[18:04:42] <mambrus> Could be HW. This is the oldest one of mine.
[18:04:53] <jynik> mambrus: Could you re-confirm with FW v1.8.0
[18:05:12] <Arlee> No luck. Still getting the messages from osmocom_fft.
[18:05:15] <mambrus> sure
[18:05:47] <jynik> mambrus: If it happens again, grab a verbose log... perhaps we'll see things snowballing into catastrophe
[18:05:55] <Arlee> Last line is: ImportError: libbladeRF.so.0: cannot open shared object file: No such file or directory
[18:06:19] <mambrus> jynik: okidok
[18:06:25] <jont> What does ls -l /usr/local/lib/libblade* show you now exactly?
[18:09:02] <Arlee> Two lines: lrwxrwxrwx 1 root root 15 Jan  2  2014 /usr/local/lib/libbladeRF.so -> libbladeRF.so.0
[18:09:17] <Arlee> And this: lrwxrwxrwx 1 root root 30 Feb 11 11:56 /usr/local/lib/libbladeRF.so.0 -> /usr/local/lib/libbladeRF.so.1
[18:09:43] <mambrus> jynik: The last tag in gitlog is FW 1.7.1. Are you sure about changing to fw v1.8.0?
[18:10:34] <jont> Arlee: Oh damn, sorry.. i saw you talking about upgrading libbladeRF the other day so thought there'd be 2 files there
[18:11:18] <jont> Arlee: So now you should rm -f /usr/local/lib/libblade*  .. and then download the latest libbladeRF, install it, and you'll be fine
[18:13:43] <Arlee> Okay. I removed the libraries.
[18:15:51] <Arlee> Would the simplest way to do that be to get the latest PPA?
[18:17:02] <jont> I don't know.. the way that i know will work is manually downloading the latest libbladeRF source
[18:20:03] <jynik> mambrus: Yikes, good catch. 0b466c6aed should have been tagged. Will do that now
[18:20:30] <jynik> mambrus: Just pull v1.8.0 from nuand.com/fx3 for now
[18:20:34] * mambrus Has automation scripts that depends on it ;-)
[18:20:52] <jynik> Yeah, that was a royal fsckup on my end, sorry
[18:40:49] <mambrus> jynik: Issue is persistent.
[18:42:14] <mambrus> jynik: Rebuild fresh from HEAD and run with newest fw/fpga. Not much in logs either. I'll scrub off what's not needed an pastebin for you.
[18:43:28] <jynik> No need to scrub
[18:44:59] <jynik> mambrus: The next few days are going to be a bit hectic for me, so apologies if I don't have a good sense of what to debug right away
[18:45:31] <jynik> mambrus: If it was working fine previously, the "What changed?" game is certainly a good place to start...
[18:46:04] <mambrus> jynik: np. This is not critical (yet). Seems to be affecting one individual blade. The others work fine.
[18:47:00] <jynik> Worth isolating... those warnings seemed to indicate there was a failure in the firmware...
[18:52:27] <mambrus> jynik: http://pastebin.com/s3NBiPts
[18:53:37] <mambrus> kernel: [1222956.174308] traps: bladeRF-cli[20107] trap divide error ip:7f0e29457177 sp:7fff3499fcf0 error:0 in libbladeRF.so.1[7f0e2944b000+2c000] <-- Seems relevant
[18:53:59] <mambrus> Basically the only thing I can find that seems missplaced.
[18:54:00] <jynik> There was a different error/warning in your previous log...
[18:54:35] <jynik> Could you run the CLI in GDB and grab a backtrace where the divide-by-zero is occuring
[18:54:45] <jynik> perhaps print out the locals in that stack frame
[18:54:53] <jynik> We can tuck that into a bug report for me to look at more
[18:55:11] <mambrus> Ok
[18:55:39] <mambrus> Will need to rebuild with symbols. just a minute...
[18:56:22] <mambrus> What was the other warning you noticed btw?
[18:56:44] <jynik> mambrus: Well, whenever you have time. When you get to the stack frame that causes this, try to print out some of the variables and states that are used...
[18:57:26] <mambrus> jynik: Better now than ever or it will never be done ;-)
[18:58:01] <jynik> mambrus: Your first one printed some warnings about the FX3 firmware returning error values. Perhaps that's just because this got the thing into a buggered state...
[18:59:04] <mambrus> Youre right... those dissapeared after rebuild/reflash...
[18:59:12] <mambrus> seems related though.
[19:00:00] <jynik> I thought that's what you were focusing on initially. But yeah, whatever parameters the cal routine is coming back with is yielding an edge case that's clearly not handled properly.
[19:00:06] <jynik> So I want to see what those values are
[19:03:58] <jynik> Btw, nice little attenuator kit on minicircuits: http://www.minicircuits.com/pdfs/K1-VAT+.pdf
[19:04:13] <jynik> http://www.minicircuits.com/products/DesignerKits.shtml
[19:04:42] <l-fy> http://en.wikipedia.org/wiki/Software_defined_mobile_network
[19:16:41] <mambrus> jynik: Seems calibration is running for a while (table generation). In case it's in a outer loop it will take forever to get to the point interactively. Here's bt on the point of faliure via a crashdump instead: http://pastebin.com/yxN5pBBJ
[19:17:41] <mambrus> Unless you find something obvious I can tuck dump in together with symbols and file an issue.
[19:20:57] <jynik> I'll have to sit down while looking at the code a bit more, so feel free to open the issue
[19:21:04] <mambrus> ok