summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
blob: 98732518543e539e601866a4beda3666732408db (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
/*
 * Copyright 2019 Advanced Micro Devices, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 *
 */

#include "amdgpu_ras_eeprom.h"
#include "amdgpu.h"
#include "amdgpu_ras.h"
#include <linux/bits.h>
#include "atom.h"
#include "amdgpu_eeprom.h"
#include "amdgpu_atomfirmware.h"
#include <linux/debugfs.h>
#include <linux/uaccess.h>

#define EEPROM_I2C_MADDR_VEGA20         0x0
#define EEPROM_I2C_MADDR_ARCTURUS       0x40000
#define EEPROM_I2C_MADDR_ARCTURUS_D342  0x0
#define EEPROM_I2C_MADDR_SIENNA_CICHLID 0x0
#define EEPROM_I2C_MADDR_ALDEBARAN      0x0

/*
 * The 2 macros bellow represent the actual size in bytes that
 * those entities occupy in the EEPROM memory.
 * RAS_TABLE_RECORD_SIZE is different than sizeof(eeprom_table_record) which
 * uses uint64 to store 6b fields such as retired_page.
 */
#define RAS_TABLE_HEADER_SIZE   20
#define RAS_TABLE_RECORD_SIZE   24

/* Table hdr is 'AMDR' */
#define RAS_TABLE_HDR_VAL       0x414d4452
#define RAS_TABLE_VER           0x00010000

/* Bad GPU tag ‘BADG’ */
#define RAS_TABLE_HDR_BAD       0x42414447

/* Assume 2-Mbit size EEPROM and take up the whole space. */
#define RAS_TBL_SIZE_BYTES      (256 * 1024)
#define RAS_TABLE_START         0
#define RAS_HDR_START           RAS_TABLE_START
#define RAS_RECORD_START        (RAS_HDR_START + RAS_TABLE_HEADER_SIZE)
#define RAS_MAX_RECORD_COUNT    ((RAS_TBL_SIZE_BYTES - RAS_TABLE_HEADER_SIZE) \
				 / RAS_TABLE_RECORD_SIZE)

/* Given a zero-based index of an EEPROM RAS record, yields the EEPROM
 * offset off of RAS_TABLE_START.  That is, this is something you can
 * add to control->i2c_address, and then tell I2C layer to read
 * from/write to there. _N is the so called absolute index,
 * because it starts right after the table header.
 */
#define RAS_INDEX_TO_OFFSET(_C, _N) ((_C)->ras_record_offset + \
				     (_N) * RAS_TABLE_RECORD_SIZE)

#define RAS_OFFSET_TO_INDEX(_C, _O) (((_O) - \
				      (_C)->ras_record_offset) / RAS_TABLE_RECORD_SIZE)

/* Given a 0-based relative record index, 0, 1, 2, ..., etc., off
 * of "fri", return the absolute record index off of the end of
 * the table header.
 */
#define RAS_RI_TO_AI(_C, _I) (((_I) + (_C)->ras_fri) % \
			      (_C)->ras_max_record_count)

#define RAS_NUM_RECS(_tbl_hdr)  (((_tbl_hdr)->tbl_size - \
				  RAS_TABLE_HEADER_SIZE) / RAS_TABLE_RECORD_SIZE)

#define to_amdgpu_device(x) (container_of(x, struct amdgpu_ras, eeprom_control))->adev

static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
{
	return  adev->asic_type == CHIP_VEGA20 ||
		adev->asic_type == CHIP_ARCTURUS ||
		adev->asic_type == CHIP_SIENNA_CICHLID ||
		adev->asic_type == CHIP_ALDEBARAN;
}

static bool __get_eeprom_i2c_addr_arct(struct amdgpu_device *adev,
				       struct amdgpu_ras_eeprom_control *control)
{
	struct atom_context *atom_ctx = adev->mode_info.atom_context;

	if (!control || !atom_ctx)
		return false;

	if (strnstr(atom_ctx->vbios_version,
	            "D342",
		    sizeof(atom_ctx->vbios_version)))
		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS_D342;
	else
		control->i2c_address = EEPROM_I2C_MADDR_ARCTURUS;

	return true;
}

static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
				  struct amdgpu_ras_eeprom_control *control)
{
	u8 i2c_addr;

	if (!control)
		return false;

	if (amdgpu_atomfirmware_ras_rom_addr(adev, &i2c_addr)) {
		/* The address given by VBIOS is an 8-bit, wire-format
		 * address, i.e. the most significant byte.
		 *
		 * Normalize it to a 19-bit EEPROM address. Remove the
		 * device type identifier and make it a 7-bit address;
		 * then make it a 19-bit EEPROM address. See top of
		 * amdgpu_eeprom.c.
		 */
		i2c_addr = (i2c_addr & 0x0F) >> 1;
		control->i2c_address = ((u32) i2c_addr) << 16;

