Subject: Re: trap.c patch about longjmp
To: None <>
From: Gordon W. Ross <>
List: port-sun3
Date: 03/14/1996 15:10:35
> Date: Wed, 13 Mar 1996 12:32:52 +0200
> From: Jari Kokko <>

> I compiled the current kernel yesterday, but had to patch trap.c to do
> it. There is a code segment in the function trap like
> 	if (nofault)
> 		longjmp(nofault);
> This gives one warning, about the type of the parameter; and one
> error, about the second parameter missing.
> I added 1 as the second parameter and compiled without -Werror.
> Who should I tell about this? How can this bug, if it is one, be
> present, the compiler spots it every time?

Thanks for the report.  We decided the prototype was wrong.
The kernel implementation of longjmp does not take a second
argument. (longjmp always causes setjmp to return one.)