_debug("\t.set push\n");
pr_debug("\t.set noreorder\n");
for (i = 0; i < count; i++)
pr_debug("\t.word\t0x%08x\t\t# %p\n", handler[i], &handler[i]);
pr_debug("\t.set\tpop\n");
pr_debug("\tEND(%s)\n", symbol);
}
#ifdef CONFIG_64BIT
# define GET_CONTEXT(buf, reg) UASM_i_MFC0(buf, reg, C0_XCONTEXT)
#else
# define GET_CONTEXT(buf, reg) UASM_i_MFC0(buf, reg, C0_CONTEXT)
#endif
/* The worst case length of the handler is around 18 instructions for
* R3000-style TLBs and up to 63 instructions for R4000-style TLBs.
* Maximum space available is 32 instructions for R3000 and 64
* instructions for R4000.
*
* We deliberately chose a buffer size of 128, so we won't scribble
* over anything important on overflow before we panic.
*/
static u32 tlb_handler[128];
/* simply assume worst case size for labels and relocs */
static struct uasm_label labels[128];
static struct uasm_reloc relocs[128];
static int check_for_high_segbits;
static bool fill_includes_sw_bits;
static unsigned int kscratch_used_mask;
static inline int __maybe_unused c0_kscratch(void)
{
return 31;
}
static int allocate_kscratch(void)
{
int r;
unsigned int a = cpu_data[0].kscratch_mask & ~kscratch_used_mask;
r = ffs(a);
if (r == 0)
return -1;
r--; /* make it zero based */
kscratch_used_mask |= (1 << r);
return r;
}
static int scratch_reg;
int pgd_reg;
EXPORT_SYMBOL_GPL(pgd_reg);
enum vmalloc64_mode {not_refill, refill_scratch, refill_noscratch};
static struct work_registers build_get_work_registers(u32 **p)
{
struct work_registers r;
if (scratch_reg >= 0) {
/* Save in CPU local C0_KScratch? */
UASM_i_MTC0(p, 1, c0_kscratch(), scratch_reg);
r.r1 = GPR_K0;
r.r2 = GPR_K1;
r.r3 = GPR_AT;
return r;
}
if (num_possible_cpus() > 1) {
/* Get smp_processor_id */
UASM_i_CPUID_MFC0(p, GPR_K0, SMP_CPUID_REG);
UASM_i_SRL_SAFE(p, GPR_K0, GPR_K0, SMP_CPUID_REGSHIFT);
/* handler_reg_save index in GPR_K0 */
UASM_i_SLL(p, GPR_K0, GPR_K0, ilog2(sizeof(struct tlb_reg_save)));
UASM_i_LA(p, GPR_K1, (long)&handler_reg_save);
UASM_i_ADDU(p, GPR_K0, GPR_K0, GPR_K1);
} else {
UASM_i_LA(p, GPR_K0, (long)&handler_reg_save);
}
/* GPR_K0 now points to save area, save $1 and $2 */
UASM_i_SW(p, 1, offsetof(struct tlb_reg_save, a), GPR_K0);
UASM_i_SW(p, 2, offsetof(struct tlb_reg_save, b), GPR_K0);
r.r1 = GPR_K1;
r.r2 = 1;
r.r3 = 2;
return r;
}
static void build_restore_work_registers(u32 **p)
{
if (scratch_reg >= 0) {
uasm_i_ehb(p);
UASM_i_MFC0(p, 1, c0_kscratch(), scratch_reg);
return;
}
/* GPR_K0 already points to save area, restore $1 and $2 */
UASM_i_LW(p, 1, offsetof(struct tlb_reg_save, a), GPR_K0);
UASM_i_LW(p, 2, offsetof(struct tlb_reg_save, b), GPR_K0);
}
#ifndef CONFIG_MIPS_PGD_C0_CONTEXT
/*
* CONFIG_MIPS_PGD_C0_CONTEXT implies 64 bit and lack of pgd_current,
* we cannot do r3000 under these circumstances.
*
* The R3000 TLB handler is simple.