		return true;
	}

	switch (adev->asic_type) {
	case CHIP_VEGA20:
		control->i2c_address = EEPROM_I2C_MADDR_VEGA20;
		break;

	case CHIP_ARCTURUS:
		return __get_eeprom_i2c_addr_arct(adev, control);

	case CHIP_SIENNA_CICHLID:
		control->i2c_address = EEPROM_I2C_MADDR_SIENNA_CICHLID;
		break;

	case CHIP_ALDEBARAN:
		control->i2c_address = EEPROM_I2C_MADDR_ALDEBARAN;
		break;

	default:
		return false;
	}

	return true;
}

static void
__encode_table_header_to_buf(struct amdgpu_ras_eeprom_table_header *hdr,
			     unsigned char *buf)
{
	u32 *pp = (uint32_t *)buf;

	pp[0] = cpu_to_le32(hdr->header);
	pp[1] = cpu_to_le32(hdr->version);
	pp[2] = cpu_to_le32(hdr->first_rec_offset);
	pp[3] = cpu_to_le32(hdr->tbl_size);
	pp[4] = cpu_to_le32(hdr->checksum);
}

static void
__decode_table_header_from_buf(struct amdgpu_ras_eeprom_table_header *hdr,
			       unsigned char *buf)
{
	u32 *pp = (uint32_t *)buf;

	hdr->header	      = le32_to_cpu(pp[0]);
	hdr->version	      = le32_to_cpu(pp[1]);
	hdr->first_rec_offset = le32_to_cpu(pp[2]);
	hdr->tbl_size	      = le32_to_cpu(pp[3]);
	hdr->checksum	      = le32_to_cpu(pp[4]);
}

static int __write_table_header(struct amdgpu_ras_eeprom_control *control)
{
	u8 buf[RAS_TABLE_HEADER_SIZE];
	struct amdgpu_device *adev = to_amdgpu_device(control);
	int res;

	memset(buf, 0, sizeof(buf));
	__encode_table_header_to_buf(&control->tbl_hdr, buf);

	/* i2c may be unstable in gpu reset */
	down_read(&adev->reset_sem);
	res = amdgpu_eeprom_write(&adev->pm.smu_i2c,
				  control->i2c_address +
				  control->ras_header_offset,
				  buf, RAS_TABLE_HEADER_SIZE);
	up_read(&adev->reset_sem);

	if (res < 0) {
		DRM_ERROR("Failed to write EEPROM table header:%d", res);
	} else if (res < RAS_TABLE_HEADER_SIZE) {
		DRM_ERROR("Short write:%d out of %d\n",
			  res, RAS_TABLE_HEADER_SIZE);
		res = -EIO;
	} else {
		res = 0;
	}

	return res;
}

static u8 __calc_hdr_byte_sum(const struct amdgpu_ras_eeprom_control *control)
{
	int ii;
	u8  *pp, csum;
	size_t sz;

	/* Header checksum, skip checksum field in the calculation */
	sz = sizeof(control->tbl_hdr) - sizeof(control->tbl_hdr.checksum);
	pp = (u8 *) &control->tbl_hdr;
	csum = 0;
	for (ii = 0; ii < sz; ii++, pp++)
		csum += *pp;

	return csum;
}

static int amdgpu_ras_eeprom_correct_header_tag(
	struct amdgpu_ras_eeprom_control *control,
	uint32_t header)
{
	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
	u8 *hh;
	int res;
	u8 csum;

	csum = -hdr->checksum;

	hh = (void *) &hdr->header;
	csum -= (hh[0] + hh[1] + hh[2] + hh[3]);
	hh = (void *) &header;
	csum += hh[0] + hh[1] + hh[2] + hh[3];
	csum = -csum;
	mutex_lock(&control->ras_tbl_mutex);
	hdr->header = header;
	hdr->checksum = csum;
	res = __write_table_header(control);
	mutex_unlock(&control->ras_tbl_mutex);

	return res;
}

/**
 * amdgpu_ras_eeprom_reset_table -- Reset the RAS EEPROM table
 * @control: pointer to control structure
 *
 * Reset the contents of the header of the RAS EEPROM table.
 * Return 0 on success, -errno on error.
 */
int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
{
	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
	u8 csum;
	int res;

	mutex_lock(&control->ras_tbl_mutex);

