[Mono-devel-list] Mono segfaults with updated kernel?? 2.6.10

jrodman at mono-list.spamportal.net jrodman at mono-list.spamportal.net
Thu Jan 27 19:01:04 EST 2005


On Thu, Jan 20, 2005 at 07:47:05PM +0000, Paul wrote:
> Hi,
>  
> > >   Could you post the output of mono --version ?
> 
> > surfboard:~ # mono --version
> > Mono JIT compiler version 1.0.5, (C) 2002-2004 Novell, Inc and Contributors. www.go-mono.com
> >         TLS:           normal
> >         GC:            Included Boehm (with typed GC)
> >         SIGSEGV      : normal
> >         Globalization: ICU
> 
> It's possible that something in glib has changed with the new kernel,
> but that really is stretching things.
> 
> Personally, I think it's just a memory map problem. Try rolling your own
> from either branch and see if that works.

Okay, at this point I've tried a few things:

The following configurations segfault instantly:

	2.6.10 built with SuSE-like .config file on SuSE Linux Enterprise
	Server 9 runtime with 1.0.5 binaries for SuSE

	2.6.10 built with Knoppix-like .config file on SLES9 runtime
	with 1.0.5 binaries for SuSE

	2.6.10 built with SuSE-like .config file on Knoppix runtime
	(chroot) with Debian 1.0.5 binaries

	2.6.10 built with Knoppix-like .config on Knoppix runtime
	(CD-boot) with Debian 1.0.5 binaries

	2.6.10 built with SuSE-like .config file on SLES9 with
	hand-built vanilla make, make install binaries for mono 1.0.5,
	which results in mint, but not mono, this is an amd64 build.
	Mint segfaults.

	2.6.10 built with  Knoppix-like .config on Knoppix runtime
	(CD-boot) with hand-built vanilla mak-e, make install binaries
	for mono 1.0.5.  This is an x86 build, as all the others are.

ALL of the above scenarios result in segmentation faults.  I would
appreciate some assistance in narrowing down the source of this problem
from the mono side of things.

What does work:

	Hand-built Mono 1.1.3 for SLES9 on SLES9 for amd64.

I'm using gcc (GCC) 3.3.3 (SuSE Linux).  The kernel is 100% vanilla.

Please advise.

-josh



More information about the Mono-devel-list mailing list