On Mon, Jun 06, 2005 at 05:51:21PM +0200, Wolfgang Rohdewald wrote: > That works nicely so far. Should I try to find out what signal was interrupting? > How could I do that? I know almost nothing about kernel programming. Hm, I know vdr is using signals to "wake up" threads which are blocked waiting for I/O. The question is if there's more code in the driver which causes problems when interrupted. (This means someone needs to grep for "_interruptible" and check if the interrupt is handled correctly.) BlitBitmap() looks suspicous, too, in other cases it is OK to just return -ERESTARTSYS so the call will be retried. I guess this would also make sense for LoadBitmap(), but not for BlitBitmap() because the OSD_SetBlock command needs to call them both in proper sequence. Johannes
