[Openchrome-users] CN700 problems.

[Christiaan van Dijk]

Hi,

I'm not 100% sure if this is a openchrome issue but I'm running low on 
ideas. I'm running SuSE 10.1 on a Jetway mini-ITX board (CN700 chipset, 
C7 CPU, 1gig mem). I installed the openchrome and DRI drivers without 
any problems, the driver indentifies the system as a VM800 and further 
on in the log as a VT3344. Which probably is the same thing but is 
confusing (CN700=P4M800CE=VT3344=VM800 :-\ ).
The problem; all kernel and xorg parts load without any problem, logs 
show correct loading of the drivers,DRI and XvMC. When using the system 
performance is very low, extremely slow screen response and very slow 
starting of applications while cpu load seems to be low. Playing a DivX 
file with Mplayer seems to go fine but when I start some disc I/O 
(search for files) the entire thing locks up, right now it locks up when 
just playing the file. Only the xorg part seems to lock, the kernel is 
still running in the background, only no keyboard/mouse/screen :-( .
If I'm not loading the DRI in xorg.conf everything seems to work fine, 
at least performance looks normal. However when playing a DivX file 
there seem to be situations in which the entire thing slows down and 
starts dropping a lot of frames. Pausing/restarting can solve this and 
once running there seem to be no further hickups, no other apps running 
and also no log messages. The entire XvMC part does work when DRI is 
enabled, tested with mythTV.

What did I try; Disabled vertical interrupt, disabled DMA (all/for 2D), 
disabled cpu clock switching, searched for remains of older DRI driver 
parts, looked through the log files without success.
At first I could generate weird DMA/IRQ error messages on the HDD in the 
log files when doing background disc I/O and heavy screen access. This 
seems to have disappeared with the latest openchrome version or the 
message is not getting saved any more. I don't think it's a drive issue 
but can't be 100% sure (new drive).

Some log parts without the DRI loaded:
...
agpgart: Detected VIA P4M800CE chipset
agpgart: AGP aperture is 128M @ 0xe8000000
...

...
(II) VIA: driver for VIA chipsets: CLE266, KM400/KN400, K8M800,
        PM800/PM880/CN400, VM800
(II) Primary Device is: PCI 01:00:0
(--) Chipset VM800 found
(!!) VIA Technologies does not support or endorse this driver in any way.
(!!) For support, please refer to http://www.openchrome.org/ or
(!!) your X vendor.
(!!) (development build, at svn revision  255)
...
(==) VIA(0): No default TV output signal type is set.
(--) VIA(0): Chipset: "VM800"
(II) VIA(0): VIAMapMMIO
(--) VIA(0): mapping MMIO @ 0xfb000000 with size 0x9000
(--) VIA(0): mapping BitBlt MMIO @ 0xfb200000 with size 0x10000
(II) VIA(0): vgaHWGetIOBase: hwp->IOBase is 0x03d0, hwp->PIOOffset is 0x0000
(==) VIA(0): Will not print VGA Registers.
(==) VIA(0): Will not scan I2C buses.
(==) VIA(0): Chipset Rev.: 0
(WW) VIA(0): Manufacturer plainly copied main PCI ids to Subsystem/Card ids.
(--) VIA(0): Detected VIA VT3344 (VM800) - EPIA EN.
(II) VIA(0): ...Finished parsing config file options.
(II) VIA(0): Detected MemClk 7
(II) VIA(0): ViaGetMemoryBandwidth
...

Right now I'm stuck, I get the impression there's some kind of DMA 
conflict. What can I do to get any  more information on what's going on? 
I searched for ways to extend the information logged by the drivers but 
it's not clear to me how to do this. Any help would be welcome.

Regds,
Christiaan van Dijk.