*/
static void build_r3000_tlb_refill_handler(void)
{
long pgdc = (long)pgd_current;
u32 *p;
memset(tlb_handler, 0, sizeof(tlb_handler));
p = tlb_handler;
uasm_i_mfc0(&p, GPR_K0, C0_BADVADDR);
uasm_i_lui(&p, GPR_K1, uasm_rel_hi(pgdc)); /* cp0 delay */
uasm_i_lw(&p, GPR_K1, uasm_rel_lo(pgdc), GPR_K1);
uasm_i_srl(&p, GPR_K0, GPR_K0, 22); /* load delay */
uasm_i_sll(&p, GPR_K0, GPR_K0, 2);
uasm_i_addu(&p, GPR_K1, GPR_K1, GPR_K0);
uasm_i_mfc0(&p, GPR_K0, C0_CONTEXT);
uasm_i_lw(&p, GPR_K1, 0, GPR_K1); /* cp0 delay */
uasm_i_andi(&p, GPR_K0, GPR_K0, 0xffc); /* load delay */
uasm_i_addu(&p, GPR_K1, GPR_K1, GPR_K0);
uasm_i_lw(&p, GPR_K0, 0, GPR_K1);
uasm_i_nop(&p); /* load delay */
uasm_i_mtc0(&p, GPR_K0, C0_ENTRYLO0);
uasm_i_mfc0(&p, GPR_K1, C0_EPC); /* cp0 delay */
uasm_i_tlbwr(&p); /* cp0 delay */
uasm_i_jr(&p, GPR_K1);
uasm_i_rfe(&p); /* branch delay */
if (p > tlb_handler + 32)
panic("TLB refill handler space exceeded");
pr_debug("Wrote TLB refill handler (%u instructions).\n",
(unsigned int)(p - tlb_handler));
memcpy((void *)ebase, tlb_handler, 0x80);
local_flush_icache_range(ebase, ebase + 0x80);
dump_handler("r3000_tlb_refill", (u32 *)ebase, (u32 *)(ebase + 0x80));
}
#endif /* CONFIG_MIPS_PGD_C0_CONTEXT */
/*
* The R4000 TLB handler is much more complicated. We have two
* consecutive handler areas with 32 instructions space each.
* Since they aren't used at the same time, we can overflow in the
* other one.To keep things simple, we first assume linear space,
* then we relocate it to the final handler layout as needed.
*/
static u32 final_handler[64];
/*
* Hazards
*
* From the IDT errata for the QED RM5230 (Nevada), processor revision 1.0:
* 2. A timing hazard exists for the TLBP instruction.
*
* stalling_instruction
* TLBP
*
* The JTLB is being read for the TLBP throughout the stall generated by the
* previous instruction. This is not really correct as the stalling instruction
* can modify the address used to access the JTLB. The failure symptom is that
* the TLBP instruction will use an address created for the stalling instruction
* and not the address held in C0_ENHI and thus report the wrong results.
*
* The software work-around is to not allow the instruction preceding the TLBP
* to stall - make it an NOP or some other instruction guaranteed not to stall.
*
* Errata 2 will not be fixed. This errata is also on the R5000.
*
* As if we MIPS hackers wouldn't know how to nop pipelines happy ...
*/
static void __maybe_unused build_tlb_probe_entry(u32 **p)
{
switch (current_cpu_type()) {
/* Found by experiment: R4600 v2.0/R4700 needs this, too. */
case CPU_R4600:
case CPU_R4700:
case CPU_R5000:
case CPU_NEVADA:
uasm_i_nop(p);
uasm_i_tlbp(p);
break;
default:
uasm_i_tlbp(p);
break;
}
}
void build_tlb_write_entry(u32 **p, struct uasm_label **l,
struct uasm_reloc **r,
enum tlb_write_entry wmode)
{
void(*tlbw)(u32 **) = NULL;
switch (wmode) {
case tlb_random: tlbw = uasm_i_tlbwr; break;
case tlb_indexed: tlbw = uasm_i_tlbwi; break;
}
if (cpu_has_mips_r2_r6) {
if (cpu_has_mips_r2_exec_hazard)
uasm_i_ehb(p);
tlbw(p);
return;
}
switch (current_cpu_type()) {
case CPU_R4000PC:
case CPU_R4000SC:
case CPU_R4000MC:
case CPU_R4400PC:
case CPU_R4400SC:
case CPU_R4400MC:
/*
* This branch uses up a mtc0 hazard nop slot and saves
* two nops after the tlbw instruction.
