[pxc] status

[Alessandro Rubini]

I've fixed the /tmp/bla.pgm problem and the spinlock problem reported
by Chanop.

I made a CVS snapshot available as ftp://ar.linux.it/pub/pxc/pxc-latest .
(a link to /pub/pxc/CVS-snapshots/pxc-snapshot-010914.tar.gz )

I'm currently working on PXC200A support (the new Imagenation device,
based on the Bt878 instead of the 848 chip). I'll look in other issues
as well.

Chanop:

> With pxc200F on 2 different boxes, one is ABit VP6 with via apollo
> chipset and the other is Gigabyte GA6BXD with BX chipset. I've got oops
> on VP6 but a plain lockup on GA6BXD.

However, the oops you reported is a whatchdog oops, like the computer
has internal checks for CPU stalls.
 
> With pxc200L, after applying the following patch
> [...]
> The driver works on both motherboards, however, I have to set the BIOS
> to MPS 1.1 instead of MPS 1.4 to get the interrupt to happen. e.g. pxc
> driver reports interupt 10 while /proc/pci reports interupt 19 etc....
> , hence, could be the source of the problem I experienced.

Hmm.... I'm not up to date about this kind of stuff. Can someone send
a pointer to more information?

> With pxc200L reading /dev/pxc0[H]pgm is fine. /dev/pxc0[H]ppm give a
> swap blue/red bands?? /dev/pxc0{rgb,bgr} are still problems for me.

Oh well... I don't use color :)

You are right, when I folded pnmread() into read() I forgot the conversion
step. Here (cvs log main.c):

   ----------------------------
   revision 1.30
   date: 2001/05/23 17:48:57;  author: rubini;  state: Exp;  lines: +378 -171
   implemented multiple buffers, folded pnmread() into read(), initial poll, spinlocks

(after 0.29-beta)

Just re-did the snapshot with this fix in.

/alessandro