Date: Sat, 25 Nov 1995 02:59:21 GMT From: Mark Valentine <mark@linus.demon.co.uk> To: Julian Elischer <julian@ref.tfs.com>, joerg_wunsch@uriah.heep.sax.de Cc: freebsd-scsi@freebsd.org Subject: Re: kern/820: scsi tape problems Message-ID: <199511250259.CAA00258@linus.demon.co.uk> In-Reply-To: Mark Valentine's message of Nov 14, 12:43am
next in thread | raw e-mail | index | archive | help
A little more on the SCSI tape timeout problems I reported in the PR
reproduced below...
After noticing that an ASPI version of GNU tar under DOS would eventually
get past the troublesome spot on my tape, I grubbed around in st.c and
bumped up a couple of hard-wired timeouts passed to scsi_scsi_cmd() -
there were two which looked likely, both set to 100000 - I added a
significant zero to both.
This allowed to me to successfully read the tape (yes, my tape drive
really did take >100s to read that block!).
It looks like the code should recover at this point - tar indeed gets
an I/O error, but the aha_timeout() sets a 4 second timeout on the abort,
and this seems to end me up in a call to Debugger("1542"), having erroneously
diagnosed my adapter as frozen...
Mark.
> >Number: 820
> >Category: kern
> >Synopsis: scsi tape problems
> >Confidential: no
> >Severity: serious
> >Priority: medium
> >Responsible: freebsd-bugs
> >State: open
> >Class: sw-bug
> >Submitter-Id: current-users
> >Arrival-Date: Sun Nov 12 15:50:01 PST 1995
> >Last-Modified: Mon Nov 13 00:00:01 PST 1995
> >Originator: Mark Valentine
> >Organization:
> >Release: FreeBSD 2.2-CURRENT i386
> >Environment:
>
> FreeBSD 2.2-CURRENT #2: Sun Nov 12 19:29:29 GMT 1995
> mark@linus.demon.co.uk:/usr/src/sys/compile/LINUS
> CPU: i486 DX2 (486-class CPU)
> Origin = "GenuineIntel" Id = 0x435 Stepping=5
> Features=0x3<FPU,VME>
> real memory = 16777216 (16384K bytes)
> avail memory = 14868480 (14520K bytes)
> Probing for devices on the ISA bus:
> sc0 at 0x60-0x6f irq 1 on motherboard
> sc0: VGA color <16 virtual consoles, flags=0x0>
> sio0 at 0x3f8-0x3ff irq 4 on isa
> sio0: type 16450
> sio1 at 0x2f8-0x2ff irq 3 on isa
> sio1: type 16550A
> sio2 not found at 0x3e8
> lpt0 at 0x378-0x37f irq 7 on isa
> lpt0: Interrupt-driven port
> lp0: TCP/IP capable interface
> pca0 on motherboard
> pca0: PC speaker audio driver
> fdc0 at 0x3f0-0x3f7 irq 6 drq 2 on isa
> fdc0: NEC 72065B
> fd0: 1.44MB 3.5in
> wdc0 at 0x1f0-0x1f7 irq 14 on isa
> wdc0: unit 0 (wd0): <Maxtor 7850 AV>
> wd0: 814MB (1667232 sectors), 1654 cyls, 16 heads, 63 S/T, 512 B/S
> aha0 at 0x334-0x337 irq 11 drq 5 on isa
> aha0 waiting for scsi devices to settle
> (aha0:0:0): "QUANTUM PD1050iS 3072" type 0 fixed SCSI 2
> sd0(aha0:0:0): Direct-Access 1003MB (2055096 512 byte sectors)
> sd0(aha0:0:0): with 2448 cyls, 12 heads, and an average 69 sectors/track
> (aha0:4:0): "WANGTEK 5525ES SCSI M74H 11" type 1 removable SCSI 1
> st0(aha0:4:0): Sequential-Access density code 0xf, 512-byte blocks, write-enabled
> (aha0:6:0): "TOSHIBA CD-ROM XM-3401TA 2873" type 5 removable SCSI 2
> cd0(aha0:6:0): CD-ROM
> cd0(aha0:6:0): NOT READY asc:3a,0 Medium not present
> can't get the size
>
> npx0 on motherboard
> npx0: INT 16 interface
> sb0 at 0x220 irq 5 drq 1 on isa
> sb0: <MV Jazz16 3.1>
> opl0 at 0x388 on isa
> opl0: <Yamaha OPL-3 FM>
> uart0 at 0x330 irq 9 on isa
> uart0: <6850 Midi Interface>
> joy0 at 0x201 on isa
> joy0: joystick
> RTC BIOS diagnostic error 16<memory_size,invalid_time>
> new masks: bio c0004840, tty c003009a, net c003009a
> WARNING: / was not properly dismounted.
>
> [The Adaptec is a 1542C (an early one, but all devices are internal) -
> SCSI_DELAY=15.]
>
> >Description:
>
> (1) I had a system hang while reading a QIC-120 tape (co-inciding with
> the timeout shown below).
>
> Nov 12 20:02:54 linus /kernel: st0(aha0:4:0): MEDIUM ERROR info:11 field replaceable unit: 1
> Nov 12 20:18:21 linus /kernel: st0(aha0:4:0): timed out
>
> The first error seems to be unrelated, since it didn't show up on
> subsequent attempts, but the problem occurs repeatably at the same
> place on the tape (about two thirds through).
>
> The tape error isn't a simple retry - the head shuffles about on
> the tape considerably. The tape wasn't written on this drive.
> (Could this be a tape drive problem?)
>
> On my third attempt I came out of X, and this time the system timed
> out and trapped into DDB ("adapter not taking commands.. frozen?!").
>
> (2) The system doesn't like to be rebooted while the tape is rewinding;
> SCSI_DELAY=15 is nowhere near enough here! The system panics in the
> usual way after a SCSI timeout.
>
> Is this just because the probe isn't clever enough to delay talking
> to the SCSI device, or should it be able to find out what it needs
> from the device even while rewinding?
>
> >How-To-Repeat:
>
> Happens at the same position on my tape each time.
>
> >Fix:
>
> >Audit-Trail:
>
> From: J Wunsch <j@uriah.heep.sax.de>
> To: mark@linus.demon.co.uk
> Cc: FreeBSD-gnats-submit@FreeBSD.org
> Subject: Re: kern/820: scsi tape problems
> Date: Mon, 13 Nov 1995 08:47:42 +0100 (MET)
>
> As Mark Valentine wrote:
>
> > Nov 12 20:02:54 linus /kernel: st0(aha0:4:0): MEDIUM ERROR info:11 field replaceable unit: 1
>
> This is clearly a problem with your tape (and/or drive).
>
> > Nov 12 20:18:21 linus /kernel: st0(aha0:4:0): timed out
>
> This looks like a subsequent problem in the driver (since your tape
> drive did lengthy attempts to recover from the above error). I've
> once been discussing the problems of adapter timeouts with Peter
> Dufault. I think we came to the conclusion to make something else
> than now, but i forgot what it was. :-)
>
> --
> cheers, J"org
>
> joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE
> Never trust an operating system you don't have sources for. ;-)
> >Unformatted:
--
"Tigers will do ANYTHING for a tuna fish sandwich."
"We're kind of stupid that way." *munch* *munch*
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511250259.CAA00258>