Subject: Re: FreeBSD's /dev on NetBSD
To: None <firstname.lastname@example.org>
From: Nathan J. Williams <email@example.com>
Date: 08/15/2005 12:59:42
Hans Petter Selasky <firstname.lastname@example.org> writes:
> On Sunday 14 August 2005 03:00, Jason Thorpe wrote:
> > On Aug 13, 2005, at 3:09 PM, Hans Petter Selasky wrote:
> > > I've heard that NetBSD is very against mutexes. Is this true? Will
> > > NetBSD
> > > always use one lock for the whole kernel?
> > Where did you hear that?
> I think it was an article at "http://www.bsdnews.com/" or some mailing list.
> I cannot find the article right now, but there was a comparison between
> FreeBSD 5/6 and NetBSD, where NetBSD was found to be faster, and then there
> was some disscussion whether this had to do with mutexes. And then someone
> wrote that NetBSD did not use mutexes like on FreeBSD and that this made the
> system faster. But someone else said that it was extra checking in the kernel
> that made FreeBSD perform so bad in this comparison ...
This sounds like it might have been contrasting userland thread
implementations; NetBSD's user thread mutexes don't get help from the
kernel, which is a performance advantage. It doesn't have anything to
do with the use of locking inside the kernel, though.