	hdr->header = RAS_TABLE_HDR_VAL;
	hdr->version = RAS_TABLE_VER;
	hdr->first_rec_offset = RAS_RECORD_START;
	hdr->tbl_size = RAS_TABLE_HEADER_SIZE;

	csum = __calc_hdr_byte_sum(control);
	csum = -csum;
	hdr->checksum = csum;
	res = __write_table_header(control);

	control->ras_num_recs = 0;
	control->ras_fri = 0;

	amdgpu_ras_debugfs_set_ret_size(control);

	mutex_unlock(&control->ras_tbl_mutex);

	return res;
}

static void
__encode_table_record_to_buf(struct amdgpu_ras_eeprom_control *control,
			     struct eeprom_table_record *record,
			     unsigned char *buf)
{
	__le64 tmp = 0;
	int i = 0;

	/* Next are all record fields according to EEPROM page spec in LE foramt */
	buf[i++] = record->err_type;

	buf[i++] = record->bank;

	tmp = cpu_to_le64(record->ts);
	memcpy(buf + i, &tmp, 8);
	i += 8;

	tmp = cpu_to_le64((record->offset & 0xffffffffffff));
	memcpy(buf + i, &tmp, 6);
	i += 6;

	buf[i++] = record->mem_channel;
	buf[i++] = record->mcumc_id;

	tmp = cpu_to_le64((record->retired_page & 0xffffffffffff));
	memcpy(buf + i, &tmp, 6);
}

static void
__decode_table_record_from_buf(struct amdgpu_ras_eeprom_control *control,
			       struct eeprom_table_record *record,
			       unsigned char *buf)
{
	__le64 tmp = 0;
	int i =  0;

	/* Next are all record fields according to EEPROM page spec in LE foramt */
	record->err_type = buf[i++];

	record->bank = buf[i++];

	memcpy(&tmp, buf + i, 8);
	record->ts = le64_to_cpu(tmp);
	i += 8;

	memcpy(&tmp, buf + i, 6);
	record->offset = (le64_to_cpu(tmp) & 0xffffffffffff);
	i += 6;

	record->mem_channel = buf[i++];
	record->mcumc_id = buf[i++];

	memcpy(&tmp, buf + i,  6);
	record->retired_page = (le64_to_cpu(tmp) & 0xffffffffffff);
}

bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev)
{
	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);

	if (!__is_ras_eeprom_supported(adev))
		return false;

	/* skip check eeprom table for VEGA20 Gaming */
	if (!con)
		return false;
	else
		if (!(con->features & BIT(AMDGPU_RAS_BLOCK__UMC)))
			return false;

	if (con->eeprom_control.tbl_hdr.header == RAS_TABLE_HDR_BAD) {
		dev_warn(adev->dev, "This GPU is in BAD status.");
		dev_warn(adev->dev, "Please retire it or set a larger "
			 "threshold value when reloading driver.\n");
		return true;
	}

	return false;
}

/**
 * __amdgpu_ras_eeprom_write -- write indexed from buffer to EEPROM
 * @control: pointer to control structure
 * @buf: pointer to buffer containing data to write
 * @fri: start writing at this index
 * @num: number of records to write
 *
 * The caller must hold the table mutex in @control.
 * Return 0 on success, -errno otherwise.
 */
static int __amdgpu_ras_eeprom_write(struct amdgpu_ras_eeprom_control *control,
				     u8 *buf, const u32 fri, const u32 num)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	u32 buf_size;
	int res;

	/* i2c may be unstable in gpu reset */
	down_read(&adev->reset_sem);
	buf_size = num * RAS_TABLE_RECORD_SIZE;
	res = amdgpu_eeprom_write(&adev->pm.smu_i2c,
				  control->i2c_address +
				  RAS_INDEX_TO_OFFSET(control, fri),
				  buf, buf_size);
	up_read(&adev->reset_sem);
	if (res < 0) {
		DRM_ERROR("Writing %d EEPROM table records error:%d",
			  num, res);
	} else if (res < buf_size) {
		/* Short write, return error.
		 */
		DRM_ERROR("Wrote %d records out of %d",
			  res / RAS_TABLE_RECORD_SIZE, num);
		res = -EIO;
	} else {
		res = 0;
	}

	return res;
}

static int
amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
			       struct eeprom_table_record *record,
			       const u32 num)
{
	u32 a, b, i;
	u8 *buf, *pp;
	int res;

	buf = kcalloc(num, RAS_TABLE_RECORD_SIZE, GFP_KERNEL);
	if (!buf)
		return -ENOMEM;

	/* Encode all of them in one go.
	 */
	pp = buf;
	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
		__encode_table_record_to_buf(control, &record[i], pp);

