diff options
author | Bryan O'Sullivan <bos@pathscale.com> | 2007-03-15 17:44:55 -0400 |
---|---|---|
committer | Roland Dreier <rolandd@cisco.com> | 2007-04-18 23:20:55 -0400 |
commit | 8ec1077b35359c973f4b1de7c516be570a6df495 (patch) | |
tree | 21ba8b8099034a35a1fb4f8bf301223b1be4af73 /drivers/infiniband/hw/ipath/ipath_intr.c | |
parent | 6f5c407460bba332d6bee52e19f2305539395511 (diff) |
IB/ipath: Change packet problems vs chip errors handling and reporting
Some types of packet errors are moderately common with longer IB
cables and large clusters, and are not reported with prints by other
IB HCA drivers. This suppresses those messages unless the new
__IPATH_ERRPKTDBG bit is set in ipath_debug. Reporting of temporarily
disabled frequent error interrupts was also made clearer
We also distinguish between chip errors, and bad packets sent or
received in the wording of the messages.
Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Bryan O'Sullivan <bryan.osullivan@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Diffstat (limited to 'drivers/infiniband/hw/ipath/ipath_intr.c')
-rw-r--r-- | drivers/infiniband/hw/ipath/ipath_intr.c | 57 |
1 files changed, 40 insertions, 17 deletions
diff --git a/drivers/infiniband/hw/ipath/ipath_intr.c b/drivers/infiniband/hw/ipath/ipath_intr.c index 72b9e279d19..037b8e27642 100644 --- a/drivers/infiniband/hw/ipath/ipath_intr.c +++ b/drivers/infiniband/hw/ipath/ipath_intr.c | |||
@@ -403,10 +403,13 @@ static void handle_supp_msgs(struct ipath_devdata *dd, | |||
403 | * happens so often we never want to count it. | 403 | * happens so often we never want to count it. |
404 | */ | 404 | */ |
405 | if (dd->ipath_lasterror & ~INFINIPATH_E_IBSTATUSCHANGED) { | 405 | if (dd->ipath_lasterror & ~INFINIPATH_E_IBSTATUSCHANGED) { |
406 | ipath_decode_err(msg, sizeof msg, dd->ipath_lasterror & | 406 | int iserr; |
407 | ~INFINIPATH_E_IBSTATUSCHANGED); | 407 | iserr = ipath_decode_err(msg, sizeof msg, |
408 | dd->ipath_lasterror & | ||
409 | ~INFINIPATH_E_IBSTATUSCHANGED); | ||
408 | if (dd->ipath_lasterror & | 410 | if (dd->ipath_lasterror & |
409 | ~(INFINIPATH_E_RRCVEGRFULL | INFINIPATH_E_RRCVHDRFULL)) | 411 | ~(INFINIPATH_E_RRCVEGRFULL | |
412 | INFINIPATH_E_RRCVHDRFULL | INFINIPATH_E_PKTERRS)) | ||
410 | ipath_dev_err(dd, "Suppressed %u messages for " | 413 | ipath_dev_err(dd, "Suppressed %u messages for " |
411 | "fast-repeating errors (%s) (%llx)\n", | 414 | "fast-repeating errors (%s) (%llx)\n", |
412 | supp_msgs, msg, | 415 | supp_msgs, msg, |
@@ -420,8 +423,13 @@ static void handle_supp_msgs(struct ipath_devdata *dd, | |||
420 | * them. So only complain about these at debug | 423 | * them. So only complain about these at debug |
421 | * level. | 424 | * level. |
422 | */ | 425 | */ |
423 | ipath_dbg("Suppressed %u messages for %s\n", | 426 | if (iserr) |
424 | supp_msgs, msg); | 427 | ipath_dbg("Suppressed %u messages for %s\n", |
428 | supp_msgs, msg); | ||
429 | else | ||
430 | ipath_cdbg(ERRPKT, | ||
431 | "Suppressed %u messages for %s\n", | ||
432 | supp_msgs, msg); | ||
425 | } | 433 | } |
426 | } | 434 | } |
427 | } | 435 | } |
@@ -462,7 +470,7 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
462 | { | 470 | { |
463 | char msg[512]; | 471 | char msg[512]; |
464 | u64 ignore_this_time = 0; | 472 | u64 ignore_this_time = 0; |
465 | int i; | 473 | int i, iserr = 0; |
466 | int chkerrpkts = 0, noprint = 0; | 474 | int chkerrpkts = 0, noprint = 0; |
467 | unsigned supp_msgs; | 475 | unsigned supp_msgs; |
468 | 476 | ||
@@ -502,6 +510,7 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
502 | } | 510 | } |
503 | 511 | ||
504 | if (supp_msgs == 250000) { | 512 | if (supp_msgs == 250000) { |
513 | int s_iserr; | ||
505 | /* | 514 | /* |
