Subject: port-sparc/13031: 5/24/01 -current: SCSI errors render system unusable
To: None <>
From: None <>
List: netbsd-bugs
Date: 05/24/2001 10:26:44
>Number:         13031
>Category:       port-sparc
>Synopsis:       attemping to mount CD renders system unusable: SCSI reset loop
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-sparc-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu May 24 08:26:00 PDT 2001
>Originator:     Rob Misiak
>Release:        5/24/01
5/24 -current on Sparc 5, Toshiba XM-4101B CDROM (this drive is hacked
to support 512 byte blocks so that it can be bootable, as described
in the Sun CDROM FAQ at

dmesg output:
  NetBSD 1.5V (MERCURY) #17: Thu May 24 08:49:42 CDT 2001
  mainbus0 (root): SUNW,SPARCstation-5
  cpu0 at mainbus0: MB86904 @ 85 MHz, on-chip FPU
  cpu0: 16K instruction (32 b/l), 8K data (16 b/l): cache enabled
  dma0 at sbus0 slot 5 offset 0x8400000: dma rev 2
  esp0 at dma0 slot 5 offset 0x8800000 level 4: ESP200, 40MHz, SCSI ID 7
  scsibus0 at esp0: 8 targets, 8 luns per target
  cd0 at scsibus0 target 4 lun 0: \<IBM, CDRM00201     !F, 0724\> SCSI2 5/cdrom removable
  cd0: sync (248.0ns offset 15), 8-bit (4.032MB/s) transfers

(I've left a copy of the entire dmesg output at

System: NetBSD mercury 1.5V NetBSD 1.5V (MERCURY) #17: Thu May 24 08:49:42 CDT 2001 rob@mercury:/usr/obj/sys/arch/sparc/MERCURY sparc
Architecture: sparc
Machine: sparc

I want to state first that this is my first time building -current
on this box.  The CDROM worked with little problems (occasional timeout
warnings, but never had a problem mounting/reading from it) using 1.5
and 1.5.1_beta*.

After issuing the mount command, the system hangs for about 15 seconds.
Then a warning is printed to the console stating esp0 timed out, and
sync negotiations are being disabled.  After that, the following two
lines are printed repeatedly in an endless loop:

   esp0: invalid state: 6esp0: resetting scsi bus
   esp0: SCSI bus reset

Doing a L1-A and issuing the PROM 'sync' command only returns to the
endless loop of printing these errors, and no core dump is done.

   mount -t cd9660 /dev/cd0a /mnt