	/* a, first record index to write into.
	 * b, last record index to write into.
	 * a = first index to read (fri) + number of records in the table,
	 * b = a + @num - 1.
	 * Let N = control->ras_max_num_record_count, then we have,
	 * case 0: 0 <= a <= b < N,
	 *   just append @num records starting at a;
	 * case 1: 0 <= a < N <= b,
	 *   append (N - a) records starting at a, and
	 *   append the remainder,  b % N + 1, starting at 0.
	 * case 2: 0 <= fri < N <= a <= b, then modulo N we get two subcases,
	 * case 2a: 0 <= a <= b < N
	 *   append num records starting at a; and fix fri if b overwrote it,
	 *   and since a <= b, if b overwrote it then a must've also,
	 *   and if b didn't overwrite it, then a didn't also.
	 * case 2b: 0 <= b < a < N
	 *   write num records starting at a, which wraps around 0=N
	 *   and overwrite fri unconditionally. Now from case 2a,
	 *   this means that b eclipsed fri to overwrite it and wrap
	 *   around 0 again, i.e. b = 2N+r pre modulo N, so we unconditionally
	 *   set fri = b + 1 (mod N).
	 * Now, since fri is updated in every case, except the trivial case 0,
	 * the number of records present in the table after writing, is,
	 * num_recs - 1 = b - fri (mod N), and we take the positive value,
	 * by adding an arbitrary multiple of N before taking the modulo N
	 * as shown below.
	 */
	a = control->ras_fri + control->ras_num_recs;
	b = a + num  - 1;
	if (b < control->ras_max_record_count) {
		res = __amdgpu_ras_eeprom_write(control, buf, a, num);
	} else if (a < control->ras_max_record_count) {
		u32 g0, g1;

		g0 = control->ras_max_record_count - a;
		g1 = b % control->ras_max_record_count + 1;
		res = __amdgpu_ras_eeprom_write(control, buf, a, g0);
		if (res)
			goto Out;
		res = __amdgpu_ras_eeprom_write(control,
						buf + g0 * RAS_TABLE_RECORD_SIZE,
						0, g1);
		if (res)
			goto Out;
		if (g1 > control->ras_fri)
			control->ras_fri = g1 % control->ras_max_record_count;
	} else {
		a %= control->ras_max_record_count;
		b %= control->ras_max_record_count;

		if (a <= b) {
			/* Note that, b - a + 1 = num. */
			res = __amdgpu_ras_eeprom_write(control, buf, a, num);
			if (res)
				goto Out;
			if (b >= control->ras_fri)
				control->ras_fri = (b + 1) % control->ras_max_record_count;
		} else {
			u32 g0, g1;

			/* b < a, which means, we write from
			 * a to the end of the table, and from
			 * the start of the table to b.
			 */
			g0 = control->ras_max_record_count - a;
			g1 = b + 1;
			res = __amdgpu_ras_eeprom_write(control, buf, a, g0);
			if (res)
				goto Out;
			res = __amdgpu_ras_eeprom_write(control,
							buf + g0 * RAS_TABLE_RECORD_SIZE,
							0, g1);
			if (res)
				goto Out;
			control->ras_fri = g1 % control->ras_max_record_count;
		}
	}
	control->ras_num_recs = 1 + (control->ras_max_record_count + b
				     - control->ras_fri)
		% control->ras_max_record_count;
Out:
	kfree(buf);
	return res;
}

static int
amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
	u8 *buf, *pp, csum;
	u32 buf_size;
	int res;

	/* Modify the header if it exceeds.
	 */
	if (amdgpu_bad_page_threshold != 0 &&
	    control->ras_num_recs >= ras->bad_page_cnt_threshold) {
		dev_warn(adev->dev,
			"Saved bad pages %d reaches threshold value %d\n",
			control->ras_num_recs, ras->bad_page_cnt_threshold);
		control->tbl_hdr.header = RAS_TABLE_HDR_BAD;
	}

	control->tbl_hdr.version = RAS_TABLE_VER;
	control->tbl_hdr.first_rec_offset = RAS_INDEX_TO_OFFSET(control, control->ras_fri);
	control->tbl_hdr.tbl_size = RAS_TABLE_HEADER_SIZE + control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
	control->tbl_hdr.checksum = 0;

	buf_size = control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
	buf = kcalloc(control->ras_num_recs, RAS_TABLE_RECORD_SIZE, GFP_KERNEL);
	if (!buf) {
		DRM_ERROR("allocating memory for table of size %d bytes failed\n",
			  control->tbl_hdr.tbl_size);
		res = -ENOMEM;
		goto Out;
	}

	down_read(&adev->reset_sem);
	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
				 control->i2c_address +
				 control->ras_record_offset,
				 buf, buf_size);
	up_read(&adev->reset_sem);
	if (res < 0) {
		DRM_ERROR("EEPROM failed reading records:%d\n",
			  res);
		goto Out;
	} else if (res < buf_size) {
		DRM_ERROR("EEPROM read %d out of %d bytes\n",
			  res, buf_size);
		res = -EIO;
		goto Out;
	}

