1 // SPDX-License-Identifier: GPL-2.0+
2 /* Copyright (c) 2016-2017 Hisilicon Limited. */
3
4 #include <linux/sched/clock.h>
5
6 #include "hclge_err.h"
7
8 static const struct hclge_hw_error hclge_imp_tcm_ecc_int[] = {
9 {
10 .int_msk = BIT(1),
11 .msg = "imp_itcm0_ecc_mbit_err",
12 .reset_level = HNAE3_NONE_RESET
13 }, {
14 .int_msk = BIT(3),
15 .msg = "imp_itcm1_ecc_mbit_err",
16 .reset_level = HNAE3_NONE_RESET
17 }, {
18 .int_msk = BIT(5),
19 .msg = "imp_itcm2_ecc_mbit_err",
20 .reset_level = HNAE3_NONE_RESET
21 }, {
22 .int_msk = BIT(7),
23 .msg = "imp_itcm3_ecc_mbit_err",
24 .reset_level = HNAE3_NONE_RESET
25 }, {
26 .int_msk = BIT(9),
27 .msg = "imp_dtcm0_mem0_ecc_mbit_err",
28 .reset_level = HNAE3_NONE_RESET
29 }, {
30 .int_msk = BIT(11),
31 .msg = "imp_dtcm0_mem1_ecc_mbit_err",
32 .reset_level = HNAE3_NONE_RESET
33 }, {
34 .int_msk = BIT(13),
35 .msg = "imp_dtcm1_mem0_ecc_mbit_err",
36 .reset_level = HNAE3_NONE_RESET
37 }, {
38 .int_msk = BIT(15),
39 .msg = "imp_dtcm1_mem1_ecc_mbit_err",
40 .reset_level = HNAE3_NONE_RESET
41 }, {
42 .int_msk = BIT(17),
43 .msg = "imp_itcm4_ecc_mbit_err",
44 .reset_level = HNAE3_NONE_RESET
45 }, {
46 /* sentinel */
47 }
48 };
49
50 static const struct hclge_hw_error hclge_cmdq_nic_mem_ecc_int[] = {
51 {
52 .int_msk = BIT(1),
53 .msg = "cmdq_nic_rx_depth_ecc_mbit_err",
54 .reset_level = HNAE3_NONE_RESET
55 }, {
56 .int_msk = BIT(3),
57 .msg = "cmdq_nic_tx_depth_ecc_mbit_err",
58 .reset_level = HNAE3_NONE_RESET
59 }, {
60 .int_msk = BIT(5),
61 .msg = "cmdq_nic_rx_tail_ecc_mbit_err",
62 .reset_level = HNAE3_NONE_RESET
63 }, {
64 .int_msk = BIT(7),
65 .msg = "cmdq_nic_tx_tail_ecc_mbit_err",
66 .reset_level = HNAE3_NONE_RESET
67 }, {
68 .int_msk = BIT(9),
69 .msg = "cmdq_nic_rx_head_ecc_mbit_err",
70 .reset_level = HNAE3_NONE_RESET
71 }, {
72 .int_msk = BIT(11),
73 .msg = "cmdq_nic_tx_head_ecc_mbit_err",
74 .reset_level = HNAE3_NONE_RESET
75 }, {
76 .int_msk = BIT(13),
77 .msg = "cmdq_nic_rx_addr_ecc_mbit_err",
78 .reset_level = HNAE3_NONE_RESET
79 }, {
80 .int_msk = BIT(15),
81 .msg = "cmdq_nic_tx_addr_ecc_mbit_err",
82 .reset_level = HNAE3_NONE_RESET
83 }, {
84 .int_msk = BIT(17),
85 .msg = "cmdq_rocee_rx_depth_ecc_mbit_err",
86 .reset_level = HNAE3_NONE_RESET
87 }, {
88 .int_msk = BIT(19),
89 .msg = "cmdq_rocee_tx_depth_ecc_mbit_err",
90 .reset_level = HNAE3_NONE_RESET
91 }, {
92 .int_msk = BIT(21),
93 .msg = "cmdq_rocee_rx_tail_ecc_mbit_err",
94 .reset_level = HNAE3_NONE_RESET
95 }, {
96 .int_msk = BIT(23),
97 .msg = "cmdq_rocee_tx_tail_ecc_mbit_err",
98 .reset_level = HNAE3_NONE_RESET
99 }, {
100 .int_msk = BIT(25),
101 .msg = "cmdq_rocee_rx_head_ecc_mbit_err",
102 .reset_level = HNAE3_NONE_RESET
103 }, {
104 .int_msk = BIT(27),
105 .msg = "cmdq_rocee_tx_head_ecc_mbit_err",
106 .reset_level = HNAE3_NONE_RESET
107 }, {
108 .int_msk = BIT(29),
109 .msg = "cmdq_rocee_rx_addr_ecc_mbit_err",
110 .reset_level = HNAE3_NONE_RESET
111 }, {
112 .int_msk = BIT(31),
113 .msg = "cmdq_rocee_tx_addr_ecc_mbit_err",
114 .reset_level = HNAE3_NONE_RESET
115 }, {
116 /* sentinel */
117 }
118 };
119
120 static const struct hclge_hw_error hclge_tqp_int_ecc_int[] = {
121 {
122 .int_msk = BIT(6),
123 .msg = "tqp_int_cfg_even_ecc_mbit_err",
124 .reset_level = HNAE3_NONE_RESET
125 }, {
126 .int_msk = BIT(7),
127 .msg = "tqp_int_cfg_odd_ecc_mbit_err",
128 .reset_level = HNAE3_NONE_RESET
129 }, {
130 .int_msk = BIT(8),
131 .msg = "tqp_int_ctrl_even_ecc_mbit_err",
132 .reset_level = HNAE3_NONE_RESET
133 }, {
134 .int_msk = BIT(9),
135 .msg = "tqp_int_ctrl_odd_ecc_mbit_err",
136 .reset_level = HNAE3_NONE_RESET
137 }, {
138 .int_msk = BIT(10),
139 .msg = "tx_que_scan_int_ecc_mbit_err",
140 .reset_level = HNAE3_NONE_RESET
141 }, {
142 .int_msk = BIT(11),
143 .msg = "rx_que_scan_int_ecc_mbit_err",
144 .reset_level = HNAE3_NONE_RESET
145 }, {
146 /* sentinel */
147 }
148 };
149
150 static const struct hclge_hw_error hclge_msix_sram_ecc_int[] = {
151 {
152 .int_msk = BIT(1),
153 .msg = "msix_nic_ecc_mbit_err",
154 .reset_level = HNAE3_NONE_RESET
155 }, {
156 .int_msk = BIT(3),
157 .msg = "msix_rocee_ecc_mbit_err",
158 .reset_level = HNAE3_NONE_RESET
159 }, {
160 /* sentinel */
161 }
162 };
163
164 static const struct hclge_hw_error hclge_igu_int[] = {
165 {
166 .int_msk = BIT(0),
167 .msg = "igu_rx_buf0_ecc_mbit_err",
168 .reset_level = HNAE3_GLOBAL_RESET
169 }, {
170 .int_msk = BIT(2),
171 .msg = "igu_rx_buf1_ecc_mbit_err",
172 .reset_level = HNAE3_GLOBAL_RESET
173 }, {
174 /* sentinel */
175 }
176 };
177
178 static const struct hclge_hw_error hclge_igu_egu_tnl_int[] = {
179 {
180 .int_msk = BIT(0),
181 .msg = "rx_buf_overflow",
182 .reset_level = HNAE3_GLOBAL_RESET
183 }, {
184 .int_msk = BIT(1),
185 .msg = "rx_stp_fifo_overflow",
186 .reset_level = HNAE3_GLOBAL_RESET
187 }, {
188 .int_msk = BIT(2),
189 .msg = "rx_stp_fifo_underflow",
190 .reset_level = HNAE3_GLOBAL_RESET
191 }, {
192 .int_msk = BIT(3),
193 .msg = "tx_buf_overflow",
194 .reset_level = HNAE3_GLOBAL_RESET
195 }, {
196 .int_msk = BIT(4),
197 .msg = "tx_buf_underrun",
198 .reset_level = HNAE3_GLOBAL_RESET
199 }, {
200 .int_msk = BIT(5),
201 .msg = "rx_stp_buf_overflow",
202 .reset_level = HNAE3_GLOBAL_RESET
203 }, {
204 /* sentinel */
205 }
206 };
207
208 static const struct hclge_hw_error hclge_ncsi_err_int[] = {
209 {
210 .int_msk = BIT(1),
211 .msg = "ncsi_tx_ecc_mbit_err",
212 .reset_level = HNAE3_NONE_RESET
213 }, {
214 /* sentinel */
215 }
216 };
217
218 static const struct hclge_hw_error hclge_ppp_mpf_abnormal_int_st1[] = {
219 {
220 .int_msk = BIT(0),
221 .msg = "vf_vlan_ad_mem_ecc_mbit_err",
222 .reset_level = HNAE3_GLOBAL_RESET
223 }, {
224 .int_msk = BIT(1),
225 .msg = "umv_mcast_group_mem_ecc_mbit_err",
226 .reset_level = HNAE3_GLOBAL_RESET
227 }, {
228 .int_msk = BIT(2),
229 .msg = "umv_key_mem0_ecc_mbit_err",
230 .reset_level = HNAE3_GLOBAL_RESET
231 }, {
232 .int_msk = BIT(3),
233 .msg = "umv_key_mem1_ecc_mbit_err",
234 .reset_level = HNAE3_GLOBAL_RESET
235 }, {
236 .int_msk = BIT(4),
237 .msg = "umv_key_mem2_ecc_mbit_err",
238 .reset_level = HNAE3_GLOBAL_RESET
239 }, {
240 .int_msk = BIT(5),
241 .msg = "umv_key_mem3_ecc_mbit_err",
242 .reset_level = HNAE3_GLOBAL_RESET
243 }, {
244 .int_msk = BIT(6),
245 .msg = "umv_ad_mem_ecc_mbit_err",
246 .reset_level = HNAE3_GLOBAL_RESET
247 }, {
248 .int_msk = BIT(7),
249 .msg = "rss_tc_mode_mem_ecc_mbit_err",
250 .reset_level = HNAE3_GLOBAL_RESET
251 }, {
252 .int_msk = BIT(8),
253 .msg = "rss_idt_mem0_ecc_mbit_err",
254 .reset_level = HNAE3_GLOBAL_RESET
255 }, {
256 .int_msk = BIT(9),
257 .msg = "rss_idt_mem1_ecc_mbit_err",
258 .reset_level = HNAE3_GLOBAL_RESET
259 }, {
260 .int_msk = BIT(10),
261 .msg = "rss_idt_mem2_ecc_mbit_err",
262 .reset_level = HNAE3_GLOBAL_RESET
263 }, {
264 .int_msk = BIT(11),
265 .msg = "rss_idt_mem3_ecc_mbit_err",
266 .reset_level = HNAE3_GLOBAL_RESET
267 }, {
268 .int_msk = BIT(12),
269 .msg = "rss_idt_mem4_ecc_mbit_err",
270 .reset_level = HNAE3_GLOBAL_RESET
271 }, {
272 .int_msk = BIT(13),
273 .msg = "rss_idt_mem5_ecc_mbit_err",
274 .reset_level = HNAE3_GLOBAL_RESET
275 }, {
276 .int_msk = BIT(14),
277 .msg = "rss_idt_mem6_ecc_mbit_err",
278 .reset_level = HNAE3_GLOBAL_RESET
279 }, {
280 .int_msk = BIT(15),
281 .msg = "rss_idt_mem7_ecc_mbit_err",
282 .reset_level = HNAE3_GLOBAL_RESET
283 }, {
284 .int_msk = BIT(16),
285 .msg = "rss_idt_mem8_ecc_mbit_err",
286 .reset_level = HNAE3_GLOBAL_RESET
287 }, {
288 .int_msk = BIT(17),
289 .msg = "rss_idt_mem9_ecc_mbit_err",
290 .reset_level = HNAE3_GLOBAL_RESET
291 }, {
292 .int_msk = BIT(18),
293 .msg = "rss_idt_mem10_ecc_mbit_err",
294 .reset_level = HNAE3_GLOBAL_RESET
295 }, {
296 .int_msk = BIT(19),
297 .msg = "rss_idt_mem11_ecc_mbit_err",
298 .reset_level = HNAE3_GLOBAL_RESET
299 }, {
300 .int_msk = BIT(20),
301 .msg = "rss_idt_mem12_ecc_mbit_err",
302 .reset_level = HNAE3_GLOBAL_RESET
303 }, {
304 .int_msk = BIT(21),
305 .msg = "rss_idt_mem13_ecc_mbit_err",
306 .reset_level = HNAE3_GLOBAL_RESET
307 }, {
308 .int_msk = BIT(22),
309 .msg = "rss_idt_mem14_ecc_mbit_err",
310 .reset_level = HNAE3_GLOBAL_RESET
311 }, {
312 .int_msk = BIT(23),
313 .msg = "rss_idt_mem15_ecc_mbit_err",
314 .reset_level = HNAE3_GLOBAL_RESET
315 }, {
316 .int_msk = BIT(24),
317 .msg = "port_vlan_mem_ecc_mbit_err",
318 .reset_level = HNAE3_GLOBAL_RESET
319 }, {
320 .int_msk = BIT(25),
321 .msg = "mcast_linear_table_mem_ecc_mbit_err",
322 .reset_level = HNAE3_GLOBAL_RESET
323 }, {
324 .int_msk = BIT(26),
325 .msg = "mcast_result_mem_ecc_mbit_err",
326 .reset_level = HNAE3_GLOBAL_RESET
327 }, {
328 .int_msk = BIT(27),
329 .msg = "flow_director_ad_mem0_ecc_mbit_err",
330 .reset_level = HNAE3_GLOBAL_RESET
331 }, {
332 .int_msk = BIT(28),
333 .msg = "flow_director_ad_mem1_ecc_mbit_err",
334 .reset_level = HNAE3_GLOBAL_RESET
335 }, {
336 .int_msk = BIT(29),
337 .msg = "rx_vlan_tag_memory_ecc_mbit_err",
338 .reset_level = HNAE3_GLOBAL_RESET
339 }, {
340 .int_msk = BIT(30),
341 .msg = "Tx_UP_mapping_config_mem_ecc_mbit_err",
342 .reset_level = HNAE3_GLOBAL_RESET
343 }, {
344 /* sentinel */
345 }
346 };
347
348 static const struct hclge_hw_error hclge_ppp_pf_abnormal_int[] = {
349 {
350 .int_msk = BIT(0),
351 .msg = "tx_vlan_tag_err",
352 .reset_level = HNAE3_NONE_RESET
353 }, {
354 .int_msk = BIT(1),
355 .msg = "rss_list_tc_unassigned_queue_err",
356 .reset_level = HNAE3_NONE_RESET
357 }, {
358 /* sentinel */
359 }
360 };
361
362 static const struct hclge_hw_error hclge_ppp_mpf_abnormal_int_st3[] = {
363 {
364 .int_msk = BIT(0),
365 .msg = "hfs_fifo_mem_ecc_mbit_err",
366 .reset_level = HNAE3_GLOBAL_RESET
367 }, {
368 .int_msk = BIT(1),
369 .msg = "rslt_descr_fifo_mem_ecc_mbit_err",
370 .reset_level = HNAE3_GLOBAL_RESET
371 }, {
372 .int_msk = BIT(2),
373 .msg = "tx_vlan_tag_mem_ecc_mbit_err",
374 .reset_level = HNAE3_GLOBAL_RESET
375 }, {
376 .int_msk = BIT(3),
377 .msg = "FD_CN0_memory_ecc_mbit_err",
378 .reset_level = HNAE3_GLOBAL_RESET
379 }, {
380 .int_msk = BIT(4),
381 .msg = "FD_CN1_memory_ecc_mbit_err",
382 .reset_level = HNAE3_GLOBAL_RESET
383 }, {
384 .int_msk = BIT(5),
385 .msg = "GRO_AD_memory_ecc_mbit_err",
386 .reset_level = HNAE3_GLOBAL_RESET
387 }, {
388 /* sentinel */
389 }
390 };
391
392 static const struct hclge_hw_error hclge_tm_sch_rint[] = {
393 {
394 .int_msk = BIT(1),
395 .msg = "tm_sch_ecc_mbit_err",
396 .reset_level = HNAE3_GLOBAL_RESET
397 }, {
398 .int_msk = BIT(2),
