Hi, On Wed, 22 Jun 2005, Johannes Stezenbach wrote: >> after some time of investigation the no-data problem of the >> flexcop-pci-driver (f.k.a skystar2-driver) is fixed (partially at least). >> >> Short problem description: after some time stressing the card with >> (re)tuning (e.g. as a slave card in VDR doing EPG/channels-scan) the card >> stops to producing interrupts and thus delivers not data anymore. > > Did you find a problem and fixed it, or is this a workaround? I assume it is just a workaround, but it can be _the_ fix ;). See below. > Could you give a short description of what the patch does > and what "block 300" is? Register block 0x300 affects single-PID/group-PID and MAC-filtering. The problem was, that after some time of stressing the card with (re)tuning to a (new) frequency simply no more DMA activities took place. Reloading the driver solved the problem, because the Flexcop was reinitialized. I now tried to find out, which part of the initialization fixed it. After some testing (+ finding and curing other problems) it turned out, that reinitializing register block 0x3xx made it work again. So I simply added the reset to the start/stop-data-reception-function. My affected system is now doing scans for about 6 hours. Without the reset it took 1 hour at most to have to no-data situation. A workaround, because I cannot test the signal-loss-issue (losing/regaining signal without restarting data reception), but it can be that this issue is solved as well (with the help of the smaller changes I made). We have to wait for user comments. Patrick. -- Mail: patrick.boettcher@xxxxxxx WWW: http://www.wi-bw.tfh-wildau.de/~pboettch/