*/
uasm_bgezl_hazard(p, r, hazard_instance);
tlbw(p);
uasm_bgezl_label(l, p, hazard_instance);
hazard_instance++;
uasm_i_nop(p);
break;
case CPU_R4600:
case CPU_R4700:
uasm_i_nop(p);
tlbw(p);
uasm_i_nop(p);
break;
case CPU_R5000:
case CPU_NEVADA:
uasm_i_nop(p); /* QED specifies 2 nops hazard */
uasm_i_nop(p); /* QED specifies 2 nops hazard */
tlbw(p);
break;
case CPU_R4300:
case CPU_5KC:
case CPU_TX49XX:
case CPU_PR4450:
uasm_i_nop(p);
tlbw(p);
break;
case CPU_R10000:
case CPU_R12000:
case CPU_R14000:
case CPU_R16000:
case CPU_4KC:
case CPU_4KEC:
case CPU_M14KC:
case CPU_M14KEC:
case CPU_SB1:
case CPU_SB1A:
case CPU_4KSC:
case CPU_20KC:
case CPU_25KF:
case CPU_BMIPS32:
case CPU_BMIPS3300:
case CPU_BMIPS4350:
case CPU_BMIPS4380:
case CPU_BMIPS5000:
case CPU_LOONGSON2EF:
case CPU_LOONGSON64:
case CPU_R5500:
if (m4kc_tlbp_war())
uasm_i_nop(p);
fallthrough;
case CPU_ALCHEMY:
tlbw(p);
break;
case CPU_RM7000:
uasm_i_nop(p);
uasm_i_nop(p);
uasm_i_nop(p);
uasm_i_nop(p);
tlbw(p);
break;
case CPU_XBURST:
tlbw(p);
uasm_i_nop(p);
break;
default:
panic("No TLB refill handler yet (CPU type: %d)",
current_cpu_type());
break;
}
}
EXPORT_SYMBOL_GPL(build_tlb_write_entry);
static __maybe_unused void build_convert_pte_to_entrylo(u32 **p,
unsigned int reg)
{
if (_PAGE_GLOBAL_SHIFT == 0) {
/* pte_t is already in EntryLo format */
return;
}
if (cpu_has_rixi && _PAGE_NO_EXEC != 0) {
if (fill_includes_sw_bits) {
UASM_i_ROTR(p, reg, reg, ilog2(_PAGE_GLOBAL));
} else {
UASM_i_SRL(p, reg, reg, ilog2(_PAGE_NO_EXEC));
UASM_i_ROTR(p, reg, reg,
ilog2(_PAGE_GLOBAL) - ilog2(_PAGE_NO_EXEC));
}
} else {
#ifdef CONFIG_PHYS_ADDR_T_64BIT
uasm_i_dsrl_safe(p, reg, reg, ilog2(_PAGE_GLOBAL));
#else
UASM_i_SRL(p, reg, reg, ilog2(_PAGE_GLOBAL));
#endif
}
}
#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT
static void build_restore_pagemask(u32 **p, struct uasm_reloc **r,
unsigned int tmp, enum label_id lid,
int restore_scratch)
{
if (restore_scratch) {
/*
* Ensure the MFC0 below observes the value written to the
* KScratch register by the prior MTC0.
