**** BEGIN LOGGING AT Mon Feb 06 03:00:01 2017
Feb 06 05:05:05 <minnesotags> zmatt; That's ok, I removed a "for/done" loop from the script and it solved the problem. I have no idea why running a script in a "screen" session would make that fail, but don't care if I have a workaround.
Feb 06 05:09:57 <zmatt> can't you just run it as service instead?
Feb 06 05:10:21 <zmatt> and trust me, for-loops are not somehow a problem in a screen session :P
Feb 06 05:11:32 <zmatt> (but obviously I can't guess the problem without seeing what you did and what went wrong)
Feb 06 06:47:05 <navidr> Our university gave me 1500$, and they want to buy iot stuff for lab, I am C++ programmer, and I am not very familiar with boards and stuff . What do you think should I buy ?
Feb 06 06:51:12 <LetoThe2nd> navidr: a network enabled coffee maker, of course. then you have Internet, a Thing, and COFFEEE! all in one! yay!
Feb 06 06:51:26 <navidr> :))))
Feb 06 06:51:51 <navidr> I am looking for something like iot starter package .
Feb 06 06:52:07 <LetoThe2nd> navidr: seriously, do you expect us to give you a meaningful answer when your requirements and curriculum are totally unknown to us?
Feb 06 06:52:35 <LetoThe2nd> navidr: rather ask the teachers/professors what they thing might deepen the understanding of their topics.
Feb 06 06:52:44 <navidr> LetoThe2nd: I know it is ridiclous question, but some kind of starting point for IOT board would be great .
Feb 06 06:52:46 <LetoThe2nd> s/thing/think/
Feb 06 06:53:59 <LetoThe2nd> navidr: even that request is ridiculous, as everything with ethernet or bluetooth or wifi or nfc connectivity totally qualifies as a "IoT board"
Feb 06 06:54:55 <LetoThe2nd> navidr: really. do not think about what to buy, think about what to teach. and once you know what to teach, look for a way to do that.
Feb 06 06:56:14 <navidr> LetoThe2nd: LTE extensions for raspberry pi and beaglebone are different ? I want to buy a sim cart cape (is that what they call it), but i don't know it is same for raspberry pi and beaglebone or not
Feb 06 06:56:29 <LetoThe2nd> navidr: some ideas about what to teach: MQTT, Bluetooth, Low Power, Bare metal programming, Mesh networking,...
Feb 06 06:56:44 <LetoThe2nd> navidr: no idea, i never used a rpi or lte.
Feb 06 06:57:35 <navidr> LetoThe2nd: thank you
Feb 06 07:07:01 <navidr> LetoThe2nd : " Bare metal programming" can you explain about this a little bit ?
Feb 06 07:09:44 <LetoThe2nd> navidr: as opposed to "programming on a full blown operating system"
Feb 06 07:10:04 <LetoThe2nd> navidr: think arduino (not exactly in my opinion, but good enough for the moment)
Feb 06 07:11:56 <navidr> LetoThe2nd: In arduino, we don't have any os/kernel under our program ?
Feb 06 07:12:22 <LetoThe2nd> navidr: usually not.
Feb 06 07:12:50 <navidr> LetoThe2nd: Cool, what language do we program ? c
Feb 06 07:12:58 <LetoThe2nd> navidr: c, usually
Feb 06 07:41:00 <zmatt> I discovered that almost certainly my oven runs a baremetal application on an am335x :)
Feb 06 07:41:25 <zmatt> with use of the gpu, interesting enough (that combination is not supported by TI, at least publicly/officially)
Feb 06 07:42:07 * LetoThe2nd prefers heavy metal over bare metal.
Feb 06 07:47:14 <zmatt> bare metal on a SoC like this is kinda neat
Feb 06 07:47:53 <zmatt> being able to use those resources freely without a meddlesome kernel who wants you to play by its rules :)
Feb 06 07:48:22 <LetoThe2nd> depends (TM)
Feb 06 07:48:59 <zmatt> of course there are downsides too :P
Feb 06 07:49:12 <zmatt> linux is simply quite convenient
Feb 06 08:13:46 <navidr> What they call credit card reader ? something like a POS , But I want to read the low level data on the card
Feb 06 08:21:07 <xl0> Hi, people. Would you know, what's the status of the linux kernel on bb-xm?
Feb 06 08:36:42 <xl0> I'm confused, why is stuff called omap3, when it's DaVinchi and not omap?
Feb 06 08:52:59 <zmatt> it's an omap3
Feb 06 08:54:44 <zmatt> the dm37xx is the same SoC as omap36xx
Feb 06 08:55:56 <zmatt> "davinci" and "sitara" are meaningless marketing terms, they don't correspond to any technical characteristics
Feb 06 09:02:45 * KotH hugs zmatt, then goes back to sleep
Feb 06 09:31:02 <Guest83876> Hello, I just flashed a new debian image onto my beaglebone black wireless. When I tried to use connmanctl to connect to my wifi, I use scan and get"Error /net/connman/technology/wifi: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
Feb 06 09:31:02 <Guest83876> " Any ideas on how to fix this?
