Subject: Re: bad system call
To: Jim Wise <jwise@draga.com>
From: None <mcmahill@mtl.mit.edu>
List: netbsd-users
Date: 04/25/1999 16:18:16
On Sun, 25 Apr 1999, Jim Wise wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> 
> On Sun, 25 Apr 1999 mcmahill@mtl.mit.edu wrote:
> 
> >Can anyone shed some light on what this message means:
> [...]
> >Program received signal SIGSYS, Bad system call.
> >0x204d1db4 in shmget () at shmget:2
> >shmget:2: No such file or directory.
> >(gdb) quit
> >The program is running.  Quit anyway (and kill it)? (y or n) y
> 
> Do you have:
> 
> 	options         SYSVSEM         # System V-like semaphores
> 
> in your kernel config?  The code in question seems to be getting a
> SIGSYS while trying to call shmget(2).  If you _do_ have the above in
> your kernel config, you may want to use ktrace(1) to get a better idea
> where it's failing.
hmmmm.  Thats interesting.  I do not have that option.  The interesting
part is that from /usr/src/sys/arch (1.3.3 source tree), 
grep SYSVSEM */conf/GENERIC
shows SYSVSEM listed on the GENERIC config files for 13 ports (including
all the other systems I ever use), but grep SYSVSEM pmax/conf/* shows
nothing (not even commented out SYSVSEM entries).  

I'll try adding that.  Is that something which should be in GENERIC pmax
kernels?


> 
> Also, you may want to post questions like this to
> netbsd-users@netbsd.org or netbsd-help@netbsd.org, as it is not specific
> to NetBSD/pmax, and someone who has seen the same thing under another
> port may be able to help you as well...
actually I thought it was pmax specific since my other NetBSD machines
didn't exhibit that behaviour, but in light of your information, it may
well not be a pmax issue (other that why that option isn't in the GENERIC
config file).

Thanks for the help

-Dan