*/
if (scratch_reg >= 0)
uasm_i_ehb(p);
/* Reset default page size */
if (PM_DEFAULT_MASK >> 16) {
uasm_i_lui(p, tmp, PM_DEFAULT_MASK >> 16);
uasm_i_ori(p, tmp, tmp, PM_DEFAULT_MASK & 0xffff);
uasm_i_mtc0(p, tmp, C0_PAGEMASK);
uasm_il_b(p, r, lid);
} else if (PM_DEFAULT_MASK) {
uasm_i_ori(p, tmp, 0, PM_DEFAULT_MASK);
uasm_i_mtc0(p, tmp, C0_PAGEMASK);
uasm_il_b(p, r, lid);
} else {
uasm_i_mtc0(p, 0, C0_PAGEMASK);
uasm_il_b(p, r, lid);
}
if (scratch_reg >= 0)
UASM_i_MFC0(p, 1, c0_kscratch(), scratch_reg);
else
UASM_i_LW(p, 1, scratchpad_offset(0), 0);
} else {
/* Reset default page size */
if (PM_DEFAULT_MASK >> 16) {
uasm_i_lui(p, tmp, PM_DEFAULT_MASK >> 16);
uasm_i_ori(p, tmp, tmp, PM_DEFAULT_MASK & 0xffff);
uasm_il_b(p, r, lid);
uasm_i_mtc0(p, tmp, C0_PAGEMASK);
} else if (PM_DEFAULT_MASK) {
uasm_i_ori(p, tmp, 0, PM_DEFAULT_MASK);
uasm_il_b(p, r, lid);
uasm_i_mtc0(p, tmp, C0_PAGEMASK);
} else {
uasm_il_b(p, r, lid);
uasm_i_mtc0(p, 0, C0_PAGEMASK);
}
}
}
static void build_huge_tlb_write_entry(u32 **p, struct uasm_label **l,
struct uasm_reloc **r,
unsigned int tmp,
enum tlb_write_entry wmode,
int restore_scratch)
{
/* Set huge page tlb entry size */
uasm_i_lui(p, tmp, PM_HUGE_MASK >> 16);
uasm_i_ori(p, tmp, tmp, PM_HUGE_MASK & 0xffff);
uasm_i_mtc0(p, tmp, C0_PAGEMASK);
build_tlb_write_entry(p, l, r, wmode);
build_restore_pagemask(p, r, tmp, label_leave, restore_scratch);
}
/*
* Check if Huge PTE is present, if so then jump to LABEL.
*/
static void
build_is_huge_pte(u32 **p, struct uasm_reloc **r, unsigned int tmp,
unsigned int pmd, int lid)
{
UASM_i_LW(p, tmp, 0, pmd);
if (use_bbit_insns()) {
uasm_il_bbit1(p, r, tmp, ilog2(_PAGE_HUGE), lid);
} else {
uasm_i_andi(p, tmp, tmp, _PAGE_HUGE);
uasm_il_bnez(p, r, tmp, lid);
}
}
static void build_huge_update_entries(u32 **p, unsigned int pte,
unsigned int tmp)
{
int small_sequence;
/*
* A huge PTE describes an area the size of the
* configured huge page size. This is twice the
* of the large TLB entry size we intend to use.
* A TLB entry half the size of the configured
* huge page size is configured into entrylo0
* and entrylo1 to cover the contiguous huge PTE
* address space.