Feb 06 09:39:49 <NTQ> Hi there. I own a Beaglebone Green and installed Debian Linux on the NAND Flash. Is there a tutorial which shows me how to access the GPIO pins in manner that I can use them in full speed, for example with the PRU unit?
Feb 06 09:49:04 <KotH> NTQ: google for "PRU tutorial"
Feb 06 09:51:37 <xl0> zmatt: Thanks for clarifying. What are the differences between the two?
Feb 06 10:35:04 <zmatt> xl0: between which two?
Feb 06 10:43:15 <xl0> zmatt: dm3730 and a corresponding omap
Feb 06 11:09:02 <lore4> hi folks
Feb 06 11:14:29 <KotH> lo lore4
Feb 06 11:29:49 <lore4> KotH, hi
Feb 06 11:29:56 <lore4> gtg lunch
Feb 06 11:30:25 <KotH> en guete!
Feb 06 13:00:33 <zmatt> xl0: some efuse bits at the very least ;)
Feb 06 13:02:59 <zmatt> and the omap36xx will have some specific stuff for handset vendors, like the interface to the modem chipset
Feb 06 13:03:10 <zmatt> secure boot
Feb 06 13:03:14 <zmatt> etc
Feb 06 13:05:57 <lore4> hi folks
Feb 06 13:07:07 <gquere> Hello. I'm looking at this bytestring in a memory I dumped, what could its purpose be ? 0.1.2.3.4.5.6.7.8.9.:.;.<.=.>.?.....@`A.B.C.D.E`F.G.H.I.J.K.L.M.N.O...P.Q.R.S`TdUhVlWtXxY.Z. ?
Feb 06 13:07:18 <zmatt> lore4: I looked very briefly at clpru... it produces pretty bad code hehe
Feb 06 13:08:49 <zmatt> gquere: consider looking at a hex dump instead, since this is clearly not text but perhaps a sequence of somewhat-but-not-quite-sequential 16-bit numbers
Feb 06 13:09:47 <gquere> i'm looking at the hex dump I pasted the readable string because it'd be easier to recognize if you've seen it before
Feb 06 13:10:06 <lore4> zmatt, hi! What? O.o
Feb 06 13:15:53 <zmatt> gquere: except it's not "readable"
Feb 06 13:17:07 <gquere> it's at least recognizable :p
Feb 06 13:17:39 <zmatt> lore4: I tested a variety of one-line functions to see what code it produces (at highest optimization level)
Feb 06 13:18:16 <zeekhuge> was about to say that there are various optimization levels.
Feb 06 13:20:31 <zmatt> I know :P  and I even still knew many of the compiler options (it's very similar to TI's other compilers such as for the c6x dsps)
Feb 06 13:21:20 <paolo-45> hi
Feb 06 13:22:05 <paolo-45> i need one Beagle Drone to use in flight with my Garmin Virb XE
Feb 06 13:22:50 <lore4> zmatt zeekhuge, so you mean cgt produced code?
Feb 06 13:22:55 <lore4> *do
Feb 06 13:23:03 <paolo-45> it must stay in flight 6 hours
Feb 06 13:23:35 <zmatt> lore4: eh yes, what else could I have meant?
Feb 06 13:23:40 <lore4> paolo-45, O.o?
Feb 06 13:24:00 <lore4> zmatt, I guessed as I did not see any context :P
Feb 06 13:24:09 <paolo-45> i saw Atlantik Solar, it was in flight 81 hours
Feb 06 13:26:06 <paolo-45> i need modules for: Bluetooth, GPS+GLONASS, 4G(LTE), wi-fi for beagle board
Feb 06 13:26:34 <zeekhuge> yeah .. IIRC, '-k' option instructs the clpru to keep the assembly code file that it generates.
Feb 06 13:27:06 <paolo-45> i need a Beagle to use for a drone.
Feb 06 13:28:49 <zmatt> lore4: for example  bool lt0( int a ) {  return a < 0;  }  can be done in two instructions (logical shift right 31, return)
Feb 06 13:29:01 <zmatt> lore4: clpru manages to need 8 instructions for it
Feb 06 13:29:45 <zeekhuge> oh.. its that bad !
Feb 06 13:30:13 <zmatt> using 3 instructions to set bit 31 of a register is pretty amazing too :P
Feb 06 13:33:30 <zeekhuge> and that is what clpru did ?