399 .msg = "tm_sch_port_shap_sub_fifo_wr_err",
400 .reset_level = HNAE3_GLOBAL_RESET
401 }, {
402 .int_msk = BIT(3),
403 .msg = "tm_sch_port_shap_sub_fifo_rd_err",
404 .reset_level = HNAE3_GLOBAL_RESET
405 }, {
406 .int_msk = BIT(4),
407 .msg = "tm_sch_pg_pshap_sub_fifo_wr_err",
408 .reset_level = HNAE3_GLOBAL_RESET
409 }, {
410 .int_msk = BIT(5),
411 .msg = "tm_sch_pg_pshap_sub_fifo_rd_err",
412 .reset_level = HNAE3_GLOBAL_RESET
413 }, {
414 .int_msk = BIT(6),
415 .msg = "tm_sch_pg_cshap_sub_fifo_wr_err",
416 .reset_level = HNAE3_GLOBAL_RESET
417 }, {
418 .int_msk = BIT(7),
419 .msg = "tm_sch_pg_cshap_sub_fifo_rd_err",
420 .reset_level = HNAE3_GLOBAL_RESET
421 }, {
422 .int_msk = BIT(8),
423 .msg = "tm_sch_pri_pshap_sub_fifo_wr_err",
424 .reset_level = HNAE3_GLOBAL_RESET
425 }, {
426 .int_msk = BIT(9),
427 .msg = "tm_sch_pri_pshap_sub_fifo_rd_err",
428 .reset_level = HNAE3_GLOBAL_RESET
429 }, {
430 .int_msk = BIT(10),
431 .msg = "tm_sch_pri_cshap_sub_fifo_wr_err",
432 .reset_level = HNAE3_GLOBAL_RESET
433 }, {
434 .int_msk = BIT(11),
435 .msg = "tm_sch_pri_cshap_sub_fifo_rd_err",
436 .reset_level = HNAE3_GLOBAL_RESET
437 }, {
438 .int_msk = BIT(12),
439 .msg = "tm_sch_port_shap_offset_fifo_wr_err",
440 .reset_level = HNAE3_GLOBAL_RESET
441 }, {
442 .int_msk = BIT(13),
443 .msg = "tm_sch_port_shap_offset_fifo_rd_err",
444 .reset_level = HNAE3_GLOBAL_RESET
445 }, {
446 .int_msk = BIT(14),
447 .msg = "tm_sch_pg_pshap_offset_fifo_wr_err",
448 .reset_level = HNAE3_GLOBAL_RESET
449 }, {
450 .int_msk = BIT(15),
451 .msg = "tm_sch_pg_pshap_offset_fifo_rd_err",
452 .reset_level = HNAE3_GLOBAL_RESET
453 }, {
454 .int_msk = BIT(16),
455 .msg = "tm_sch_pg_cshap_offset_fifo_wr_err",
456 .reset_level = HNAE3_GLOBAL_RESET
457 }, {
458 .int_msk = BIT(17),
459 .msg = "tm_sch_pg_cshap_offset_fifo_rd_err",
460 .reset_level = HNAE3_GLOBAL_RESET
461 }, {
462 .int_msk = BIT(18),
463 .msg = "tm_sch_pri_pshap_offset_fifo_wr_err",
464 .reset_level = HNAE3_GLOBAL_RESET
465 }, {
466 .int_msk = BIT(19),
467 .msg = "tm_sch_pri_pshap_offset_fifo_rd_err",
468 .reset_level = HNAE3_GLOBAL_RESET
469 }, {
470 .int_msk = BIT(20),
471 .msg = "tm_sch_pri_cshap_offset_fifo_wr_err",
472 .reset_level = HNAE3_GLOBAL_RESET
473 }, {
474 .int_msk = BIT(21),
475 .msg = "tm_sch_pri_cshap_offset_fifo_rd_err",
476 .reset_level = HNAE3_GLOBAL_RESET
477 }, {
478 .int_msk = BIT(22),
479 .msg = "tm_sch_rq_fifo_wr_err",
480 .reset_level = HNAE3_GLOBAL_RESET
481 }, {
482 .int_msk = BIT(23),
483 .msg = "tm_sch_rq_fifo_rd_err",
484 .reset_level = HNAE3_GLOBAL_RESET
485 }, {
486 .int_msk = BIT(24),
487 .msg = "tm_sch_nq_fifo_wr_err",
488 .reset_level = HNAE3_GLOBAL_RESET
489 }, {
490 .int_msk = BIT(25),
491 .msg = "tm_sch_nq_fifo_rd_err",
492 .reset_level = HNAE3_GLOBAL_RESET
493 }, {
494 .int_msk = BIT(26),
495 .msg = "tm_sch_roce_up_fifo_wr_err",
496 .reset_level = HNAE3_GLOBAL_RESET
497 }, {
498 .int_msk = BIT(27),
499 .msg = "tm_sch_roce_up_fifo_rd_err",
500 .reset_level = HNAE3_GLOBAL_RESET
501 }, {
502 .int_msk = BIT(28),
503 .msg = "tm_sch_rcb_byte_fifo_wr_err",
504 .reset_level = HNAE3_GLOBAL_RESET
505 }, {
506 .int_msk = BIT(29),
507 .msg = "tm_sch_rcb_byte_fifo_rd_err",
508 .reset_level = HNAE3_GLOBAL_RESET
509 }, {
510 .int_msk = BIT(30),
511 .msg = "tm_sch_ssu_byte_fifo_wr_err",
512 .reset_level = HNAE3_GLOBAL_RESET
513 }, {
514 .int_msk = BIT(31),
515 .msg = "tm_sch_ssu_byte_fifo_rd_err",
516 .reset_level = HNAE3_GLOBAL_RESET
517 }, {
518 /* sentinel */
519 }
520 };
521
522 static const struct hclge_hw_error hclge_qcn_fifo_rint[] = {
523 {
524 .int_msk = BIT(0),
525 .msg = "qcn_shap_gp0_sch_fifo_rd_err",
526 .reset_level = HNAE3_GLOBAL_RESET
527 }, {
528 .int_msk = BIT(1),
529 .msg = "qcn_shap_gp0_sch_fifo_wr_err",
530 .reset_level = HNAE3_GLOBAL_RESET
531 }, {
532 .int_msk = BIT(2),
533 .msg = "qcn_shap_gp1_sch_fifo_rd_err",
534 .reset_level = HNAE3_GLOBAL_RESET
535 }, {
536 .int_msk = BIT(3),
537 .msg = "qcn_shap_gp1_sch_fifo_wr_err",
538 .reset_level = HNAE3_GLOBAL_RESET
539 }, {
540 .int_msk = BIT(4),
541 .msg = "qcn_shap_gp2_sch_fifo_rd_err",
542 .reset_level = HNAE3_GLOBAL_RESET
543 }, {
544 .int_msk = BIT(5),
545 .msg = "qcn_shap_gp2_sch_fifo_wr_err",
546 .reset_level = HNAE3_GLOBAL_RESET
547 }, {
548 .int_msk = BIT(6),
549 .msg = "qcn_shap_gp3_sch_fifo_rd_err",
550 .reset_level = HNAE3_GLOBAL_RESET
551 }, {
552 .int_msk = BIT(7),
553 .msg = "qcn_shap_gp3_sch_fifo_wr_err",
554 .reset_level = HNAE3_GLOBAL_RESET
555 }, {
556 .int_msk = BIT(8),
557 .msg = "qcn_shap_gp0_offset_fifo_rd_err",
558 .reset_level = HNAE3_GLOBAL_RESET
559 }, {
560 .int_msk = BIT(9),
561 .msg = "qcn_shap_gp0_offset_fifo_wr_err",
562 .reset_level = HNAE3_GLOBAL_RESET
563 }, {
564 .int_msk = BIT(10),
565 .msg = "qcn_shap_gp1_offset_fifo_rd_err",
566 .reset_level = HNAE3_GLOBAL_RESET
567 }, {
568 .int_msk = BIT(11),
569 .msg = "qcn_shap_gp1_offset_fifo_wr_err",
570 .reset_level = HNAE3_GLOBAL_RESET
571 }, {
572 .int_msk = BIT(12),
573 .msg = "qcn_shap_gp2_offset_fifo_rd_err",
574 .reset_level = HNAE3_GLOBAL_RESET
575 }, {
576 .int_msk = BIT(13),
577 .msg = "qcn_shap_gp2_offset_fifo_wr_err",
578 .reset_level = HNAE3_GLOBAL_RESET
579 }, {
580 .int_msk = BIT(14),
581 .msg = "qcn_shap_gp3_offset_fifo_rd_err",
582 .reset_level = HNAE3_GLOBAL_RESET
583 }, {
584 .int_msk = BIT(15),
585 .msg = "qcn_shap_gp3_offset_fifo_wr_err",
586 .reset_level = HNAE3_GLOBAL_RESET
587 }, {
588 .int_msk = BIT(16),
589 .msg = "qcn_byte_info_fifo_rd_err",
590 .reset_level = HNAE3_GLOBAL_RESET
591 }, {
592 .int_msk = BIT(17),
593 .msg = "qcn_byte_info_fifo_wr_err",
594 .reset_level = HNAE3_GLOBAL_RESET
595 }, {
596 /* sentinel */
597 }
598 };
599
600 static const struct hclge_hw_error hclge_qcn_ecc_rint[] = {
601 {
602 .int_msk = BIT(1),
603 .msg = "qcn_byte_mem_ecc_mbit_err",
604 .reset_level = HNAE3_GLOBAL_RESET
605 }, {
606 .int_msk = BIT(3),
607 .msg = "qcn_time_mem_ecc_mbit_err",
608 .reset_level = HNAE3_GLOBAL_RESET
609 }, {
610 .int_msk = BIT(5),
611 .msg = "qcn_fb_mem_ecc_mbit_err",
612 .reset_level = HNAE3_GLOBAL_RESET
613 }, {
614 .int_msk = BIT(7),
615 .msg = "qcn_link_mem_ecc_mbit_err",
616 .reset_level = HNAE3_GLOBAL_RESET
617 }, {
618 .int_msk = BIT(9),
619 .msg = "qcn_rate_mem_ecc_mbit_err",
620 .reset_level = HNAE3_GLOBAL_RESET
621 }, {
622 .int_msk = BIT(11),
623 .msg = "qcn_tmplt_mem_ecc_mbit_err",
624 .reset_level = HNAE3_GLOBAL_RESET
625 }, {
626 .int_msk = BIT(13),
627 .msg = "qcn_shap_cfg_mem_ecc_mbit_err",
628 .reset_level = HNAE3_GLOBAL_RESET
629 }, {
630 .int_msk = BIT(15),
631 .msg = "qcn_gp0_barrel_mem_ecc_mbit_err",
632 .reset_level = HNAE3_GLOBAL_RESET
633 }, {
634 .int_msk = BIT(17),
635 .msg = "qcn_gp1_barrel_mem_ecc_mbit_err",
636 .reset_level = HNAE3_GLOBAL_RESET
637 }, {
638 .int_msk = BIT(19),
639 .msg = "qcn_gp2_barrel_mem_ecc_mbit_err",
640 .reset_level = HNAE3_GLOBAL_RESET
641 }, {
642 .int_msk = BIT(21),
643 .msg = "qcn_gp3_barral_mem_ecc_mbit_err",
644 .reset_level = HNAE3_GLOBAL_RESET
645 }, {
646 /* sentinel */
647 }
648 };
649
650 static const struct hclge_hw_error hclge_mac_afifo_tnl_int[] = {
651 {
652 .int_msk = BIT(0),
653 .msg = "egu_cge_afifo_ecc_1bit_err",
654 .reset_level = HNAE3_NONE_RESET
655 }, {
656 .int_msk = BIT(1),
657 .msg = "egu_cge_afifo_ecc_mbit_err",
658 .reset_level = HNAE3_GLOBAL_RESET
659 }, {
660 .int_msk = BIT(2),
661 .msg = "egu_lge_afifo_ecc_1bit_err",
662 .reset_level = HNAE3_NONE_RESET
663 }, {
664 .int_msk = BIT(3),
665 .msg = "egu_lge_afifo_ecc_mbit_err",
666 .reset_level = HNAE3_GLOBAL_RESET
667 }, {
668 .int_msk = BIT(4),
669 .msg = "cge_igu_afifo_ecc_1bit_err",
670 .reset_level = HNAE3_NONE_RESET
671 }, {
672 .int_msk = BIT(5),
673 .msg = "cge_igu_afifo_ecc_mbit_err",
674 .reset_level = HNAE3_GLOBAL_RESET
675 }, {
676 .int_msk = BIT(6),
677 .msg = "lge_igu_afifo_ecc_1bit_err",
678 .reset_level = HNAE3_NONE_RESET
679 }, {
680 .int_msk = BIT(7),
681 .msg = "lge_igu_afifo_ecc_mbit_err",
682 .reset_level = HNAE3_GLOBAL_RESET
683 }, {
684 .int_msk = BIT(8),
685 .msg = "cge_igu_afifo_overflow_err",
686 .reset_level = HNAE3_GLOBAL_RESET
687 }, {
688 .int_msk = BIT(9),
689 .msg = "lge_igu_afifo_overflow_err",
690 .reset_level = HNAE3_GLOBAL_RESET
691 }, {
692 .int_msk = BIT(10),
693 .msg = "egu_cge_afifo_underrun_err",
694 .reset_level = HNAE3_GLOBAL_RESET
695 }, {
696 .int_msk = BIT(11),
697 .msg = "egu_lge_afifo_underrun_err",
698 .reset_level = HNAE3_GLOBAL_RESET
699 }, {
700 .int_msk = BIT(12),
701 .msg = "egu_ge_afifo_underrun_err",
702 .reset_level = HNAE3_GLOBAL_RESET
703 }, {
704 .int_msk = BIT(13),
705 .msg = "ge_igu_afifo_overflow_err",
706 .reset_level = HNAE3_GLOBAL_RESET
707 }, {
708 /* sentinel */
709 }
710 };
711
712 static const struct hclge_hw_error hclge_ppu_mpf_abnormal_int_st2[] = {
713 {
714 .int_msk = BIT(13),
715 .msg = "rpu_rx_pkt_bit32_ecc_mbit_err",
716 .reset_level = HNAE3_GLOBAL_RESET
717 }, {
718 .int_msk = BIT(14),
719 .msg = "rpu_rx_pkt_bit33_ecc_mbit_err",
720 .reset_level = HNAE3_GLOBAL_RESET
721 }, {
722 .int_msk = BIT(15),
723 .msg = "rpu_rx_pkt_bit34_ecc_mbit_err",
724 .reset_level = HNAE3_GLOBAL_RESET
725 }, {
726 .int_msk = BIT(16),
727 .msg = "rpu_rx_pkt_bit35_ecc_mbit_err",
728 .reset_level = HNAE3_GLOBAL_RESET
729 }, {
730 .int_msk = BIT(17),
731 .msg = "rcb_tx_ring_ecc_mbit_err",
732 .reset_level = HNAE3_GLOBAL_RESET
733 }, {
734 .int_msk = BIT(18),
735 .msg = "rcb_rx_ring_ecc_mbit_err",
736 .reset_level = HNAE3_GLOBAL_RESET
737 }, {
738 .int_msk = BIT(19),
739 .msg = "rcb_tx_fbd_ecc_mbit_err",
740 .reset_level = HNAE3_GLOBAL_RESET
741 }, {
742 .int_msk = BIT(20),
743 .msg = "rcb_rx_ebd_ecc_mbit_err",
744 .reset_level = HNAE3_GLOBAL_RESET
745 }, {
746 .int_msk = BIT(21),
747 .msg = "rcb_tso_info_ecc_mbit_err",
748 .reset_level = HNAE3_GLOBAL_RESET
749 }, {
750 .int_msk = BIT(22),
751 .msg = "rcb_tx_int_info_ecc_mbit_err",
752 .reset_level = HNAE3_GLOBAL_RESET
753 }, {
754 .int_msk = BIT(23),
755 .msg = "rcb_rx_int_info_ecc_mbit_err",
756 .reset_level = HNAE3_GLOBAL_RESET
757 }, {
758 .int_msk = BIT(24),
759 .msg = "tpu_tx_pkt_0_ecc_mbit_err",
760 .reset_level = HNAE3_GLOBAL_RESET
761 }, {
762 .int_msk = BIT(25),
763 .msg = "tpu_tx_pkt_1_ecc_mbit_err",
764 .reset_level = HNAE3_GLOBAL_RESET
765 }, {
766 .int_msk = BIT(26),
767 .msg = "rd_bus_err",
768 .reset_level = HNAE3_GLOBAL_RESET
769 }, {
770 .int_msk = BIT(27),
771 .msg = "wr_bus_err",