	/* Recalc the checksum.
	 */
	csum = 0;
	for (pp = buf; pp < buf + buf_size; pp++)
		csum += *pp;

	csum += __calc_hdr_byte_sum(control);
	/* avoid sign extension when assigning to "checksum" */
	csum = -csum;
	control->tbl_hdr.checksum = csum;
	res = __write_table_header(control);
Out:
	kfree(buf);
	return res;
}

/**
 * amdgpu_ras_eeprom_append -- append records to the EEPROM RAS table
 * @control: pointer to control structure
 * @record: array of records to append
 * @num: number of records in @record array
 *
 * Append @num records to the table, calculate the checksum and write
 * the table back to EEPROM. The maximum number of records that
 * can be appended is between 1 and control->ras_max_record_count,
 * regardless of how many records are already stored in the table.
 *
 * Return 0 on success or if EEPROM is not supported, -errno on error.
 */
int amdgpu_ras_eeprom_append(struct amdgpu_ras_eeprom_control *control,
			     struct eeprom_table_record *record,
			     const u32 num)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	int res;

	if (!__is_ras_eeprom_supported(adev))
		return 0;

	if (num == 0) {
		DRM_ERROR("will not append 0 records\n");
		return -EINVAL;
	} else if (num > control->ras_max_record_count) {
		DRM_ERROR("cannot append %d records than the size of table %d\n",
			  num, control->ras_max_record_count);
		return -EINVAL;
	}

	mutex_lock(&control->ras_tbl_mutex);

	res = amdgpu_ras_eeprom_append_table(control, record, num);
	if (!res)
		res = amdgpu_ras_eeprom_update_header(control);
	if (!res)
		amdgpu_ras_debugfs_set_ret_size(control);

	mutex_unlock(&control->ras_tbl_mutex);
	return res;
}

/**
 * __amdgpu_ras_eeprom_read -- read indexed from EEPROM into buffer
 * @control: pointer to control structure
 * @buf: pointer to buffer to read into
 * @fri: first record index, start reading at this index, absolute index
 * @num: number of records to read
 *
 * The caller must hold the table mutex in @control.
 * Return 0 on success, -errno otherwise.
 */
static int __amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
				    u8 *buf, const u32 fri, const u32 num)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	u32 buf_size;
	int res;

	/* i2c may be unstable in gpu reset */
	down_read(&adev->reset_sem);
	buf_size = num * RAS_TABLE_RECORD_SIZE;
	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
				 control->i2c_address +
				 RAS_INDEX_TO_OFFSET(control, fri),
				 buf, buf_size);
	up_read(&adev->reset_sem);
	if (res < 0) {
		DRM_ERROR("Reading %d EEPROM table records error:%d",
			  num, res);
	} else if (res < buf_size) {
		/* Short read, return error.
		 */
		DRM_ERROR("Read %d records out of %d",
			  res / RAS_TABLE_RECORD_SIZE, num);
		res = -EIO;
	} else {
		res = 0;
	}

	return res;
}

/**
 * amdgpu_ras_eeprom_read -- read EEPROM
 * @control: pointer to control structure
 * @record: array of records to read into
 * @num: number of records in @record
 *
 * Reads num records from the RAS table in EEPROM and
 * writes the data into @record array.
 *
 * Returns 0 on success, -errno on error.
 */
int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
			   struct eeprom_table_record *record,
			   const u32 num)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	int i, res;
	u8 *buf, *pp;
	u32 g0, g1;

	if (!__is_ras_eeprom_supported(adev))
		return 0;

	if (num == 0) {
		DRM_ERROR("will not read 0 records\n");
		return -EINVAL;
	} else if (num > control->ras_num_recs) {
		DRM_ERROR("too many records to read:%d available:%d\n",
			  num, control->ras_num_recs);
		return -EINVAL;
	}

	buf = kcalloc(num, RAS_TABLE_RECORD_SIZE, GFP_KERNEL);
	if (!buf)
		return -ENOMEM;

