tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Apeasing diagnostic assertions after ddb "sync" command

I tested some things with sparc64 crashdumps today and noticed an inconsistent
behaviour: if you break into ddb and say "reboot 0x104" the system will most
likely not just dump, but trigger DIAGNOSTIC/LOCKDEBUG assertions (e.g.
when aquiring mutexes due to still being in interrupt context). This assertion
will panic, and drop back to ddb. Now repeating the same "reboot 0x104"
will work - since "panicstr" is set.

I would like to change this with the following patch.

Any objections?

Index: db_command.c
RCS file: /cvsroot/src/sys/ddb/db_command.c,v
retrieving revision 1.112
diff -c -u -r1.112 db_command.c
--- db_command.c        13 Dec 2007 02:45:11 -0000      1.112
+++ db_command.c        17 Jan 2008 11:12:51 -0000
@@ -1361,6 +1361,7 @@
         * called from cpu_reboot.
        db_recover = 0;
+       panicstr = "dump forced via kernel debugger";
        cpu_reboot(RB_DUMP, NULL);

Home | Main Index | Thread Index | Old Index