772 .reset_level = HNAE3_GLOBAL_RESET
773 }, {
774 .int_msk = BIT(28),
775 .msg = "reg_search_miss",
776 .reset_level = HNAE3_GLOBAL_RESET
777 }, {
778 .int_msk = BIT(29),
779 .msg = "rx_q_search_miss",
780 .reset_level = HNAE3_NONE_RESET
781 }, {
782 .int_msk = BIT(30),
783 .msg = "ooo_ecc_err_detect",
784 .reset_level = HNAE3_NONE_RESET
785 }, {
786 .int_msk = BIT(31),
787 .msg = "ooo_ecc_err_multpl",
788 .reset_level = HNAE3_GLOBAL_RESET
789 }, {
790 /* sentinel */
791 }
792 };
793
794 static const struct hclge_hw_error hclge_ppu_mpf_abnormal_int_st3[] = {
795 {
796 .int_msk = BIT(4),
797 .msg = "gro_bd_ecc_mbit_err",
798 .reset_level = HNAE3_GLOBAL_RESET
799 }, {
800 .int_msk = BIT(5),
801 .msg = "gro_context_ecc_mbit_err",
802 .reset_level = HNAE3_GLOBAL_RESET
803 }, {
804 .int_msk = BIT(6),
805 .msg = "rx_stash_cfg_ecc_mbit_err",
806 .reset_level = HNAE3_GLOBAL_RESET
807 }, {
808 .int_msk = BIT(7),
809 .msg = "axi_rd_fbd_ecc_mbit_err",
810 .reset_level = HNAE3_GLOBAL_RESET
811 }, {
812 /* sentinel */
813 }
814 };
815
816 static const struct hclge_hw_error hclge_ppu_pf_abnormal_int[] = {
817 {
818 .int_msk = BIT(0),
819 .msg = "over_8bd_no_fe",
820 .reset_level = HNAE3_FUNC_RESET
821 }, {
822 .int_msk = BIT(1),
823 .msg = "tso_mss_cmp_min_err",
824 .reset_level = HNAE3_NONE_RESET
825 }, {
826 .int_msk = BIT(2),
827 .msg = "tso_mss_cmp_max_err",
828 .reset_level = HNAE3_NONE_RESET
829 }, {
830 .int_msk = BIT(3),
831 .msg = "tx_rd_fbd_poison",
832 .reset_level = HNAE3_FUNC_RESET
833 }, {
834 .int_msk = BIT(4),
835 .msg = "rx_rd_ebd_poison",
836 .reset_level = HNAE3_FUNC_RESET
837 }, {
838 .int_msk = BIT(5),
839 .msg = "buf_wait_timeout",
840 .reset_level = HNAE3_NONE_RESET
841 }, {
842 /* sentinel */
843 }
844 };
845
846 static const struct hclge_hw_error hclge_ssu_com_err_int[] = {
847 {
848 .int_msk = BIT(0),
849 .msg = "buf_sum_err",
850 .reset_level = HNAE3_NONE_RESET
851 }, {
852 .int_msk = BIT(1),
853 .msg = "ppp_mb_num_err",
854 .reset_level = HNAE3_NONE_RESET
855 }, {
856 .int_msk = BIT(2),
857 .msg = "ppp_mbid_err",
858 .reset_level = HNAE3_GLOBAL_RESET
859 }, {
860 .int_msk = BIT(3),
861 .msg = "ppp_rlt_mac_err",
862 .reset_level = HNAE3_GLOBAL_RESET
863 }, {
864 .int_msk = BIT(4),
865 .msg = "ppp_rlt_host_err",
866 .reset_level = HNAE3_GLOBAL_RESET
867 }, {
868 .int_msk = BIT(5),
869 .msg = "cks_edit_position_err",
870 .reset_level = HNAE3_GLOBAL_RESET
871 }, {
872 .int_msk = BIT(6),
873 .msg = "cks_edit_condition_err",
874 .reset_level = HNAE3_GLOBAL_RESET
875 }, {
876 .int_msk = BIT(7),
877 .msg = "vlan_edit_condition_err",
878 .reset_level = HNAE3_GLOBAL_RESET
879 }, {
880 .int_msk = BIT(8),
881 .msg = "vlan_num_ot_err",
882 .reset_level = HNAE3_GLOBAL_RESET
883 }, {
884 .int_msk = BIT(9),
885 .msg = "vlan_num_in_err",
886 .reset_level = HNAE3_GLOBAL_RESET
887 }, {
888 /* sentinel */
889 }
890 };
891
892 #define HCLGE_SSU_MEM_ECC_ERR(x) \
893 { \
894 .int_msk = BIT(x), \
895 .msg = "ssu_mem" #x "_ecc_mbit_err", \
896 .reset_level = HNAE3_GLOBAL_RESET \
897 }
898
899 static const struct hclge_hw_error hclge_ssu_mem_ecc_err_int[] = {
900 HCLGE_SSU_MEM_ECC_ERR(0),
901 HCLGE_SSU_MEM_ECC_ERR(1),
902 HCLGE_SSU_MEM_ECC_ERR(2),
903 HCLGE_SSU_MEM_ECC_ERR(3),
904 HCLGE_SSU_MEM_ECC_ERR(4),
905 HCLGE_SSU_MEM_ECC_ERR(5),
906 HCLGE_SSU_MEM_ECC_ERR(6),
907 HCLGE_SSU_MEM_ECC_ERR(7),
908 HCLGE_SSU_MEM_ECC_ERR(8),
909 HCLGE_SSU_MEM_ECC_ERR(9),
910 HCLGE_SSU_MEM_ECC_ERR(10),
911 HCLGE_SSU_MEM_ECC_ERR(11),
912 HCLGE_SSU_MEM_ECC_ERR(12),
913 HCLGE_SSU_MEM_ECC_ERR(13),
914 HCLGE_SSU_MEM_ECC_ERR(14),
915 HCLGE_SSU_MEM_ECC_ERR(15),
916 HCLGE_SSU_MEM_ECC_ERR(16),
917 HCLGE_SSU_MEM_ECC_ERR(17),
918 HCLGE_SSU_MEM_ECC_ERR(18),
919 HCLGE_SSU_MEM_ECC_ERR(19),
920 HCLGE_SSU_MEM_ECC_ERR(20),
921 HCLGE_SSU_MEM_ECC_ERR(21),
922 HCLGE_SSU_MEM_ECC_ERR(22),
923 HCLGE_SSU_MEM_ECC_ERR(23),
924 HCLGE_SSU_MEM_ECC_ERR(24),
925 HCLGE_SSU_MEM_ECC_ERR(25),
926 HCLGE_SSU_MEM_ECC_ERR(26),
927 HCLGE_SSU_MEM_ECC_ERR(27),
928 HCLGE_SSU_MEM_ECC_ERR(28),
929 HCLGE_SSU_MEM_ECC_ERR(29),
930 HCLGE_SSU_MEM_ECC_ERR(30),
931 HCLGE_SSU_MEM_ECC_ERR(31),
932 { /* sentinel */ }
933 };
934
935 static const struct hclge_hw_error hclge_ssu_port_based_err_int[] = {
936 {
937 .int_msk = BIT(0),
938 .msg = "roc_pkt_without_key_port",
939 .reset_level = HNAE3_FUNC_RESET
940 }, {
941 .int_msk = BIT(1),
942 .msg = "tpu_pkt_without_key_port",
943 .reset_level = HNAE3_GLOBAL_RESET
944 }, {
945 .int_msk = BIT(2),
946 .msg = "igu_pkt_without_key_port",
947 .reset_level = HNAE3_GLOBAL_RESET
948 }, {
949 .int_msk = BIT(3),
950 .msg = "roc_eof_mis_match_port",
951 .reset_level = HNAE3_GLOBAL_RESET
952 }, {
953 .int_msk = BIT(4),
954 .msg = "tpu_eof_mis_match_port",
955 .reset_level = HNAE3_GLOBAL_RESET
956 }, {
957 .int_msk = BIT(5),
958 .msg = "igu_eof_mis_match_port",
959 .reset_level = HNAE3_GLOBAL_RESET
960 }, {
961 .int_msk = BIT(6),
962 .msg = "roc_sof_mis_match_port",
963 .reset_level = HNAE3_GLOBAL_RESET
964 }, {
965 .int_msk = BIT(7),
966 .msg = "tpu_sof_mis_match_port",
967 .reset_level = HNAE3_GLOBAL_RESET
968 }, {
969 .int_msk = BIT(8),
970 .msg = "igu_sof_mis_match_port",
971 .reset_level = HNAE3_GLOBAL_RESET
972 }, {
973 .int_msk = BIT(11),
974 .msg = "ets_rd_int_rx_port",
975 .reset_level = HNAE3_GLOBAL_RESET
976 }, {
977 .int_msk = BIT(12),
978 .msg = "ets_wr_int_rx_port",
979 .reset_level = HNAE3_GLOBAL_RESET
980 }, {
981 .int_msk = BIT(13),
982 .msg = "ets_rd_int_tx_port",
983 .reset_level = HNAE3_GLOBAL_RESET
984 }, {
985 .int_msk = BIT(14),
986 .msg = "ets_wr_int_tx_port",
987 .reset_level = HNAE3_GLOBAL_RESET
988 }, {
989 /* sentinel */
990 }
991 };
992
993 static const struct hclge_hw_error hclge_ssu_fifo_overflow_int[] = {
994 {
995 .int_msk = BIT(0),
996 .msg = "ig_mac_inf_int",
997 .reset_level = HNAE3_GLOBAL_RESET
998 }, {
999 .int_msk = BIT(1),
1000 .msg = "ig_host_inf_int",
1001 .reset_level = HNAE3_GLOBAL_RESET
1002 }, {
1003 .int_msk = BIT(2),
1004 .msg = "ig_roc_buf_int",
1005 .reset_level = HNAE3_GLOBAL_RESET
1006 }, {
1007 .int_msk = BIT(3),
1008 .msg = "ig_host_data_fifo_int",
1009 .reset_level = HNAE3_GLOBAL_RESET
1010 }, {
1011 .int_msk = BIT(4),
1012 .msg = "ig_host_key_fifo_int",
1013 .reset_level = HNAE3_GLOBAL_RESET
1014 }, {
1015 .int_msk = BIT(5),
1016 .msg = "tx_qcn_fifo_int",
1017 .reset_level = HNAE3_GLOBAL_RESET
1018 }, {
1019 .int_msk = BIT(6),
1020 .msg = "rx_qcn_fifo_int",
1021 .reset_level = HNAE3_GLOBAL_RESET
1022 }, {
1023 .int_msk = BIT(7),
1024 .msg = "tx_pf_rd_fifo_int",
1025 .reset_level = HNAE3_GLOBAL_RESET
1026 }, {
1027 .int_msk = BIT(8),
1028 .msg = "rx_pf_rd_fifo_int",
1029 .reset_level = HNAE3_GLOBAL_RESET
1030 }, {
1031 .int_msk = BIT(9),
1032 .msg = "qm_eof_fifo_int",
1033 .reset_level = HNAE3_GLOBAL_RESET
1034 }, {
1035 .int_msk = BIT(10),
1036 .msg = "mb_rlt_fifo_int",
1037 .reset_level = HNAE3_GLOBAL_RESET
1038 }, {
1039 .int_msk = BIT(11),
1040 .msg = "dup_uncopy_fifo_int",
1041 .reset_level = HNAE3_GLOBAL_RESET
1042 }, {
1043 .int_msk = BIT(12),
1044 .msg = "dup_cnt_rd_fifo_int",
1045 .reset_level = HNAE3_GLOBAL_RESET
1046 }, {
1047 .int_msk = BIT(13),
1048 .msg = "dup_cnt_drop_fifo_int",
1049 .reset_level = HNAE3_GLOBAL_RESET
1050 }, {
1051 .int_msk = BIT(14),
1052 .msg = "dup_cnt_wrb_fifo_int",
1053 .reset_level = HNAE3_GLOBAL_RESET
1054 }, {
1055 .int_msk = BIT(15),
1056 .msg = "host_cmd_fifo_int",
1057 .reset_level = HNAE3_GLOBAL_RESET
1058 }, {
1059 .int_msk = BIT(16),
1060 .msg = "mac_cmd_fifo_int",
1061 .reset_level = HNAE3_GLOBAL_RESET
1062 }, {
1063 .int_msk = BIT(17),
1064 .msg = "host_cmd_bitmap_empty_int",
1065 .reset_level = HNAE3_GLOBAL_RESET
1066 }, {
1067 .int_msk = BIT(18),
1068 .msg = "mac_cmd_bitmap_empty_int",
1069 .reset_level = HNAE3_GLOBAL_RESET
1070 }, {
1071 .int_msk = BIT(19),
1072 .msg = "dup_bitmap_empty_int",
1073 .reset_level = HNAE3_GLOBAL_RESET
1074 }, {
1075 .int_msk = BIT(20),
1076 .msg = "out_queue_bitmap_empty_int",
1077 .reset_level = HNAE3_GLOBAL_RESET
1078 }, {
1079 .int_msk = BIT(21),
1080 .msg = "bank2_bitmap_empty_int",
1081 .reset_level = HNAE3_GLOBAL_RESET
1082 }, {
1083 .int_msk = BIT(22),
1084 .msg = "bank1_bitmap_empty_int",
1085 .reset_level = HNAE3_GLOBAL_RESET
1086 }, {
1087 .int_msk = BIT(23),
1088 .msg = "bank0_bitmap_empty_int",
1089 .reset_level = HNAE3_GLOBAL_RESET
1090 }, {
1091 /* sentinel */
1092 }
1093 };
1094
1095 static const struct hclge_hw_error hclge_ssu_ets_tcg_int[] = {
1096 {
1097 .int_msk = BIT(0),
1098 .msg = "ets_rd_int_rx_tcg",
1099 .reset_level = HNAE3_GLOBAL_RESET
1100 }, {
1101 .int_msk = BIT(1),
1102 .msg = "ets_wr_int_rx_tcg",
1103 .reset_level = HNAE3_GLOBAL_RESET
1104 }, {
1105 .int_msk = BIT(2),
1106 .msg = "ets_rd_int_tx_tcg",
1107 .reset_level = HNAE3_GLOBAL_RESET
1108 }, {
1109 .int_msk = BIT(3),
1110 .msg = "ets_wr_int_tx_tcg",
1111 .reset_level = HNAE3_GLOBAL_RESET
1112 }, {
1113 /* sentinel */
1114 }
1115 };
1116
1117 static const struct hclge_hw_error hclge_ssu_port_based_pf_int[] = {
1118 {
1119 .int_msk = BIT(0),
1120 .msg = "roc_pkt_without_key_port",
1121 .reset_level = HNAE3_FUNC_RESET
1122 }, {
1123 .int_msk = BIT(9),
1124 .msg = "low_water_line_err_port",
1125 .reset_level = HNAE3_NONE_RESET
1126 }, {
1127 .int_msk = BIT(10),
1128 .msg = "hi_water_line_err_port",
1129 .reset_level = HNAE3_GLOBAL_RESET
1130 }, {
1131 /* sentinel */
1132 }
1133 };
1134
1135 static const struct hclge_hw_error hclge_rocee_qmm_ovf_err_int[] = {
1136 {
1137 .int_msk = 0,
1138 .msg = "rocee qmm ovf: sgid invalid err"
1139 }, {
1140 .int_msk = 0x4,
1141 .msg = "rocee qmm ovf: sgid ovf err"
1142 }, {
1143 .int_msk = 0x8,
1144 .msg = "rocee qmm ovf: smac invalid err"
1145 }, {
1146 .int_msk = 0xC,
1147 .msg = "rocee qmm ovf: smac ovf err"
1148 }, {
1149 .int_msk = 0x10,
1150 .msg = "rocee qmm ovf: cqc invalid err"
1151 }, {
1152 .int_msk = 0x11,
1153 .msg = "rocee qmm ovf: cqc ovf err"
1154 }, {
1155 .int_msk = 0x12,
1156 .msg = "rocee qmm ovf: cqc hopnum err"
1157 }, {
1158 .int_msk = 0x13,
1159 .msg = "rocee qmm ovf: cqc ba0 err"
1160 }, {
1161 .int_msk = 0x14,
1162 .msg = "rocee qmm ovf: srqc invalid err"
1163 }, {
1164 .int_msk = 0x15,
1165 .msg = "rocee qmm ovf: srqc ovf err"
1166 }, {
1167 .int_msk = 0x16,
1168 .msg = "rocee qmm ovf: srqc hopnum err"
1169 }, {
1170 .int_msk = 0x17,
1171 .msg = "rocee qmm ovf: srqc ba0 err"
1172 }, {
1173 .int_msk = 0x18,
1174 .msg = "rocee qmm ovf: mpt invalid err"
