git: 05cf85775b2a - stable/14 - mpi3mr: Handle Insufficient Power Fault Code

From: Warner Losh <imp_at_FreeBSD.org>
Date: Wed, 30 Apr 2025 17:21:52 UTC
The branch stable/14 has been updated by imp:

URL: https://cgit.FreeBSD.org/src/commit/?id=05cf85775b2a120211ce5530acffd3b335543593

commit 05cf85775b2a120211ce5530acffd3b335543593
Author:     Chandrakanth patil <chandrakanth.patil@broadcom.com>
AuthorDate: 2025-04-27 23:35:23 +0000
Commit:     Warner Losh <imp@FreeBSD.org>
CommitDate: 2025-04-30 17:05:53 +0000

    mpi3mr: Handle Insufficient Power Fault Code
    
    The driver now checks for insufficient power faults during the load
    phase and immediately fails initialization instead of retrying.
    
    Additionally, if an insufficient power fault is detected by the watchdog
    after the controller is up, the controller is marked as unrecoverable
    instead of triggering a reset.
    
    This improves fault handling and avoids unnecessary recovery attempts
    in low-power conditions.
    
    Reviewed by:    ssaxena, imp
    Differential Revision:  https://reviews.freebsd.org/D49747
    
    (cherry picked from commit 116c8b18a2b5278df0d1982a683193c3dba6f30c)
---
 sys/dev/mpi3mr/mpi3mr.c | 50 ++++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 39 insertions(+), 11 deletions(-)

diff --git a/sys/dev/mpi3mr/mpi3mr.c b/sys/dev/mpi3mr/mpi3mr.c
index bfb032a86c5b..1611f2c4aceb 100644
--- a/sys/dev/mpi3mr/mpi3mr.c
+++ b/sys/dev/mpi3mr/mpi3mr.c
@@ -2785,18 +2785,37 @@ retry_init:
         }
 
         if (ioc_state != MRIOC_STATE_RESET) {
-                mpi3mr_print_fault_info(sc);
-		 mpi3mr_dprint(sc, MPI3MR_ERROR, "issuing soft reset to bring to reset state\n");
-                 retval = mpi3mr_issue_reset(sc,
-                     MPI3_SYSIF_HOST_DIAG_RESET_ACTION_SOFT_RESET,
-                     MPI3MR_RESET_FROM_BRINGUP);
-                if (retval) {
-                        mpi3mr_dprint(sc, MPI3MR_ERROR,
-                            "%s :Failed to soft reset IOC, error 0x%d\n",
-                            __func__, retval);
+		if (ioc_state == MRIOC_STATE_FAULT) {
+			mpi3mr_print_fault_info(sc);
+
+			U32 fault = mpi3mr_regread(sc, MPI3_SYSIF_FAULT_OFFSET) &
+						   MPI3_SYSIF_FAULT_CODE_MASK;
+			if (fault == MPI3_SYSIF_FAULT_CODE_INSUFFICIENT_PCI_SLOT_POWER)
+				mpi3mr_dprint(sc, MPI3MR_INFO,
+					      "controller faulted due to insufficient power. "
+					      "try by connecting it in a different slot\n");
+				goto err;
+
+			U32 host_diagnostic;
+			timeout = MPI3_SYSIF_DIAG_SAVE_TIMEOUT * 10;
+			do {
+				host_diagnostic = mpi3mr_regread(sc, MPI3_SYSIF_HOST_DIAG_OFFSET);
+				if (!(host_diagnostic & MPI3_SYSIF_HOST_DIAG_SAVE_IN_PROGRESS))
+					break;
+				DELAY(100 * 1000);
+			} while (--timeout);
+		}
+		mpi3mr_dprint(sc, MPI3MR_ERROR, "issuing soft reset to bring to reset state\n");
+		retval = mpi3mr_issue_reset(sc,
+		     MPI3_SYSIF_HOST_DIAG_RESET_ACTION_SOFT_RESET,
+		     MPI3MR_RESET_FROM_BRINGUP);
+		if (retval) {
+			mpi3mr_dprint(sc, MPI3MR_ERROR,
+			    "%s :Failed to soft reset IOC, error 0x%d\n",
+			    __func__, retval);
 			goto err_retry;
-                }
-        }
+		}
+	}
 
 	ioc_state = mpi3mr_get_iocstate(sc);
 
@@ -3166,6 +3185,15 @@ mpi3mr_watchdog_thread(void *arg)
 				sc->unrecoverable = 1;
 				break;
 			}
+
+			if (fault == MPI3_SYSIF_FAULT_CODE_INSUFFICIENT_PCI_SLOT_POWER) {
+				mpi3mr_dprint(sc, MPI3MR_INFO,
+					      "controller faulted due to insufficient power, marking"
+					      " controller as unrecoverable\n");
+				sc->unrecoverable = 1;
+				break;
+			}
+
 			if ((fault == MPI3_SYSIF_FAULT_CODE_DIAG_FAULT_RESET)
 			    || (fault == MPI3_SYSIF_FAULT_CODE_SOFT_RESET_IN_PROGRESS)
 			    || (sc->reset_in_progress))