From mboxrd@z Thu Jan 1 00:00:00 1970 Received: with ECARTIS (v1.0.0; list linux-mips); Thu, 05 Aug 2004 21:16:54 +0100 (BST) Received: from web50806.mail.yahoo.com ([IPv6:::ffff:206.190.38.115]:17751 "HELO web50806.mail.yahoo.com") by linux-mips.org with SMTP id ; Thu, 5 Aug 2004 21:16:50 +0100 Message-ID: <20040805201643.6422.qmail@web50806.mail.yahoo.com> Received: from [65.204.143.11] by web50806.mail.yahoo.com via HTTP; Thu, 05 Aug 2004 13:16:43 PDT Date: Thu, 5 Aug 2004 13:16:43 -0700 (PDT) From: G H Subject: Re: do_ri failure in cache flushing routines To: Pete Popov Cc: linux-mips@linux-mips.org In-Reply-To: <411277BD.7070108@mvista.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-470146373-1091737003=:5841" Return-Path: X-Envelope-To: <"|/home/ecartis/ecartis -s linux-mips"> (uid 0) X-Orcpt: rfc822;linux-mips@linux-mips.org Original-Recipient: rfc822;linux-mips@linux-mips.org X-archive-position: 5606 X-ecartis-version: Ecartis v1.0.0 Sender: linux-mips-bounce@linux-mips.org Errors-to: linux-mips-bounce@linux-mips.org X-original-sender: giles67@yahoo.com Precedence: bulk X-list: linux-mips --0-470146373-1091737003=:5841 Content-Type: text/plain; charset=us-ascii At the moment I don't have the board set up for using kgdb and it's complicated by the fact that we only have one serial console port. But I am looking into setting it up for kgdb now. As far as stressing the system, it doesn't have enough resources ( disk space ) to be able to compile the kernel, but we did write a simple program that would stress the system by spawning multiple threads, each one performing floating point calculations. With this test, top reported a load average of over 400 and we have seen no failure so far. Pete Popov wrote: G H wrote: > I've not had much response to this question so I would like to > rephrase it : > > Can anyone think of any possible scenario where do_ri could occur in > blast_icache32() ?? > > Is this possibly a cache synchronisation problem ?? > Could be a hardware memory glitch. I would use kgdb to put a breakpoint there and see what the data in memory looks like when this happens -- look for memory corruption, etc. Pete > TIA > > >While testing out an amd au1500 based board I have been getting " > do_ri " exceptions >that always occur in the cache flushing routines. > More often than not in >blast_icache_32(). > > >So far this has mainly happened after running the board for days on > end while running >multiple telnet sessions to it. It has sometimes ( > quite rarely ) happened after a few >hours to a day of multiple telnet > session use. > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > --------------------------------- Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish. --0-470146373-1091737003=:5841 Content-Type: text/html; charset=us-ascii
At the moment I don't have the board set up for using kgdb and it's complicated by the fact that we only have one serial console port. But I am looking into setting it up for kgdb now.
 
As far as stressing the system, it doesn't have enough resources ( disk space ) to be able to compile the kernel, but we did write a simple program that would stress the system by spawning multiple threads, each one performing floating point calculations. With this test, top reported a load average of over 400 and we have seen no failure so far.

Pete Popov <ppopov@mvista.com> wrote:
G H wrote:

> I've not had much response to this question so I would like to
> rephrase it :
>
> Can anyone think of any possible scenario where do_ri could occur in
> blast_icache32() ??
>
> Is this possibly a cache synchronisation problem ??
>

Could be a hardware memory glitch. I would use kgdb to put a breakpoint
there and see what the data in memory looks like when this happens --
look for memory corruption, etc.

Pete

> TIA
>
> >While testing out an amd au1500 based board I have been getting "
> do_ri " exceptions >that always occur in the cache flushing routines.
> More often than not in >blast_icache_32().
>
> >So far this has mainly happened after running the board for days on
> end while running >multiple telnet sessions to it. It has sometimes (
> quite rarely ) happened after a few >hours to a day of multiple telnet
> session use.
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>

Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish. --0-470146373-1091737003=:5841--