1175 }, {
1176 .int_msk = 0x19,
1177 .msg = "rocee qmm ovf: mpt ovf err"
1178 }, {
1179 .int_msk = 0x1A,
1180 .msg = "rocee qmm ovf: mpt hopnum err"
1181 }, {
1182 .int_msk = 0x1B,
1183 .msg = "rocee qmm ovf: mpt ba0 err"
1184 }, {
1185 .int_msk = 0x1C,
1186 .msg = "rocee qmm ovf: qpc invalid err"
1187 }, {
1188 .int_msk = 0x1D,
1189 .msg = "rocee qmm ovf: qpc ovf err"
1190 }, {
1191 .int_msk = 0x1E,
1192 .msg = "rocee qmm ovf: qpc hopnum err"
1193 }, {
1194 .int_msk = 0x1F,
1195 .msg = "rocee qmm ovf: qpc ba0 err"
1196 }, {
1197 /* sentinel */
1198 }
1199 };
1200
1201 static const struct hclge_hw_module_id hclge_hw_module_id_st[] = {
1202 {
1203 .module_id = MODULE_NONE,
1204 .msg = "MODULE_NONE"
1205 }, {
1206 .module_id = MODULE_BIOS_COMMON,
1207 .msg = "MODULE_BIOS_COMMON"
1208 }, {
1209 .module_id = MODULE_GE,
1210 .msg = "MODULE_GE"
1211 }, {
1212 .module_id = MODULE_IGU_EGU,
1213 .msg = "MODULE_IGU_EGU"
1214 }, {
1215 .module_id = MODULE_LGE,
1216 .msg = "MODULE_LGE"
1217 }, {
1218 .module_id = MODULE_NCSI,
1219 .msg = "MODULE_NCSI"
1220 }, {
1221 .module_id = MODULE_PPP,
1222 .msg = "MODULE_PPP"
1223 }, {
1224 .module_id = MODULE_QCN,
1225 .msg = "MODULE_QCN"
1226 }, {
1227 .module_id = MODULE_RCB_RX,
1228 .msg = "MODULE_RCB_RX"
1229 }, {
1230 .module_id = MODULE_RTC,
1231 .msg = "MODULE_RTC"
1232 }, {
1233 .module_id = MODULE_SSU,
1234 .msg = "MODULE_SSU"
1235 }, {
1236 .module_id = MODULE_TM,
1237 .msg = "MODULE_TM"
1238 }, {
1239 .module_id = MODULE_RCB_TX,
1240 .msg = "MODULE_RCB_TX"
1241 }, {
1242 .module_id = MODULE_TXDMA,
1243 .msg = "MODULE_TXDMA"
1244 }, {
1245 .module_id = MODULE_MASTER,
1246 .msg = "MODULE_MASTER"
1247 }, {
1248 .module_id = MODULE_HIMAC,
1249 .msg = "MODULE_HIMAC"
1250 }, {
1251 .module_id = MODULE_ROCEE_TOP,
1252 .msg = "MODULE_ROCEE_TOP"
1253 }, {
1254 .module_id = MODULE_ROCEE_TIMER,
1255 .msg = "MODULE_ROCEE_TIMER"
1256 }, {
1257 .module_id = MODULE_ROCEE_MDB,
1258 .msg = "MODULE_ROCEE_MDB"
1259 }, {
1260 .module_id = MODULE_ROCEE_TSP,
1261 .msg = "MODULE_ROCEE_TSP"
1262 }, {
1263 .module_id = MODULE_ROCEE_TRP,
1264 .msg = "MODULE_ROCEE_TRP"
1265 }, {
1266 .module_id = MODULE_ROCEE_SCC,
1267 .msg = "MODULE_ROCEE_SCC"
1268 }, {
1269 .module_id = MODULE_ROCEE_CAEP,
1270 .msg = "MODULE_ROCEE_CAEP"
1271 }, {
1272 .module_id = MODULE_ROCEE_GEN_AC,
1273 .msg = "MODULE_ROCEE_GEN_AC"
1274 }, {
1275 .module_id = MODULE_ROCEE_QMM,
1276 .msg = "MODULE_ROCEE_QMM"
1277 }, {
1278 .module_id = MODULE_ROCEE_LSAN,
1279 .msg = "MODULE_ROCEE_LSAN"
1280 }
1281 };
1282
1283 static const struct hclge_hw_type_id hclge_hw_type_id_st[] = {
1284 {
1285 .type_id = NONE_ERROR,
1286 .msg = "none_error"
1287 }, {
1288 .type_id = FIFO_ERROR,
1289 .msg = "fifo_error"
1290 }, {
1291 .type_id = MEMORY_ERROR,
1292 .msg = "memory_error"
1293 }, {
1294 .type_id = POISON_ERROR,
1295 .msg = "poison_error"
1296 }, {
1297 .type_id = MSIX_ECC_ERROR,
1298 .msg = "msix_ecc_error"
1299 }, {
1300 .type_id = TQP_INT_ECC_ERROR,
1301 .msg = "tqp_int_ecc_error"
1302 }, {
1303 .type_id = PF_ABNORMAL_INT_ERROR,
1304 .msg = "pf_abnormal_int_error"
1305 }, {
1306 .type_id = MPF_ABNORMAL_INT_ERROR,
1307 .msg = "mpf_abnormal_int_error"
1308 }, {
1309 .type_id = COMMON_ERROR,
1310 .msg = "common_error"
1311 }, {
1312 .type_id = PORT_ERROR,
1313 .msg = "port_error"
1314 }, {
1315 .type_id = ETS_ERROR,
1316 .msg = "ets_error"
1317 }, {
1318 .type_id = NCSI_ERROR,
1319 .msg = "ncsi_error"
1320 }, {
1321 .type_id = GLB_ERROR,
1322 .msg = "glb_error"
1323 }, {
1324 .type_id = LINK_ERROR,
1325 .msg = "link_error"
1326 }, {
1327 .type_id = PTP_ERROR,
1328 .msg = "ptp_error"
1329 }, {
1330 .type_id = ROCEE_NORMAL_ERR,
1331 .msg = "rocee_normal_error"
1332 }, {
1333 .type_id = ROCEE_OVF_ERR,
1334 .msg = "rocee_ovf_error"
1335 }, {
1336 .type_id = ROCEE_BUS_ERR,
1337 .msg = "rocee_bus_error"
1338 },
1339 };
1340
hclge_log_error(struct device * dev,char * reg,const struct hclge_hw_error * err,u32 err_sts,unsigned long * reset_requests)1341 static void hclge_log_error(struct device *dev, char *reg,
1342 const struct hclge_hw_error *err,
1343 u32 err_sts, unsigned long *reset_requests)
1344 {
1345 while (err->msg) {
1346 if (err->int_msk & err_sts) {
1347 dev_err(dev, "%s %s found [error status=0x%x]\n",
1348 reg, err->msg, err_sts);
1349 if (err->reset_level &&
1350 err->reset_level != HNAE3_NONE_RESET)
1351 set_bit(err->reset_level, reset_requests);
1352 }
1353 err++;
1354 }
1355 }
1356
1357 /* hclge_cmd_query_error: read the error information
1358 * @hdev: pointer to struct hclge_dev
1359 * @desc: descriptor for describing the command
1360 * @cmd: command opcode
1361 * @flag: flag for extended command structure
1362 *
1363 * This function query the error info from hw register/s using command
1364 */
hclge_cmd_query_error(struct hclge_dev * hdev,struct hclge_desc * desc,u32 cmd,u16 flag)1365 static int hclge_cmd_query_error(struct hclge_dev *hdev,
1366 struct hclge_desc *desc, u32 cmd, u16 flag)
1367 {
1368 struct device *dev = &hdev->pdev->dev;
1369 int desc_num = 1;
1370 int ret;
1371
1372 hclge_cmd_setup_basic_desc(&desc[0], cmd, true);
1373 if (flag) {
1374 desc[0].flag |= cpu_to_le16(flag);
1375 hclge_cmd_setup_basic_desc(&desc[1], cmd, true);
1376 desc_num = 2;
1377 }
1378
1379 ret = hclge_cmd_send(&hdev->hw, &desc[0], desc_num);
1380 if (ret)
1381 dev_err(dev, "query error cmd failed (%d)\n", ret);
1382
1383 return ret;
1384 }
1385
hclge_clear_mac_tnl_int(struct hclge_dev * hdev)1386 static int hclge_clear_mac_tnl_int(struct hclge_dev *hdev)
1387 {
1388 struct hclge_desc desc;
1389
1390 hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_CLEAR_MAC_TNL_INT, false);
1391 desc.data[0] = cpu_to_le32(HCLGE_MAC_TNL_INT_CLR);
1392
1393 return hclge_cmd_send(&hdev->hw, &desc, 1);
1394 }
1395
hclge_config_common_hw_err_int(struct hclge_dev * hdev,bool en)1396 static int hclge_config_common_hw_err_int(struct hclge_dev *hdev, bool en)
1397 {
1398 struct device *dev = &hdev->pdev->dev;
1399 struct hclge_desc desc[2];
1400 int ret;
1401
1402 /* configure common error interrupts */
1403 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_COMMON_ECC_INT_CFG, false);
1404 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
1405 hclge_cmd_setup_basic_desc(&desc[1], HCLGE_COMMON_ECC_INT_CFG, false);
1406
1407 if (en) {
1408 desc[0].data[0] = cpu_to_le32(HCLGE_IMP_TCM_ECC_ERR_INT_EN);
1409 desc[0].data[2] = cpu_to_le32(HCLGE_CMDQ_NIC_ECC_ERR_INT_EN |
1410 HCLGE_CMDQ_ROCEE_ECC_ERR_INT_EN);
1411 desc[0].data[3] = cpu_to_le32(HCLGE_IMP_RD_POISON_ERR_INT_EN);
1412 desc[0].data[4] = cpu_to_le32(HCLGE_TQP_ECC_ERR_INT_EN |
1413 HCLGE_MSIX_SRAM_ECC_ERR_INT_EN);
1414 desc[0].data[5] = cpu_to_le32(HCLGE_IMP_ITCM4_ECC_ERR_INT_EN);
1415 }
1416
1417 desc[1].data[0] = cpu_to_le32(HCLGE_IMP_TCM_ECC_ERR_INT_EN_MASK);
1418 desc[1].data[2] = cpu_to_le32(HCLGE_CMDQ_NIC_ECC_ERR_INT_EN_MASK |
1419 HCLGE_CMDQ_ROCEE_ECC_ERR_INT_EN_MASK);
1420 desc[1].data[3] = cpu_to_le32(HCLGE_IMP_RD_POISON_ERR_INT_EN_MASK);
1421 desc[1].data[4] = cpu_to_le32(HCLGE_TQP_ECC_ERR_INT_EN_MASK |
1422 HCLGE_MSIX_SRAM_ECC_ERR_INT_EN_MASK);
1423 desc[1].data[5] = cpu_to_le32(HCLGE_IMP_ITCM4_ECC_ERR_INT_EN_MASK);
1424
1425 ret = hclge_cmd_send(&hdev->hw, &desc[0], 2);
1426 if (ret)
1427 dev_err(dev,
1428 "fail(%d) to configure common err interrupts\n", ret);
1429
1430 return ret;
1431 }
1432
hclge_config_ncsi_hw_err_int(struct hclge_dev * hdev,bool en)1433 static int hclge_config_ncsi_hw_err_int(struct hclge_dev *hdev, bool en)
1434 {
1435 struct device *dev = &hdev->pdev->dev;
1436 struct hclge_desc desc;
1437 int ret;
1438
1439 if (hdev->ae_dev->dev_version < HNAE3_DEVICE_VERSION_V2)
1440 return 0;
1441
1442 /* configure NCSI error interrupts */
1443 hclge_cmd_setup_basic_desc(&desc, HCLGE_NCSI_INT_EN, false);
1444 if (en)
1445 desc.data[0] = cpu_to_le32(HCLGE_NCSI_ERR_INT_EN);
1446
1447 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1448 if (ret)
1449 dev_err(dev,
1450 "fail(%d) to configure NCSI error interrupts\n", ret);
1451
1452 return ret;
1453 }
1454
hclge_config_igu_egu_hw_err_int(struct hclge_dev * hdev,bool en)1455 static int hclge_config_igu_egu_hw_err_int(struct hclge_dev *hdev, bool en)
1456 {
1457 struct device *dev = &hdev->pdev->dev;
1458 struct hclge_desc desc;
1459 int ret;
1460
1461 /* configure IGU,EGU error interrupts */
1462 hclge_cmd_setup_basic_desc(&desc, HCLGE_IGU_COMMON_INT_EN, false);
1463 desc.data[0] = cpu_to_le32(HCLGE_IGU_ERR_INT_TYPE);
1464 if (en)
1465 desc.data[0] |= cpu_to_le32(HCLGE_IGU_ERR_INT_EN);
1466
1467 desc.data[1] = cpu_to_le32(HCLGE_IGU_ERR_INT_EN_MASK);
1468
1469 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1470 if (ret) {
1471 dev_err(dev,
1472 "fail(%d) to configure IGU common interrupts\n", ret);
1473 return ret;
1474 }
1475
1476 hclge_cmd_setup_basic_desc(&desc, HCLGE_IGU_EGU_TNL_INT_EN, false);
1477 if (en)
1478 desc.data[0] = cpu_to_le32(HCLGE_IGU_TNL_ERR_INT_EN);
1479
1480 desc.data[1] = cpu_to_le32(HCLGE_IGU_TNL_ERR_INT_EN_MASK);
1481
1482 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1483 if (ret) {
1484 dev_err(dev,
1485 "fail(%d) to configure IGU-EGU TNL interrupts\n", ret);
1486 return ret;
1487 }
1488
1489 ret = hclge_config_ncsi_hw_err_int(hdev, en);
1490
1491 return ret;
1492 }
1493
hclge_config_ppp_error_interrupt(struct hclge_dev * hdev,u32 cmd,bool en)1494 static int hclge_config_ppp_error_interrupt(struct hclge_dev *hdev, u32 cmd,
1495 bool en)
1496 {
1497 struct device *dev = &hdev->pdev->dev;
1498 struct hclge_desc desc[2];
1499 int ret;
1500
1501 /* configure PPP error interrupts */
1502 hclge_cmd_setup_basic_desc(&desc[0], cmd, false);
1503 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
1504 hclge_cmd_setup_basic_desc(&desc[1], cmd, false);
1505
1506 if (cmd == HCLGE_PPP_CMD0_INT_CMD) {
1507 if (en) {
1508 desc[0].data[0] =
1509 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT0_EN);
1510 desc[0].data[1] =
1511 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT1_EN);
1512 desc[0].data[4] = cpu_to_le32(HCLGE_PPP_PF_ERR_INT_EN);
1513 }
1514
1515 desc[1].data[0] =
1516 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT0_EN_MASK);
1517 desc[1].data[1] =
1518 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT1_EN_MASK);
1519 if (hdev->ae_dev->dev_version >= HNAE3_DEVICE_VERSION_V2)
1520 desc[1].data[2] =
1521 cpu_to_le32(HCLGE_PPP_PF_ERR_INT_EN_MASK);
