Pauli Borodulin wrote: > [...] > The problem is: In the beginning the driver is able to communicate with > cu1216 using I2C. For example, the driver is able to check that the > demodulator is there using cu1216_readreg(). As soon as DMA is started (you > get "mantis start feed & dma" in your kernel log), all i2c reading and > writing fails. Every read returns 0. I have not yet found out why. You can > verify this by adding > [...] I made a quick and dirty hack and commented out the contents of the functions mantis_dma.c/mantis_dma_start() and mantis_dma.c/mantis_dma_stop() and now I2C seems to works ok. This will totally break the driver, so it's not a fix. I did it because I was curios whether it would make any difference, and it did. I then tried uncommenting lines in mantis_dma_start() one by one starting from the beginning, and found out that the line mmwrite(MANTIS_GPIF_RDWRN, MANTIS_GPIF_ADDR); was the one that triggered the bug. I have no knowledge on the hardware in question (I don't have any datasheets etc.), so I have no idea what I should do next. There's also some bugs in the cu1216 driver, but it's no use to debug them before the DMA(?) problem is fixed. Br, -- Pauli Borodulin <pauli@xxxxxxxxxxxx> gsm. +358 (0) 44 5555 381 http://pauli.borodulin.fi
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ linux-dvb mailing list linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb