[1016] in linux-scsi channel archive
Re: Fujitsu MO M2513A 2048 bytes/sector SCSI problem ?
daemon@ATHENA.MIT.EDU (Eric Youngdale)
Sun Nov 24 20:38:50 1996
Date: Sun, 24 Nov 1996 20:36:53 -0500
From: Eric Youngdale <eric@sub2317.jic.com>
To: rkiendl@btp2x5.phy.uni-bayreuth.de
CC: linux-scsi@vger.rutgers.edu
In-reply-to: <Pine.HPP.3.91.961124131719.16862B@btp2x5.phy.uni-bayreuth.de>
(message from Robert Kiendl on Sun, 24 Nov 1996 13:21:28 +0100 (MET))
>>Studiing kernel scsidisk-driver sd.c I found that the current driver doesn't
>>support more than 1024 bytes/sector. Do you have/find a solution for this
>>problem ?
>
> I beleive Eric Youngdale has some prototype patches to implement this:
>
> ftp://sub2317.jic.com/pub/scsi/2048-1.diff
>
Correct (sort of). I flushed the old versions of the patches,
so there is only one diff file, 2048.diff. Someone sent me patches for
the msdos filesystem so that it also works with 2k block discs. I have
the diffs for that in a seperate file in the same directory. The diffs
are against the 2.1.8 kernel - anything 2.0 and beyond should work without
too much trouble.
I also took the opportunity to clean out a lot of cobwebs in the
iso9660 filesystem. It used to allow the user to set the blocksize that the
filesystem would use independently of the natural extent size for the disc. I
fixed things so that the filesystem forced a blocksize for the device that was
the same as the extent size, and was able to remove a lot of moldy code which
would try and piece together 2 1024 byte blocks to make a 2K sector.
I would like to get a few positive reports that the patches that
I have work out of the box (and I need to get Werner to look over the msdos
patches) before I was going to send them to Linus.
-Eric