1522 } else if (cmd == HCLGE_PPP_CMD1_INT_CMD) {
1523 if (en) {
1524 desc[0].data[0] =
1525 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT2_EN);
1526 desc[0].data[1] =
1527 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT3_EN);
1528 }
1529
1530 desc[1].data[0] =
1531 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT2_EN_MASK);
1532 desc[1].data[1] =
1533 cpu_to_le32(HCLGE_PPP_MPF_ECC_ERR_INT3_EN_MASK);
1534 }
1535
1536 ret = hclge_cmd_send(&hdev->hw, &desc[0], 2);
1537 if (ret)
1538 dev_err(dev, "fail(%d) to configure PPP error intr\n", ret);
1539
1540 return ret;
1541 }
1542
hclge_config_ppp_hw_err_int(struct hclge_dev * hdev,bool en)1543 static int hclge_config_ppp_hw_err_int(struct hclge_dev *hdev, bool en)
1544 {
1545 int ret;
1546
1547 ret = hclge_config_ppp_error_interrupt(hdev, HCLGE_PPP_CMD0_INT_CMD,
1548 en);
1549 if (ret)
1550 return ret;
1551
1552 ret = hclge_config_ppp_error_interrupt(hdev, HCLGE_PPP_CMD1_INT_CMD,
1553 en);
1554
1555 return ret;
1556 }
1557
hclge_config_tm_hw_err_int(struct hclge_dev * hdev,bool en)1558 static int hclge_config_tm_hw_err_int(struct hclge_dev *hdev, bool en)
1559 {
1560 struct device *dev = &hdev->pdev->dev;
1561 struct hclge_desc desc;
1562 int ret;
1563
1564 /* configure TM SCH hw errors */
1565 hclge_cmd_setup_basic_desc(&desc, HCLGE_TM_SCH_ECC_INT_EN, false);
1566 if (en)
1567 desc.data[0] = cpu_to_le32(HCLGE_TM_SCH_ECC_ERR_INT_EN);
1568
1569 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1570 if (ret) {
1571 dev_err(dev, "fail(%d) to configure TM SCH errors\n", ret);
1572 return ret;
1573 }
1574
1575 /* configure TM QCN hw errors */
1576 hclge_cmd_setup_basic_desc(&desc, HCLGE_TM_QCN_MEM_INT_CFG, false);
1577 desc.data[0] = cpu_to_le32(HCLGE_TM_QCN_ERR_INT_TYPE);
1578 if (en) {
1579 desc.data[0] |= cpu_to_le32(HCLGE_TM_QCN_FIFO_INT_EN);
1580 desc.data[1] = cpu_to_le32(HCLGE_TM_QCN_MEM_ERR_INT_EN);
1581 }
1582
1583 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1584 if (ret)
1585 dev_err(dev,
1586 "fail(%d) to configure TM QCN mem errors\n", ret);
1587
1588 return ret;
1589 }
1590
hclge_config_mac_err_int(struct hclge_dev * hdev,bool en)1591 static int hclge_config_mac_err_int(struct hclge_dev *hdev, bool en)
1592 {
1593 struct device *dev = &hdev->pdev->dev;
1594 struct hclge_desc desc;
1595 int ret;
1596
1597 /* configure MAC common error interrupts */
1598 hclge_cmd_setup_basic_desc(&desc, HCLGE_MAC_COMMON_INT_EN, false);
1599 if (en)
1600 desc.data[0] = cpu_to_le32(HCLGE_MAC_COMMON_ERR_INT_EN);
1601
1602 desc.data[1] = cpu_to_le32(HCLGE_MAC_COMMON_ERR_INT_EN_MASK);
1603
1604 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
1605 if (ret)
1606 dev_err(dev,
1607 "fail(%d) to configure MAC COMMON error intr\n", ret);
1608
1609 return ret;
1610 }
1611
hclge_config_mac_tnl_int(struct hclge_dev * hdev,bool en)1612 int hclge_config_mac_tnl_int(struct hclge_dev *hdev, bool en)
1613 {
1614 struct hclge_desc desc;
1615
1616 hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_MAC_TNL_INT_EN, false);
1617 if (en)
1618 desc.data[0] = cpu_to_le32(HCLGE_MAC_TNL_INT_EN);
1619 else
1620 desc.data[0] = 0;
1621
1622 desc.data[1] = cpu_to_le32(HCLGE_MAC_TNL_INT_EN_MASK);
1623
1624 return hclge_cmd_send(&hdev->hw, &desc, 1);
1625 }
1626
hclge_config_ppu_error_interrupts(struct hclge_dev * hdev,u32 cmd,bool en)1627 static int hclge_config_ppu_error_interrupts(struct hclge_dev *hdev, u32 cmd,
1628 bool en)
1629 {
1630 struct device *dev = &hdev->pdev->dev;
1631 struct hclge_desc desc[2];
1632 int desc_num = 1;
1633 int ret;
1634
1635 /* configure PPU error interrupts */
1636 if (cmd == HCLGE_PPU_MPF_ECC_INT_CMD) {
1637 hclge_cmd_setup_basic_desc(&desc[0], cmd, false);
1638 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
1639 hclge_cmd_setup_basic_desc(&desc[1], cmd, false);
1640 if (en) {
1641 desc[0].data[0] =
1642 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT0_EN);
1643 desc[0].data[1] =
1644 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT1_EN);
1645 desc[1].data[3] =
1646 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT3_EN);
1647 desc[1].data[4] =
1648 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT2_EN);
1649 }
1650
1651 desc[1].data[0] =
1652 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT0_EN_MASK);
1653 desc[1].data[1] =
1654 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT1_EN_MASK);
1655 desc[1].data[2] =
1656 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT2_EN_MASK);
1657 desc[1].data[3] |=
1658 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT3_EN_MASK);
1659 desc_num = 2;
1660 } else if (cmd == HCLGE_PPU_MPF_OTHER_INT_CMD) {
1661 hclge_cmd_setup_basic_desc(&desc[0], cmd, false);
1662 if (en)
1663 desc[0].data[0] =
1664 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT2_EN2);
1665
1666 desc[0].data[2] =
1667 cpu_to_le32(HCLGE_PPU_MPF_ABNORMAL_INT2_EN2_MASK);
1668 } else if (cmd == HCLGE_PPU_PF_OTHER_INT_CMD) {
1669 hclge_cmd_setup_basic_desc(&desc[0], cmd, false);
1670 if (en)
1671 desc[0].data[0] =
1672 cpu_to_le32(HCLGE_PPU_PF_ABNORMAL_INT_EN);
1673
1674 desc[0].data[2] =
1675 cpu_to_le32(HCLGE_PPU_PF_ABNORMAL_INT_EN_MASK);
1676 } else {
1677 dev_err(dev, "Invalid cmd to configure PPU error interrupts\n");
1678 return -EINVAL;
1679 }
1680
1681 ret = hclge_cmd_send(&hdev->hw, &desc[0], desc_num);
1682
1683 return ret;
1684 }
1685
hclge_config_ppu_hw_err_int(struct hclge_dev * hdev,bool en)1686 static int hclge_config_ppu_hw_err_int(struct hclge_dev *hdev, bool en)
1687 {
1688 struct device *dev = &hdev->pdev->dev;
1689 int ret;
1690
1691 ret = hclge_config_ppu_error_interrupts(hdev, HCLGE_PPU_MPF_ECC_INT_CMD,
1692 en);
1693 if (ret) {
1694 dev_err(dev, "fail(%d) to configure PPU MPF ECC error intr\n",
1695 ret);
1696 return ret;
1697 }
1698
1699 ret = hclge_config_ppu_error_interrupts(hdev,
1700 HCLGE_PPU_MPF_OTHER_INT_CMD,
1701 en);
1702 if (ret) {
1703 dev_err(dev, "fail(%d) to configure PPU MPF other intr\n", ret);
1704 return ret;
1705 }
1706
1707 ret = hclge_config_ppu_error_interrupts(hdev,
1708 HCLGE_PPU_PF_OTHER_INT_CMD, en);
1709 if (ret)
1710 dev_err(dev, "fail(%d) to configure PPU PF error interrupts\n",
1711 ret);
1712 return ret;
1713 }
1714
hclge_config_ssu_hw_err_int(struct hclge_dev * hdev,bool en)1715 static int hclge_config_ssu_hw_err_int(struct hclge_dev *hdev, bool en)
1716 {
1717 struct device *dev = &hdev->pdev->dev;
1718 struct hclge_desc desc[2];
1719 int ret;
1720
1721 /* configure SSU ecc error interrupts */
1722 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_SSU_ECC_INT_CMD, false);
1723 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
1724 hclge_cmd_setup_basic_desc(&desc[1], HCLGE_SSU_ECC_INT_CMD, false);
1725 if (en) {
1726 desc[0].data[0] = cpu_to_le32(HCLGE_SSU_1BIT_ECC_ERR_INT_EN);
1727 desc[0].data[1] =
1728 cpu_to_le32(HCLGE_SSU_MULTI_BIT_ECC_ERR_INT_EN);
1729 desc[0].data[4] = cpu_to_le32(HCLGE_SSU_BIT32_ECC_ERR_INT_EN);
1730 }
1731
1732 desc[1].data[0] = cpu_to_le32(HCLGE_SSU_1BIT_ECC_ERR_INT_EN_MASK);
1733 desc[1].data[1] = cpu_to_le32(HCLGE_SSU_MULTI_BIT_ECC_ERR_INT_EN_MASK);
1734 desc[1].data[2] = cpu_to_le32(HCLGE_SSU_BIT32_ECC_ERR_INT_EN_MASK);
1735
1736 ret = hclge_cmd_send(&hdev->hw, &desc[0], 2);
1737 if (ret) {
1738 dev_err(dev,
1739 "fail(%d) to configure SSU ECC error interrupt\n", ret);
1740 return ret;
1741 }
1742
1743 /* configure SSU common error interrupts */
1744 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_SSU_COMMON_INT_CMD, false);
1745 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
1746 hclge_cmd_setup_basic_desc(&desc[1], HCLGE_SSU_COMMON_INT_CMD, false);
1747
1748 if (en) {
1749 if (hdev->ae_dev->dev_version >= HNAE3_DEVICE_VERSION_V2)
1750 desc[0].data[0] =
1751 cpu_to_le32(HCLGE_SSU_COMMON_INT_EN);
1752 else
1753 desc[0].data[0] =
1754 cpu_to_le32(HCLGE_SSU_COMMON_INT_EN & ~BIT(5));
1755 desc[0].data[1] = cpu_to_le32(HCLGE_SSU_PORT_BASED_ERR_INT_EN);
1756 desc[0].data[2] =
1757 cpu_to_le32(HCLGE_SSU_FIFO_OVERFLOW_ERR_INT_EN);
1758 }
1759
1760 desc[1].data[0] = cpu_to_le32(HCLGE_SSU_COMMON_INT_EN_MASK |
1761 HCLGE_SSU_PORT_BASED_ERR_INT_EN_MASK);
1762 desc[1].data[1] = cpu_to_le32(HCLGE_SSU_FIFO_OVERFLOW_ERR_INT_EN_MASK);
1763
1764 ret = hclge_cmd_send(&hdev->hw, &desc[0], 2);
1765 if (ret)
1766 dev_err(dev,
1767 "fail(%d) to configure SSU COMMON error intr\n", ret);
1768
1769 return ret;
1770 }
1771
1772 /* hclge_query_bd_num: query number of buffer descriptors
1773 * @hdev: pointer to struct hclge_dev
1774 * @is_ras: true for ras, false for msix
1775 * @mpf_bd_num: number of main PF interrupt buffer descriptors
1776 * @pf_bd_num: number of not main PF interrupt buffer descriptors
1777 *
1778 * This function querys number of mpf and pf buffer descriptors.
1779 */
hclge_query_bd_num(struct hclge_dev * hdev,bool is_ras,u32 * mpf_bd_num,u32 * pf_bd_num)1780 static int hclge_query_bd_num(struct hclge_dev *hdev, bool is_ras,
1781 u32 *mpf_bd_num, u32 *pf_bd_num)
1782 {
1783 struct device *dev = &hdev->pdev->dev;
1784 u32 mpf_min_bd_num, pf_min_bd_num;
1785 enum hclge_opcode_type opcode;
1786 struct hclge_desc desc_bd;
1787 int ret;
1788
1789 if (is_ras) {
1790 opcode = HCLGE_QUERY_RAS_INT_STS_BD_NUM;
1791 mpf_min_bd_num = HCLGE_MPF_RAS_INT_MIN_BD_NUM;
1792 pf_min_bd_num = HCLGE_PF_RAS_INT_MIN_BD_NUM;
1793 } else {
1794 opcode = HCLGE_QUERY_MSIX_INT_STS_BD_NUM;
1795 mpf_min_bd_num = HCLGE_MPF_MSIX_INT_MIN_BD_NUM;
1796 pf_min_bd_num = HCLGE_PF_MSIX_INT_MIN_BD_NUM;
1797 }
1798
1799 hclge_cmd_setup_basic_desc(&desc_bd, opcode, true);
1800 ret = hclge_cmd_send(&hdev->hw, &desc_bd, 1);
1801 if (ret) {
1802 dev_err(dev, "fail(%d) to query msix int status bd num\n",
1803 ret);
1804 return ret;
1805 }
1806
1807 *mpf_bd_num = le32_to_cpu(desc_bd.data[0]);
1808 *pf_bd_num = le32_to_cpu(desc_bd.data[1]);
1809 if (*mpf_bd_num < mpf_min_bd_num || *pf_bd_num < pf_min_bd_num) {
1810 dev_err(dev, "Invalid bd num: mpf(%u), pf(%u)\n",
1811 *mpf_bd_num, *pf_bd_num);
1812 return -EINVAL;
1813 }
1814
1815 return 0;
1816 }
1817
1818 /* hclge_handle_mpf_ras_error: handle all main PF RAS errors
1819 * @hdev: pointer to struct hclge_dev
1820 * @desc: descriptor for describing the command
1821 * @num: number of extended command structures
1822 *
1823 * This function handles all the main PF RAS errors in the
1824 * hw register/s using command.