Feb 06 13:34:04 <zmatt> yeah, here's a nice example of a few problems: https://hastebin.com/gedogomere.cpp
Feb 06 13:35:23 <lore4> zmatt, 300% of overhead than asm
Feb 06 13:35:26 <zmatt> due to integer promotion rules of C, the type of *x + 1 is int
Feb 06 13:35:59 <zmatt> hence it feels the need to do an arithmetic shift right (apparently range analysis is absent or broken)
Feb 06 13:36:42 <zmatt> and it sets bit 31 using three instructions (fill, lsl, or)
Feb 06 13:38:59 <zeekhuge> lore4: it would be great if you could reply that on the ml. https://groups.google.com/forum/#!topic/beagleboard-gsoc/ypiSl_2kQyk
Feb 06 13:39:54 <zmatt> https://hastebin.com/mosonoxugi.cpp
Feb 06 13:40:46 <lore4> zeekhuge, roger that! I thought you wrote me an email :P
Feb 06 13:42:50 <lore4> zmatt, is it not enought 1 QBLE + 1 move 0 to a register for mosonoxugi?
Feb 06 13:43:08 <lore4> i.e. 2 instructions
Feb 06 13:43:23 <zmatt> lsr r14.w0, r14.w0, 31
Feb 06 13:43:24 <zmatt> return
Feb 06 13:43:30 <zmatt> eh
Feb 06 13:43:35 <zmatt> lsr r14, r14, 31
Feb 06 13:44:17 <zmatt> sorry, my brain reads 'w'ord and always thinks 32-bit (never mind that it makes no sense to use subfield syntax for that, brain had a hiccup I guess :P )
Feb 06 13:46:47 <lore4> lol
Feb 06 13:48:43 <lore4> why lsr r14, r14, 31 to do a compare?
Feb 06 13:49:04 <zmatt> extract sign-bit
Feb 06 13:49:56 <zmatt> which is what (x < 0) does
Feb 06 13:50:33 <lore4> uhm.. roger that... I tought to use qblt
Feb 06 13:51:05 <lore4> but probably it doesn't know how to distinguish with signed
Feb 06 13:51:05 <zmatt> that requires 3 instructions though
Feb 06 13:51:25 <zmatt> you would test  r14.b3 >= 0x80
Feb 06 13:51:43 <zmatt> nothing in pru natively supports signed integers
Feb 06 13:52:12 <zmatt> which is why C is not great for this, since it rather steers you toward using signed ints
Feb 06 13:53:33 <lore4> ouch! On the C/C++ optimization guide, there is wrote that 32 bit  forloops has somehow some overhead compared with 16 bit loops
Feb 06 14:00:35 <zmatt> https://hastebin.com/umesalefos.cpp   this one is funny too
Feb 06 14:00:58 <zmatt> implements it as  a ? 1 : 0   followed by a pointless MOV that suggests their register allocator sucks
Feb 06 14:08:18 <lore4> zmatt: oh my gosh... my 146 lines example resulted in 12k lines of asm
Feb 06 14:12:42 <zeekhuge> a lot of lines in the generated ASM are just comments and other such things.
Feb 06 14:12:52 <zeekhuge> if i remember correctly ..
Feb 06 14:13:22 <lore4> zeekhuge, I see several "dwattr" lines... but I ignore what they do
Feb 06 14:13:32 <lore4> googling
Feb 06 14:15:48 <zmatt> you can get rid of those with --symdebug:none
Feb 06 14:19:05 <lore4> ty
Feb 06 14:20:46 <lore4> uhm.. done but it seems that thay are still there O.o
Feb 06 14:21:38 <lore4> ok, now it worked
Feb 06 14:21:46 <lore4> 460 lines..
Feb 06 14:21:53 <lore4> 20% of overhead
Feb 06 14:21:56 <lore4> 200%
Feb 06 14:22:39 <zmatt> you shouldn't measure in lines anyhow... you can do a lot more with a line of C than a line of asm
Feb 06 14:23:44 <KotH> and then there is perl
Feb 06 14:25:14 <lore4> zmatt, you are right, but it is a rough kpi ;)
Feb 06 14:25:38 <lore4> I am also counting in C and ASM empty lines, lines of comments, etc.
Feb 06 14:33:43 <Akos> Hi! Is there a method for protecting intellectual property? Like blowing readout fuses on microcontroller or encrypted memory with read-protected password.
Feb 06 14:34:33 <zmatt> Akos: not in GP (General Purpose) devices, you'd an HS device (High Security)
Feb 06 14:34:45 <zmatt> you'd need to contact TI for more info
Feb 06 15:14:52 <gquere> errr ... how do I find pin 1 on a chip that doesn't have a dot on its package ?
Feb 06 15:16:25 <lore4> it is wrote on the pcb
Feb 06 15:16:33 <lore4> ?
