Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 27 Aug 2003 20:00:20 +0200 (CEST)
From:      Soren Schmidt <sos@spider.deepcore.dk>
To:        Sean Chittenden <sean@chittenden.org>
Cc:        freebsd-current@FreeBSD.ORG
Subject:   Re: ata crash when disk io saturated on laptop...
Message-ID:  <200308271800.h7RI0Kgh014678@spider.deepcore.dk>
In-Reply-To: <20030827173505.GA72378@perrin.nxad.com>

next in thread | previous in thread | raw e-mail | index | archive | help
It seems Sean Chittenden wrote:
> I haven't been able to get any kind of a stack trace, but if I run two
> 'cvs up's at the same time on my laptop (one of /usr/src, the other of
> /usr/ports), my laptop eventually locks up in what looks like the disk
> thrashing continually.  When I pull the plug and restart, I am only
> able to find this in the logs:
> 
> Aug 27 10:05:44 localhost kernel: ad0: READ command timeout tag=0 serv=0 - resetting
> Aug 27 10:05:44 localhost kernel: ata0: resetting devices ..
> Aug 27 10:05:44 localhost kernel: done
> [boot msg follows]
> 
> I can let my laptop sit for up to 5min grinding away, but caps lock
> doesn't work, acpi doesn't work, and my drive just cranks away
> resetting itself as fast as it can (resetting itself frequently
> given the sounds) in a nasty unbreakable loop.  The HDD is a few
> months old and has been pretty battle proven to date, though I'd never
> done anything quite this disk intensive.  It doesn't happen instantly
> either, it takes a good 5min or so before the drive begins to stumble.
> The relevant dmesg info:
> 
> atapci0: <Intel PIIX4 UDMA33 controller> port 0x860-0x86f at device 7.1 on pci0
> ata0: at 0x1f0 irq 14 on atapci0
> ata1: at 0x170 irq 15 on atapci0
> ad0: 38154MB <IC25N040ATCS04-0> [77520/16/63] at ata0-master UDMA33
> 
> hw.ata.ata_dma: 1
> hw.ata.wc: 1
> hw.ata.tags: 0
> hw.ata.atapi_dma: 0
> 
> What's more odd, however, is that I can't reproduce this if I'm not
> listening to MP3s, it's like the extra interrupts from the snd card
> push it over the edge in terms of latency and the kernel improperly
> resets ad0 or resets it incorrectly.
> 
> Soren, is this something that would be addressed in your recent ATAng
> commit? :-/

It is propably not worse at least, it does sound a bit like busmastering
trouble to me if it only happens when sound is playing...

-Søren



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200308271800.h7RI0Kgh014678>