1825 */
hclge_handle_mpf_ras_error(struct hclge_dev * hdev,struct hclge_desc * desc,int num)1826 static int hclge_handle_mpf_ras_error(struct hclge_dev *hdev,
1827 struct hclge_desc *desc,
1828 int num)
1829 {
1830 struct hnae3_ae_dev *ae_dev = hdev->ae_dev;
1831 struct device *dev = &hdev->pdev->dev;
1832 __le32 *desc_data;
1833 u32 status;
1834 int ret;
1835
1836 /* query all main PF RAS errors */
1837 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_CLEAR_MPF_RAS_INT,
1838 true);
1839 ret = hclge_cmd_send(&hdev->hw, &desc[0], num);
1840 if (ret) {
1841 dev_err(dev, "query all mpf ras int cmd failed (%d)\n", ret);
1842 return ret;
1843 }
1844
1845 /* log HNS common errors */
1846 status = le32_to_cpu(desc[0].data[0]);
1847 if (status)
1848 hclge_log_error(dev, "IMP_TCM_ECC_INT_STS",
1849 &hclge_imp_tcm_ecc_int[0], status,
1850 &ae_dev->hw_err_reset_req);
1851
1852 status = le32_to_cpu(desc[0].data[1]);
1853 if (status)
1854 hclge_log_error(dev, "CMDQ_MEM_ECC_INT_STS",
1855 &hclge_cmdq_nic_mem_ecc_int[0], status,
1856 &ae_dev->hw_err_reset_req);
1857
1858 if ((le32_to_cpu(desc[0].data[2])) & BIT(0))
1859 dev_warn(dev, "imp_rd_data_poison_err found\n");
1860
1861 status = le32_to_cpu(desc[0].data[3]);
1862 if (status)
1863 hclge_log_error(dev, "TQP_INT_ECC_INT_STS",
1864 &hclge_tqp_int_ecc_int[0], status,
1865 &ae_dev->hw_err_reset_req);
1866
1867 status = le32_to_cpu(desc[0].data[4]);
1868 if (status)
1869 hclge_log_error(dev, "MSIX_ECC_INT_STS",
1870 &hclge_msix_sram_ecc_int[0], status,
1871 &ae_dev->hw_err_reset_req);
1872
1873 /* log SSU(Storage Switch Unit) errors */
1874 desc_data = (__le32 *)&desc[2];
1875 status = le32_to_cpu(*(desc_data + 2));
1876 if (status)
1877 hclge_log_error(dev, "SSU_ECC_MULTI_BIT_INT_0",
1878 &hclge_ssu_mem_ecc_err_int[0], status,
1879 &ae_dev->hw_err_reset_req);
1880
1881 status = le32_to_cpu(*(desc_data + 3)) & BIT(0);
1882 if (status) {
1883 dev_err(dev, "SSU_ECC_MULTI_BIT_INT_1 ssu_mem32_ecc_mbit_err found [error status=0x%x]\n",
1884 status);
1885 set_bit(HNAE3_GLOBAL_RESET, &ae_dev->hw_err_reset_req);
1886 }
1887
1888 status = le32_to_cpu(*(desc_data + 4)) & HCLGE_SSU_COMMON_ERR_INT_MASK;
1889 if (status)
1890 hclge_log_error(dev, "SSU_COMMON_ERR_INT",
1891 &hclge_ssu_com_err_int[0], status,
1892 &ae_dev->hw_err_reset_req);
1893
1894 /* log IGU(Ingress Unit) errors */
1895 desc_data = (__le32 *)&desc[3];
1896 status = le32_to_cpu(*desc_data) & HCLGE_IGU_INT_MASK;
1897 if (status)
1898 hclge_log_error(dev, "IGU_INT_STS",
1899 &hclge_igu_int[0], status,
1900 &ae_dev->hw_err_reset_req);
1901
1902 /* log PPP(Programmable Packet Process) errors */
1903 desc_data = (__le32 *)&desc[4];
1904 status = le32_to_cpu(*(desc_data + 1));
1905 if (status)
1906 hclge_log_error(dev, "PPP_MPF_ABNORMAL_INT_ST1",
1907 &hclge_ppp_mpf_abnormal_int_st1[0], status,
1908 &ae_dev->hw_err_reset_req);
1909
1910 status = le32_to_cpu(*(desc_data + 3)) & HCLGE_PPP_MPF_INT_ST3_MASK;
1911 if (status)
1912 hclge_log_error(dev, "PPP_MPF_ABNORMAL_INT_ST3",
1913 &hclge_ppp_mpf_abnormal_int_st3[0], status,
1914 &ae_dev->hw_err_reset_req);
1915
1916 /* log PPU(RCB) errors */
1917 desc_data = (__le32 *)&desc[5];
1918 status = le32_to_cpu(*(desc_data + 1));
1919 if (status) {
1920 dev_err(dev,
1921 "PPU_MPF_ABNORMAL_INT_ST1 rpu_rx_pkt_ecc_mbit_err found\n");
1922 set_bit(HNAE3_GLOBAL_RESET, &ae_dev->hw_err_reset_req);
1923 }
1924
1925 status = le32_to_cpu(*(desc_data + 2));
1926 if (status)
1927 hclge_log_error(dev, "PPU_MPF_ABNORMAL_INT_ST2",
1928 &hclge_ppu_mpf_abnormal_int_st2[0], status,
1929 &ae_dev->hw_err_reset_req);
1930
1931 status = le32_to_cpu(*(desc_data + 3)) & HCLGE_PPU_MPF_INT_ST3_MASK;
1932 if (status)
1933 hclge_log_error(dev, "PPU_MPF_ABNORMAL_INT_ST3",
1934 &hclge_ppu_mpf_abnormal_int_st3[0], status,
1935 &ae_dev->hw_err_reset_req);
1936
1937 /* log TM(Traffic Manager) errors */
1938 desc_data = (__le32 *)&desc[6];
1939 status = le32_to_cpu(*desc_data);
1940 if (status)
1941 hclge_log_error(dev, "TM_SCH_RINT",
1942 &hclge_tm_sch_rint[0], status,
1943 &ae_dev->hw_err_reset_req);
1944
1945 /* log QCN(Quantized Congestion Control) errors */
1946 desc_data = (__le32 *)&desc[7];
1947 status = le32_to_cpu(*desc_data) & HCLGE_QCN_FIFO_INT_MASK;
1948 if (status)
1949 hclge_log_error(dev, "QCN_FIFO_RINT",
1950 &hclge_qcn_fifo_rint[0], status,
1951 &ae_dev->hw_err_reset_req);
1952
1953 status = le32_to_cpu(*(desc_data + 1)) & HCLGE_QCN_ECC_INT_MASK;
1954 if (status)
1955 hclge_log_error(dev, "QCN_ECC_RINT",
1956 &hclge_qcn_ecc_rint[0], status,
1957 &ae_dev->hw_err_reset_req);
1958
1959 /* log NCSI errors */
1960 desc_data = (__le32 *)&desc[9];
1961 status = le32_to_cpu(*desc_data) & HCLGE_NCSI_ECC_INT_MASK;
1962 if (status)
1963 hclge_log_error(dev, "NCSI_ECC_INT_RPT",
1964 &hclge_ncsi_err_int[0], status,
1965 &ae_dev->hw_err_reset_req);
1966
1967 /* clear all main PF RAS errors */
1968 hclge_comm_cmd_reuse_desc(&desc[0], false);
1969 ret = hclge_cmd_send(&hdev->hw, &desc[0], num);
1970 if (ret)
1971 dev_err(dev, "clear all mpf ras int cmd failed (%d)\n", ret);
1972
1973 return ret;
1974 }
1975
1976 /* hclge_handle_pf_ras_error: handle all PF RAS errors
1977 * @hdev: pointer to struct hclge_dev
1978 * @desc: descriptor for describing the command
1979 * @num: number of extended command structures
1980 *
1981 * This function handles all the PF RAS errors in the
1982 * hw registers using command.
1983 */
hclge_handle_pf_ras_error(struct hclge_dev * hdev,struct hclge_desc * desc,int num)1984 static int hclge_handle_pf_ras_error(struct hclge_dev *hdev,
1985 struct hclge_desc *desc,
1986 int num)
1987 {
1988 struct hnae3_ae_dev *ae_dev = hdev->ae_dev;
1989 struct device *dev = &hdev->pdev->dev;
1990 __le32 *desc_data;
1991 u32 status;
1992 int ret;
1993
1994 /* query all PF RAS errors */
1995 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_CLEAR_PF_RAS_INT,
1996 true);
1997 ret = hclge_cmd_send(&hdev->hw, &desc[0], num);
1998 if (ret) {
1999 dev_err(dev, "query all pf ras int cmd failed (%d)\n", ret);
2000 return ret;
2001 }
2002
2003 /* log SSU(Storage Switch Unit) errors */
2004 status = le32_to_cpu(desc[0].data[0]);
2005 if (status)
2006 hclge_log_error(dev, "SSU_PORT_BASED_ERR_INT",
2007 &hclge_ssu_port_based_err_int[0], status,
2008 &ae_dev->hw_err_reset_req);
2009
2010 status = le32_to_cpu(desc[0].data[1]);
2011 if (status)
2012 hclge_log_error(dev, "SSU_FIFO_OVERFLOW_INT",
2013 &hclge_ssu_fifo_overflow_int[0], status,
2014 &ae_dev->hw_err_reset_req);
2015
2016 status = le32_to_cpu(desc[0].data[2]);
2017 if (status)
2018 hclge_log_error(dev, "SSU_ETS_TCG_INT",
2019 &hclge_ssu_ets_tcg_int[0], status,
2020 &ae_dev->hw_err_reset_req);
2021
2022 /* log IGU(Ingress Unit) EGU(Egress Unit) TNL errors */
2023 desc_data = (__le32 *)&desc[1];
2024 status = le32_to_cpu(*desc_data) & HCLGE_IGU_EGU_TNL_INT_MASK;
2025 if (status)
2026 hclge_log_error(dev, "IGU_EGU_TNL_INT_STS",
2027 &hclge_igu_egu_tnl_int[0], status,
2028 &ae_dev->hw_err_reset_req);
2029
2030 /* log PPU(RCB) errors */
2031 desc_data = (__le32 *)&desc[3];
2032 status = le32_to_cpu(*desc_data) & HCLGE_PPU_PF_INT_RAS_MASK;
2033 if (status) {
2034 hclge_log_error(dev, "PPU_PF_ABNORMAL_INT_ST0",
2035 &hclge_ppu_pf_abnormal_int[0], status,
2036 &ae_dev->hw_err_reset_req);
2037 hclge_report_hw_error(hdev, HNAE3_PPU_POISON_ERROR);
2038 }
2039
2040 /* clear all PF RAS errors */
2041 hclge_comm_cmd_reuse_desc(&desc[0], false);
2042 ret = hclge_cmd_send(&hdev->hw, &desc[0], num);
2043 if (ret)
2044 dev_err(dev, "clear all pf ras int cmd failed (%d)\n", ret);
2045
2046 return ret;
2047 }
2048
hclge_handle_all_ras_errors(struct hclge_dev * hdev)2049 static int hclge_handle_all_ras_errors(struct hclge_dev *hdev)
2050 {
2051 u32 mpf_bd_num, pf_bd_num, bd_num;
2052 struct hclge_desc *desc;
2053 int ret;
2054
2055 /* query the number of registers in the RAS int status */
2056 ret = hclge_query_bd_num(hdev, true, &mpf_bd_num, &pf_bd_num);
2057 if (ret)
2058 return ret;
2059
2060 bd_num = max_t(u32, mpf_bd_num, pf_bd_num);
2061 desc = kcalloc(bd_num, sizeof(struct hclge_desc), GFP_KERNEL);
2062 if (!desc)
2063 return -ENOMEM;
2064
2065 /* handle all main PF RAS errors */
2066 ret = hclge_handle_mpf_ras_error(hdev, desc, mpf_bd_num);
2067 if (ret) {
2068 kfree(desc);
2069 return ret;
2070 }
2071 memset(desc, 0, bd_num * sizeof(struct hclge_desc));
2072
2073 /* handle all PF RAS errors */
2074 ret = hclge_handle_pf_ras_error(hdev, desc, pf_bd_num);
2075 kfree(desc);
2076
2077 return ret;
2078 }
2079
hclge_log_rocee_axi_error(struct hclge_dev * hdev)2080 static int hclge_log_rocee_axi_error(struct hclge_dev *hdev)
2081 {
2082 struct device *dev = &hdev->pdev->dev;
2083 struct hclge_desc desc[3];
2084 int ret;
2085
2086 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
2087 true);
2088 hclge_cmd_setup_basic_desc(&desc[1], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
2089 true);
2090 hclge_cmd_setup_basic_desc(&desc[2], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
2091 true);
2092 desc[0].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
2093 desc[1].flag |= cpu_to_le16(HCLGE_COMM_CMD_FLAG_NEXT);
2094
2095 ret = hclge_cmd_send(&hdev->hw, &desc[0], 3);
2096 if (ret) {
2097 dev_err(dev, "failed(%d) to query ROCEE AXI error sts\n", ret);
2098 return ret;
2099 }
2100
2101 dev_err(dev, "AXI1: %08X %08X %08X %08X %08X %08X\n",
2102 le32_to_cpu(desc[0].data[0]), le32_to_cpu(desc[0].data[1]),
2103 le32_to_cpu(desc[0].data[2]), le32_to_cpu(desc[0].data[3]),
2104 le32_to_cpu(desc[0].data[4]), le32_to_cpu(desc[0].data[5]));
2105 dev_err(dev, "AXI2: %08X %08X %08X %08X %08X %08X\n",
2106 le32_to_cpu(desc[1].data[0]), le32_to_cpu(desc[1].data[1]),
2107 le32_to_cpu(desc[1].data[2]), le32_to_cpu(desc[1].data[3]),
2108 le32_to_cpu(desc[1].data[4]), le32_to_cpu(desc[1].data[5]));
2109 dev_err(dev, "AXI3: %08X %08X %08X %08X\n",
2110 le32_to_cpu(desc[2].data[0]), le32_to_cpu(desc[2].data[1]),
2111 le32_to_cpu(desc[2].data[2]), le32_to_cpu(desc[2].data[3]));
2112
2113 return 0;
2114 }
2115
hclge_log_rocee_ecc_error(struct hclge_dev * hdev)2116 static int hclge_log_rocee_ecc_error(struct hclge_dev *hdev)
2117 {
2118 struct device *dev = &hdev->pdev->dev;
2119 struct hclge_desc desc[2];
2120 int ret;
2121
2122 ret = hclge_cmd_query_error(hdev, &desc[0],
2123 HCLGE_QUERY_ROCEE_ECC_RAS_INFO_CMD,
2124 HCLGE_COMM_CMD_FLAG_NEXT);
2125 if (ret) {
2126 dev_err(dev, "failed(%d) to query ROCEE ECC error sts\n", ret);
2127 return ret;
2128 }
2129
2130 dev_err(dev, "ECC1: %08X %08X %08X %08X %08X %08X\n",
2131 le32_to_cpu(desc[0].data[0]), le32_to_cpu(desc[0].data[1]),
2132 le32_to_cpu(desc[0].data[2]), le32_to_cpu(desc[0].data[3]),
2133 le32_to_cpu(desc[0].data[4]), le32_to_cpu(desc[0].data[5]));
2134 dev_err(dev, "ECC2: %08X %08X %08X\n", le32_to_cpu(desc[1].data[0]),
2135 le32_to_cpu(desc[1].data[1]), le32_to_cpu(desc[1].data[2]));
2136
2137 return 0;
2138 }
2139
hclge_log_rocee_ovf_error(struct hclge_dev * hdev)2140 static int hclge_log_rocee_ovf_error(struct hclge_dev *hdev)
2141 {
2142 struct device *dev = &hdev->pdev->dev;
2143 struct hclge_desc desc[2];
2144 int ret;
2145
2146 /* read overflow error status */
2147 ret = hclge_cmd_query_error(hdev, &desc[0], HCLGE_ROCEE_PF_RAS_INT_CMD,
2148 0);
2149 if (ret) {
2150 dev_err(dev, "failed(%d) to query ROCEE OVF error sts\n", ret);
2151 return ret;
2152 }
2153
2154 /* log overflow error */
2155 if (le32_to_cpu(desc[0].data[0]) & HCLGE_ROCEE_OVF_ERR_INT_MASK) {
2156 const struct hclge_hw_error *err;
2157 u32 err_sts;
2158
2159 err = &hclge_rocee_qmm_ovf_err_int[0];
2160 err_sts = HCLGE_ROCEE_OVF_ERR_TYPE_MASK &
2161 le32_to_cpu(desc[0].data[0]);
2162 while (err->msg) {
2163 if (err->int_msk == err_sts) {
2164 dev_err(dev, "%s [error status=0x%x] found\n",
2165 err->msg,
2166 le32_to_cpu(desc[0].data[0]));
2167 break;
2168 }
2169 err++;
2170 }
2171 }
2172
2173 if (le32_to_cpu(desc[0].data[1]) & HCLGE_ROCEE_OVF_ERR_INT_MASK) {
2174 dev_err(dev, "ROCEE TSP OVF [error status=0x%x] found\n",
2175 le32_to_cpu(desc[0].data[1]));
2176 }
2177
