This bug was initially created as a copy of Bug #1748051

I am copying this bug because: 



Description of problem:

The /lib/udev/rules.d/40-redhat.rules file defines the following rules for "memory" subsystem:

-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
# Memory hotadd request
SUBSYSTEM!="memory", GOTO="memory_hotplug_end"
ACTION!="add", GOTO="memory_hotplug_end"
PROGRAM="/bin/uname -p", RESULT=="s390*", GOTO="memory_hotplug_end"

ENV{.state}="online"
PROGRAM="/bin/systemd-detect-virt", RESULT=="none", ENV{.state}="online_movable"
ATTR{state}=="offline", ATTR{state}="$env{.state}"

LABEL="memory_hotplug_end"
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------

On system that have many Memory devices (in some cases, I can see up to 1000 or 12000 memory devices!!!), these rules will spawn tons of "uname -p" / "systemd-detect-virt" processes which will kill the system.

This issue affects multiple packages as well, for example trace-cmd ("/lib/udev/rules.d/98-trace-cmd.rules") and kexec-tools ("/lib/udev/rules.d/98-kexec.rules").


Version-Release number of selected component (if applicable):

systemd-219-67.el7_7.1


How reproducible:

Always


Steps to Reproduce:
1. Enable debugging in udev: "udev.log-priority=debug" while booting

2. Check the number of times a command executed

  # journalctl -b | egrep '(PROGRAM|RUN)' | cut -f2- -d"'" | sort | uniq -c | sort -nr


Actual results:

    168 /usr/bin/systemctl is-active trace-cmd.service' /usr/lib/udev/rules.d/98-trace-cmd.rules:1
    124 kmod load $env{MODALIAS}' /usr/lib/udev/rules.d/80-drivers.rules:5
     32 /bin/uname -p' /usr/lib/udev/rules.d/40-redhat.rules:9
     32 /bin/systemd-detect-virt' /usr/lib/udev/rules.d/40-redhat.rules:12
    ...

Expected results:

  1 call to "systemd-detect-virt"
  1 call to "uname -p"

Additional info:

This was observed on a system with 16 memory devices (2GB VM with 1 CPU).
On large physical systems, it can be just crazy.