[KLUG Members] Re: FC2 Test 1, X, & > 2.6.1

[Adam Tauno Williams]

> > Ok, so now I have something resembling a working distribution on FC2t1 &
> > kernel 2.6.1.  Interesting that if I attempt to use one of their newer
> > kernels X doesn't run, but dies with -
> > [drm] Initialized i830 1.3.2 20021108 on minor 0
> > mtrr: base(0xe8020000) is not aligned on a size(0x180000) boundary
> Ow. Not seen that one before. Still, X could run without MTRRs.

Pheeew, then I'm not alone.  I've seen some X wackiness before, but nothing that
looked like this.

> This is a "standard" bug. The kernel people want X to use the kernel to
> manipulate the keyboard. I see them all the time.

And I whole-heartedly agree.  Direct acces to hardware = BAD!  Compare NT 3.51
which used real privilage levels with the wreck that was NT 4.0.  And they did
it because direct-access was so much faster (and in those days NT 3.51 was
glacial).  I certainly hope we don't make the same compromise.

> > [drm:i830_wait_ring] *ERROR* space: 131056 wanted 131064
> > [drm:i830_wait_ring] *ERROR* lockup
> Not seen this either. (I have a Voodoo 3000.) Looks like some byte-level
> problems in FC2 and 2.6.1.

2.6.1 is working,  it is there update to 2.6.5-* that fails.  The thought that
the Fedora RPMs are whacked is pretty easy to believe at this point.

This is a generic i810/815 video & audio chip,  supported since time immemorium.