**** BEGIN LOGGING AT Sun Jan 24 02:59:57 2021 Jan 24 13:14:50 q/q Jan 24 19:28:42 good evening all Jan 24 19:29:44 oh it's you Jan 24 19:30:33 oh hi! Jan 24 19:31:14 mru: thanks to your recommendation for the x15 that i'm here ;-) Jan 24 19:31:32 i already have a bbb, so the x15 is something i'm gonna look into Jan 24 22:34:49 oof, the recent revision of the TRM for the j721e (am752x/dra82x/tda4vm) is now 320 MB and 18155 pages Jan 24 22:35:34 the size increase seems rather disproportionate to its increase in pages Jan 24 22:36:06 try running it through a pdf cleaner Jan 24 22:36:30 got a link to the pdf? Jan 24 22:36:49 https://www.ti.com/lit/ug/spruil1b/spruil1b.zip is the zipped pdf Jan 24 22:43:35 the cleaner is taking its time Jan 24 22:43:38 and 2G ram Jan 24 22:44:08 lol Jan 24 22:52:35 still running Jan 24 22:52:41 maybe it's stuck Jan 24 22:52:53 trying another tool Jan 24 22:53:36 now they're both using 2G ram and 100% cpu Jan 24 22:54:50 looks like the new c7x dsp subsystem (unsurprisingly) has some fun memory coherency bugs Jan 24 22:55:45 (I say unsurprisingly since whenever something new is designed with memory coherency, there invariably seem to be exciting memory coherency bugs in the first few revisions) Jan 24 22:55:54 oh, I'm not surprised Jan 24 22:56:19 these are software changes or would this show up in later hw Jan 24 22:58:29 I mean, fixing the hardware bugs will take hardware revisions... software may be able to work around the issues of course, possibly with imposing restrictions on use (e.g. avoid certain types of DMA transfers) Jan 24 22:59:32 or nasty performance penalties Jan 24 23:01:47 using revision 1.0 of any complex chip comes with risks ;) Jan 24 23:02:48 got the pdf down to 223M Jan 24 23:03:11 the funniest (unless you're the one affected) I've seen was one where certain power-up sequencing could result in the eFuse controller being clocked without being reset (or prior to being reset), potentially causing it to proceed to burn a random efuse... new dice-roll each power-up :D Jan 24 23:04:02 sounds like some fun debugging Jan 24 23:04:17 who checks for that Jan 24 23:04:18 I mean, I guess that's a significant reduction, but that's still huge compared to the 79M of the previous revision of this TRM Jan 24 23:04:26 command to shrink that pdf: qpdf --stream-data=compress --compress-streams=y --decode-level=specialized --recompress-flate --compression-level=9 --object-streams=generate --remove-unreferenced-resources=yes SPRUIL1B.pdf out.pdf Jan 24 23:04:27 sounds like a nightmare trying to chase that down Jan 24 23:04:44 not really, this sounds pretty easy to pin down Jan 24 23:05:20 like, if your devices suddenly lose features or reside in a lower speed bin Jan 24 23:05:43 figuring that an efuse was burned is one thing Jan 24 23:05:51 figuring out why seems harder Jan 24 23:06:42 maybe, I don't feel like I can estimate that part Jan 24 23:20:29 what do entry level people do in this field Jan 24 23:20:30 debugging Jan 24 23:20:42 embedded systems role Jan 24 23:59:14 Anyone have any advice or pointers for a (iot) cellular modem for the BBGW. I was looking at the Hologram Nova/SIM. Wondering what's working for you all. Jan 25 01:59:00 was there a way to fix the annoying wl18xx driver crashes on the beaglebone black/green wireless? Jan 25 01:59:26 have you confirmed that that's what's happening? Jan 25 02:02:20 unsure of how to.. I'm assuming that's what causing it since I get disconnected from ssh and the wifi no longer shows on my pc Jan 25 02:29:28 you'll need to log in via usb and check logs Jan 25 02:30:05 how do I check logs? Jan 25 02:30:28 journalctl Jan 25 02:32:26 journalctl -o short-monotonic to get timestamps since boot rather than date+time since your beaglebone isn't connected to the internet (I presume?) hence won't have the correct date/time :P **** ENDING LOGGING AT Mon Jan 25 02:59:57 2021