*/
small_sequence = (HPAGE_SIZE >> 7) < 0x10000;
/* We can clobber tmp. It isn't used after this.*/
if (!small_sequence)
uasm_i_lui(p, tmp, HPAGE_SIZE >> (7 + 16));
build_convert_pte_to_entrylo(p, pte);
UASM_i_MTC0(p, pte, C0_ENTRYLO0); /* load it */
/* convert to entrylo1 */
if (small_sequence)
UASM_i_ADDIU(p, pte, pte, HPAGE_SIZE >> 7);
else
UASM_i_ADDU(p, pte, pte, tmp);
UASM_i_MTC0(p, pte, C0_ENTRYLO1); /* load it */
}
static void build_huge_handler_tail(u32 **p, struct uasm_reloc **r,
struct uasm_label **l,
unsigned int pte,
unsigned int ptr,
unsigned int flush)
{
#ifdef CONFIG_SMP
UASM_i_SC(p, pte, 0, ptr);
uasm_il_beqz(p, r, pte, label_tlb_huge_update);
UASM_i_LW(p, pte, 0, ptr); /* Needed because SC killed our PTE */
#else
UASM_i_SW(p, pte, 0, ptr);
#endif
if (cpu_has_ftlb && flush) {
BUG_ON(!cpu_has_tlbinv);
UASM_i_MFC0(p, ptr, C0_ENTRYHI);
uasm_i_ori(p, ptr, ptr, MIPS_ENTRYHI_EHINV);
UASM_i_MTC0(p, ptr, C0_ENTRYHI);
build_tlb_write_entry(p, l, r, tlb_indexed);
uasm_i_xori(p, ptr, ptr, MIPS_ENTRYHI_EHINV);
UASM_i_MTC0(p, ptr, C0_ENTRYHI);
build_huge_update_entries(p, pte, ptr);
build_huge_tlb_write_entry(p, l, r, pte, tlb_random, 0);
return;
}
build_huge_update_entries(p, pte, ptr);
build_huge_tlb_write_entry(p, l, r, pte, tlb_indexed, 0);
}
#endif /* CONFIG_MIPS_HUGE_TLB_SUPPORT */
#ifdef CONFIG_64BIT
/*
* TMP and PTR are scratch.
* TMP will be clobbered, PTR will hold the pmd entry.
*/
void build_get_pmde64(u32 **p, struct uasm_label **l, struct uasm_reloc **r,
unsigned int tmp, unsigned int ptr)
{
#ifndef CONFIG_MIPS_PGD_C0_CONTEXT
long pgdc = (long)pgd_current;
#endif
/*
* The vmalloc handling is not in the hotpath.
*/
uasm_i_dmfc0(p, tmp, C0_BADVADDR);
if (check_for_high_segbits) {
/*
* The kernel currently implicitly assumes that the
* MIPS SEGBITS parameter for the processor is
* (PGDIR_SHIFT+PGDIR_BITS) or less, and will never
* allocate virtual addresses outside the maximum
* range for SEGBITS = (PGDIR_SHIFT+PGDIR_BITS). But
* that doesn't prevent user code from accessing the
* higher xuseg addresses. Here, we make sure that
* everything but the lower xuseg addresses goes down
* the module_alloc/vmalloc path.
*/
uasm_i_dsrl_safe(p, ptr, tmp, PGDIR_SHIFT + PGD_TABLE_ORDER + PAGE_SHIFT - 3);
uasm_il_bnez(p, r, ptr, label_vmalloc);
} else {
uasm_il_bltz(p, r, tmp, label_vmalloc);
}
/* No uasm_i_nop needed here, since the next insn doesn't touch TMP. */
if (pgd_reg != -1) {
/* pgd is in pgd_reg */
if (cpu_has_ldpte)
UASM_i_MFC0(p, ptr, C0_PWBASE);
else
UASM_i_MFC0(p, ptr, c0_kscratch(), pgd_reg);
} else {
#if defined(CONFIG_MIPS_PGD_C0_CONTEXT)
/*
* &pgd << 11 stored in CONTEXT [23..63].
