From owner-freebsd-current  Tue Jun  9 18:23:22 1998
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id SAA18169
          for freebsd-current-outgoing; Tue, 9 Jun 1998 18:23:22 -0700 (PDT)
          (envelope-from owner-freebsd-current@FreeBSD.ORG)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.15.68.22])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA18155;
          Tue, 9 Jun 1998 18:23:14 -0700 (PDT)
          (envelope-from bde@godzilla.zeta.org.au)
Received: (from bde@localhost)
	by godzilla.zeta.org.au (8.8.7/8.8.7) id LAA06267;
	Wed, 10 Jun 1998 11:23:04 +1000
Date: Wed, 10 Jun 1998 11:23:04 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199806100123.LAA06267@godzilla.zeta.org.au>
To: ache@nagual.pp.ru, bde@FreeBSD.ORG
Subject: Re: PIPE_BUF?
Cc: current@FreeBSD.ORG
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

>>From sys/syslimits.h:
>
>#define PIPE_BUF                  512   /* max bytes for atomic pipe 
>writes */				   ^^^
>
>Is it still true? If not, we should change it to _real_ value since some
>software (like apache) depends on it.

Yes.  In 2.2, it really shouldn't be defined, since the max differs for
ordinary pipes (they use PIPE_MAX = 512) and fifos (they use the default
socket sb_lowat = MCLBYTES = 2048).  However, applications can't tell
that a too-small value is too small, since the system is not required to
gratuitously make larger writes _non_ atomic.  A smaller value is just
more efficient for writers than don't care about atomicicity (since the
system is less constrained), and less efficient and more restrictive
for writers that do care (since they must send tinygrams).

Bruce

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