Package: release.debian.org Severity: normal Tags: trixie X-Debbugs-Cc: [email protected] Control: affects -1 + src:mdadm User: [email protected] Usertags: pu
Hello, I have uploaded mdadm 4.4-11+deb13u1 for proposal to trixie [ Reason ] As reported on #1115497, there is a memory leak in the mdmonitor daemon. After a few weeks of operation, this can eat a few gigabytes. This is a regression over oldstable. [ Impact ] Administrators of servers using mdmonitor have to restart the service from times to times to drop the memory consumption. [ Tests ] This is commited upstream, and confirmed on the three servers where I noticed the issue. [ Risks ] The code is a mere added unref, and commited upstream. [ Checklist ] [X] *all* changes are documented in the d/changelog [X] I reviewed all changes and I approve them [X] attach debdiff against the package in (old)stable [X] the issue is verified as fixed in unstable [ Changes ] This adds an unref call corresponding to the udev_monitor_receive_device call whose returned pointer is not stored. Samuel
diff -Nru mdadm-4.4/debian/changelog mdadm-4.4/debian/changelog --- mdadm-4.4/debian/changelog 2025-04-25 18:24:19.000000000 +0200 +++ mdadm-4.4/debian/changelog 2026-08-23 21:40:00.000000000 +0200 @@ -1,3 +1,11 @@ +mdadm (4.4-11+deb13u1) trixie; urgency=medium + + * Non-Maintainer Upload. + * Adding patch from Peter Mann <[email protected]> to fix memory leak in + mdadm --monitor --scan (Closes: #1115497). + + -- Samuel Thibault <[email protected]> Sun, 23 Aug 2026 21:40:00 +0200 + mdadm (4.4-11) sid; urgency=medium [ Daniel Baumann ] diff -Nru mdadm-4.4/debian/patches/debian/0015-monitor-memory-leak.patch mdadm-4.4/debian/patches/debian/0015-monitor-memory-leak.patch --- mdadm-4.4/debian/patches/debian/0015-monitor-memory-leak.patch 1970-01-01 01:00:00.000000000 +0100 +++ mdadm-4.4/debian/patches/debian/0015-monitor-memory-leak.patch 2026-08-23 21:40:00.000000000 +0200 @@ -0,0 +1,39 @@ +Author: Peter Mann <[email protected]> +Description: Memory leak in mdadm --monitor --scan (Closes: #1115497). + At mdadm/udev.c:153: + ... + if (udev_monitor_receive_device(udev_monitor)) + return UDEV_STATUS_SUCCESS; /* event detected */ + ... + . + According to libudev docs: + On success, udev_monitor_receive_device() returns a pointer to a newly referenced device that was received via the monitor. The caller is responsible to drop this reference when done. + . + As you can see, the reference to the device never gets dropped. + We put together a quick patch, which seems to have fixed the issue or at least substantially reduced the amount of leaks. + . + Since we have no experience with the codebase, we have no idea about the implications of these changes. + We would appreciate if someone took a closer look. + . + Additionaly as a workaround, passing MDADM_NO_UDEV=1 env to mdadm stops the leaks as well, since it bypasses the leaking codepath. +Forwarded: https://github.com/md-raid-utilities/mdadm/pull/197 + +diff -Naurp mdadm.orig/udev.c mdadm/udev.c +--- mdadm.orig/udev.c ++++ mdadm/udev.c +@@ -149,9 +149,13 @@ enum udev_status udev_wait_for_events(in + tv.tv_sec = seconds; + tv.tv_usec = 0; + +- if (select(fd + 1, &readfds, NULL, NULL, &tv) > 0 && FD_ISSET(fd, &readfds)) +- if (udev_monitor_receive_device(udev_monitor)) ++ if (select(fd + 1, &readfds, NULL, NULL, &tv) > 0 && FD_ISSET(fd, &readfds)) { ++ struct udev_device *dev = udev_monitor_receive_device(udev_monitor); ++ if (dev) { ++ udev_device_unref(dev); + return UDEV_STATUS_SUCCESS; /* event detected */ ++ } ++ } + return UDEV_STATUS_TIMEOUT; + } + #endif diff -Nru mdadm-4.4/debian/patches/series mdadm-4.4/debian/patches/series --- mdadm-4.4/debian/patches/series 2025-03-09 18:03:14.000000000 +0100 +++ mdadm-4.4/debian/patches/series 2026-08-23 21:40:00.000000000 +0200 @@ -10,4 +10,5 @@ debian/0011-systemd-directory.patch debian/0012-bin-directory.patch debian/0013-xmalloc-ftbfs.patch +debian/0015-monitor-memory-leak.patch upstream/0001-manpage-remove-bitmap.patch
