Date: Wed, 11 Sep 2002 07:09:09 +0200 From: Grigori Goronzy <greg@chown.ath.cx> To: freebsd-stable@FreeBSD.org Subject: problem with ATA tagged queueing Message-ID: <20020911070909.26191379.greg@chown.ath.cx>
next in thread | raw e-mail | index | archive | help
I'm using two harddisks: ad0: 14664MB <IBM-DJNA-351520> [29795/16/63] at ata0-master UDMA66 ad1: 78533MB <IC35L080AVVA07-0> [159560/16/63] at ata0-slave UDMA100 when I switch on tagged queuing by setting hw.ata.tags="1" in loader.conf, the second disk gets tags(the first can't do this). It firstly works, but after some time(2 days or so) I get some strange messages, and the first disk is resetted to udma33, the second to pio4: Sep 7 23:04:53 duron /kernel: ad1: READ command timeout tag=0 serv=1 - resettin g Sep 7 23:04:53 duron /kernel: ad1: invalidating queued requests Sep 7 23:04:53 duron /kernel: ata0: resetting devices .. ad0: DMA limited to UD MA33, non-ATA66 cable or device Sep 7 23:04:53 duron /kernel: ad1: invalidating queued requests Sep 7 23:04:53 duron /kernel: done Sep 7 23:04:53 duron /kernel: ad1: no request for tag=0 Sep 7 23:04:53 duron /kernel: ad1: invalidating queued requests Sep 7 23:05:03 duron /kernel: ad1: READ command timeout tag=0 serv=1 - resettin g [...] (these messages repeat very often) But, the first strangeness, I'm using a ATA66(80pol) cable. I tried two different cables so far. I'm using FreeBSD duron.hell 4.6.2-RELEASE FreeBSD 4.6.2-RELEASE #0: Mon Sep 2 16:23:46 CEST 2002 root@duron.hell:/usr/src/sys/compile/GENERIC i386 My motherboard is an ECS K7VZA with an VIA chipset atapci0: <VIA 82C686 ATA100 controller> port 0xc000-0xc00f at device 7.1 on pci0 atapci0: Correcting VIA config for southbridge data corruption bug I heard there are some tagge queueing related bugs in RELENG_4_6.. is this true? greg 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?20020911070909.26191379.greg>