*/
UASM_i_MFC0(p, ptr, C0_CONTEXT);
/* Clear lower 23 bits of context. */
uasm_i_dins(p, ptr, 0, 0, 23);
/* insert bit[63:59] of CAC_BASE into bit[11:6] of ptr */
uasm_i_ori(p, ptr, ptr, ((u64)(CAC_BASE) >> 53));
uasm_i_drotr(p, ptr, ptr, 11);
#elif defined(CONFIG_SMP)
UASM_i_CPUID_MFC0(p, ptr, SMP_CPUID_REG);
uasm_i_dsrl_safe(p, ptr, ptr, SMP_CPUID_PTRSHIFT);
UASM_i_LA_mostly(p, tmp, pgdc);
uasm_i_daddu(p, ptr, ptr, tmp);
uasm_i_dmfc0(p, tmp, C0_BADVADDR);
uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr);
#else
UASM_i_LA_mostly(p, ptr, pgdc);
uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr);
#endif
}
uasm_l_vmalloc_done(l, *p);
/* get pgd offset in bytes */
uasm_i_dsrl_safe(p, tmp, tmp, PGDIR_SHIFT - 3);
uasm_i_andi(p, tmp, tmp, (PTRS_PER_PGD - 1)<<3);
uasm_i_daddu(p, ptr, ptr, tmp); /* add in pgd offset */
#ifndef __PAGETABLE_PUD_FOLDED
uasm_i_dmfc0(p, tmp, C0_BADVADDR); /* get faulting address */
uasm_i_ld(p, ptr, 0, ptr); /* get pud pointer */
uasm_i_dsrl_safe(p, tmp, tmp, PUD_SHIFT - 3); /* get pud offset in bytes */
uasm_i_andi(p, tmp, tmp, (PTRS_PER_PUD - 1) << 3);
uasm_i_daddu(p, ptr, ptr, tmp); /* add in pud offset */
#endif
#ifndef __PAGETABLE_PMD_FOLDED
uasm_i_dmfc0(p, tmp, C0_BADVADDR); /* get faulting address */
uasm_i_ld(p, ptr, 0, ptr); /* get pmd pointer */
uasm_i_dsrl_safe(p, tmp, tmp, PMD_SHIFT-3); /* get pmd offset in bytes */
uasm_i_andi(p, tmp, tmp, (PTRS_PER_PMD - 1)<<3);
uasm_i_daddu(p, ptr, ptr, tmp); /* add in pmd offset */
#endif
}
EXPORT_SYMBOL_GPL(build_get_pmde64);
/*
* BVADDR is the faulting address, PTR is scratch.
* PTR will hold the pgd for vmalloc.
*/
static void
build_get_pgd_vmalloc64(u32 **p, struct uasm_label **l, struct uasm_reloc **r,
unsigned int bvaddr, unsigned int ptr,
enum vmalloc64_mode mode)
{
long swpd = (long)swapper_pg_dir;
int single_insn_swpd;
int did_vmalloc_branch = 0;
single_insn_swpd = uasm_in_compat_space_p(swpd) && !uasm_rel_lo(swpd);
uasm_l_vmalloc(l, *p);
if (mode != not_refill && check_for_high_segbits) {
if (single_insn_swpd) {
uasm_il_bltz(p, r, bvaddr, label_vmalloc_done);
uasm_i_lui(p, ptr, uasm_rel_hi(swpd));
did_vmalloc_branch = 1;
/* fall through */
} else {
uasm_il_bgez(p, r, bvaddr, label_large_segbits_fault);
}
}
if (!did_vmalloc_branch) {
if (single_insn_swpd) {
uasm_il_b(p, r, label_vmalloc_done);
uasm_i_lui(p, ptr, uasm_rel_hi(swpd));
} else {
UASM_i_LA_mostly(p, ptr, swpd);
uasm_il_b(p, r, label_vmalloc_done);
if (uasm_in_compat_space_p(swpd))
uasm_i_addiu(p, ptr, ptr, uasm_rel_lo(swpd));
else
uasm_i_daddiu(p, ptr, ptr, uasm_rel_lo(swpd));
}
}
if (mode != not_refill && check_for_high_segbits) {
uasm_l_large_segbits_fault(l, *p);
if (mode == refill_scratch && scratch_reg >= 0)
uasm_i_ehb(p);
/*
* We get here if we are an xsseg address, or if we are
* an xuseg address above (PGDIR_SHIFT+PGDIR_BITS) boundary.
*
* Ignoring xsseg (assume disabled so would generate
* (address errors?), the only remaining possibility
* is the upper xuseg addresses. On processors with
* TLB_SEGBITS <= PGDIR_SHIFT+PGDIR_BITS, these
* addresses would have taken an address error. We try
* to mimic that here by taking a load/istream page
* fault.