506 | * It's not entirely reasonable assuming that the errors set | 515 | * It's not entirely reasonable assuming that the errors set |
507 | * in the last clear period are all responsible for the | 516 | * in the last clear period are all responsible for the |
@@ -511,17 +520,17 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
511 | dd->ipath_maskederrs |= dd->ipath_lasterror | errs; | 520 | dd->ipath_maskederrs |= dd->ipath_lasterror | errs; |
512 | ipath_write_kreg(dd, dd->ipath_kregs->kr_errormask, | 521 | ipath_write_kreg(dd, dd->ipath_kregs->kr_errormask, |
513 | ~dd->ipath_maskederrs); | 522 | ~dd->ipath_maskederrs); |
514 | ipath_decode_err(msg, sizeof msg, | 523 | s_iserr = ipath_decode_err(msg, sizeof msg, |
515 | (dd->ipath_maskederrs & ~dd-> | 524 | (dd->ipath_maskederrs & ~dd-> |
516 | ipath_ignorederrs)); | 525 | ipath_ignorederrs)); |
517 | 526 | ||
518 | if ((dd->ipath_maskederrs & ~dd->ipath_ignorederrs) & | 527 | if ((dd->ipath_maskederrs & ~dd->ipath_ignorederrs) & |
519 | ~(INFINIPATH_E_RRCVEGRFULL | INFINIPATH_E_RRCVHDRFULL)) | 528 | ~(INFINIPATH_E_RRCVEGRFULL | |
520 | ipath_dev_err(dd, "Disabling error(s) %llx because " | 529 | INFINIPATH_E_RRCVHDRFULL | INFINIPATH_E_PKTERRS)) |
521 | "occurring too frequently (%s)\n", | 530 | ipath_dev_err(dd, "Temporarily disabling " |
522 | (unsigned long long) | 531 | "error(s) %llx reporting; too frequent (%s)\n", |
523 | (dd->ipath_maskederrs & | 532 | (unsigned long long) (dd->ipath_maskederrs & |
524 | ~dd->ipath_ignorederrs), msg); | 533 | ~dd->ipath_ignorederrs), msg); |
525 | else { | 534 | else { |
526 | /* | 535 | /* |
527 | * rcvegrfull and rcvhdrqfull are "normal", | 536 | * rcvegrfull and rcvhdrqfull are "normal", |
@@ -530,8 +539,15 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
530 | * processing them. So only complain about | 539 | * processing them. So only complain about |
531 | * these at debug level. | 540 | * these at debug level. |
532 | */ | 541 | */ |
533 | ipath_dbg("Disabling frequent queue full errors " | 542 | if (s_iserr) |
534 | "(%s)\n", msg); | 543 | ipath_dbg("Temporarily disabling reporting " |
544 | "too frequent queue full errors (%s)\n", | ||
545 | msg); | ||
546 | else | ||
547 | ipath_cdbg(ERRPKT, | ||
548 | "Temporarily disabling reporting too" | ||
549 | " frequent packet errors (%s)\n", | ||
550 | msg); | ||
535 | } | 551 | } |
536 | 552 | ||
537 | /* | 553 | /* |
@@ -589,6 +605,8 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
589 | ipath_stats.sps_crcerrs++; | 605 | ipath_stats.sps_crcerrs++; |
590 | chkerrpkts = 1; | 606 | chkerrpkts = 1; |
591 | } | 607 | } |
608 | iserr = errs & ~(E_SUM_PKTERRS | INFINIPATH_E_PKTERRS); | ||
609 | |||
592 | 610 | ||
593 | /* | 611 | /* |
594 | * We don't want to print these two as they happen, or we can make | 612 | * We don't want to print these two as they happen, or we can make |
@@ -677,8 +695,13 @@ static int handle_errors(struct ipath_devdata *dd, ipath_err_t errs) | |||
677 | *dd->ipath_statusp &= ~IPATH_STATUS_IB_CONF; | 695 | *dd->ipath_statusp &= ~IPATH_STATUS_IB_CONF; |
678 | } | 696 | } |
679 | 697 | ||
680 | if (!noprint && *msg) | 698 | if (!noprint && *msg) { |
681 | ipath_dev_err(dd, "%s error\n", msg); | 699 | if (iserr) |
700 | ipath_dev_err(dd, "%s error\n", msg); | ||
701 | else | ||
702 | dev_info(&dd->pcidev->dev, "%s packet problems\n", | ||
703 | msg); | ||
704 | } | ||
682 | if (dd->ipath_state_wanted & dd->ipath_flags) { | 705 | if (dd->ipath_state_wanted & dd->ipath_flags) { |
683 | ipath_cdbg(VERBOSE, "driver wanted state %x, iflags now %x, " | 706 | ipath_cdbg(VERBOSE, "driver wanted state %x, iflags now %x, " |
684 | "waking\n", dd->ipath_state_wanted, | 707 | "waking\n", dd->ipath_state_wanted, |