Subject: Re: CVS commit: src/sys/compat/linux/common
To: None <source-changes@netbsd.org>
From: Christos Zoulas <christos@astron.com>
List: source-changes
Date: 05/09/2007 21:45:45
In article <20734.1178746348@splode.eterna.com.au>,
matthew green  <mrg@eterna.com.au> wrote:

>note how there is a compile-time option in the quoted code to set
>this?  LINUX_UNAME_ARCH.  what i tried to do was make this sysctlable
>but the code i commited is broken (it worked on my test, but i'm not
>sure exactly what i did when transferring it somewhere i could run
>cvs ci from), and the attempt i made so far to make it use a
>dynamically allocated value failed...
>
>
>i need to set this (to "i686") for a system that expects it (since
>that is what "uname -m" says under linux.)

I understand that, but I am confused about one thing. Are we talking
about the linux uname -m returning i686, or our uname? I looked at
the linux sysctl mib and they don't have a corresponding sysctl for
"machine", "arch" or equivalent. On linux uname uses uname(2), so
the linux uname should already work. Why do we need our uname to
return what linux returns?

uname({sysname="Linux", nodename="lunix.astron.com", release="2.6.18", version="#1 SMP PREEMPT Fri Sep 29 14:06:04 GMT 2006", machine="x86_64"}) = 0

christos