Subject: Re: problems with bcopy compiling -current
To: None <port-pmax@netbsd.org>
From: Toru Nishimura <nisimura@is.aist-nara.ac.jp>
List: port-pmax
Date: 08/22/1998 13:11:11
> Doing some code scanning, the offending snippet in trap.c looks like this:
>
>        else if (pcBetween(bcopy, setrunqueue)) {
>                subr = (unsigned) bcopy;
>        }
>
> bcopy is a #define in sys/systm.h:
>
> #define bcopy(src, dst, len)    memcpy(dst, src, len)
>
> Shouldn't the define take care of replacing bcopy with memcpy?

I think the original code fragment itself is brain damaged.  It tries
to do primitive 'hand crafted backtrace' when the kernel gets paniced. 
The back tracing relies on code order in locore.S.  I propose to remove
the entire code segment from trap.c to eliminate further disturburance.

Tohru Nishimura
Nara Institute of Science and Technology