rtl28xx Leadtek

Hi,

I have been trying to get support going for a Leadtek WinFast DTV2000DS Plus (usbid 0413:6f12)

Christian Dale (cc:d) posted a patch a while ago and since that time there have been a few patches from Olle and others that have significantly improved things.

However I am still having issues with the card; sometimes it does not scan any of the multiplexes I look at with it, sometimes it does some of them and sometimes it pulls out some program ids on one mulitplex but not the others in the scan. The behaviour is different between the two tuners on the card as well.

I've written a test script that turns on various kernel dynamic debug entries and tries to scan, using the old dvb-apps 'scan' (1.1.1+rev1500) and the latest dvbv5-scan from git.

Would anyone on the list interested in helping please take a look at the attached logs of the run I did? It includes the script output and what was happening in syslog at each step.

There should be no signal issues, as the same test script on the same system with the same coax input works perfectly when I swap the Leadtek for a DVICO card using the cx23885 driver. Both drivers are built from the media_build git.

My questions:
- are there debug items that could be turned on or off to help diagnose the issues?
- are there some perf traces that could be run that would be helpful?

Kind regards
Vince

In case it matters here, I have these cards and am using the driver built from
git clone git://linuxtv.org/media_build.git
git clone [email protected]:jaredquinn/DVB-Realtek-RTL2832U.git

I get rather reliable tuning with hardly any of the old problems of zero length recordings of fails to tune some channels.

I run old fedora 19 though, and things may have deteriorated since? Last time I needed to build was Jan 31 for kernel 3.14.27-100.fc19.x86_64. Is this driver included with the kernel these days?

cheers
Eyal

Eyal Lebedinsky ([email protected])

Thanks for your reply Eyal.
There is certainly an in-kernel driver, which I am trying to use. I don't know what relation it has to the github code, which looks to have useful information on registers and other magic numbers that might be needed for proper operation.

Antti, would you care to comment on whether the github code linked above is useful at all for your work on the realtek drivers?

Cheers
Vince