Report pool suspended due to MMP

When the pool is suspended, record whether it was due to an I/O error or
due to MMP writes failing to succeed within the required time.

Change spa_suspended from uint8_t to zio_suspend_reason_t to store the
reason.

When userspace queries pool status via spa_tryimport(), report the
reason the pool was suspended in a new key,
ZPOOL_CONFIG_SUSPENDED_REASON.

In libzfs, when interpreting the returned config nvlist, report
suspension due to MMP with a new pool status enum value,
ZPOOL_STATUS_IO_FAILURE_MMP.

In status_callback(), which generates and emits the message when 'zpool
status' is executed, add a case to print an appropriate message for the
new pool status enum value.

Reviewed-by: George Melikov <[email protected]>
Reviewed-by: Giuseppe Di Natale <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Tony Hutter <[email protected]>
Signed-off-by: Olaf Faaland <[email protected]>
Closes #7296

4 files changed, 12 insertions, 8 deletions

diff --git a/module/zfs/mmp.c b/module/zfs/mmp.c
index a08e0864e..fc8346dc3 100644
--- a/module/zfs/mmp.c
+++ b/module/zfs/mmp.c
@@ -519,7 +519,7 @@ mmp_thread(spa_t *spa)
 			    "succeeded in over %llus; suspending pool",
 			    spa_name(spa),
 			    NSEC2SEC(start - mmp->mmp_last_write));
-			zio_suspend(spa, NULL);
+			zio_suspend(spa, NULL, ZIO_SUSPEND_MMP);
 		}
 
 		if (multihost && !suspended)
diff --git a/module/zfs/spa.c b/module/zfs/spa.c
index a7a2f6281..561f4d04b 100644
--- a/module/zfs/spa.c
+++ b/module/zfs/spa.c
@@ -3778,10 +3778,14 @@ spa_get_stats(const char *name, nvlist_t **config,
 			    ZPOOL_CONFIG_ERRCOUNT,
 			    spa_get_errlog_size(spa)) == 0);
 
-			if (spa_suspended(spa))
+			if (spa_suspended(spa)) {
 				VERIFY(nvlist_add_uint64(*config,
 				    ZPOOL_CONFIG_SUSPENDED,
 				    spa->spa_failmode) == 0);
+				VERIFY(nvlist_add_uint64(*config,
+				    ZPOOL_CONFIG_SUSPENDED_REASON,
+				    spa->spa_suspended) == 0);
+			}
 
 			spa_add_spares(spa, *config);
 			spa_add_l2cache(spa, *config);
@@ -6969,7 +6973,7 @@ spa_sync(spa_t *spa, uint64_t txg)
 
 		if (error == 0)
 			break;
-		zio_suspend(spa, NULL);
+		zio_suspend(spa, NULL, ZIO_SUSPEND_IOERR);
 		zio_resume_wait(spa);
 	}
 	dmu_tx_commit(tx);
diff --git a/module/zfs/spa_misc.c b/module/zfs/spa_misc.c
index 3787e010f..e92c39482 100644
--- a/module/zfs/spa_misc.c
+++ b/module/zfs/spa_misc.c
@@ -1691,7 +1691,7 @@ spa_get_failmode(spa_t *spa)
 boolean_t
 spa_suspended(spa_t *spa)
 {
-	return (spa->spa_suspended);
+	return (spa->spa_suspended != ZIO_SUSPEND_NONE);
 }
 
 uint64_t
diff --git a/module/zfs/zio.c b/module/zfs/zio.c
index cd0a473e0..9a465e1be 100644
--- a/module/zfs/zio.c
+++ b/module/zfs/zio.c
@@ -1864,7 +1864,7 @@ zio_reexecute(zio_t *pio)
 }
 
 void
-zio_suspend(spa_t *spa, zio_t *zio)
+zio_suspend(spa_t *spa, zio_t *zio, zio_suspend_reason_t reason)
 {
 	if (spa_get_failmode(spa) == ZIO_FAILURE_MODE_PANIC)
 		fm_panic("Pool '%s' has encountered an uncorrectable I/O "
@@ -1883,7 +1883,7 @@ zio_suspend(spa_t *spa, zio_t *zio)
 		    ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
 		    ZIO_FLAG_GODFATHER);
 
-	spa->spa_suspended = B_TRUE;
+	spa->spa_suspended = reason;
 
 	if (zio != NULL) {
 		ASSERT(!(zio->io_flags & ZIO_FLAG_GODFATHER));
@@ -1906,7 +1906,7 @@ zio_resume(spa_t *spa)
 	 * Reexecute all previously suspended i/o.
 	 */
 	mutex_enter(&spa->spa_suspend_lock);
-	spa->spa_suspended = B_FALSE;
+	spa->spa_suspended = ZIO_SUSPEND_NONE;
 	cv_broadcast(&spa->spa_suspend_cv);
 	pio = spa->spa_suspend_zio_root;
 	spa->spa_suspend_zio_root = NULL;
@@ -3975,7 +3975,7 @@ zio_done(zio_t *zio)
 			 * We'd fail again if we reexecuted now, so suspend
 			 * until conditions improve (e.g. device comes online).
 			 */
-			zio_suspend(zio->io_spa, zio);
+			zio_suspend(zio->io_spa, zio, ZIO_SUSPEND_IOERR);
 		} else {
 			/*
 			 * Reexecution is potentially a huge amount of work.