This is merely a historical archive of years 2008-2021, before the migration to mailman3.
A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/osmocom-sdr@lists.osmocom.org/.
Steve Markgraf steve at steve-m.deHi, sorry for the delay, I was on vacation. On 28.08.2013 13:55, keenerd wrote: > Overhauled scanning in rtl_fm. This fixed a number of issues, where > either scanning did not work at all and some brokenness on OpenBSD. > It also meant giving up on the async API entirely. Merged the patch, thanks. > Released a whole new application, rtl_power. This is something I'd > had bouncing around on my hard drive for many months, half-unwritten > due to over-complexity. It'll generate CSV files of arbitrarily > slow/wide/detailed FFTs. Also a little visualization script[2] that > won't choke when you give it 50,000 columns of data. I'll take a look at it, although I'm not sure if we should pick up yet another application in the rtl-sdr repository, or if it's better if you just maintain it outside of the tree (like dump1090 for example). Any opinions? > Second was the discovery[5] that scanning had gotten a whole lot > slower since my original experiments with rtl_fm. 85mS to retune is > unacceptably slow. Since the r820t driver is the slowest, I started > looking into figuring out what is going on. Still figuring that out > but I have done basic cleanups to the code (removing redundancy, > particularly around register writes) and the driver is 15% shorter > with no modifications to code behavior. Yeah, the r820t driver is really a mess. I've seen that the v4l-guys have rewritten the driver [1], my goal for the next months is to use that for rtl-sdr, merge back our gain code and make it thread-safe. > I would not mind talking to Steve or one of the other devs on IRC > about some of the more ambiguous aspects of the driver. (Like, why > some reg writes are backed up to R828_Arry[] and others aren't. And > is that array thread safe at all? If not I'll have to do some very > large changes to support multiple dongles in rtl_power.) According some recent tests by patchvonbraun the array isn't thread-safe. Regards, Steve [1] http://git.linuxtv.org/linux-2.6.git/blob/HEAD:/drivers/media/tuners/r820t.c