Date: Thu, 28 Jun 2001 08:36:06 +0200 (CEST) From: Søren Schmidt <sos@freebsd.dk> To: Andrew Reilly <a.reilly@lake.com.au> Cc: freebsd-stable@freebsd.org, Soren Schmidt <sos@freebsd.org> Subject: Re: Undesirable behaviour of burncd erase Message-ID: <200106280637.f5S6bQJ78303@freebsd.dk> In-Reply-To: <20010627221236.17737.qmail@areilly.bpc-users.org> "from Andrew Reilly at Jun 28, 2001 08:12:36 am"
next in thread | previous in thread | raw e-mail | index | archive | help
It seems Andrew Reilly wrote: > Hi all, but particularly Soren, Hi :) > I run > FreeBSD 4.3-STABLE #5: Sat Jun 16 11:45:32 EST 2001 > root@gurney.reilly.home:/usr/obj/usr/src/sys/GURNEY > > acd0: CD-RW <RICOH DVD/CDRW MP9120> at ata1-slave using UDMA33 > ad0: 6149MB <QUANTUM FIREBALL EX6.4A> [13328/15/63] at ata0-master UDMA33 > ad1: 6149MB <QUANTUM FIREBALL EX6.4A> [13328/15/63] at ata1-master UDMA33 > > Over the next few seconds my X session locked-up hard, apart from > the mouse, which could still move. I was also unable to connect > to the box over my LAN, using ssh. > > Since the CD erase operation still seemed to be happening, I > decided to just let it go for a while, to see if that would stop. > Sure enough, many minutes later (I didn't time it), the disk > ejected itself, and my X session restarted itself, probably > because I had been madly pressing ctrl-alt-backspace while things > were locked up. So the system did not actually crash at all, but > it became unuseful for a very long time. Afterwards, the logs and > the console were seen to contain these messages: > > > swap_pager: indefinite wait buffer: device: #ad/0x30009, blkno: 7936, size: 36864 > > I'm guessing that this is not expected behaviour. Is there > anything that I can do to stop it from happening, besides just not > using the burncd erase command? Should I have some part of my > system configured differently? Well, its a "feature" of how things work on an ATA channel :) The problem is that the erase command takes *a long time* and the ATAPI device claims the channel during the entire operation the way things are done in -stable. However, in -current I've implemented a way for the device to release the channel during long operations, so that the system can get at the disk during the erase. Now the problem is that that part of the spec is not read by most firmware writers, so I've had a hell of a time getting it to work on as many devices as possible, and thats why it isn't in stable yet. -Søren To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200106280637.f5S6bQJ78303>