Ok, I finally found the reason for the video corruptions with the new
newport console some reported to me. It happens only, when there are
less than 1024 scanlines on the screen. My fix is to enable the
faster scrolling only with 1024 scanline screen modes. This slows
down scrolling, and I hope there is a better solution (but I doubt it).
It looks like video rams get only refreshed, when they are displayed.
Is this true ? I've tried enabling the vram refresh in the config
register, but that didn't change anything. Is there are a way to avoid
losing the content of offscreen scanlines ?
As the newport problem is mostly solved, I'll try to get the scsi fixed.
I'm now able to reproduce a complete lockup with my DAT drive. While
looking for the reason, I've got GIO fifo full interrupts (INT2 local
interrupt 0). Can someone explain, when these interrupts occur ?
This device has completely bogus header. Compaq scores again :-|
It's a host bridge, but it should be called ghost bridge instead ;^)
[Martin `MJ' Mares on linux-kernel]