Date: Sat, 11 Oct 1997 09:16:05 +0200 From: Stefan Esser <se@FreeBSD.ORG> To: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de> Cc: scsi@FreeBSD.ORG Subject: Re: Trouble with dump on ncr Message-ID: <19971011091605.32335@mi.uni-koeln.de> In-Reply-To: <19971010212337.WB21205@uriah.heep.sax.de>; from J Wunsch on Fri, Oct 10, 1997 at 09:23:37PM %2B0200 References: <199710100624.KAA06724@asteroid.mgt.msk.ru> <19971010212337.WB21205@uriah.heep.sax.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On 1997-10-10 21:23 +0200, J Wunsch <j@uriah.heep.sax.de> wrote: > Re-read the freebsd-scsi archives... and then either disable tagged > command queuing (``options FAILSAFE''), or alternatively remove all > calls to scsi_start_unit() in /sys/scsi/sd.c. > > Btw., i think we are safe to make the latter change anyway. As i've > already mentioned in a private mail to Justin, i've recently proven > that the calls to scsi_start_unit() are completely useless and don't > work. (I had taken down a disk manually, and later tried to > implicitly start it by opening it again.) Thus, by removing these > calls, we won't break anything that's not already broken. > > Anybody objecting? I'm all for the removal of the scsi_start_unit() calls. The start unit command should instead be issued from the recovery code in the generic SCSI layer, IMHO, whenever the return status from the SCSI card driver indicates a drive has (been) stopped in an attempt to recover from that situation. Regards, STefan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19971011091605.32335>