	/* Determine how many records to read, from the first record
	 * index, fri, to the end of the table, and from the beginning
	 * of the table, such that the total number of records is
	 * @num, and we handle wrap around when fri > 0 and
	 * fri + num > RAS_MAX_RECORD_COUNT.
	 *
	 * First we compute the index of the last element
	 * which would be fetched from each region,
	 * g0 is in [fri, fri + num - 1], and
	 * g1 is in [0, RAS_MAX_RECORD_COUNT - 1].
	 * Then, if g0 < RAS_MAX_RECORD_COUNT, the index of
	 * the last element to fetch, we set g0 to _the number_
	 * of elements to fetch, @num, since we know that the last
	 * indexed to be fetched does not exceed the table.
	 *
	 * If, however, g0 >= RAS_MAX_RECORD_COUNT, then
	 * we set g0 to the number of elements to read
	 * until the end of the table, and g1 to the number of
	 * elements to read from the beginning of the table.
	 */
	g0 = control->ras_fri + num - 1;
	g1 = g0 % control->ras_max_record_count;
	if (g0 < control->ras_max_record_count) {
		g0 = num;
		g1 = 0;
	} else {
		g0 = control->ras_max_record_count - control->ras_fri;
		g1 += 1;
	}

	mutex_lock(&control->ras_tbl_mutex);
	res = __amdgpu_ras_eeprom_read(control, buf, control->ras_fri, g0);
	if (res)
		goto Out;
	if (g1) {
		res = __amdgpu_ras_eeprom_read(control,
					       buf + g0 * RAS_TABLE_RECORD_SIZE,
					       0, g1);
		if (res)
			goto Out;
	}

	res = 0;

	/* Read up everything? Then transform.
	 */
	pp = buf;
	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
		__decode_table_record_from_buf(control, &record[i], pp);
Out:
	kfree(buf);
	mutex_unlock(&control->ras_tbl_mutex);

	return res;
}

uint32_t amdgpu_ras_eeprom_max_record_count(void)
{
	return RAS_MAX_RECORD_COUNT;
}

static ssize_t
amdgpu_ras_debugfs_eeprom_size_read(struct file *f, char __user *buf,
				    size_t size, loff_t *pos)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
	struct amdgpu_ras_eeprom_control *control = ras ? &ras->eeprom_control : NULL;
	u8 data[50];
	int res;

	if (!size)
		return size;

	if (!ras || !control) {
		res = snprintf(data, sizeof(data), "Not supported\n");
	} else {
		res = snprintf(data, sizeof(data), "%d bytes or %d records\n",
			       RAS_TBL_SIZE_BYTES, control->ras_max_record_count);
	}

	if (*pos >= res)
		return 0;

	res -= *pos;
	res = min_t(size_t, res, size);

	if (copy_to_user(buf, &data[*pos], res))
		return -EFAULT;

	*pos += res;

	return res;
}

const struct file_operations amdgpu_ras_debugfs_eeprom_size_ops = {
	.owner = THIS_MODULE,
	.read = amdgpu_ras_debugfs_eeprom_size_read,
	.write = NULL,
	.llseek = default_llseek,
};

static const char *tbl_hdr_str = " Signature    Version  FirstOffs       Size   Checksum\n";
static const char *tbl_hdr_fmt = "0x%08X 0x%08X 0x%08X 0x%08X 0x%08X\n";
#define tbl_hdr_fmt_size (5 * (2+8) + 4 + 1)
static const char *rec_hdr_str = "Index  Offset ErrType Bank/CU          TimeStamp      Offs/Addr MemChl MCUMCID    RetiredPage\n";
static const char *rec_hdr_fmt = "%5d 0x%05X %7s    0x%02X 0x%016llX 0x%012llX   0x%02X    0x%02X 0x%012llX\n";
#define rec_hdr_fmt_size (5 + 1 + 7 + 1 + 7 + 1 + 7 + 1 + 18 + 1 + 14 + 1 + 6 + 1 + 7 + 1 + 14 + 1)

static const char *record_err_type_str[AMDGPU_RAS_EEPROM_ERR_COUNT] = {
	"ignore",
	"re",
	"ue",
};

static loff_t amdgpu_ras_debugfs_table_size(struct amdgpu_ras_eeprom_control *control)
{
	return strlen(tbl_hdr_str) + tbl_hdr_fmt_size +
		strlen(rec_hdr_str) + rec_hdr_fmt_size * control->ras_num_recs;
}

void amdgpu_ras_debugfs_set_ret_size(struct amdgpu_ras_eeprom_control *control)
{
	struct amdgpu_ras *ras = container_of(control, struct amdgpu_ras,
					      eeprom_control);
	struct dentry *de = ras->de_ras_eeprom_table;

	if (de)
		d_inode(de)->i_size = amdgpu_ras_debugfs_table_size(control);
}

static ssize_t amdgpu_ras_debugfs_table_read(struct file *f, char __user *buf,
					     size_t size, loff_t *pos)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
	struct amdgpu_ras_eeprom_control *control = &ras->eeprom_control;
	const size_t orig_size = size;
	int res = -EFAULT;
	size_t data_len;

	mutex_lock(&control->ras_tbl_mutex);

