Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 27 Apr 2012 21:12:19 +0300
From:      Alexander Motin <mav@FreeBSD.org>
To:        John Baldwin <jhb@freebsd.org>
Cc:        "svn-src-head@freebsd.org" <svn-src-head@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "src-committers@freebsd.org" <src-committers@freebsd.org>, Scott Long <scott4long@yahoo.com>
Subject:   Re: svn commit: r234603 - head/sys/geom/raid
Message-ID:  <4F9AE183.7060209@FreeBSD.org>
In-Reply-To: <201204271338.37236.jhb@freebsd.org>
References:  <201204231304.q3ND43Yg098748@svn.freebsd.org> <201204270745.16844.jhb@freebsd.org> <1335547643.283.YahooMailNeo@web45713.mail.sp1.yahoo.com> <201204271338.37236.jhb@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 04/27/12 20:38, John Baldwin wrote:
> On Friday, April 27, 2012 1:27:23 pm Scott Long wrote:
>>
>> ----- Original Message -----
>>> From: John Baldwin<jhb@freebsd.org>
>>> To: Alexander Motin<mav@freebsd.org>
>>> Cc: src-committers@freebsd.org; svn-src-all@freebsd.org; svn-src-head@freebsd.org
>>> Sent: Friday, April 27, 2012 5:45 AM
>>> Subject: Re: svn commit: r234603 - head/sys/geom/raid
>>>
>>> On Monday, April 23, 2012 9:04:03 am Alexander Motin wrote:
>>>>   Author: mav
>>>>   Date: Mon Apr 23 13:04:02 2012
>>>>   New Revision: 234603
>>>>   URL: http://svn.freebsd.org/changeset/base/234603
>>>>
>>>>   Log:
>>>>     Add names for all primary RAID levels defined by DDF 2.0 specification.
>>>>
>>>>   Modified:
>>>>     head/sys/geom/raid/g_raid.c
>>>>     head/sys/geom/raid/g_raid.h
>>>>     head/sys/geom/raid/tr_raid1.c
>>>>     head/sys/geom/raid/tr_raid1e.c
>>>
>>> We should probably add a separate header to hold DDF constants.  graid isn't
>>> the only place that uses them (e.g. mfi(4) uses it to describe volumes, so
>>> mfiutil(8) has its own DDF constants as well in mfiutil.h).
>>>
>>> --
>>
>> You mean src/sys/dev/ata/ata-raid-ddf.h?
>
> Yes.
>
>> That said, I trust DDF to be neither universal nor standard, and it's probably a futile micro-optimization to try too hard at this.  At the very
> least, leave MFI alone.

Standard or not, DDF is the only public specification and it is quite 
universal from point of reviewing possible configurations. That's why I 
took it as reference instead of reinventing own numbers. Same time 
metadata modules in graid may do own translation and theoretically 
support some other configurations, that's why I am not exactly 
comfortable with idea of merging these lists and strictly tying to DDF.

> Hmm, LSI claims that MFI uses the constant values (but not necessarily the
> structures from DDF).  Certainly the primary RAID type in an mfi(4) volume
> uses the same constants as both g_raid.h and ata-raid-ddf.h.

ata-raid-ddf.h as all ataraid(4) is not actively used now and should 
leave. Instead I am now writing DDF module for graid and going to 
resurrect that file there with minor updates. These two files could be 
merged, but I am not sure it worth to be done either, as it is temporary 
coexistence.

-- 
Alexander Motin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4F9AE183.7060209>