Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 26 Jun 2001 13:07:00 -0400
From:      "Brian F. Feldman" <green@FreeBSD.org>
To:        Poul-Henning Kamp <phk@critter.freebsd.dk>
Cc:        "Brian F. Feldman" <green@FreeBSD.org>, "Jonathan Lemon <jlemon@flugsvamp.com>    Alfred Perlstein" <bright@sneakerz.org>, Mike Silbersack <silby@silby.com>, Matt Dillon <dillon@earth.backplane.com>, Mike Silbersack <silby@FreeBSD.org>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org, jlemon@FreeBSD.org, bmilekic@FreeBSD.org
Subject:   Re: cvs commit: src/sys/netinet tcp_input.c tcp_output.c tcp_subr.c tcp_timer.c tcp_usrreq.c tcp_var.h 
Message-ID:  <200106261707.f5QH70k41274@green.bikeshed.org>
In-Reply-To: Message from Poul-Henning Kamp <phk@critter.freebsd.dk>  of "Tue, 26 Jun 2001 13:39:45 %2B0200." <71068.993555585@critter> 

next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp <phk@critter.freebsd.dk> wrote:
> In message <200106242141.f5OLfc176777@green.bikeshed.org>, "Brian F. Feldman" w
> rites:
> 
> >> bzero seems to be optimized for large areas, perhaps it would help
> >> malloc some if we used some alternative zero'ing function for small
> >> allocations with M_ZERO set?
> >
> >That's pretty pointless; M_ZERO is _supposed_ to eventually be providing 
> >pre-zeroed memory, which should remove that bzero in the general case, 
> >anyway.
> 
> Well, M_ZERO was designed to remove a lot of distinct calls to bzero
> in the hope that we could:
> 
> A) collect statistics showing if demand for zeroed RAM is big enough
>    to persue B)
> 
> B) Implement prezeroed RAM as an optimization.
> 
> Neither of these precludes an optimization of the current 
> implementation of M_ZERO

Well, I don't get exactly how it would be optimized more than it is now 
because it won't be able to take advantage of these "smaller" bzero()s... 
unless... what about making malloc() an inline that checks M_ZERO and uses 
the new constant-bzero() on sufficiently small sizes after calling malloc 
without the M_ZERO?  I'm pretty certain GCC would optimize that fine, and 
that would buy us the faster-constant-sized-bzero back from the M_ZERO 
optimization.

-- 
 Brian Fundakowski Feldman           \  FreeBSD: The Power to Serve!  /
 green@FreeBSD.org                    `------------------------------'



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200106261707.f5QH70k41274>