Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 06 Oct 2003 22:13:15 -0400
From:      Joe Marcus Clarke <marcus@marcuscom.com>
To:        Jay Cornwall <jay@evilrealms.net>
Cc:        FreeBSD GNOME Users <gnome@freebsd.org>
Subject:   Re: FreeBSD Current lang/mono runtime problems
Message-ID:  <1065492795.33346.0.camel@shumai.marcuscom.com>
In-Reply-To: <3F820179.80100@evilrealms.net>
References:  <3F81E292.8070308@evilrealms.net> <1065477237.748.95.camel@gyros>  <3F820179.80100@evilrealms.net>

next in thread | previous in thread | raw e-mail | index | archive | help

--=-Y1HHymU/3YY455ioEz2W
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Mon, 2003-10-06 at 19:57, Jay Cornwall wrote:
> Joe Marcus Clarke wrote:
>=20
> >>[jay@python mono]$ mcs --version
> >>Mono C# compiler version 0.27.0.0
> >>
> >>And never finishes. I tried to get an strace, but that seems to have=20
> >>issues with a missing /proc filesystem. Is there any more useful=20
> >>information I can obtain under FreeBSD to help debug this?
>=20
> > Mono doesn't work on FreeBSD.  The problem seems to be in the garbage
> > collector.  If you build mono without GC, it should work okay (but neve=
r
> > release memory).
>=20
> Ah, hmm.
>=20
> > Honestly, gnome@ doesn't really want mono since no one here is really a
> > C# user.  We are keeping it since one of our developers, Juli Mallett,
> > has been busy, but wants to get back into Mono.  In the meantime, if yo=
u
> > want to look at improving boehm-gc on FreeBSD, I'm sure your
> > contributions will not be turned down.
>=20
> Well I've had a brief look tonight, but free time is a little on the=20
> short side at the moment.
>=20
> It seems a segfault occurs inside mono-0.28/mono/mini.c:7291,=20
> sigusr1_signal_handler(). In order to identify the thread (through=20
> mono_current_thread()), mono/metadata/threads.c stores the MonoThread *=20
> in a thread-local structure using pthread_setspecific in previously=20
> called methods of mono/metadata/threads.c.

I've fixed this a while ago, but I don't have the patch in the ports
tree yet.  Now I'm not getting the hang, but I have another problem.=20
I'm trying to fix that now.

Joe

>=20
> Dropping in a few debug statements shows that a thread is entering=20
> sigusr1_signal_handler() that hasn't previously had its MonoThread * set=20
> (it may not even be a MonoThread, for all I know). That's the direct=20
> reason for the crash - to find the indirect reason, i.e. why this thread=20
> is entering without its MonoThread * set, will take me a bit longer.
>=20
> Enough technobabble, I'll try and find the bug and fix it!
--=20
PGP Key : http://www.marcuscom.com/pgp.asc

--=-Y1HHymU/3YY455ioEz2W
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQA/giE7b2iPiv4Uz4cRAnAeAKCXzNCVqNA4GFaCsoBkYuEXAgdFnwCfZEcD
4OljFTFC8dH9p9DZVdw5JvU=
=7jJ6
-----END PGP SIGNATURE-----

--=-Y1HHymU/3YY455ioEz2W--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1065492795.33346.0.camel>