Date: Sun, 23 Nov 1997 10:22:05 +0100 (MET) From: "Helmut F. Wirth" <hfwirth@ping.at> To: FreeBSD-gnats-submit@FreeBSD.ORG Subject: i386/5128: Adaptec 2940U Timeouts with QUANTUM disk Message-ID: <199711230922.KAA00668@atlantis.ping.at> Resent-Message-ID: <199711230930.BAA00907@hub.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 5128
>Category: i386
>Synopsis: Adaptec 2940U Timeouts with QUANTUM disk
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: freebsd-bugs
>State: open
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Sun Nov 23 01:30:01 PST 1997
>Last-Modified:
>Originator: Helmut F. Wirth
>Organization:
private
>Release: FreeBSD 3.0-CURRENT i386
>Environment:
ASUS P55T2P4, Pentium 200 MMX, Adaptec 2940 Bios Ver.1.21
64MB Memory, 1x Quantum Atlas 2GB, 1x IBM 1GB, 1xQuantum Atlas 2GB
Plextor CDROM, HP 6020i Worm, Archive Viper Tape, Jaz SCSI
Version FreeBSD-current, last update today via CTM from
ftp.freebsd.org
(Last ctm upgrade included is now src-cur.3139.gz)
The newer QUANTUM ATLAS (sd2) has *upgraded* firmware, for the older
QUANTUM ATLAS such an upgrade does not exist. All disks worked well
up to the last change in the AHC script inside the driver.
The following is the relevant part of the boot sequence:
Nov 23 09:42:58 atlantis /kernel: ahc0: <Adaptec 2940 Ultra SCSI host adapter> rev 0x00 int a irq 10 on pci0.11.0
Nov 23 09:42:58 atlantis /kernel: ahc0: aic7880 Single Channel, SCSI Id=7, 16 SCBs
Nov 23 09:42:58 atlantis /kernel: ahc0: waiting for scsi devices to settle
Nov 23 09:42:58 atlantis /kernel: scbus0 at ahc0 bus 0
Nov 23 09:42:58 atlantis /kernel: sd0 at scbus0 target 0 lun 0
Nov 23 09:42:58 atlantis /kernel: sd0: <Quantum XP32150 81HB> type 0 fixed SCSI 2
Nov 23 09:42:58 atlantis /kernel: sd0: Direct-Access 2050MB (4199760 512 byte sectors)
Nov 23 09:42:58 atlantis /kernel: sd1 at scbus0 target 1 lun 0
Nov 23 09:42:58 atlantis /kernel: sd1: <IBM DPES-31080 S31Q> type 0 fixed SCSI 2
Nov 23 09:42:58 atlantis /kernel: sd1: Direct-Access 1034MB (2118144 512 byte sectors)
Nov 23 09:42:58 atlantis /kernel: sd2 at scbus0 target 2 lun 0
Nov 23 09:42:58 atlantis /kernel: sd2: <Quantum XP32150W L915> type 0 fixed SCSI 2
Nov 23 09:42:58 atlantis /kernel: sd2: Direct-Access 2151MB (4406960 512 byte sectors)
Nov 23 09:42:58 atlantis /kernel: sd3 at scbus0 target 3 lun 0
Nov 23 09:42:58 atlantis /kernel: sd3: <iomega jaz 1GB H.72> type 0 removable SCSI 2
Nov 23 09:42:58 atlantis /kernel: sd3: Direct-Access
Nov 23 09:42:58 atlantis /kernel: sd3: ILLEGAL REQUEST asc:24,0 Invalid field in CDB
Nov 23 09:42:58 atlantis /kernel: sd3 could not mode sense (4). Using ficticious geometry
Nov 23 09:42:58 atlantis /kernel:
Nov 23 09:42:58 atlantis /kernel: sd3: NOT READY asc:3a,0 Medium not present
Nov 23 09:42:58 atlantis /kernel: sd3: could not get size
Nov 23 09:42:58 atlantis /kernel: 0MB (0 512 byte sectors)
Nov 23 09:42:58 atlantis /kernel: cd0 at scbus0 target 4 lun 0