*/
if (IS_ENABLED(CONFIG_CPU_LOONGSON3_WORKAROUNDS))
uasm_i_sync(p, 0);
UASM_i_LA(p, ptr, (unsigned long)tlb_do_page_fault_0);
uasm_i_jr(p, ptr);
if (mode == refill_scratch) {
if (scratch_reg >= 0)
UASM_i_MFC0(p, 1, c0_kscratch(), scratch_reg);
else
UASM_i_LW(p, 1, scratchpad_offset(0), 0);
} else {
uasm_i_nop(p);
}
}
}
#else /* !CONFIG_64BIT */
/*
* TMP and PTR are scratch.
* TMP will be clobbered, PTR will hold the pgd entry.
*/
void build_get_pgde32(u32 **p, unsigned int tmp, unsigned int ptr)
{
if (pgd_reg != -1) {
/* pgd is in pgd_reg */
uasm_i_mfc0(p, ptr, c0_kscratch(), pgd_reg);
uasm_i_mfc0(p, tmp, C0_BADVADDR); /* get faulting address */
} else {
long pgdc = (long)pgd_current;
/* 32 bit SMP has smp_processor_id() stored in CONTEXT. */
#ifdef CONFIG_SMP
uasm_i_mfc0(p, ptr, SMP_CPUID_REG);
UASM_i_LA_mostly(p, tmp, pgdc);
uasm_i_srl(p, ptr, ptr, SMP_CPUID_PTRSHIFT);
uasm_i_addu(p, ptr, tmp, ptr);
#else
UASM_i_LA_mostly(p, ptr, pgdc);
#endif
uasm_i_mfc0(p, tmp, C0_BADVADDR); /* get faulting address */
uasm_i_lw(p, ptr, uasm_rel_lo(pgdc), ptr);
}
uasm_i_srl(p, tmp, tmp, PGDIR_SHIFT); /* get pgd only bits */
uasm_i_sll(p, tmp, tmp, PGD_T_LOG2);
uasm_i_addu(p, ptr, ptr, tmp); /* add in pgd offset */
}
EXPORT_SYMBOL_GPL(build_get_pgde32);
#endif /* !CONFIG_64BIT */
static void build_adjust_context(u32 **p, unsigned int ctx)
{
unsigned int shift = 4 - (PTE_T_LOG2 + 1) + PAGE_SHIFT - 12;
unsigned int mask = (PTRS_PER_PTE / 2 - 1) << (PTE_T_LOG2 + 1);
if (shift)
UASM_i_SRL(p, ctx, ctx, shift);
uasm_i_andi(p, ctx, ctx, mask);
}
void build_get_ptep(u32 **p, unsigned int tmp, unsigned int ptr)
{
/*
* Bug workaround for the Nevada. It seems as if under certain
* circumstances the move from cp0_context might produce a
* bogus result when the mfc0 instruction and its consumer are
* in a different cacheline or a load instruction, probably any
* memory reference, is between them.
*/
switch (current_cpu_type()) {
case CPU_NEVADA:
UASM_i_LW(p, ptr, 0, ptr);
GET_CONTEXT(p, tmp); /* get context reg */
break;
default:
GET_CONTEXT(p, tmp); /* get context reg */
UASM_i_LW(p, ptr, 0, ptr);
break;
}
build_adjust_context(p, tmp);
UASM_i_ADDU(p, ptr, ptr, tmp); /* add in offset */
}
EXPORT_SYMBOL_GPL(build_get_ptep);
void build_update_entries(u32 **p, unsigned int tmp, unsigned int ptep)
{
int pte_off_even = 0;
int pte_off_odd = sizeof(pte_t);
#if defined(CONFIG_CPU_MIPS32) && defined(CONFIG_PHYS_ADDR_T_64BIT)
/* The low 32 bits of EntryLo is stored in pte_high */
pte_off_even += offsetof(pte_t, pte_high);
pte_off_odd += offsetof(pte_t, pte_high);
#endif