Date: 16 Mar 98 12:00:09 +0100 From: leifn@image.dk (Leif Neland) To: freebsd-scsi@FreeBSD.ORG Subject: Re: spin-up at start giving timeout. Message-ID: <724_9803161715@swimsuit.swimsuit.roskildebc.dk> References: <692_9803132215@swimsuit.swimsuit.roskildebc.dk> <199803160038.RAA07073@narnia.plutotech.com> <19980316083756.59368@uriah.heep.sax.de>
next in thread | previous in thread | raw e-mail | index | archive | help
At 16 Mar 98 08:37:56 J Wunsch wrote regarding Re: spin-up at start giving timeout. >> The problem is that a timeout on the command is used at all. In CAM, >> if a device is detected as needing to be given a "start unit", the >> command is given with the immediate flag set, and the device is polled >> once second to see if it has come ready. JW> But as i understand, the old code has no other chance than JW> timing it out, has it? (At least, not without rewriting the JW> mess, which is surely out of the question, given that the CAM JW> code is `around the corner'.) I can only say that after I increased the timeout of scsi_start_unit() to 30000 ms, it works here. I put a printf before and after scsi_start_unit, and I can see scsi_start_unit simply waits until the disk has spun up. This delay is hardcoded in scsi_base.c; I had the idea the "be pessimistic about joe scsi device" had an effect here, but it doesn't. So I'll put it back at 5 seconds again. Leif Neland leifn@image.dk --- |Fidonet: Leif Neland 2:234/49 |Internet: leifn@image.dk To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?724_9803161715>