Nov 23 09:42:58 atlantis /kernel: cd0: <PLEXTOR CD-ROM PX-12CS 1.00> type 5 removable SCSI 2
Nov 23 09:42:58 atlantis /kernel: cd0: CD-ROM cd present [267521 x 2048 byte records]
Nov 23 09:42:59 atlantis /kernel: ahc0:A:5: refuses synchronous negotiation. Using asynchronous transfers
Nov 23 09:42:59 atlantis /kernel: worm0 at scbus0 target 5 lun 0
Nov 23 09:42:59 atlantis /kernel: worm0: <HP CD-Writer 6020 1.07> type 5 removable SCSI 2
Nov 23 09:42:59 atlantis /kernel: worm0: Write-Once
Nov 23 09:42:59 atlantis /kernel: st0 at scbus0 target 6 lun 0
Nov 23 09:42:59 atlantis /kernel: st0: <HP HP35480A T603> type 1 removable SCSI 2
Nov 23 09:42:59 atlantis /kernel: st0: Sequential-Access density code 0x13, drive empty
The disk sd1 is for WIN95 only and is never accessed under FreeBSD.
The JAZ drive sd3 is always without disk, when I boot the system.
Mounting it does not change the behaviour.
I have currently AHC_ALLOW_MEMIO, AHC_TAGENABLE and AHC_SCBPAGING_ENABLE
*disabled*, i.e. they are not in my kernel configuration.
>Description:
The error is not new to me: Under heavy disk load (see below)
the following occurs:
Nov 23 09:44:31 atlantis /kernel: sd0: SCB 0x1 - timed out while idle, LASTPHASE == 0x1, SCSISIGI == 0x0
Nov 23 09:44:44 atlantis /kernel: SEQADDR = 0x8 SCSISEQ = 0x12 SSTAT0 = 0x5 SSTAT1 = 0xa
Nov 23 09:44:44 atlantis /kernel: sd0: Queueing an Abort SCB
Nov 23 09:44:44 atlantis /kernel: sd0: Abort Message Sent
Nov 23 09:44:44 atlantis /kernel: sd0: SCB 1 - Abort Completed.
Nov 23 09:44:44 atlantis /kernel: sd0: no longer in timeout
Nov 23 09:44:44 atlantis /kernel: sd0: SCB 0x1 - timed out while idle, LASTPHASE == 0x1, SCSISIGI == 0x0
Nov 23 09:44:46 atlantis /kernel: SEQADDR = 0x6 SCSISEQ = 0x12 SSTAT0 = 0x5 SSTAT1 = 0xa
Nov 23 09:44:46 atlantis /kernel: sd0: Queueing an Abort SCB
Nov 23 09:44:46 atlantis /kernel: sd0: Abort Message Sent
Nov 23 09:44:46 atlantis /kernel: sd0: SCB 1 - Abort Completed.
Nov 23 09:44:46 atlantis /kernel: sd0: no longer in timeout
The behaviour is a bit improved against earlier occurences of this
bug: The system is often able to recover and does not hang. However
I do not know what happens to the block in the aborted SCB.
Important:
I some cases the system does not recover (sd0 and sd2 are my swap
disks too) and hangs. I this case a *hard reset* does not suffice to
bring the system up again!!! The disk seems to be in an special
faulty state. Using the reset button (i.e. hard reset) boots the
system, the disk reports normal while the ADAPTEC bios boots, but
FreeBSD loads only the boot block, the system refuses to load and
hangs. I have to power cycle the system, then all is well again.
>How-To-Repeat:
The error occurs with some disk load, in my case especially when
building something large, like FreeBSD make world or an X11 build.
>Fix:
Unknown
>Audit-Trail:
>Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199711230922.KAA00668>