Re: Substantial COMPAT_LINUX changes in netbsd-5?

[christos%zoulas.com@localhost (Christos Zoulas)]

On Jan 20,  8:40pm, pooka%NetBSD.org@localhost (Antti Kantee) wrote:
-- Subject: Re: Substantial COMPAT_LINUX changes in netbsd-5?

| On Wed, Jan 20, 2010 at 10:41:52AM -0500, Christos Zoulas wrote:
| > On Jan 20,  4:20pm, he%nordu.net@localhost (Havard Eidnes) wrote:
| > -- Subject: Re: Substantial COMPAT_LINUX changes in netbsd-5?
| > 
| > | Managed to get that.  When the backup process got stuck, I typed
| > | "df" out of old habit, and it also got stuck:
| > | 
| > | mail-server% ps axl | egrep tstile\|PID
| > |  UID   PID  PPID  CPU PRI NI   VSZ   RSS WCHAN   STAT TTY       TIME 
COMMAND
| > |    0 24173     1  195 117  0 17860  4876 tstile  D    ?      0:00.03 
bpbkar -r 
| > |    0 19422 26923    0 117  0  2824   796 tstile  D+   ttyp5  0:00.00 df 
| > | mail-server% 
| > | 
| > | Broke into DDB on the console, and here's the two stack backtraces
| > | for these two processes:
| > 
| > Can you try LK_CANRECURSE in nfs_root()? Not that it is a real fix? Can you
| > try dumping who owns the locks for the nfs root vnode?
| 
| Shouldn't LOCKDEBUG (or one of smaller hammers) catch locking-against-myself?

It should and it does I think; this is why we should dump who owns the locks.

christos