fix mdadm blocking caused by incorrect stop of inactive MD devices during shutdown

backport-shutdown-get-only-active-md-arrays.patch

@@ -0,0 +1,72 @@
+From 3a3b022d2cc112803ea7b9beea98bbcad110368a Mon Sep 17 00:00:00 2001
+From: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
+Date: Tue, 29 Mar 2022 12:49:54 +0200
+Subject: [PATCH] shutdown: get only active md arrays.
+
+Current md_list_get() implementation filters all block devices, started from
+"md*". This is ambiguous because list could contain:
+- partitions created upon md device (mdXpY)
+- external metadata container- specific type of md array.
+
+For partitions there is no issue, because they aren't handle STOP_ARRAY
+ioctl sent later. It generates misleading errors only.
+
+Second case is more problematic because containers are not locked in kernel.
+They are stopped even if container member array is active. For that reason
+reboot or shutdown flow could be blocked because metadata manager cannot be
+restarted after switch root on shutdown.
+
+Add filters to remove partitions and containers from md_list. Partitions
+can be excluded by DEVTYPE. Containers are determined by MD_LEVEL
+property, we are excluding all with "container" value.
+
+Signed-off-by: Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>
+
+Conflict:NA
+Reference:https://github.com/systemd/systemd/commit/3a3b022d2cc112803ea7b9beea98bbcad110368a
+
+---
+ src/shutdown/umount.c | 18 +++++++++++++++++-
+ 1 file changed, 17 insertions(+), 1 deletion(-)
+
+diff --git a/src/shutdown/umount.c b/src/shutdown/umount.c
+index 3e9e241499..820aa8e286 100644
+--- a/src/shutdown/umount.c
++++ b/src/shutdown/umount.c
+@@ -352,9 +352,14 @@ static int md_list_get(MountPoint **head) {
+         if (r < 0)
+                 return r;
+ 
++        /* Filter out partitions. */
++        r = sd_device_enumerator_add_match_property(e, "DEVTYPE", "disk");
++        if (r < 0)
++                return r;
++
+         FOREACH_DEVICE(e, d) {
+                 _cleanup_free_ char *p = NULL;
+-                const char *dn;
++                const char *dn, *md_level;
+                 MountPoint *m;
+                 dev_t devnum;
+ 
+@@ -362,6 +367,17 @@ static int md_list_get(MountPoint **head) {
+                     sd_device_get_devname(d, &dn) < 0)
+                         continue;
+ 
++                r = sd_device_get_property_value(d, "MD_LEVEL", &md_level);
++                if (r < 0) {
++                        log_warning_errno(r, "Failed to get MD_LEVEL property for %s, ignoring: %m", dn);
++                        continue;
++                }
++
++                /* MD "containers" are a special type of MD devices, used for external metadata.
++                 * Since it doesn't provide RAID functionality in itself we don't need to stop it. */
++                if (streq(md_level, "container"))
++                        continue;
++
+                 p = strdup(dn);
+                 if (!p)
+                         return -ENOMEM;
+-- 
+2.33.0
+

systemd.spec

@@ -21,7 +21,7 @@
 Name:           systemd
 Url:            https://www.freedesktop.org/wiki/Software/systemd
 Version:        249
-Release:        73
+Release:        74
 License:        MIT and LGPLv2+ and GPLv2+
 Summary:        System and Service Manager
 
@@ -636,6 +636,7 @@ Patch6587:      backport-basic-fix-overflow-detection-in-sigbus_pop.patch
 Patch6588:      backport-busctl-avoid-asserting-on-NULL-message.patch
 Patch6589:      backport-resolve-don-t-add-sockets-to-the-graveyard-on-shutdo.patch
 Patch6590:      backport-sd-journal-check-sd-event-state-before-setting-up-po.patch
+Patch6591:      backport-shutdown-get-only-active-md-arrays.patch
 
 Patch9001:      update-rtc-with-system-clock-when-shutdown.patch
 Patch9002:      udev-add-actions-while-rename-netif-failed.patch
@@ -2140,6 +2141,9 @@ grep -q -E '^KEYMAP="?fi-latin[19]"?' /etc/vconsole.conf 2>/dev/null &&
 %{_libdir}/security/pam_systemd.so
 
 %changelog
+* Sat Mar 16 2024 huyubiao <huyubiao@huawei.com> - 249-74
+- fix mdadm blocking caused by incorrect stop of inactive MD devices during shutdown
+
 * Mon Mar 4 2024 huyubiao <huyubiao@huawei.com> - 249-73
 - move the architecture patches after Patch9800 (excluding Patch9800)
   delete the duplicate systemd-coredump.socket