kernel: power off on panic

Instead of just rebooting, which may lead to an infinite loop, try to
resolve the situation by just powering off the VM - it can be restarted
any time anyway.

Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>

diff --git a/src/patches/kernel/0004-PBS-restore-halt-machine-on-kernel-panic.patch b/src/patches/kernel/0004-PBS-restore-halt-machine-on-kernel-panic.patch
new file mode 100644 (file)
index 0000000..d79833f
--- /dev/null
+++ b/src/patches/kernel/0004-PBS-restore-halt-machine-on-kernel-panic.patch
@@ -0,0 +1,32 @@
+From 7222e7424aab957f63b98853ea9fb30eec83666e Mon Sep 17 00:00:00 2001
+From: Stefan Reiter <s.reiter@proxmox.com>
+Date: Mon, 3 May 2021 11:13:10 +0200
+Subject: [PATCH] PBS-restore: halt machine on kernel panic
+
+Otherwise we might get into a loop where the user-space watchdog never
+has time to start, and thus the VM will run forever. Still not an idea
+options, since the kernel might hang and not panic, but better than
+nothing, and at least solves the out-of-memory forever looping.
+
+Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
+---
+ kernel/panic.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/kernel/panic.c b/kernel/panic.c
+index 332736a72a58..56339ae5165c 100644
+--- a/kernel/panic.c
++++ b/kernel/panic.c
+@@ -325,6 +325,9 @@ void panic(const char *fmt, ...)
+               }
+       }
+       if (panic_timeout != 0) {
++              /* PBS restore: stop machine on panic, let host deal with it */
++              machine_power_off();
++
+               /*
+                * This will not be a clean reboot, with everything
+                * shutting down.  But if there is a chance of
+-- 
+2.20.1
+