[ltp] Its not heat, its not memory... what is it?

[John Hunt]

--=-CqstFtNwUjjJ9bqeLygD
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

CVS version of your video driver? Hmm...unless it locks up with a stable
one, I don't think you can rule everything out, unless you know exactly
what that driver is doing...which you might do! I don't know for sure.

I wouldn't be so sure that theres no way a kernel-land program can't
cause a lock up. Yes unlikely, impossible, no.

Having said this, I'm no guru.

John.

On Wed, 2005-08-24 at 09:40 -0400, David A. Desrosiers wrote:
> > Anyway, I think it's more than likely your graphics card drivers.=20
> > Could you be a bit more specific about what causes it to lock up, or=20
> > is it entirely random?
>=20
>  	Nothing in kernel-land should cause a hardware lockup. A=20
> kernel crash, yes. A lockup, absolutely no way. That being said...
>=20
> > Have you checked all your logs for various things and stuff? dmesg=20
> > etc..?
>=20
>  	Yes, nothing at all out of the ordinary, and I log pretty=20
> pendantically as part of my development work anyway on this machine.
>=20
> > If it's an ati graphics card on that model, make sure you're using=20
> > the open source drivers and not the proprietary ones, ati make=20
> > rubbish drivers for Linux.
>=20
>  	I'm using the 7/20/2005 CVS tree of Xorg/Mesa/r300/drm. It is=20
> the only version of this driver that works properly with DRM. Anything=20
> else and I get pathetic fps.
>=20
> > What distro are you using? What kernel etc? I don't think you can=20
> > call it a hardware issue until you're using something really rock=20
> > solid like Debian woody with a 2.4 kernel.
>=20
>  	2.6.12.4, 2.6.12.5 on Debian Unstable/Experimental most of the=20
> time. I could certainly try a 2.4 kernel, but I seriously suspect that=20
> as being the problem.
>=20
>  	I don't use distro kernels, so any wacky Debian-broken patches=20
> are out of the equation here. I've tried all three elevators, changing=20
> mtrr and some other related pokes, all to no avail.
>=20
>  	I'll try chasing down the video driver issue a bit more and=20
> see if I can really exclude that from the issue.
>=20
>  	If it was a video driver though, I'd expect it to lock up=20
> under a _much_ more intense GL application than these two examples.
>=20
>=20
> David A. Desrosiers
> desrod@gnu-designs.com
> http://gnu-designs.com
--=20
John Hunt <jrh@geodata.soton.ac.uk>
GeoData Institute

--=-CqstFtNwUjjJ9bqeLygD
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDDIxufmjF/242LDgRAkPwAJwNySoOL+JrLvfKGQrDCcwri/extACgmnBN
H3ha0EX5IkWdXdg+S9yBqs0=
=gq8I
-----END PGP SIGNATURE-----

--=-CqstFtNwUjjJ9bqeLygD--