Subject: Re: LFS and big files
To: Rhialto <rhialto@falu.nl>
From: Steven M. Bellovin <smb@cs.columbia.edu>
List: current-users
Date: 01/05/2007 23:58:30
On Sat, 6 Jan 2007 04:47:10 +0100
Rhialto <rhialto@falu.nl> wrote:

> On Wed 03 Jan 2007 at 17:49:05 -0500, Steven M. Bellovin wrote:
> > Note that sysutils/memtestplus can be loaded via the standard NetBSD
> > boot loader or Grub.  It can be compiled to use a serial console,
> > too; the two together mean that I install it on all of my machines,
> > including (of course) every remote server.
> 
> But:
> 
> $ sudo pkg_comp build sysutils/memtest86
> PKG_COMP ==> Mounting sandboxed filesystems
> PKG_COMP ==> Installing new `pkg-vulnerabilities' file
> PKG_COMP ==> Checking if pkg_install is up to date
> PKG_COMP ==> Building and installing sysutils/memtest86
> WARNING: Skipping memtest86-3.2:
> WARNING: memtest86-3.2 is not available for NetBSD-3.0-x86_64
> ===> do-clean [gmake-3.81] ===> Cleaning for gmake-3.81
> ===> do-clean [digest-20060826] ===> Cleaning for digest-20060826
> ===> do-clean [memtest86-3.2] ===> Cleaning for memtest86-3.2
> PKG_COMP ==> Unmounting sandboxed filesystems
> 
> Is the bootloader already in 64-bit mode such that it can't run
> memtest86 anymore? (I'm not going to try since I don't want to reboot
> my production machine). Maybe the restriction can be loosened a bit.
> 
First, note that there is a difference between memtest86 and
memtestplus.  I recommend the latter (and am listed as the maintainer
of it).  Second, I do not have any 64-bit machines to test on.  Given
that it's chipset-dependent, I'm quite loathe to assume anything about
its behavior.  It's easy enough to test, though -- just change the
ONLY_FOR_PLATFORM line in the Makefile, and let me know what happens.

		--Steve Bellovin, http://www.cs.columbia.edu/~smb