Date: Sun, 8 Oct 1995 14:53:10 -0400 From: Mark Starner <starner@clark.net> To: "'bugs@freebsd.org'" <bugs@freebsd.org> Cc: "'moriya@ifi.unicamp.br'" <moriya@ifi.unicamp.br> Subject: Bug accessing /dev/cd1 Message-ID: <01BA958D.CA1F82A0@ras2.HAN.UnisysGSG.COM>
next in thread | raw e-mail | index | archive | help
Was a fix ever found for this bug? I had the same problem, and sent my system info to
(what were then) the appropriate people. I just loaded up the current snapshot
and have the same problem... a panic as below, at the same place when accessing cd1.
(actually it may be worse, it used to never crash with cd0, now it does sometimes)
FYI, I am running with a AHA-3940, a NEC 3xi and Plextor 4-Plex CD.
Mark
>From: Alexandre Moriya - esp <moriya@ifi.unicamp.br>
Subject: KERNEL BUG !!!!
To: bugs@FreeBSD.org
Date sent: Fri, 11 Aug 1995 11:25:24 -0300 (BSC)
Copies to: hackers@FreeBSD.org, questions@FreeBSD.org,
moriya@ifi.unicamp.br (Alexandre Moriya - esp)
THIS mail has been sent to {bugs,hackers,questions}@FreeBSD.org
Hi all,
I don't if it is the right place for this question. If it isn't,
could you please send it to the right people ?
I believe it is a KERNEL BUG.
I have the following configuration:
1 PC/FreeBSD-2.0.5
1 Seven CD Tower (SCSI Express Tower / MDI)
1 Adaptec 1542CF ISA-to-SCSI adaptor
The 1542CF adaptor is configured as the kernel messages bellow says:
Aug 9 12:54:30 by-tor /kernel: aha0: AHA-1542CF BIOS v2.01-VB.0, enabling mai
lbox, enabling residuals
Aug 9 12:54:30 by-tor /kernel: aha0: reading board settings, dma=5 int=11
(bus speed defaulted)
Aug 9 12:54:30 by-tor /kernel: aha0 at 0x330-0x333 irq 11 drq 5 on isa
Aug 9 12:54:30 by-tor /kernel: aha0 waiting for scsi devices to settle
Aug 9 12:54:30 by-tor /kernel: (aha0:0:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:30 by-tor /kernel: cd0(aha0:0:0): CD-ROM
Aug 9 12:54:30 by-tor /kernel: cd0(aha0:0:0): NOT READY asc:3a,0 Medium
not present
Aug 9 12:54:30 by-tor /kernel: can't get the size
Aug 9 12:54:31 by-tor /kernel:
Aug 9 12:54:31 by-tor /kernel: (aha0:1:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:31 by-tor /kernel: cd1(aha0:1:0): CD-ROM
Aug 9 12:54:31 by-tor /kernel: cd1(aha0:1:0): NOT READY asc:3a,0 Medium
not present
Aug 9 12:54:31 by-tor /kernel: can't get the size
Aug 9 12:54:31 by-tor /kernel:
Aug 9 12:54:31 by-tor /kernel: (aha0:2:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:32 by-tor /kernel: cd2(aha0:2:0): CD-ROM
Aug 9 12:54:32 by-tor /kernel: cd2(aha0:2:0): NOT READY asc:3a,0 Medium
not present
Aug 9 12:54:32 by-tor /kernel: can't get the size
Aug 9 12:54:32 by-tor /kernel:
Aug 9 12:54:32 by-tor /kernel: (aha0:3:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:32 by-tor /kernel: cd3(aha0:3:0): CD-ROM cd present.[183255 x 204
8 byte records]
Aug 9 12:54:32 by-tor /kernel: (aha0:4:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:33 by-tor /kernel: cd4(aha0:4:0): CD-ROM
Aug 9 12:54:33 by-tor /kernel: cd4(aha0:4:0): NOT READY asc:3a,0 Medium
not present
Aug 9 12:54:33 by-tor /kernel: can't get the size
Aug 9 12:54:33 by-tor /kernel:
Aug 9 12:54:33 by-tor /kernel: (aha0:5:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:33 by-tor /kernel: cd5(aha0:5:0): CD-ROM
Aug 9 12:54:33 by-tor /kernel: cd5(aha0:5:0): NOT READY asc:3a,0 Medium
not present
Aug 9 12:54:33 by-tor /kernel: can't get the size
Aug 9 12:54:33 by-tor /kernel:
Aug 9 12:54:34 by-tor /kernel: (aha0:6:0): "MDI SE53CD 0925" type 5 removable
SCSI 2
Aug 9 12:54:34 by-tor /kernel: cd6(aha0:6:0): CD-ROM
Aug 9 12:54:34 by-tor /kernel: cd6(aha0:6:0): NOT READY asc:3a,0 Medium not p
resent
Aug 9 12:54:34 by-tor /kernel: can't get the size
Aug 9 12:54:34 by-tor /kernel:
My problem is that I mounted a CD-ROM from the first device (/dev/cd0a)
in the directory /0 using "mount_cd9660 /dev/cd0a /0" or by a corresponding
line in the /fstab.