2178 if (le32_to_cpu(desc[0].data[2]) & HCLGE_ROCEE_OVF_ERR_INT_MASK) {
2179 dev_err(dev, "ROCEE SCC OVF [error status=0x%x] found\n",
2180 le32_to_cpu(desc[0].data[2]));
2181 }
2182
2183 return 0;
2184 }
2185
2186 static enum hnae3_reset_type
hclge_log_and_clear_rocee_ras_error(struct hclge_dev * hdev)2187 hclge_log_and_clear_rocee_ras_error(struct hclge_dev *hdev)
2188 {
2189 enum hnae3_reset_type reset_type = HNAE3_NONE_RESET;
2190 struct device *dev = &hdev->pdev->dev;
2191 struct hclge_desc desc[2];
2192 unsigned int status;
2193 int ret;
2194
2195 /* read RAS error interrupt status */
2196 ret = hclge_cmd_query_error(hdev, &desc[0],
2197 HCLGE_QUERY_CLEAR_ROCEE_RAS_INT, 0);
2198 if (ret) {
2199 dev_err(dev, "failed(%d) to query ROCEE RAS INT SRC\n", ret);
2200 /* reset everything for now */
2201 return HNAE3_GLOBAL_RESET;
2202 }
2203
2204 status = le32_to_cpu(desc[0].data[0]);
2205 if (status & HCLGE_ROCEE_AXI_ERR_INT_MASK) {
2206 if (status & HCLGE_ROCEE_RERR_INT_MASK)
2207 dev_err(dev, "ROCEE RAS AXI rresp error\n");
2208
2209 if (status & HCLGE_ROCEE_BERR_INT_MASK)
2210 dev_err(dev, "ROCEE RAS AXI bresp error\n");
2211
2212 reset_type = HNAE3_FUNC_RESET;
2213
2214 hclge_report_hw_error(hdev, HNAE3_ROCEE_AXI_RESP_ERROR);
2215
2216 ret = hclge_log_rocee_axi_error(hdev);
2217 if (ret)
2218 return HNAE3_GLOBAL_RESET;
2219 }
2220
2221 if (status & HCLGE_ROCEE_ECC_INT_MASK) {
2222 dev_err(dev, "ROCEE RAS 2bit ECC error\n");
2223 reset_type = HNAE3_GLOBAL_RESET;
2224
2225 ret = hclge_log_rocee_ecc_error(hdev);
2226 if (ret)
2227 return HNAE3_GLOBAL_RESET;
2228 }
2229
2230 if (status & HCLGE_ROCEE_OVF_INT_MASK) {
2231 ret = hclge_log_rocee_ovf_error(hdev);
2232 if (ret) {
2233 dev_err(dev, "failed(%d) to process ovf error\n", ret);
2234 /* reset everything for now */
2235 return HNAE3_GLOBAL_RESET;
2236 }
2237 }
2238
2239 /* clear error status */
2240 hclge_comm_cmd_reuse_desc(&desc[0], false);
2241 ret = hclge_cmd_send(&hdev->hw, &desc[0], 1);
2242 if (ret) {
2243 dev_err(dev, "failed(%d) to clear ROCEE RAS error\n", ret);
2244 /* reset everything for now */
2245 return HNAE3_GLOBAL_RESET;
2246 }
2247
2248 return reset_type;
2249 }
2250
hclge_config_rocee_ras_interrupt(struct hclge_dev * hdev,bool en)2251 int hclge_config_rocee_ras_interrupt(struct hclge_dev *hdev, bool en)
2252 {
2253 struct device *dev = &hdev->pdev->dev;
2254 struct hclge_desc desc;
2255 int ret;
2256
2257 if (hdev->ae_dev->dev_version < HNAE3_DEVICE_VERSION_V2 ||
2258 !hnae3_dev_roce_supported(hdev))
2259 return 0;
2260
2261 hclge_cmd_setup_basic_desc(&desc, HCLGE_CONFIG_ROCEE_RAS_INT_EN, false);
2262 if (en) {
2263 /* enable ROCEE hw error interrupts */
2264 desc.data[0] = cpu_to_le32(HCLGE_ROCEE_RAS_NFE_INT_EN);
2265 desc.data[1] = cpu_to_le32(HCLGE_ROCEE_RAS_CE_INT_EN);
2266
2267 hclge_log_and_clear_rocee_ras_error(hdev);
2268 }
2269 desc.data[2] = cpu_to_le32(HCLGE_ROCEE_RAS_NFE_INT_EN_MASK);
2270 desc.data[3] = cpu_to_le32(HCLGE_ROCEE_RAS_CE_INT_EN_MASK);
2271
2272 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
2273 if (ret)
2274 dev_err(dev, "failed(%d) to config ROCEE RAS interrupt\n", ret);
2275
2276 return ret;
2277 }
2278
hclge_handle_rocee_ras_error(struct hnae3_ae_dev * ae_dev)2279 static void hclge_handle_rocee_ras_error(struct hnae3_ae_dev *ae_dev)
2280 {
2281 struct hclge_dev *hdev = ae_dev->priv;
2282 enum hnae3_reset_type reset_type;
2283
2284 if (test_bit(HCLGE_STATE_RST_HANDLING, &hdev->state))
2285 return;
2286
2287 reset_type = hclge_log_and_clear_rocee_ras_error(hdev);
2288 if (reset_type != HNAE3_NONE_RESET)
2289 set_bit(reset_type, &ae_dev->hw_err_reset_req);
2290 }
2291
2292 static const struct hclge_hw_blk hw_blk[] = {
2293 {
2294 .msk = BIT(0),
2295 .name = "IGU_EGU",
2296 .config_err_int = hclge_config_igu_egu_hw_err_int,
2297 }, {
2298 .msk = BIT(1),
2299 .name = "PPP",
2300 .config_err_int = hclge_config_ppp_hw_err_int,
2301 }, {
2302 .msk = BIT(2),
2303 .name = "SSU",
2304 .config_err_int = hclge_config_ssu_hw_err_int,
2305 }, {
2306 .msk = BIT(3),
2307 .name = "PPU",
2308 .config_err_int = hclge_config_ppu_hw_err_int,
2309 }, {
2310 .msk = BIT(4),
2311 .name = "TM",
2312 .config_err_int = hclge_config_tm_hw_err_int,
2313 }, {
2314 .msk = BIT(5),
2315 .name = "COMMON",
2316 .config_err_int = hclge_config_common_hw_err_int,
2317 }, {
2318 .msk = BIT(8),
2319 .name = "MAC",
2320 .config_err_int = hclge_config_mac_err_int,
2321 }, {
2322 /* sentinel */
2323 }
2324 };
2325
hclge_config_all_msix_error(struct hclge_dev * hdev,bool enable)2326 static void hclge_config_all_msix_error(struct hclge_dev *hdev, bool enable)
2327 {
2328 u32 reg_val;
2329
2330 reg_val = hclge_read_dev(&hdev->hw, HCLGE_PF_OTHER_INT_REG);
2331
2332 if (enable)
2333 reg_val |= BIT(HCLGE_VECTOR0_ALL_MSIX_ERR_B);
2334 else
2335 reg_val &= ~BIT(HCLGE_VECTOR0_ALL_MSIX_ERR_B);
2336
2337 hclge_write_dev(&hdev->hw, HCLGE_PF_OTHER_INT_REG, reg_val);
2338 }
2339
hclge_config_nic_hw_error(struct hclge_dev * hdev,bool state)2340 int hclge_config_nic_hw_error(struct hclge_dev *hdev, bool state)
2341 {
2342 const struct hclge_hw_blk *module = hw_blk;
2343 int ret = 0;
2344
2345 hclge_config_all_msix_error(hdev, state);
2346
2347 while (module->name) {
2348 if (module->config_err_int) {
2349 ret = module->config_err_int(hdev, state);
2350 if (ret)
2351 return ret;
2352 }
2353 module++;
2354 }
2355
2356 return ret;
2357 }
2358
hclge_handle_hw_ras_error(struct hnae3_ae_dev * ae_dev)2359 pci_ers_result_t hclge_handle_hw_ras_error(struct hnae3_ae_dev *ae_dev)
2360 {
2361 struct hclge_dev *hdev = ae_dev->priv;
2362 struct device *dev = &hdev->pdev->dev;
2363 u32 status;
2364
2365 if (!test_bit(HCLGE_STATE_SERVICE_INITED, &hdev->state)) {
2366 dev_err(dev,
2367 "Can't recover - RAS error reported during dev init\n");
2368 return PCI_ERS_RESULT_NONE;
2369 }
2370
2371 status = hclge_read_dev(&hdev->hw, HCLGE_RAS_PF_OTHER_INT_STS_REG);
2372 if (status & HCLGE_RAS_REG_NFE_MASK ||
2373 status & HCLGE_RAS_REG_ROCEE_ERR_MASK)
2374 ae_dev->hw_err_reset_req = 0;
2375 else
2376 goto out;
2377
2378 /* Handling Non-fatal HNS RAS errors */
2379 if (status & HCLGE_RAS_REG_NFE_MASK) {
2380 dev_err(dev,
2381 "HNS Non-Fatal RAS error(status=0x%x) identified\n",
2382 status);
2383 hclge_handle_all_ras_errors(hdev);
2384 }
2385
2386 /* Handling Non-fatal Rocee RAS errors */
2387 if (hdev->ae_dev->dev_version >= HNAE3_DEVICE_VERSION_V2 &&
2388 status & HCLGE_RAS_REG_ROCEE_ERR_MASK) {
2389 dev_err(dev, "ROCEE Non-Fatal RAS error identified\n");
2390 hclge_handle_rocee_ras_error(ae_dev);
2391 }
2392
2393 if (ae_dev->hw_err_reset_req)
2394 return PCI_ERS_RESULT_NEED_RESET;
2395
2396 out:
2397 return PCI_ERS_RESULT_RECOVERED;
2398 }
2399
hclge_clear_hw_msix_error(struct hclge_dev * hdev,struct hclge_desc * desc,bool is_mpf,u32 bd_num)2400 static int hclge_clear_hw_msix_error(struct hclge_dev *hdev,
2401 struct hclge_desc *desc, bool is_mpf,
2402 u32 bd_num)
2403 {
2404 if (is_mpf)
2405 desc[0].opcode =
2406 cpu_to_le16(HCLGE_QUERY_CLEAR_ALL_MPF_MSIX_INT);
2407 else
2408 desc[0].opcode = cpu_to_le16(HCLGE_QUERY_CLEAR_ALL_PF_MSIX_INT);
2409
2410 desc[0].flag = cpu_to_le16(HCLGE_COMM_CMD_FLAG_NO_INTR |
2411 HCLGE_COMM_CMD_FLAG_IN);
2412
2413 return hclge_cmd_send(&hdev->hw, &desc[0], bd_num);
2414 }
2415
2416 /* hclge_query_8bd_info: query information about over_8bd_nfe_err
2417 * @hdev: pointer to struct hclge_dev
2418 * @vf_id: Index of the virtual function with error
2419 * @q_id: Physical index of the queue with error
2420 *
2421 * This function get specific index of queue and function which causes
2422 * over_8bd_nfe_err by using command. If vf_id is 0, it means error is
2423 * caused by PF instead of VF.
2424 */
hclge_query_over_8bd_err_info(struct hclge_dev * hdev,u16 * vf_id,u16 * q_id)2425 static int hclge_query_over_8bd_err_info(struct hclge_dev *hdev, u16 *vf_id,
2426 u16 *q_id)
2427 {
2428 struct hclge_query_ppu_pf_other_int_dfx_cmd *req;
2429 struct hclge_desc desc;
2430 int ret;
2431
2432 hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_PPU_PF_OTHER_INT_DFX, true);
2433 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
2434 if (ret)
2435 return ret;
2436
2437 req = (struct hclge_query_ppu_pf_other_int_dfx_cmd *)desc.data;
2438 *vf_id = le16_to_cpu(req->over_8bd_no_fe_vf_id);
2439 *q_id = le16_to_cpu(req->over_8bd_no_fe_qid);
2440
2441 return 0;
2442 }
2443
2444 /* hclge_handle_over_8bd_err: handle MSI-X error named over_8bd_nfe_err
2445 * @hdev: pointer to struct hclge_dev
2446 * @reset_requests: reset level that we need to trigger later
2447 *
2448 * over_8bd_nfe_err is a special MSI-X because it may caused by a VF, in
2449 * that case, we need to trigger VF reset. Otherwise, a PF reset is needed.
2450 */
hclge_handle_over_8bd_err(struct hclge_dev * hdev,unsigned long * reset_requests)2451 static void hclge_handle_over_8bd_err(struct hclge_dev *hdev,
2452 unsigned long *reset_requests)
2453 {
2454 struct device *dev = &hdev->pdev->dev;
2455 u16 vf_id;
2456 u16 q_id;
2457 int ret;
2458
2459 ret = hclge_query_over_8bd_err_info(hdev, &vf_id, &q_id);
2460 if (ret) {
2461 dev_err(dev, "fail(%d) to query over_8bd_no_fe info\n",
2462 ret);
2463 return;
2464 }
2465
2466 dev_err(dev, "PPU_PF_ABNORMAL_INT_ST over_8bd_no_fe found, vport(%u), queue_id(%u)\n",
2467 vf_id, q_id);
2468
2469 if (vf_id) {
2470 if (vf_id >= hdev->num_alloc_vport) {
2471 dev_err(dev, "invalid vport(%u)\n", vf_id);
2472 return;
2473 }
2474
2475 /* If we need to trigger other reset whose level is higher
2476 * than HNAE3_VF_FUNC_RESET, no need to trigger a VF reset
2477 * here.
2478 */
2479 if (*reset_requests != 0)
2480 return;
2481
2482 ret = hclge_inform_reset_assert_to_vf(&hdev->vport[vf_id]);
2483 if (ret)
2484 dev_err(dev, "inform reset to vport(%u) failed %d!\n",
2485 vf_id, ret);
2486 } else {
2487 set_bit(HNAE3_FUNC_RESET, reset_requests);
2488 }
2489 }
2490
2491 /* hclge_handle_mpf_msix_error: handle all main PF MSI-X errors
2492 * @hdev: pointer to struct hclge_dev
2493 * @desc: descriptor for describing the command
2494 * @mpf_bd_num: number of extended command structures
2495 * @reset_requests: record of the reset level that we need
2496 *
2497 * This function handles all the main PF MSI-X errors in the hw register/s
2498 * using command.
2499 */
hclge_handle_mpf_msix_error(struct hclge_dev * hdev,struct hclge_desc * desc,int mpf_bd_num,unsigned long * reset_requests)2500 static int hclge_handle_mpf_msix_error(struct hclge_dev *hdev,
2501 struct hclge_desc *desc,
2502 int mpf_bd_num,
2503 unsigned long *reset_requests)
2504 {
2505 struct device *dev = &hdev->pdev->dev;
2506 __le32 *desc_data;
2507 u32 status;
2508 int ret;
2509 /* query all main PF MSIx errors */
2510 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_CLEAR_ALL_MPF_MSIX_INT,
2511 true);
2512 ret = hclge_cmd_send(&hdev->hw, &desc[0], mpf_bd_num);
2513 if (ret) {
2514 dev_err(dev, "query all mpf msix int cmd failed (%d)\n", ret);
2515 return ret;
2516 }
2517
2518 /* log MAC errors */
2519 desc_data = (__le32 *)&desc[1];
2520 status = le32_to_cpu(*desc_data);
2521 if (status)
2522 hclge_log_error(dev, "MAC_AFIFO_TNL_INT_R",
2523 &hclge_mac_afifo_tnl_int[0], status,
2524 reset_requests);
2525
2526 /* log PPU(RCB) MPF errors */
2527 desc_data = (__le32 *)&desc[5];
2528 status = le32_to_cpu(*(desc_data + 2)) &
2529 HCLGE_PPU_MPF_INT_ST2_MSIX_MASK;
2530 if (status)
2531 dev_err(dev, "PPU_MPF_ABNORMAL_INT_ST2 rx_q_search_miss found [dfx status=0x%x\n]",
2532 status);
2533
2534 /* clear all main PF MSIx errors */
2535 ret = hclge_clear_hw_msix_error(hdev, desc, true, mpf_bd_num);
2536 if (ret)
2537 dev_err(dev, "clear all mpf msix int cmd failed (%d)\n", ret);
2538
2539 return ret;
2540 }
2541
2542 /* hclge_handle_pf_msix_error: handle all PF MSI-X errors
2543 * @hdev: pointer to struct hclge_dev
2544 * @desc: descriptor for describing the command
2545 * @mpf_bd_num: number of extended command structures
2546 * @reset_requests: record of the reset level that we need
2547 *
2548 * This function handles all the PF MSI-X errors in the hw register/s using
2549 * command.
