Subject: misc/11145: core(5) is completely wrong
To: None <gnats-bugs@gnats.netbsd.org>
From: None <nathanw@mit.edu>
List: netbsd-bugs
Date: 10/05/2000 09:06:11
>Number:         11145
>Category:       misc
>Synopsis:       core(5) is completely wrong
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    misc-bug-people
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Oct 05 09:06:01 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     Nathan J. Williams
>Release:        -current 20001004
>Organization:
	Massachvsetts Institvte of Technology
>Environment:
System: NetBSD dream.ai.mit.edu 1.5G NetBSD 1.5G (DREAM) #31: Thu Sep 28 19:06:46 EDT 2000 nathanw@dream.ai.mit.edu:/u1/nbsd/src/sys/arch/i386/compile/DREAM i386


>Description:
	The man page that purports to describe the core dump format
is completely wrong. It describes the core dump as consisting of the process's
u. area, followed by the data area and stack area, with sizes given in the 
u. area.

Cursory examination of sys/user.h shows that the referred-to sizes do not
exist; examination of coredump() in kern/kern_sig.c and sys/core.h shows
what the real format is, but it should be properly written up.
>How-To-Repeat:
	"man core". Scratch head at reference to long-defunct u. area contents.
>Fix:
	Rewrite. I might do it, but this is here in case I don't get to it.
>Release-Note:
>Audit-Trail:
>Unformatted: