76 lines
2.4 KiB
Diff
76 lines
2.4 KiB
Diff
|
From 245a547c61a9356cdb7dba0032c09ad58c17143b Mon Sep 17 00:00:00 2001
|
||
|
|
From: Mike Christie <michael.christie@oracle.com>
|
||
|
|
Date: Fri, 5 Nov 2021 16:33:20 -0500
|
||
|
|
Subject: [PATCH] iscsi sysfs: check state before onlining devs
|
||
|
|
|
||
|
|
In 5.6, the commit:
|
||
|
|
|
||
|
|
commit 0ab710458da113a71c461c4df27e7f1353d9f864
|
||
|
|
Author: Bharath Ravi <rbharath@google.com>
|
||
|
|
Date: Sat Jan 25 01:19:25 2020 -0500
|
||
|
|
|
||
|
|
scsi: iscsi: Perform connection failure entirely in kernel space
|
||
|
|
|
||
|
|
made it so the kernel can start the recovery process. This means that
|
||
|
|
after the start conn operation the kernel could set the device into the
|
||
|
|
block stated. We can then hit a race where iscsid has done start conn,
|
||
|
|
and is calling session_online_devs but the kernel has hit an issue and is
|
||
|
|
now setting the device's to blocked.
|
||
|
|
|
||
|
|
This adds a check for if the device is in the offline state before trying
|
||
|
|
to set the state to running.
|
||
|
|
---
|
||
|
|
usr/iscsi_sysfs.c | 25 ++++++++++++++++++++++++-
|
||
|
|
1 file changed, 24 insertions(+), 1 deletion(-)
|
||
|
|
|
||
|
|
diff --git a/usr/iscsi_sysfs.c b/usr/iscsi_sysfs.c
|
||
|
|
index abefde2..7bb834a 100644
|
||
|
|
--- a/usr/iscsi_sysfs.c
|
||
|
|
+++ b/usr/iscsi_sysfs.c
|
||
|
|
@@ -1929,18 +1929,41 @@ void iscsi_sysfs_set_queue_depth(void *data, int hostno, int target, int lun)
|
||
|
|
void iscsi_sysfs_set_device_online(__attribute__((unused))void *data,
|
||
|
|
int hostno, int target, int lun)
|
||
|
|
{
|
||
|
|
- char *write_buf = "running\n";
|
||
|
|
+ char *write_buf = "running\n", *state;
|
||
|
|
char id[NAME_SIZE];
|
||
|
|
int err;
|
||
|
|
|
||
|
|
snprintf(id, sizeof(id), "%d:0:%d:%d", hostno, target, lun);
|
||
|
|
log_debug(4, "online device %s", id);
|
||
|
|
|
||
|
|
+ state = sysfs_get_value(id, SCSI_SUBSYS, "state");
|
||
|
|
+ if (!state) {
|
||
|
|
+ log_error("Could not read state for LUN %s\n", id);
|
||
|
|
+ goto set_state;
|
||
|
|
+ }
|
||
|
|
+
|
||
|
|
+ if (!strcmp(state, "running"))
|
||
|
|
+ goto done;
|
||
|
|
+ /*
|
||
|
|
+ * The kernel can start to perform session level recovery cleanup
|
||
|
|
+ * any time after the conn start call, so we only want to change the
|
||
|
|
+ * state if we are in one of the offline states.
|
||
|
|
+ */
|
||
|
|
+ if (strcmp(state, "offline") && strcmp(state, "transport-offline")) {
|
||
|
|
+ log_debug(4, "Dev not offline. Skip onlining %s", id);
|
||
|
|
+ goto done;
|
||
|
|
+ }
|
||
|
|
+
|
||
|
|
+set_state:
|
||
|
|
err = sysfs_set_param(id, SCSI_SUBSYS, "state", write_buf,
|
||
|
|
strlen(write_buf));
|
||
|
|
if (err && err != EINVAL)
|
||
|
|
/* we should read the state */
|
||
|
|
log_error("Could not online LUN %d err %d.", lun, err);
|
||
|
|
+
|
||
|
|
+done:
|
||
|
|
+ if (state)
|
||
|
|
+ free(state);
|
||
|
|
}
|
||
|
|
|
||
|
|
void iscsi_sysfs_rescan_device(__attribute__((unused))void *data,
|
||
|
|
--
|
||
|
|
2.27.0
|
||
|
|
|