Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 15 Jan 2009 10:31:42 -0500
From:      John Baldwin <jhb@freebsd.org>
To:        Marius Strobl <marius@freebsd.org>
Cc:        svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org
Subject:   Re: svn commit: r186878 - head/sys/dev/mpt
Message-ID:  <200901151031.42725.jhb@freebsd.org>
In-Reply-To: <200901072152.n07Lql7h000807@svn.freebsd.org>
References:  <200901072152.n07Lql7h000807@svn.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 07 January 2009 4:52:47 pm Marius Strobl wrote:
> Author: marius
> Date: Wed Jan  7 21:52:47 2009
> New Revision: 186878
> URL: http://svn.freebsd.org/changeset/base/186878
>
> Log:
>   Make the whole initiator mode part of mpt(4) endian-clean,
>   specifically SPI controllers now also work in big-endian
>   machines and some conversions relevant for FC and SAS
>   controllers as well as support for ILP32 machines which all
>   were omitted in previous attempts are now also implemented.
>   The IOCTL-interface is intentionally left (and where needed
>   actually changed) to be completely little-endian as otherwise
>   we would have to add conversion code for every possible
>   configuration page to mpt(4), which didn't seem the right
>   thing to do, neither did converting only half of the user-
>   interface to the native byte order.
>   This change was tested on amd64 (SAS+SPI), i386 (SAS) and
>   sparc64 (SAS+SPI). Due to lack of the necessary hardware
>   the target mode code is still left to be made endian-clean.
>
>   Reviewed by:	scottl
>   MFC after:	1 month

If you wish to test the mpt_user interface on a big-endian machine you may 
use //depot/jhb/raid/usr.sbin/mptutil/...

It is likely going to need a lot of sprinkled htole*() and letoh*() though.

-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200901151031.42725.jhb>