Date: Wed, 18 Feb 1998 16:15:35 -0700 From: "Justin T. Gibbs" <gibbs@plutotech.com> To: Tom <tom@sdf.com> Cc: "Justin T. Gibbs" <gibbs@plutotech.com>, Wee Teck Ng <weeteck@eecs.umich.edu>, freebsd-scsi@FreeBSD.ORG Subject: Re: very slow scsi performance Message-ID: <199802182318.QAA01242@pluto.plutotech.com> In-Reply-To: Your message of "Wed, 18 Feb 1998 14:21:54 PST." <Pine.BSF.3.95q.980218141739.20019A-100000@misery.sdf.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> Didn't the IBM DCAS or DORS drive have some firmware bugs in regards to >tags? Are you referrring to the newer IBM Ultrastar line? The problems that have been reported against many IBM drives stem from the fact that the DQUE bit is set on many OEM models. The current SCSI system does not look at the control mode page to see this, attempts to perform tagged queuing operation anyway, and fails miserably. The IBM drives are behaving completely within the bounds of the SCSI spec. CAM handles this "problem" by honoring the DQUE bit assuming that the user knows what they are doing. > In fact, I just found a message from a DCAS sent to a freebsd list >today, that has to disable disconnection to get reliable operation with >his 2940 and his DCAS drive. It is probably the same problem I listed above. Disabling disconnection effectively prevents tags from being used. >I can't verify that, but I have few problems >with 11 Seagate drives hung off a 3940UW. The ahc driver freezes up the >system every 6 weeks of heavy 24x7 i/o though. I'm working on finishing up this 7890 support so that I can release an updated CAM snapshot. Once that happens, you might want to try CAM and see if it addresses this problem. >Tom -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802182318.QAA01242>