Date: Sat, 21 Jul 2001 20:11:53 -0700 (PDT) From: Ed Alley <alley1@llnl.gov> To: Mike Meyer <mwm@mired.org> Cc: Ed Alley <alley1@llnl.gov>, <freebsd-questions@freebsd.org> Subject: Re: Why panic when mounting bad CD? Message-ID: <20010721200752.H4193-100000@jordan.llnl.gov> In-Reply-To: <15194.16500.171698.883443@guru.mired.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I guess the way to find out is to panic the kernel and get it to coredump. I've never done it deliberately before. If anyone reading this knows how to panic the kernel without fouling up my filesystem I'd like to hear it; it would be interesting to find out just what is causing the panic, maybe the fix is simple. Ed On Sat, 21 Jul 2001, Mike Meyer wrote: > Ed Alley <alley1@llnl.gov> types: > > So fixating a disk is the same as writing and/or validating the > > file system descriptor tables for the ISO filesystem on the CD. > > I see the problem. > > I don't know that that's what fixating a disk actually does. It seems > more likely that "fixating" writes an "end of volume" mark of some > kind, and without that you just read off the end of what you are > written. That's just speculation, though. > > > Thanks for the explanation Mike, I learned alot; if I knew anything > > about ISO9660 I would try to figure something out myself. :( > > I don't know very much about it myself. I know a bit about what it's > capable of, but not any real details. > > You're welcome, > <mike > > > On Sat, 21 Jul 2001, Mike Meyer wrote: > > > > > Ed Alley <alley1@llnl.gov> types: > > > > I am running FreeBSD 4.3 with an HP 9500 ide CD-RW. > > > > > > > > My question is why does the kernel panic > > > > under these conditions? Does it get lost > > > > trying to find the disklabel or does the > > > > kernel think that the numbers it picked > > > > up are the disklabel and merrily goes > > > > off into some other world looking for > > > > the superblock? > > > > > > CD's generally use the ISO 9660 file system, not UFS. They don't have > > > superblocks or disklabels, they have volume descriptors. And yes, if a > > > file system has bogus description data in it, the system tends to > > > crash while trying to deal with it. This is true for most file > > > systems. > > > > > > > Wouldn't it be a good idea to have a > > > > magic number in with the disklabel > > > > that the kernel could test against > > > > before it believes what it finds? > > > > > > UFS has a magic number in the superblock, and it's checked. ISO 9660 > > > file systems also have a magic number in the volume descriptor, and > > > that's checked. The code in -stable check for either the ISO 9660 or > > > the ISO SIERRA magic numbers, as it understands both formats. > > > > > > Neither of these guarantee that the file system descriptor tables are > > > actually correct. Mounting a file system with a valid magic number and > > > bogus descriptor tables is generally a bad idea. It appears that an > > > unfixated CD meets that description. There may be a way to detect that > > > the CD is unfixated; if so, submitting a patch to the > > > /usr/src/sys/isofs/cd9660_vfsops.c via a PR would be appreciated. > > > > > > <mike > > > -- > > > Mike Meyer <mwm@mired.org> http://www.mired.org/home/mwm/ > > > Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information. > > > > > > > > > > > -- > Mike Meyer <mwm@mired.org> http://www.mired.org/home/mwm/ > Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information. > > 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?20010721200752.H4193-100000>