Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 18 Jul 2003 11:45:56 +0200
From:      Bernd Walter <ticso@cicely12.cicely.de>
To:        ticso@cicely.de, freebsd-alpha@freebsd.org
Subject:   Re: Problem building Mozilla
Message-ID:  <20030718094556.GE19386@cicely12.cicely.de>
In-Reply-To: <20030718031240.GF430@gsmx07.alcatel.com.au>
References:  <20030716215658.GA43409@gsmx07.alcatel.com.au> <20030717132611.GJ14530@cicely12.cicely.de> <20030718031240.GF430@gsmx07.alcatel.com.au>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Jul 18, 2003 at 01:12:40PM +1000, Peter Jeremy wrote:
> On 2003-Jul-17 15:26:12 +0200, Bernd Walter <ticso@cicely12.cicely.de> wrote:
> >On Thu, Jul 17, 2003 at 07:56:58AM +1000, Peter Jeremy wrote:
> >> My attempts to build Mozilla on a quad-CPU AS4100 running a
> >> locally-built copy of 5.1-RELEASE consistently fail, ending:
> >
> >Oh - the port was updated to 1.4
> >Downgrade the port to 5.1-RELEASE time.
> 
> Thanks for that.  mozilla-1.3.1,2 compiles and installs successfully.
> 
> Unfortunately, it doesn't run.  I get an unaligned access report,
> followed by a segmentation violation:
> pid 49966 (mozilla-bin): unaligned access: va=0x160bbb2df pc=0x160b9db2c ra=0x16200e8ac op=ldq

I see the alignment errors too.
Typically I get a  bunch at startup and that's it.
But then it works.
If you already have a ~/.mozilla - can you rename it and try again?

> (gdb) where
> #0  0x160b9eaa000 in ?? ()
> #1  0x160b9db30 in nsPref::GetDefaultIntPref(char const*, int*) ()
>    from /usr/X11R6/lib/mozilla/components/libpref.so
> #2  0x160b9db30 in nsPref::GetDefaultIntPref(char const*, int*) ()
>    from /usr/X11R6/lib/mozilla/components/libpref.so
> Error accessing memory address 0x10: Bad address.
> (gdb) 
> 
> Suspiciously, the address looks like it is shifted 8-bits left, which
> is consistent with the unaligned access being one byte below an
> aligned address.  This looks like an out-by-one error somewhere.  I'll
> try re-compiling with the defaults (rather than for ev56) and see if
> this is just a gcc bug (as I suspect).

I compiled myself with:
CFLAGS=-O2 -pipe
CPUTYPE?=ev56
The system is an ev56 AS4100 too.

-- 
B.Walter                   BWCT                http://www.bwct.de
ticso@bwct.de                                  info@bwct.de



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030718094556.GE19386>