Installing 6.0-BETA5 on Ultra30: RED State Exception

Marius Strobl marius at
Mon Oct 10 06:47:53 PDT 2005

On Sun, Oct 09, 2005 at 02:33:29PM +0200, Gheorghe Ardelean wrote:
> Hi,
> I have given to FreeBSD 6.0-BETA5 a try on my 250MHz Ultra30.
> Unfortunately there are problems with it just from the beginning.
> The first problem I have encountered was the "Release name"
> which is 6.0-BETA4. Because of this none of the distributions could be
> found. After restarting and changing this in 6.0-BETA5 (in the Options
> menu) the distributions could be found but after chunk 34 from bin I get a
> RED State Exception.
> .......
>   47834112 bytes read from base dist, chunk 34 of 41 @ 655.3 KBytes/sec.
> RED State Exception
> TL=0000.0000.0000.0005 TT=0000.0000.0000.0010
>    TPC=0000.0000.c004.4200 TnPC=0000.0000.c004.4204
> TSTATE=0000.0044.5800.1505
> TL=0000.0000.0000.0004 TT=0000.0000.0000.0010
>    TPC=0000.0000.c004.4200 TnPC=0000.0000.c004.4204
> TSTATE=0000.0044.5800.1505
> TL=0000.0000.0000.0003 TT=0000.0000.0000.0010

trap level 3-5, trap type 0x10: illegal_instruction

>    TPC=0000.0000.c004.4c60 TnPC=0000.0000.c004.4c64
> TSTATE=0000.0044.5800.1505
> TL=0000.0000.0000.0002 TT=0000.0000.0000.0063

trap level 2, trap type 0x63: corrected_ECC_error

>    TPC=0000.0000.c004.8f6c TnPC=0000.0000.c004.8f70
> TSTATE=0000.0044.5800.1605
> TL=0000.0000.0000.0001 TT=0000.0000.0000.0141

trap level 1, trap type 0x141: system call

>    TPC=0000.0000.0020.eba4 TnPC=0000.0000.0020.eba8
> TSTATE=0000.0044.0000.1204
> .......
> A FreeBSD snapshot from 27.07.04 (some 5.2?) works perfectly.
> What could case this? Any idea?

The problem is that a corrected_ECC_error trap happens while
handling another one. In general FreeBSD doesn't handle
corrected_ECC_error traps very well. Nevertheless it means
a hardware problem. Often these happen due to problems with
the electrical contacts and re-seating the CPU and the memory
modules solves them. If not you probably have a fault memory
module if these traps happen regularly. The OBP diagnostics
(vary from model to model however) and SunVTS should be able
to identify a broken memory module.
It's most likely just coincidence that you don't see them
with FreeBSD 5.2, there where no changes since July '04
which affect handling of corrected_ECC_error traps.
Btw., if you actually have a broken memory module would
you ship it to me for implementing proper handling of
corrected_ECC_error traps?


This mail was scanned by AntiVir Milter.
This product is licensed for non-commercial use.
See for details.

More information about the freebsd-sparc64 mailing list