Date: Wed, 22 Jan 2003 10:53:14 -0800 From: Nathan Kinkade <nkinkade@dsl-only.net> To: questions@FreeBSD.ORG Subject: Re: ata1 resetting Message-ID: <20030122185314.GK25795@sub21-156.member.dsl-only.net> In-Reply-To: <20030122163834.F5592-100000@voo.doo.net> References: <20030122153130.GI25795@sub21-156.member.dsl-only.net> <20030122163834.F5592-100000@voo.doo.net>
next in thread | previous in thread | raw e-mail | index | archive | help
--x+RZeZVNR8VILNfK Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Jan 22, 2003 at 04:49:40PM +0100, Marc Schneiders wrote: > On Wed, 22 Jan 2003, at 07:31 [=3DGMT-0800], Nathan Kinkade wrote: > > On Wed, Jan 22, 2003 at 07:58:31AM +0100, Marc Schneiders wrote: >=20 > [...] > > > Jan 21 19:47:39 pan /kernel: ata1: resetting devices .. ata1-slave: > > > ATA identify retries exceeded > > > Jan 21 19:47:39 pan /kernel: done > > > Jan 21 22:06:39 pan /kernel: ad0: WRITE command timeout tag=3D0 serv= =3D0 - > > > resetting > > <snip> > > > > Do you have DMA enabled on those drives when possibly they don't support > > it? >=20 > sysctl reports: hw.ata.ata_dma: 1 >=20 > > What type of ribbon cable are you using - a 40 or 80 conductor? >=20 > 40, it is a rather old motherboard, as I mentioned. >=20 > > Try setting the sysctl(8) value "hw.ata.ata_dmai" to 0 and see what res= ults > > you get. The errors you are getting look similar to ones I've seen > > where the kernel is trying to use DMA on a drive that doesn't support > > it, or on a drive that supports DMA that is using an improper 40 > > conductor cable instead of the correct 80 conductor cable. You can use > > the atacontrol(8) utilitly to find out more about the capabilities of > > your devices. For example, `atacontrol cap 0 0` should give you all > > manner of info about your primary master ATA device. >=20 > This is what it says: >=20 > ATA channel 0, Master, device ad0: >=20 > ATA/ATAPI revision 4 > device model ST34312A > serial number [secret] > firmware revision 3.09 > cylinders 8354 > heads 16 > sectors/track 63 > lba supported 8420832 sectors > lba48 not supported > dma supported > overlap not supported >=20 > Feature Support Enable Value Vendor > write cache yes yes > read ahead yes yes > dma queued no no 0/00 > SMART yes no > microcode download yes yes > security yes no > power management yes yes > advanced power management no no 0/00 > automatic acoustic management no no 0/00 0/00 >=20 > So it can do DMA, but doesn't use it?? >=20 > I have another machine with the same hardware, except for video and > ethernet cards, that does _not_ have the error messages (and the > problem). It also has just one harrdisk, and the problem machine two, > on two channels. Can that be it? What does the command `grep "ad[0-9]" /var/run/dmesg.boot` return. Specifically, look at the end of the line which defines each of your disks and you should see how the device was detected (PIO, UDMA, etc). Regarding the ribbon cable, the 80 conductor cable is really on required for UDMA mode 3 and above, although maybe recommended for all UDMA modes. In any case, as far as I know, FreeBSD should auto-detect the presence, or absence, of an 80 conductor cable and set dma modes accordingly. I'm not sure that dma/udma is causing your issues at all, but it certainly couldn't hurt to turn off dma and see if the problems go away. Nathan --=20 GPG Public Key ID: 0x4250A04C gpg --keyserver pgp.mit.edu --recv-keys 4250A04C http://63.105.21.156/gpg_nkinkade_4250A04C.asc --x+RZeZVNR8VILNfK Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE+LuiaWZYS9EJQoEwRArcrAJ9pReXEaZzygfuw+XiZLClaqrJGhACg8N8e srKpUmhd7rhDI5859fGotiU= =hynl -----END PGP SIGNATURE----- --x+RZeZVNR8VILNfK-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030122185314.GK25795>