Hi,
the struct emul and struct execsw changes are available for review
at
	ftp://ftp.netbsd.org/pub/NetBSD/misc/jdolecek/execsw.2.diff

This is probably final version before it gets committed into tree,
so please take a look and bug me if you find anything fishy.
It would be great if you could try to compile kernel with applied
patch on other archs than i386, to find out if I overlooked anything.

Changes from last version, unresolved issues:
* plugged introduced memory leak for ELF execution
* hopefully all necessary changes for individual ports done, though
  compile was tested on i386 only
* all entries in execw[] updated accordingly to new structure layout
* ecoff split similar way as elf, stuff moved from arch/alpha/alpha/machdep.c
  to compat/osf1/osf1_exec.c, and from arch/mips/mips/sys_exec.c to
  compat/ultrix/ultrix_misc.c, some functions were renamed from *_hook()
  to *_probe()
  similarily to elf, there is now one entry per emulation,
  exec_ecoff_makecmds() calls probe function - hopefully this is done
  right.
  Note this has only changed for alpha and mips, are there any
  other ports supporting ecoff?
  I'm not quite sure this does the right thing esp. wrt. osf1 and ultrix
  emulations, hopefully yes. I'm also not quite sure how is the ultrix emulation
  gets used on vax - seems like there is no way any executable can become
  running under Ultrix emulation on vax.
* compat/ibcs2/ibcs2_misc.c:ibcs2_sys_time() checked for emul_ibcs2_xout;
  I've disabled that check for now, since there is now only single
  emul_ibcs2. What is the right thing to do here ?

Jaromir
-- 
Jaromir Dolecek <jdolecek@NetBSD.org>      http://www.ics.muni.cz/~dolecek/
@@@@  Wanna a real operating system ? Go and get NetBSD, damn!  @@@@