On Thu, Aug 14, 2008 at 3:57 PM, Nicolas Will <nico@xxxxxxxxxxxx> wrote: > I cannot code or really understand the details, but could this explain > the more or less regular i2c read failures or even write failures > eventually leading to device lock-ups that we are still experiencing if > we are a bit too agressive? In regards to i2c and the dib0700, I am actually debugging two independent issues. One is the behavior I described above, where I can clearly see i2c read requests being turned into i2c writes. The second is what I would describe as "intermittent i2c read and write failures", where the dib0700 is returning "-32" to the request. For example (taken from a USB bus capture attempt to set the xc5000 signal source via the dib0700): d3cd8b00 1554060054 S Co:8:109:0 s 40 03 0000 0000 0006 6 = 03c8000d 0000 d3cd8b00 1554061321 C Co:8:109:0 -32 6 > I figured I would start with the issue that seems pretty well understood and reproducible before I start talking about the flaky/intermittent i2c read/write failures. Unfortunately I don't have an i2c bus analyzer which would be useful to better understand whether this is a failure in the dib0700, or whether the request is being passed through to the xc5000 and the xc5000 is reporting the failure. Without the dib0700 datasheet, I don't know what the status code means... Devin -- Devin J. Heitmueller http://www.devinheitmueller.com AIM: devinheitmueller _______________________________________________ linux-dvb mailing list linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