	/* We want *pos - data_len > 0, which means there's
	 * bytes to be printed from data.
	 */
	data_len = strlen(tbl_hdr_str);
	if (*pos < data_len) {
		data_len -= *pos;
		data_len = min_t(size_t, data_len, size);
		if (copy_to_user(buf, &tbl_hdr_str[*pos], data_len))
			goto Out;
		buf += data_len;
		size -= data_len;
		*pos += data_len;
	}

	data_len = strlen(tbl_hdr_str) + tbl_hdr_fmt_size;
	if (*pos < data_len && size > 0) {
		u8 data[tbl_hdr_fmt_size + 1];
		loff_t lpos;

		snprintf(data, sizeof(data), tbl_hdr_fmt,
			 control->tbl_hdr.header,
			 control->tbl_hdr.version,
			 control->tbl_hdr.first_rec_offset,
			 control->tbl_hdr.tbl_size,
			 control->tbl_hdr.checksum);

		data_len -= *pos;
		data_len = min_t(size_t, data_len, size);
		lpos = *pos - strlen(tbl_hdr_str);
		if (copy_to_user(buf, &data[lpos], data_len))
			goto Out;
		buf += data_len;
		size -= data_len;
		*pos += data_len;
	}

	data_len = strlen(tbl_hdr_str) + tbl_hdr_fmt_size + strlen(rec_hdr_str);
	if (*pos < data_len && size > 0) {
		loff_t lpos;

		data_len -= *pos;
		data_len = min_t(size_t, data_len, size);
		lpos = *pos - strlen(tbl_hdr_str) - tbl_hdr_fmt_size;
		if (copy_to_user(buf, &rec_hdr_str[lpos], data_len))
			goto Out;
		buf += data_len;
		size -= data_len;
		*pos += data_len;
	}

	data_len = amdgpu_ras_debugfs_table_size(control);
	if (*pos < data_len && size > 0) {
		u8 dare[RAS_TABLE_RECORD_SIZE];
		u8 data[rec_hdr_fmt_size + 1];
		struct eeprom_table_record record;
		int s, r;

		/* Find the starting record index
		 */
		s = *pos - strlen(tbl_hdr_str) - tbl_hdr_fmt_size -
			strlen(rec_hdr_str);
		s = s / rec_hdr_fmt_size;
		r = *pos - strlen(tbl_hdr_str) - tbl_hdr_fmt_size -
			strlen(rec_hdr_str);
		r = r % rec_hdr_fmt_size;

		for ( ; size > 0 && s < control->ras_num_recs; s++) {
			u32 ai = RAS_RI_TO_AI(control, s);
			/* Read a single record
			 */
			res = __amdgpu_ras_eeprom_read(control, dare, ai, 1);
			if (res)
				goto Out;
			__decode_table_record_from_buf(control, &record, dare);
			snprintf(data, sizeof(data), rec_hdr_fmt,
				 s,
				 RAS_INDEX_TO_OFFSET(control, ai),
				 record_err_type_str[record.err_type],
				 record.bank,
				 record.ts,
				 record.offset,
				 record.mem_channel,
				 record.mcumc_id,
				 record.retired_page);

			data_len = min_t(size_t, rec_hdr_fmt_size - r, size);
			if (copy_to_user(buf, &data[r], data_len)) {
				res = -EFAULT;
				goto Out;
			}
			buf += data_len;
			size -= data_len;
			*pos += data_len;
			r = 0;
		}
	}
	res = 0;
Out:
	mutex_unlock(&control->ras_tbl_mutex);
	return res < 0 ? res : orig_size - size;
}

static ssize_t
amdgpu_ras_debugfs_eeprom_table_read(struct file *f, char __user *buf,
				     size_t size, loff_t *pos)
{
	struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private;
	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
	struct amdgpu_ras_eeprom_control *control = ras ? &ras->eeprom_control : NULL;
	u8 data[81];
	int res;

	if (!size)
		return size;

	if (!ras || !control) {
		res = snprintf(data, sizeof(data), "Not supported\n");
		if (*pos >= res)
			return 0;

		res -= *pos;
		res = min_t(size_t, res, size);

		if (copy_to_user(buf, &data[*pos], res))
			return -EFAULT;

