Date: Fri, 22 Jan 1999 12:06:33 +1100 From: David Dawes <dawes@rf900.physics.usyd.edu.au> To: =?iso-8859-1?Q?S=F8ren_Schmidt?= <sos@freebsd.dk> Cc: current@FreeBSD.ORG Subject: Re: Promise FastTrack PCI IDE controller Message-ID: <19990122120633.A13400@rf900.physics.usyd.edu.au> In-Reply-To: =?iso-8859-1?Q?=3C199901211306=2EOAA16593=40freebsd=2Edk=3E=3B_from_S=F8?= =?iso-8859-1?Q?ren_Schmidt_on_Thu=2C_Jan_21=2C_1999_at_02:06:22PM_%2B0100?= References: <19990121234651.F5377@rf900.physics.usyd.edu.au> <199901211306.OAA16593@freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jan 21, 1999 at 02:06:22PM +0100, Søren Schmidt wrote: >It seems David Dawes wrote: >> >> That didn't improve things, but with the patch I got lots of messages >> like the following (for a disk attached to the onboard PIIX4): >> >> wd1: DMA failure, DMA status 0 >> >> (The box isn't SMP.) > >Hmm, and it is an uptodate -current system ?? >Just use this part then: It might have been related to some of the other changes I had made. With a new clean wd.c and your patch I don't get the "wd1: DMA failure, DMA status 0" messages. The Promise card doesn't behave any differently though. >> >I also run the card without its BIOS, as that tends to get in the way. >> >I think the FastTrack is just an Ultra/33 with another BIOS chip on it. >> >> I'll try pulling the BIOS tomorrow. We didn't get this card for it's >> RAID features anyway. > >That helped me in a lot if situations, but performance wont be as good >as we seem not to initialize the chip(s)... When I remove the BIOS it works OK. The initialisation is different though. For the drives on the PIIX4 I get: wdc0 at 0x1f0-0x1f7 irq 14 flags 0xa0ffa0ff on isa wdc0: unit 0 (wd0): <IBM-DTTA-351680>, DMA, 32-bit, multi-block-16 wd0: 16124MB (33022080 sectors), 32760 cyls, 16 heads, 63 S/T, 512 B/S wd0: ATA INQUIRE valid = 0007, dmamword = 0007, apio = 0003, udma = 0407 for the ones on the Promise card I get: wdc2 at 0xeff0-0xeff7 flags 0xa0ffa0ff on isa wdc2: unit 0 (wd4): <IBM-DTTA-351680>, DMA, 32-bit, multi-block-16 wd4: 16124MB (33022080 sectors), 32760 cyls, 16 heads, 63 S/T, 512 B/S wd4: ATA INQUIRE valid = 0007, dmamword = 0407, apio = 0003, udma = 0007 I also get the following warning, but only for the secondary controller on the Promise card: ide_pci: generic_dmainit efa8:0: warning, IDE controller timing not set and in the PCI probe section: ide_pci: busmaster 0 status: 60 from port: 0000ef82 ide_pci: ide0:0 has been configured for DMA by BIOS ide_pci: ide0:1 has been configured for DMA by BIOS ... ide_pci: busmaster 1 status: 00 from port: 0000ef8a The performance isn't as good, as you say. For one of these disks on the PIIX4 I get about 12 MByte/s for a large (1GB) sequential read from the raw device. I only get about 4 MByte/s for one of the same disks on the Promise controller. I presume that's because it isn't initialised to the fastest DMA mode? The performance is adequate for our needs though, so removing the BIOS is a workable solution. Thanks. David To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990122120633.A13400>