The CDROM has about 100MB of info in the root dir. When I try to "cat *" the /0 directory,
everything is fine.
Everything works well when I mount the same CDROM in the /dev/cdNa, where N=[0,5].
(I have created all devices using "cd /dev; sh MAKEDEV cdN;" where N=[0,6])
It does not made any difference if I used /dev/cdNc instead of /dev/cdNa...
The "bug" appears when I mount the CDROM in the /dev/cd6a (the last CDROM device).
I used "mount_cd9660 /dev/cd6a /0 ; cd /0 ; cat *" and a few seconds later
I received the following message:
Fatal trap 12: page fault while in kernel mode
fault virtual address = 0x10a
fault code = supervisor read, page not present
instruction pointer = 0x8:0xf011efdc
code segment = base 0x0, limit 0xfffff, type 0x1b
processor eflags = interrupt enabled, resume IOPL=0
current process = 133(cat)
interrupt mask = bio
panic: page fault
syncing disks..................... 4 4 4 4 4 4 4 4 4 giving up
Automaitc reboot in 15 seconds - press a key on the console to abort
-------
Is it easy to repair ? I took a look in the /usr/src/sys/i386/i386 dir and
found something, but nothing that lead me to any conclusion. I have seen
something concerning about this in the /usr/src/sys/scsi dir, too.
Can you help me ? Any tips ?
We here at CCJDR intend to use up to 3 "Seven CD Tower" but will
the kernel accept more than 6 devices (0-5) ? We want to use 21 !!!
Sorry for my poor english :-( If you have any doubt, ask me please.
We should have a solution to this problem as early as possible.
I believe that , if it is a kernel bug, it should be repaired before
the 2.1 release, isn't it ?
Finnaly, I'd like to thank you for the great work in the FreeBSD.
It'is working very nice !!!
BYE (Ate mais)
--
Alexandre Moriya (RA 910061)
Centro de Computacao John David Rogers (CCJDR)
Instituto de Fisica Gleb Wataghin (IFGW)
Universidade Estadual de Campinas (UNICAMP)
Campinas - SP - Brasil
e-mail:
moriya@ifi.unicamp.br
moriya@dcc.unicamp.br
--------------------
>From: J Wunsch <j@uriah.heep.sax.de>
Subject: Re: KERNEL BUG !!!!
To: moriya@ifi.unicamp.br (Alexandre Moriya - esp)
Date sent: Mon, 14 Aug 1995 20:30:30 +0200 (MET DST)
Copies to: joerg_wunsch@uriah.heep.sax.de, starner@clark.net, lenzi@mtm.ufsc.br,
davidg@freebsd.org (David Greenman)
Send reply to: joerg_wunsch@uriah.heep.sax.de (Joerg Wunsch)
As Alexandre Moriya - esp wrote:
> > > fault virtual address = 0x10a
> This number is variable. Each time the bug happens we have
> a different number.
Understandable. It's a bogus address.
> > > instruction pointer = 0x8:0xf011efdc
> OOOPS !!! The number is wrong.
> The correct number is 0xf012135c
> > nm /kernel | sort | more
> Ok. the new result is:
> f0121314 T _incore
> f01213b0 T _inmem
The bug happens in line 694 of vfs_bio.c:
686 incore(struct vnode * vp, daddr_t blkno)
687 {
688 struct buf *bp;
689 struct bufhashhdr *bh;
690
691 int s = splbio();
692
693 bh = BUFHASH(vp, blkno);
694 bp = bh->lh_first;
^^^^
Apparently, the BUFHASH macro yielded a bogus number that was an
invalid address (but not zero, since you're writing the number is
different each time).
I'm Cc'ing David Greenman, i think he's the one who would best
understand what happens here.
--
cheers, J"org
joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/
Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?01BA958D.CA1F82A0>