Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 20 Apr 1997 14:24:13 +0100
From:      James Mansion <james@wgold.demon.co.uk>
To:        "David S. Miller" <davem@jenolan.rutgers.edu>
Cc:        freebsd-hackers@freeBSD.org
Subject:   Re: Price of FreeBSD (was On Holy Wars...)
Message-ID:  <335A18FD.36C2@wgold.demon.co.uk>
References:  <199704201132.HAA05697@jenolan.caipgeneral>

next in thread | previous in thread | raw e-mail | index | archive | help
I disagree.  Clearly the sources are there, but they are not normally
compiled in.  It is not the case that a uniprocessor is simply a
particular
case of an SMP configuration - the code is quite different, and the
default build
(and versions provided from eg Red Hat) do not have the code enabled.

I'm well aware that the FreeBSD and Linux systems have SMP support if
you care
to try it out, but its not 'finished' in the way that SMP works out of
the box
with Solaris, NT, or UnixWare, for example, and as you say it is still
somewhat
'work in progress'.  (You give examples below, after all)  From my point
of view
its not done yet, and until then, its not supported.  Admitedly this
isn't
necessarily a common viewpoint in the Linux community, but I've not
enough time
to mess with version-of-the-week issues when I'm fighting to get
functionality
into my own products.

James

David S. Miller wrote:
> 
>    Date: Fri, 18 Apr 1997 09:29:08 +0100
>    From: James Mansion <james@wgold.demon.co.uk>
> 
>    There is an SMP kernel under development.  As there is one for
>    Linux.  But neither system has this as the core kernel line and
>    neither is finished or robust.
> 
> Incorrect, in 2.0.x and even more so in 2.1.x SMP is fully supported
> and in the sources for both the Sparc and Intel platform (Alpha coming
> soon).  The comment on robustness is pretty much correct although I
> use it on all my development boxes, for 2.0.x Sparc SMP is much more
> stable, for 2.1.x lately both Intel and Sparc are of similar
> stability.  We've threaded the heart of at least a few major
> subsystems of the kernel already (scheduler, interrupts, drivers, and
> wait queues), we are currently threading the vfs layer and next the
> buffer/page caches and networking should come reasonably soon after
> that.
> 
> ---------------------------------------------////
> Yow! 11.26 MB/s remote host TCP bandwidth & ////
> 199 usec remote TCP latency over 100Mb/s   ////
> ethernet.  Beat that!                     ////
> -----------------------------------------////__________  o
> David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?335A18FD.36C2>