From owner-freebsd-hackers  Wed Jul 23 19:48:19 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id TAA02536
          for hackers-outgoing; Wed, 23 Jul 1997 19:48:19 -0700 (PDT)
Received: from labinfo.iet.unipi.it (labinfo.iet.unipi.it [131.114.9.5])
          by hub.freebsd.org (8.8.5/8.8.5) with SMTP id TAA02531
          for <hackers@FreeBSD.ORG>; Wed, 23 Jul 1997 19:48:14 -0700 (PDT)
Received: from localhost (luigi@localhost) by labinfo.iet.unipi.it (8.6.5/8.6.5) id DAA25440; Thu, 24 Jul 1997 03:46:54 +0200
From: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
Message-Id: <199707240146.DAA25440@labinfo.iet.unipi.it>
Subject: Re: getting physically contiguous memory in the kernel
To: jas@flyingfox.com (Jim Shankland)
Date: Thu, 24 Jul 1997 03:46:54 +0200 (MET DST)
Cc: hackers@FreeBSD.ORG
In-Reply-To: <199707232315.QAA10162@biggusdiskus.flyingfox.com> from "Jim Shankland" at Jul 23, 97 04:14:44 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Sender: owner-freebsd-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

> Is there an easy way to get physically contiguous memory in the FreeBSD
> kernel?  The de driver calls malloc() to obtain a buffer, part of which
> must be physically contiguous; this buffer has now grown to be > 4K,
> causing the transmit descriptors to span a page boundary.  When the
> 2 pages are not physically contiguous, the driver breaks.

look at contigmalloc() or (probably better) vm_page_alloc_contig().
The latter is used in /sys/pci/brooktree848.c for the brooktree
driver and /sys/pci/meteor.c for the meteor driver. The former is
generally used for allocating buffers for use with the ISA DMA.

	Cheers
	Luigi