Date: Fri, 12 Sep 1997 01:55:14 -0600 (MDT) From: Atipa <freebsd@atipa.com> To: Greg Lehey <grog@lemis.com> Cc: FreeBSD Questions <questions@FreeBSD.ORG>, FreeBSD Hackers <hackers@FreeBSD.ORG> Subject: Re: Do *you* have problems with floppies? Message-ID: <Pine.BSF.3.91.970912013832.5749A-100000@dot.ishiboo.com> In-Reply-To: <19970912101014.37786@lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 12 Sep 1997, Greg Lehey wrote: > I've seen a lot of reports recently about problems with floppies under > FreeBSD. Now I can understand a lot of that: floppies must be the > most unreliable data storage medium I can think of, not to mention the > most expensive per byte. But I'm getting the feeling that there is > more to it than that, that possibly there's a bug in the floppy driver > and that we're blaming it on the inherent unreliability of the medium. I very consistently get floppy errors w/ FreeBSD, on systems that I know for certain have good drives, FDCs, and good media. Whenever I need to install XInside products, like Motif or AcceleratedX, I always need to untar the floppies on a Linux box and nfs export them. The confusing thing to me is that floppy support works fine during installation. I just gave up on using them on an active system. We have so much networking junk laying around that I never need floppies, or else I would be highly annoyed. > I'm looking for indications which would point towards the driver. One > of these might be: > > 1. Floppy formatted under <insert your OS here> on the same machine. Either MS-DOS format or raw TAR format. I haven't played w/ ext2 or ufs formats very much. It does not appear to be format dependent; I think it may be interrupt handling. > 2. FreeBSD runs into hardware problems with the floppy (typically > things like checksum errors). Yes. It is like the media/host tranfers are out of sync with the host/controller transfers. The expected data appears to be late, or perhaps early (doubtful). > 3. <insert your OS here> can read the entire floppy with no trouble. MS-DOG based and Linux are fully reliable. NetBSD is unrelaible. > If you can give me hard evidence of such occurrences, I'd like to hear > from you. I know that plenty of people can tell me that they've had > occurrences of (2), maybe in conjunction with (1), but unless you can > prove (3) as well, I don't want to hear from you. Case in point: the XInside. The medium is pristine, as is shown by Linux's ability to untar the floppy no problem. The machine itself is fine because DOS reads floppies fine, and sysinstall was fine. > In addition, if you have any other evidence I haven't thought of which > would also point to an error in the floppy driver, please contact me. I certainly think BIOS negotiation needs to be investigated. We use several different types of motherboards from different manufacturers, but they all have the Award BIOS in common. I had a NetBSD client who had fits with Award-based boards, but had no problem at all w/ AMI or Phoenix BIOSes. I know Award is not broken, but it may be _different_ than the BIOS that was used to test the relaibility of floppies under FreeBSD. I am happy to give more specifics if you tell me what you need. Kevin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.970912013832.5749A-100000>