		*pos += res;

		return res;
	} else {
		return amdgpu_ras_debugfs_table_read(f, buf, size, pos);
	}
}

const struct file_operations amdgpu_ras_debugfs_eeprom_table_ops = {
	.owner = THIS_MODULE,
	.read = amdgpu_ras_debugfs_eeprom_table_read,
	.write = NULL,
	.llseek = default_llseek,
};

/**
 * __verify_ras_table_checksum -- verify the RAS EEPROM table checksum
 * @control: pointer to control structure
 *
 * Check the checksum of the stored in EEPROM RAS table.
 *
 * Return 0 if the checksum is correct,
 * positive if it is not correct, and
 * -errno on I/O error.
 */
static int __verify_ras_table_checksum(struct amdgpu_ras_eeprom_control *control)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	int buf_size, res;
	u8  csum, *buf, *pp;

	buf_size = RAS_TABLE_HEADER_SIZE +
		control->ras_num_recs * RAS_TABLE_RECORD_SIZE;
	buf = kzalloc(buf_size, GFP_KERNEL);
	if (!buf) {
		DRM_ERROR("Out of memory checking RAS table checksum.\n");
		return -ENOMEM;
	}

	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
				 control->i2c_address +
				 control->ras_header_offset,
				 buf, buf_size);
	if (res < buf_size) {
		DRM_ERROR("Partial read for checksum, res:%d\n", res);
		/* On partial reads, return -EIO.
		 */
		if (res >= 0)
			res = -EIO;
		goto Out;
	}

	csum = 0;
	for (pp = buf; pp < buf + buf_size; pp++)
		csum += *pp;
Out:
	kfree(buf);
	return res < 0 ? res : csum;
}

int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
			   bool *exceed_err_limit)
{
	struct amdgpu_device *adev = to_amdgpu_device(control);
	unsigned char buf[RAS_TABLE_HEADER_SIZE] = { 0 };
	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
	int res;

	*exceed_err_limit = false;

	if (!__is_ras_eeprom_supported(adev))
		return 0;

	/* Verify i2c adapter is initialized */
	if (!adev->pm.smu_i2c.algo)
		return -ENOENT;

	if (!__get_eeprom_i2c_addr(adev, control))
		return -EINVAL;

	control->ras_header_offset = RAS_HDR_START;
	control->ras_record_offset = RAS_RECORD_START;
	control->ras_max_record_count  = RAS_MAX_RECORD_COUNT;
	mutex_init(&control->ras_tbl_mutex);

	/* Read the table header from EEPROM address */
	res = amdgpu_eeprom_read(&adev->pm.smu_i2c,
				 control->i2c_address + control->ras_header_offset,
				 buf, RAS_TABLE_HEADER_SIZE);
	if (res < RAS_TABLE_HEADER_SIZE) {
		DRM_ERROR("Failed to read EEPROM table header, res:%d", res);
		return res >= 0 ? -EIO : res;
	}

	__decode_table_header_from_buf(hdr, buf);

	control->ras_num_recs = RAS_NUM_RECS(hdr);
	control->ras_fri = RAS_OFFSET_TO_INDEX(control, hdr->first_rec_offset);

	if (hdr->header == RAS_TABLE_HDR_VAL) {
		DRM_DEBUG_DRIVER("Found existing EEPROM table with %d records",
				 control->ras_num_recs);
		res = __verify_ras_table_checksum(control);
		if (res)
			DRM_ERROR("RAS table incorrect checksum or error:%d\n",
				  res);
	} else if (hdr->header == RAS_TABLE_HDR_BAD &&
		   amdgpu_bad_page_threshold != 0) {
		res = __verify_ras_table_checksum(control);
		if (res)
			DRM_ERROR("RAS Table incorrect checksum or error:%d\n",
				  res);
		if (ras->bad_page_cnt_threshold > control->ras_num_recs) {
			/* This means that, the threshold was increased since
			 * the last time the system was booted, and now,
			 * ras->bad_page_cnt_threshold - control->num_recs > 0,
			 * so that at least one more record can be saved,
			 * before the page count threshold is reached.
			 */
			dev_info(adev->dev,
				 "records:%d threshold:%d, resetting "
				 "RAS table header signature",
				 control->ras_num_recs,
				 ras->bad_page_cnt_threshold);
			res = amdgpu_ras_eeprom_correct_header_tag(control,
								   RAS_TABLE_HDR_VAL);
		} else {
			*exceed_err_limit = true;
			dev_err(adev->dev,
				"RAS records:%d exceed threshold:%d, "
				"maybe retire this GPU?",
				control->ras_num_recs, ras->bad_page_cnt_threshold);
		}
	} else {
		DRM_INFO("Creating a new EEPROM table");

		res = amdgpu_ras_eeprom_reset_table(control);
	}

	return res < 0 ? res : 0;
}