Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Nov 2002 18:00:29 -0500 (EST)
From:      Horen <horen@ny-myth.com>
To:        Terry Lambert <tlambert2@mindspring.com>
Cc:        Eric Anholt <eta@lclark.edu>, <freebsd-current@FreeBSD.ORG>
Subject:   Re: XFree
Message-ID:  <Pine.LNX.4.44.0211081745460.1000-100000@salome.ny-myth.com>
In-Reply-To: <3DCC3300.84711FE8@mindspring.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 8 Nov 2002, Terry Lambert wrote:

> Horen wrote:
> > Now the missing facts:
> > 
> >    4.x binaries will not run on 5.0, lib.c inconsistency.
> 
> You are not running the most recent 5.x.  We can't help you, I
> think.
> 
> 
> >    Identical XFree86 sources compiled nativly on 4.6, 4.7 and 5.0.
> >    4.6-4.7: no problem
> >    5.0: logout from X blocks the graphic card, Matrox MGA 400
> >         and also Nvidia GeForce2 MX200
> 
> Different compilers.  You have failed to narrow the problem down;
> here's where it might be:
> 
> 1)	In the configuration code
> 2)	In the difference in the patches between the versions of
> 	the ports tree for X11
> 3)	In the X11 libraries' handling of threads
> 4)	In the 5.x implementation of pthreads
> 5)	In the GCC compiler optimizer
> 6)	In the GCC compiler code generator
> 7)	In the X11 source code; there are certain assumptions
> 	that the new compiler makes which it was always permissable
> 	for it to make, but which the old compiler did not make,
> 	even though it was allowed to (i.e. a revealed bug in the
> 	X11 sources)
> 8)	A bug in the AGP (/dev/agpart) code
> 9)	A bug in the ACPI handling of monitor power-down code
> 10)	A bug in the ACPI BIOS information of your computer itself
> 11)	In a mismatch between kernel and kernel module versions
> 12)	In a difference in the values of manifest constants
> 
> > I doubt, that something else than OS  related issues are the
> > cause.
> 
> Doubt it all you want; from the small amount of information you
> have provided us, it looks to me like there is a possibility
> that it's 50% non-OS-releated.
> 
> 
> > Switched back to 4.7 , will try again 5.0 in few weeks.
> 
> If it happens again, be sure and give us useful information,
> like:
> 
> o	Since Warner fixed the libc incompatabilities, have
> 	you installed the 4.7 version of X11 on 5.x, and does
> 	the problem still happen?
       
Yes, it happened, I didn't make notes exact which version
of X11 it was. 4.7 was updated on 11/1/2002
        
> o	Does it happen if ACPI isn't loaded?
Didn't test it.

> o	Does it happen with the Xvga server, or only with the
> 	accelerated server?
It happens with both server.

> o	Are you on a laptop or a desktop?

It's a desktop with following hardware configuration:

        Tyan Thunder K7 (S2462) Motherboard
        ( with onboard ATI graphic chip ) not used
        Intel MultiProcessor Specification v1.4 activated
        2 G Ram
        2 x Athlon 2000+ 
        Gr. Cards tested Matrox MGA400 Max and Nvidia GeForce2 MX200
        2 IBM  SCSI HD ( each 36 G )
        Monitor Sony GDM F520


> o	If a desktop, does turning the monitor off and back on
> 	"fix" the problem for you?

No, only a reboot.


> o	What the checkout date for your 5.x source tree?  It's
> 	unlikely that someone would be willing to spend 30 hours

Last checkout 11/7/2002.

-Horen

> 	duplicating your test environment, but you never know...
> o	The results of any discriminating tests for the possible
> 	causes, per the list, earlier in this email.
> 
> Thanks,
> -- Terry
> 



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.44.0211081745460.1000-100000>