[PATCH] - trident cyberblade FIFO status check

Laurence Darby ldarby at sdf-eu.org
Sun May 22 07:33:51 PDT 2005



On Sun, 15 May 2005, Laurence Darby wrote:

> 
> Hi,
> 
> I was getting the audio problems described here:
> www.alsa-project.org/misc/vgakills.txt
> 
> It's caused by writing to a FIFO without checking if it's full or not, and 
> if it is full, the PCI bus locks up for long enough to damage the audio 
> stream to the sound card.  I've fixed it for the trident cyberblade chip.

Hi again,

I take it that my kludge to blade_accel.c wasn't very acceptable, but 
still would really like to see this implemented, because for me it's fixed 
a major problem.  Is there a better way so it is acceptable, like making 
it configurable with an option?  It was my first ever patch, so I'm sure 
there's lots of space for improvement....

Thanks,
Laurence





More information about the xorg mailing list