Feb 06 15:16:37 <gquere> nah
Feb 06 15:16:40 <lore4> : \
Feb 06 15:17:03 <gquere> and I don't have a spare to make sure :p
Feb 06 15:23:46 <VirG> hello people
Feb 06 15:28:19 <lore4> hi VirG
Feb 06 15:31:59 <zmatt> gquere: usually there's a dot
Feb 06 15:34:35 <gquere> not for this one i'm afraid
Feb 06 16:04:56 <MathOnNapkins> sounds like a job for a datasheet
Feb 06 16:09:26 <MagicGary> hello!
Feb 06 16:11:41 <MagicGary_> Does anyone know why sometimes I cannot connect to BeagleBone's wifi?
Feb 06 16:13:48 <lore4> magicGary, hi!
Feb 06 16:27:26 <madhu> Hi all
Feb 06 16:27:55 <lore4> anyone know if PRU float are fixed-point type or not with clpru?
Feb 06 16:28:01 <madhu> did anybody boot the board using UART??
Feb 06 16:28:27 <lore4> *does anyone
Feb 06 16:31:11 <zmatt> lore4: floats are float
Feb 06 16:31:23 <zmatt> hence everything you do with them at runtime is a call to a support library
Feb 06 16:32:44 <lore4> hence, emulation?
Feb 06 16:37:20 <lore4> goody, gotta go :)
Feb 06 16:37:26 <lore4> see you all tomorrow
Feb 06 16:40:30 <zmatt> gquere: I just noticed you had said that already, sorry
Feb 06 17:21:41 <peanut_> hello
Feb 06 17:22:07 <peanut_> please help :(
Feb 06 17:30:31 <clockman> what's wrong?
Feb 06 18:31:40 <Serghei> Hello. I am looking for some information about Licensing restrictions for the beagle bone boards. Thank you.
Feb 06 18:52:48 <das> zmatt: yah it's fine I tested it one way, had a bunch of errors in dmesg
Feb 06 18:52:57 <das> flipped it, seems to work :D
Feb 06 18:53:44 <das> seems everytime I mess my 3.3V wiring I get emmc errors in dmesg
Feb 06 19:17:37 <clockman> During the boot process, I see "Using default evironment... Board: BeagleBone Black ... " But it's a BBBW. Where is it that I can manually tell it that the board is a BBBW?
Feb 06 19:19:47 <das> does it have a ethernet port?
Feb 06 19:20:20 <clockman> no. but it definitely thinks it does after booting. Once it's booted, ifconfig -a shows eth0, usb0 and lo
Feb 06 19:20:41 <clockman> but i'm definitely staring directly at a BBB-wireless
Feb 06 19:21:13 <das> i don't know then sorry, I'd install the latest image and roll with that
Feb 06 19:21:54 <clockman> hellls no. I just spent the last couple weeks getting this image perfect. And it works for most the BBBW boards that I have, just two that aren't working right
Feb 06 19:22:25 <das> ahah
Feb 06 19:23:53 <zmatt> clockman: sounds like another case of misflashed eeprom?
Feb 06 19:23:56 <MathOnNapkins> Are you positive it is actually using that image then? Not possible it's booting off removable media or some other media? Seems very strange to me that it wouldn't bring up wln0 if it's a wireless board
Feb 06 19:24:11 <zmatt> MathOnNapkins: wrong eeprom -> wrong dt -> no wireless
Feb 06 19:24:17 <MathOnNapkins> exactly
Feb 06 19:25:50 <zmatt> clockman: head -c 32 /sys/bus/nvmem/devices/0-00500/nvmem | hexdump -C
Feb 06 19:26:38 <zmatt> the first 12 bytes will be <aa 55 33 ee> "A335BNLT"
Feb 06 19:26:47 <clockman> .U3.A335BNLT....
Feb 06 19:26:51 <clockman> aa 55 33 ee
Feb 06 19:27:13 <zmatt> next 4 bytes indicate board revision or variant, and should start with "BW" for a beaglebone black wireless
Feb 06 19:27:23 <zmatt> and it looks like that's not the case for you :)
Feb 06 19:27:26 <clockman> well there's the problem right there.
Feb 06 19:27:38 <zmatt> after that comes the serial number, which should also be printable ascii
Feb 06 19:32:18 <clockman> does the am335x_evm.sh fix that? That's the only thing on Google that I'm seeing that even comes close to the problem.
Feb 06 19:33:16 <zmatt> while you could modify the eeprom yourself, this requires grounding a test point to write-unprotect the eeprom.  regardless I think it should be reported to the manufacturer
Feb 06 19:33:52 <clockman> oky doky. Will do. thanks for the help.
Feb 06 20:22:03 <Henry__> Hello
Feb 06 20:22:47 <Henry__> Does BeagleBoard works with TI Wilink8 EVM
Feb 06 20:25:56 <ahills> stop PMing people from the user list
**** ENDING LOGGING AT Tue Feb 07 03:00:01 2017