2550 */
hclge_handle_pf_msix_error(struct hclge_dev * hdev,struct hclge_desc * desc,int pf_bd_num,unsigned long * reset_requests)2551 static int hclge_handle_pf_msix_error(struct hclge_dev *hdev,
2552 struct hclge_desc *desc,
2553 int pf_bd_num,
2554 unsigned long *reset_requests)
2555 {
2556 struct device *dev = &hdev->pdev->dev;
2557 __le32 *desc_data;
2558 u32 status;
2559 int ret;
2560
2561 /* query all PF MSIx errors */
2562 hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_CLEAR_ALL_PF_MSIX_INT,
2563 true);
2564 ret = hclge_cmd_send(&hdev->hw, &desc[0], pf_bd_num);
2565 if (ret) {
2566 dev_err(dev, "query all pf msix int cmd failed (%d)\n", ret);
2567 return ret;
2568 }
2569
2570 /* log SSU PF errors */
2571 status = le32_to_cpu(desc[0].data[0]) & HCLGE_SSU_PORT_INT_MSIX_MASK;
2572 if (status)
2573 hclge_log_error(dev, "SSU_PORT_BASED_ERR_INT",
2574 &hclge_ssu_port_based_pf_int[0],
2575 status, reset_requests);
2576
2577 /* read and log PPP PF errors */
2578 desc_data = (__le32 *)&desc[2];
2579 status = le32_to_cpu(*desc_data);
2580 if (status)
2581 hclge_log_error(dev, "PPP_PF_ABNORMAL_INT_ST0",
2582 &hclge_ppp_pf_abnormal_int[0],
2583 status, reset_requests);
2584
2585 /* log PPU(RCB) PF errors */
2586 desc_data = (__le32 *)&desc[3];
2587 status = le32_to_cpu(*desc_data) & HCLGE_PPU_PF_INT_MSIX_MASK;
2588 if (status)
2589 hclge_log_error(dev, "PPU_PF_ABNORMAL_INT_ST",
2590 &hclge_ppu_pf_abnormal_int[0],
2591 status, reset_requests);
2592
2593 status = le32_to_cpu(*desc_data) & HCLGE_PPU_PF_OVER_8BD_ERR_MASK;
2594 if (status)
2595 hclge_handle_over_8bd_err(hdev, reset_requests);
2596
2597 /* clear all PF MSIx errors */
2598 ret = hclge_clear_hw_msix_error(hdev, desc, false, pf_bd_num);
2599 if (ret)
2600 dev_err(dev, "clear all pf msix int cmd failed (%d)\n", ret);
2601
2602 return ret;
2603 }
2604
hclge_handle_all_hw_msix_error(struct hclge_dev * hdev,unsigned long * reset_requests)2605 static int hclge_handle_all_hw_msix_error(struct hclge_dev *hdev,
2606 unsigned long *reset_requests)
2607 {
2608 u32 mpf_bd_num, pf_bd_num, bd_num;
2609 struct hclge_desc *desc;
2610 int ret;
2611
2612 /* query the number of bds for the MSIx int status */
2613 ret = hclge_query_bd_num(hdev, false, &mpf_bd_num, &pf_bd_num);
2614 if (ret)
2615 goto out;
2616
2617 bd_num = max_t(u32, mpf_bd_num, pf_bd_num);
2618 desc = kcalloc(bd_num, sizeof(struct hclge_desc), GFP_KERNEL);
2619 if (!desc)
2620 return -ENOMEM;
2621
2622 ret = hclge_handle_mpf_msix_error(hdev, desc, mpf_bd_num,
2623 reset_requests);
2624 if (ret)
2625 goto msi_error;
2626
2627 memset(desc, 0, bd_num * sizeof(struct hclge_desc));
2628 ret = hclge_handle_pf_msix_error(hdev, desc, pf_bd_num, reset_requests);
2629 if (ret)
2630 goto msi_error;
2631
2632 ret = hclge_handle_mac_tnl(hdev);
2633
2634 msi_error:
2635 kfree(desc);
2636 out:
2637 return ret;
2638 }
2639
hclge_handle_hw_msix_error(struct hclge_dev * hdev,unsigned long * reset_requests)2640 int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
2641 unsigned long *reset_requests)
2642 {
2643 struct device *dev = &hdev->pdev->dev;
2644
2645 if (!test_bit(HCLGE_STATE_SERVICE_INITED, &hdev->state)) {
2646 dev_err(dev,
2647 "failed to handle msix error during dev init\n");
2648 return -EAGAIN;
2649 }
2650
2651 return hclge_handle_all_hw_msix_error(hdev, reset_requests);
2652 }
2653
hclge_handle_mac_tnl(struct hclge_dev * hdev)2654 int hclge_handle_mac_tnl(struct hclge_dev *hdev)
2655 {
2656 struct hclge_mac_tnl_stats mac_tnl_stats;
2657 struct device *dev = &hdev->pdev->dev;
2658 struct hclge_desc desc;
2659 u32 status;
2660 int ret;
2661
2662 /* query and clear mac tnl interruptions */
2663 hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_QUERY_MAC_TNL_INT, true);
2664 ret = hclge_cmd_send(&hdev->hw, &desc, 1);
2665 if (ret) {
2666 dev_err(dev, "failed to query mac tnl int, ret = %d.\n", ret);
2667 return ret;
2668 }
2669
2670 status = le32_to_cpu(desc.data[0]);
2671 if (status) {
2672 /* When mac tnl interrupt occurs, we record current time and
2673 * register status here in a fifo, then clear the status. So
2674 * that if link status changes suddenly at some time, we can
2675 * query them by debugfs.
2676 */
2677 mac_tnl_stats.time = local_clock();
2678 mac_tnl_stats.status = status;
2679 kfifo_put(&hdev->mac_tnl_log, mac_tnl_stats);
2680 ret = hclge_clear_mac_tnl_int(hdev);
2681 if (ret)
2682 dev_err(dev, "failed to clear mac tnl int, ret = %d.\n",
2683 ret);
2684 }
2685
2686 return ret;
2687 }
2688
hclge_handle_all_hns_hw_errors(struct hnae3_ae_dev * ae_dev)2689 void hclge_handle_all_hns_hw_errors(struct hnae3_ae_dev *ae_dev)
2690 {
2691 struct hclge_dev *hdev = ae_dev->priv;
2692 struct device *dev = &hdev->pdev->dev;
2693 u32 mpf_bd_num, pf_bd_num, bd_num;
2694 struct hclge_desc *desc;
2695 u32 status;
2696 int ret;
2697
2698 ae_dev->hw_err_reset_req = 0;
2699 status = hclge_read_dev(&hdev->hw, HCLGE_RAS_PF_OTHER_INT_STS_REG);
2700
2701 /* query the number of bds for the MSIx int status */
2702 ret = hclge_query_bd_num(hdev, false, &mpf_bd_num, &pf_bd_num);
2703 if (ret)
2704 return;
2705
2706 bd_num = max_t(u32, mpf_bd_num, pf_bd_num);
2707 desc = kcalloc(bd_num, sizeof(struct hclge_desc), GFP_KERNEL);
2708 if (!desc)
2709 return;
2710
2711 /* Clear HNS hw errors reported through msix */
2712 memset(&desc[0].data[0], 0xFF, mpf_bd_num * sizeof(struct hclge_desc) -
2713 HCLGE_DESC_NO_DATA_LEN);
2714 ret = hclge_clear_hw_msix_error(hdev, desc, true, mpf_bd_num);
2715 if (ret) {
2716 dev_err(dev, "fail(%d) to clear mpf msix int during init\n",
2717 ret);
2718 goto msi_error;
2719 }
2720
2721 memset(&desc[0].data[0], 0xFF, pf_bd_num * sizeof(struct hclge_desc) -
2722 HCLGE_DESC_NO_DATA_LEN);
2723 ret = hclge_clear_hw_msix_error(hdev, desc, false, pf_bd_num);
2724 if (ret) {
2725 dev_err(dev, "fail(%d) to clear pf msix int during init\n",
2726 ret);
2727 goto msi_error;
2728 }
2729
2730 /* Handle Non-fatal HNS RAS errors */
2731 if (status & HCLGE_RAS_REG_NFE_MASK) {
2732 dev_err(dev, "HNS hw error(RAS) identified during init\n");
2733 hclge_handle_all_ras_errors(hdev);
2734 }
2735
2736 msi_error:
2737 kfree(desc);
2738 }
2739
hclge_find_error_source(struct hclge_dev * hdev)2740 bool hclge_find_error_source(struct hclge_dev *hdev)
2741 {
2742 u32 msix_src_flag, hw_err_src_flag;
2743
2744 msix_src_flag = hclge_read_dev(&hdev->hw, HCLGE_MISC_VECTOR_INT_STS) &
2745 HCLGE_VECTOR0_REG_MSIX_MASK;
2746
2747 hw_err_src_flag = hclge_read_dev(&hdev->hw,
2748 HCLGE_RAS_PF_OTHER_INT_STS_REG) &
2749 HCLGE_RAS_REG_ERR_MASK;
2750
2751 return msix_src_flag || hw_err_src_flag;
2752 }
2753
hclge_handle_occurred_error(struct hclge_dev * hdev)2754 void hclge_handle_occurred_error(struct hclge_dev *hdev)
2755 {
2756 struct hnae3_ae_dev *ae_dev = pci_get_drvdata(hdev->pdev);
2757
2758 if (hclge_find_error_source(hdev))
2759 hclge_handle_error_info_log(ae_dev);
2760 }
2761
2762 static void
hclge_handle_error_type_reg_log(struct device * dev,struct hclge_mod_err_info * mod_info,struct hclge_type_reg_err_info * type_reg_info)2763 hclge_handle_error_type_reg_log(struct device *dev,
2764 struct hclge_mod_err_info *mod_info,
2765 struct hclge_type_reg_err_info *type_reg_info)
2766 {
2767 #define HCLGE_ERR_TYPE_MASK 0x7F
2768 #define HCLGE_ERR_TYPE_IS_RAS_OFFSET 7
2769
2770 u8 mod_id, total_module, type_id, total_type, i, is_ras;
2771 u8 index_module = MODULE_NONE;
2772 u8 index_type = NONE_ERROR;
2773
2774 mod_id = mod_info->mod_id;
2775 type_id = type_reg_info->type_id & HCLGE_ERR_TYPE_MASK;
2776 is_ras = type_reg_info->type_id >> HCLGE_ERR_TYPE_IS_RAS_OFFSET;
2777
2778 total_module = ARRAY_SIZE(hclge_hw_module_id_st);
2779 total_type = ARRAY_SIZE(hclge_hw_type_id_st);
2780
2781 for (i = 0; i < total_module; i++) {
2782 if (mod_id == hclge_hw_module_id_st[i].module_id) {
2783 index_module = i;
2784 break;
2785 }
2786 }
2787
2788 for (i = 0; i < total_type; i++) {
2789 if (type_id == hclge_hw_type_id_st[i].type_id) {
2790 index_type = i;
2791 break;
2792 }
2793 }
2794
2795 if (index_module != MODULE_NONE && index_type != NONE_ERROR)
2796 dev_err(dev,
2797 "found %s %s, is %s error.\n",
2798 hclge_hw_module_id_st[index_module].msg,
2799 hclge_hw_type_id_st[index_type].msg,
2800 is_ras ? "ras" : "msix");
2801 else
2802 dev_err(dev,
2803 "unknown module[%u] or type[%u].\n", mod_id, type_id);
2804
2805 dev_err(dev, "reg_value:\n");
2806 for (i = 0; i < type_reg_info->reg_num; i++)
2807 dev_err(dev, "0x%08x\n", type_reg_info->hclge_reg[i]);
2808 }
2809
hclge_handle_error_module_log(struct hnae3_ae_dev * ae_dev,const u32 * buf,u32 buf_size)2810 static void hclge_handle_error_module_log(struct hnae3_ae_dev *ae_dev,
2811 const u32 *buf, u32 buf_size)
2812 {
2813 struct hclge_type_reg_err_info *type_reg_info;
2814 struct hclge_dev *hdev = ae_dev->priv;
2815 struct device *dev = &hdev->pdev->dev;
2816 struct hclge_mod_err_info *mod_info;
2817 struct hclge_sum_err_info *sum_info;
2818 u8 mod_num, err_num, i;
2819 u32 offset = 0;
2820
2821 sum_info = (struct hclge_sum_err_info *)&buf[offset++];
2822 if (sum_info->reset_type &&
2823 sum_info->reset_type != HNAE3_NONE_RESET)
2824 set_bit(sum_info->reset_type, &ae_dev->hw_err_reset_req);
2825 mod_num = sum_info->mod_num;
2826
2827 while (mod_num--) {
2828 if (offset >= buf_size) {
2829 dev_err(dev, "The offset(%u) exceeds buf's size(%u).\n",
2830 offset, buf_size);
2831 return;
2832 }
2833 mod_info = (struct hclge_mod_err_info *)&buf[offset++];
2834 err_num = mod_info->err_num;
2835
2836 for (i = 0; i < err_num; i++) {
2837 if (offset >= buf_size) {
2838 dev_err(dev,
2839 "The offset(%u) exceeds buf size(%u).\n",
2840 offset, buf_size);
2841 return;
2842 }
2843
2844 type_reg_info = (struct hclge_type_reg_err_info *)
2845 &buf[offset++];
2846 hclge_handle_error_type_reg_log(dev, mod_info,
2847 type_reg_info);
2848
2849 offset += type_reg_info->reg_num;
2850 }
2851 }
2852 }
2853
hclge_query_all_err_bd_num(struct hclge_dev * hdev,u32 * bd_num)2854 static int hclge_query_all_err_bd_num(struct hclge_dev *hdev, u32 *bd_num)
2855 {
2856 struct device *dev = &hdev->pdev->dev;
2857 struct hclge_desc desc_bd;
2858 int ret;
2859
2860 hclge_cmd_setup_basic_desc(&desc_bd, HCLGE_QUERY_ALL_ERR_BD_NUM, true);
2861 ret = hclge_cmd_send(&hdev->hw, &desc_bd, 1);
2862 if (ret) {
2863 dev_err(dev, "failed to query error bd_num, ret = %d.\n", ret);
2864 return ret;
2865 }
2866
2867 *bd_num = le32_to_cpu(desc_bd.data[0]);
2868 if (!(*bd_num)) {
2869 dev_err(dev, "The value of bd_num is 0!\n");
2870 return -EINVAL;
2871 }
2872
2873 return 0;
2874 }
2875
hclge_query_all_err_info(struct hclge_dev * hdev,struct hclge_desc * desc,u32 bd_num)2876 static int hclge_query_all_err_info(struct hclge_dev *hdev,
2877 struct hclge_desc *desc, u32 bd_num)
2878 {
2879 struct device *dev = &hdev->pdev->dev;
2880 int ret;
2881
2882 hclge_cmd_setup_basic_desc(desc, HCLGE_QUERY_ALL_ERR_INFO, true);
2883 ret = hclge_cmd_send(&hdev->hw, desc, bd_num);
2884 if (ret)
2885 dev_err(dev, "failed to query error info, ret = %d.\n", ret);
2886
2887 return ret;
2888 }
2889
hclge_handle_error_info_log(struct hnae3_ae_dev * ae_dev)2890 int hclge_handle_error_info_log(struct hnae3_ae_dev *ae_dev)
2891 {
2892 u32 bd_num, desc_len, buf_len, buf_size, i;
2893 struct hclge_dev *hdev = ae_dev->priv;
2894 struct hclge_desc *desc;
2895 __le32 *desc_data;
2896 u32 *buf;
2897 int ret;
2898
2899 ret = hclge_query_all_err_bd_num(hdev, &bd_num);
2900 if (ret)
2901 goto out;
2902
2903 desc_len = bd_num * sizeof(struct hclge_desc);
2904 desc = kzalloc(desc_len, GFP_KERNEL);
2905 if (!desc) {
2906 ret = -ENOMEM;
2907 goto out;
2908 }
2909
2910 ret = hclge_query_all_err_info(hdev, desc, bd_num);
2911 if (ret)
2912 goto err_desc;
2913
2914 buf_len = bd_num * sizeof(struct hclge_desc) - HCLGE_DESC_NO_DATA_LEN;
2915 buf_size = buf_len / sizeof(u32);
2916
2917 desc_data = kzalloc(buf_len, GFP_KERNEL);
2918 if (!desc_data) {
2919 ret = -ENOMEM;
2920 goto err_desc;
2921 }
2922
2923 buf = kzalloc(buf_len, GFP_KERNEL);
2924 if (!buf) {
2925 ret = -ENOMEM;
2926 goto err_buf_alloc;
2927 }
2928
2929 memcpy(desc_data, &desc[0].data[0], buf_len);
2930 for (i = 0; i < buf_size; i++)
2931 buf[i] = le32_to_cpu(desc_data[i]);
2932
2933 hclge_handle_error_module_log(ae_dev, buf, buf_size);
2934 kfree(buf);
2935
2936 err_buf_alloc:
2937 kfree(desc_data);
2938 err_desc:
2939 kfree(desc);
2940 out:
2941 return ret;
2942 }
2943