Re: ARM SEGV problem

[Petri Laakso <petri.laakso%asd.fi@localhost>]

On Thu, 12 Sep 2013 22:17:54 +0900 (JST)
Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost> wrote:

> From: Petri Laakso <petri.laakso%asd.fi@localhost>, Date: Wed, 11 Sep 2013 
> 22:01:09 +0300
> 
> > On Thu, 12 Sep 2013 00:02:54 +0900 (JST)
> > KIYOHARA Takashi <kiyohara%kk.iij4u.or.jp@localhost> wrote:
> > 
> >> Hi!
> >> 
> >> From: KIYOHARA Takashi <kiyohara%kk.iij4u.or.jp@localhost>
> >> Date: Wed, 11 Sep 2013 20:46:55 +0900 (JST)
> >> 
> >> > I resolved this problem in my local source maybe.
> >> > It tested on my OPENBLOCKS_A6.
> >> 
> >> I will test on OpenBlocks AX3(ARMADAXP) at today.
> >> After commit that patch.
> >> 
> >> Thanks,
> >> --
> >> kiyohara
> > 
> > Thank you! This indeed solved problems with SEGV programs at init time.
> > 
> > Now I get stuck when I try to enter multiuser.
> > 
> > WARNING: no TOD clock present
> > WARNING: using filesystem time
> > WARNING: CHECK AND RESET THE DATE!
> > init path (default /sbin/init):
> > init: trying /sbin/init
> > Wed Sep 11 18:24:32 UTC 2013
> > 
> > ... nothing happens after this.
> > I am able to enter debugger so system is not totally dead.
> > 
> > Difficult to me is that I don't know is it me or something else as I'm
> > developing system initialization code at the moment so my system can crash
> > anytime....
> > 
> > This is what I did today (I manually applied your patch after cvs update)
> > 
> > http://www.asd.fi/cgi-bin/blog.cgi/NetBSD/NetBSD_OLinuXino.1024px
> 
> Hi,
> 
> Me too on Qnap TS-110 (kirkwood machine).
> Single user mode works for me.
> 
> You can see basic dmesg in http://gnats.netbsd.org/48130 .

You have:
cpu0 at mainbus0 core 0: Sheeva 88SV131 rev 1 (ARM9E-S V5TE core)

I have:
cpu0 at mainbus0 core 0: ARM926EJ-S rev 5 (ARM9EJ-S V5TEJ core)

Could this "no working multiuser" be ARM9 related?
I have access to jtag and ddb. I'll try to get some debug data.
At least where it gets stuck.

I've been running "while true; do openssl speed; done" for at least 24h in 
single
user mode, so I guess system should be able to run init little bit further than 
timestamp.

BR
Petri Laakso