Re: PR/45162 CVS commit: src/sys/dev/raidframe

[David Holland <dholland-bugs%netbsd.org@localhost>]

The following reply was made to PR bin/45162; it has been noted by GNATS.

From: David Holland <dholland-bugs%netbsd.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: oster%netbsd.org@localhost
Subject: Re: PR/45162 CVS commit: src/sys/dev/raidframe
Date: Sat, 7 Aug 2021 02:30:04 +0000

 On Fri, Jul 23, 2021 at 10:45:01PM +0000, Greg Oster wrote:
  > From: Greg Oster <oster%netbsd.org@localhost>
  > To: gnats-bugs%netbsd.org@localhost
  > Cc: 
  > Subject: Re: PR/45162 CVS commit: src/sys/dev/raidframe
  > Date: Fri, 23 Jul 2021 16:42:18 -0600
  > 
  >  It turns out the 'fix' for this PR is not as obvious as it once seemed.
  >  
  >  RAIDframe *does* support RAID 1 with more than 2 components, but not in 
  >  the way one might expect.  Given six components: a, b, c, d, e, and f, 
  >  RAIDframe will create a 2-dimensional array, where the concatenation of 
  >    a, c, and e will be mirrored by the concatenation of b, d, and f.
 
 That seems like something that is maybe better not to inflict on users
 without warning :-)
 
  >  Where things are falling apart is with an odd number of components, 
  >  which now fails (until someone writes a n-way RAID 1 for RAIDframe).
 
 I think that means we can close the PR, which was not asking for such
 support, just complaining that it silently failed if you tried to
 configure it. Right?
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost