Re: Hitachi 500G drive issue (NetBSD 4)

[buhrow%lothlorien.nfbcal.org@localhost (Brian Buhrow)]

        Hello.  IN the 4.0-stable tree, and beyond, the wd code is supposed
to assume the lba48 quirk exists if it gets an ID not found error when
crossing the 130GB boundary.  My guess is that your drives return a
different error, which is why the wd driver isn't just doing the right
thing.  If we assume the huristic technique of detecting drives with this
bug is more general, then the correct fix would be to note the actual error
your drives return, and then testfor that error as well as the ID not found
error when crossing the 130GB boundary.  Once done, you wouldn't need to
explicitly match the make and model of the Hitachi drives which exhibit
this behavior.
-Brian
On Apr 21,  2:28pm, Ian wrote:
} Subject: Hitachi 500G drive issue (NetBSD 4)
} Hi All,
} 
} firstly, apologies if this is the wrong place to ask this.
} 
} I've recently added a RAID 5 array to my home server, which consists of 
} 4*500gig Hitachi Deskstar SATA drives.
} 
} I initially had a problem building the parity on the raid, which would 
} always fail at the same point, which I traced down to the same LBA bug 
} which affects Seagate drives at the 130G mark.
} 
} This was solved by adding a pattern match for the drive to 
} src/sys/dev/ata/wd.c to set the WD_QUIRK_FORCE_LBA48.
} 
} Firstly, is this the correct place to fix it?
} 
} Secondly, should I submit this as a change/bug request? (I can't believe 
} noone else has this issue with these drives??)
} 
} Any help much appreciated.
} 
} I'm building from source, currently tracking 4.0-STABLE, uname -a says:
} 
} (ian:/usr/src/sys/dev/ata)$ uname -a
} NetBSD thejolly.demon.co.uk 4.0_STABLE NetBSD 4.0_STABLE (spiderpig) #0: 
} Wed Mar 12 21:02:08 GMT 2008 
} root%thejolly.demon.co.uk@localhost:/usr/src/sys/arch/i386/compile/spiderpig 
i386
} 
} 
} Cheers
} 
} Ian
} 
} -- 
} Ian Clark
} 
} 
>-- End of excerpt from Ian