git: 5e9b7d0e0e0b - stable/14 - nvme: Give up when we've failed

The branch stable/14 has been updated by imp:

URL: https://cgit.FreeBSD.org/src/commit/?id=5e9b7d0e0e0bd4cba29de33569ab0ca86763f635

commit 5e9b7d0e0e0bd4cba29de33569ab0ca86763f635
Author:     Warner Losh <imp@FreeBSD.org>
AuthorDate: 2023-09-28 20:46:00 +0000
Commit:     Warner Losh <imp@FreeBSD.org>
CommitDate: 2023-09-28 21:05:14 +0000

    nvme: Give up when we've failed
    
    Normally, we poll the device every so often to see if commands have
    timed out. However, we'll go into the recovery state as part of failing
    the drive. To account for all possibilties, if we're failed when we get
    into the polling function, just stop polling: Party is over.
    
    Sponsored by:           Netflix
    
    (cherry picked from commit d95431624f934fe4740211738fc787808005b14e)
---
 sys/dev/nvme/nvme_qpair.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/sys/dev/nvme/nvme_qpair.c b/sys/dev/nvme/nvme_qpair.c
index b256c4713c8d..4e37aa0e1020 100644
--- a/sys/dev/nvme/nvme_qpair.c
+++ b/sys/dev/nvme/nvme_qpair.c
@@ -1011,6 +1011,17 @@ nvme_qpair_timeout(void *arg)
 
 	mtx_assert(&qpair->recovery, MA_OWNED);
 
+	/*
+	 * If the controller has failed, give up. We're never going to change
+	 * state from a failed controller: no further transactions are possible.
+	 * We go ahead and let the timeout expire in many cases for simplicity.
+	 */
+	if (qpair->ctrlr->is_failed) {
+		nvme_printf(ctrlr, "Controller failed, giving up\n");
+		qpair->timer_armed = false;
+		return;
+	}
+
 	switch (qpair->recovery_state) {
 	case RECOVERY_NONE:
 		/*
@@ -1094,8 +1105,8 @@ nvme_qpair_timeout(void *arg)
 		idle = false;			/* We want to keep polling */
 		break;
 	case RECOVERY_WAITING:
-		nvme_printf(ctrlr, "waiting for reset to complete\n");
-		idle = false;			/* We want to keep polling */
+		nvme_printf(ctrlr, "Waiting for reset to complete\n");
+		idle = false;		/* We want to keep polling */
 		break;
 	}