**** BEGIN LOGGING AT Wed May 17 03:00:04 2017
May 17 05:55:13 <ard001> Greetings
May 17 05:56:25 <ard001> I am Abdur Rahman Daanish.Unfortunately I couldn't apply for GSoC 2017 but still I would like to contribute to Bonescript.js
May 17 05:56:50 <ard001> Just wanted to ask how to start
May 17 05:59:12 <gquere> zmatt: not exactly sure what this implies in my case but i'll be looking at that then, thanks
May 17 06:01:07 <zmatt> gquere: note that trace doesn't actually require JTAG even
May 17 06:02:33 <zmatt> if you want some form of continuous trace, the System Trace Macrocell is able to trace continuously to hardware pins
May 17 06:03:14 <zmatt> this supports software-generated trace messages, and various hardware sources (e.g. the L3 interconnect statistics collectors), but not cpu instruction trace
May 17 06:56:47 <ard001> clear
May 17 10:58:02 <ard001> exit
May 17 11:14:12 <cyberguy> Hi, I'm about to use Marvell 88e6xxx Ethernet switch, I've seen the driver is under the linux drivers folder, but how do i configure it within the device tree to use RGMII ?
May 17 11:27:14 <cyberguy> Hi, I'm about to use Marvell 88e6xxx Ethernet switch, I've seen the driver is under the linux drivers folder, but how do i configure it within the device tree to use RGMII ?
May 17 11:49:36 <cyberguy> Hi, I'm about to use Marvell 88e6xxx Ethernet switch, I've seen the driver is under the linux drivers folder, but how do i configure it within the device tree to use RGMII ?
May 17 12:04:47 <zmatt> please don't repeat your question every 15 minutes, that's quite annoying
May 17 12:06:07 <cyberguy> zmatt: sorry, can you help with this? how do i change the device tree to use a specific driver
May 17 12:07:20 <zmatt> I'm not familiar with the Marvell 88e6xxx ethernet switch, nor do I have time right now to look it up.  just stick around and be patient, with a bit of luck someone might know or have suggestions
May 17 12:28:48 <cyberguy> zmatt: What would you suggest me to do to get that info? what to search with google...?
May 17 13:58:54 <Jacobitz> Hello
May 17 13:59:52 <Jacobitz> I have about 5 BeagleBone Greens and they all have this strange intermittent ethernet issue. I'm having trouble finding information on this by doing a Google search.
May 17 13:59:59 <Jacobitz> Has anyone heard of this before?
May 17 14:02:55 <Jacobitz> Let me provide just a bit more detail. Sometimes the ethernet lights don't light up and the Ethernet interface doesn't have an assigned IP address. Sometimes I need to reboot to get an address. I've already ruled out a bad cable or bad network. That's definitely not it.
May 17 14:11:33 <zmatt> yeah the ethernet phy can sometimes have an opinion of its own
May 17 14:12:15 <zmatt> they patched up the worst problems in software, but I still occasionally notice failure to come up right after power-on
May 17 14:12:50 <zmatt> in my experience it does seem to be influence by expansion hardware (e.g. CAPE), but I'm not sure how exactly
May 17 14:12:58 <zmatt> *influenced
May 17 14:27:13 <Jacobitz> If it could be influenced by expansion hardware it may be a power issue. I don't have any expansion hardware but having some may exacerbate the problem
May 17 14:28:11 <Jacobitz> I have seen some threads for the original beagleboard and older versions of the beaglebone black suggesting it could be a hardware issue. Can't find anything on the BeagleBone Green though.
May 17 14:34:42 <Jacobitz> Oh! There seems to be something here: https://groups.google.com/forum/#!topic/beagleboard/9mctrG26Mc8%5B176-200%5D
May 17 14:41:51 <zmatt> it might be power-related but it's fairly subtle
May 17 14:42:38 <zmatt> basically the power-on reset of the ethernet phy is very marginal
May 17 14:44:08 <Jacobitz> Seems to be a race condition between the microcontroller and phy power signals
May 17 14:44:16 <zmatt> if I were asked to get rid of this issue definitely, my gut instinct would be a proper reset-extender on the reset pin (P9.10 iirc?)
May 17 14:44:27 <zmatt> more between the pmic and phy
May 17 14:44:36 <Jacobitz> got ya
May 17 14:44:59 <zmatt> the time between power stabilized and reset-deassertion is not comfortable for the phy
May 17 14:45:21 <Jacobitz> what is a reset-extender?
May 17 14:45:45 <zmatt> a chip that keeps reset low for a certain about of time
May 17 14:45:51 <thinkfat> monoflop
May 17 14:46:31 <Jacobitz> Or a large capacitor?
May 17 14:46:47 <zmatt> a more invasive fix (i.e. pcb change) would be to reset the phy only via a gpio
May 17 14:46:48 <zmatt> no
May 17 14:46:52 <zmatt> they already did that
May 17 14:47:07 <zmatt> by some accounts it might actually be contributing to the problem
May 17 14:47:47 <zmatt> it will rise way too slow, making it quite uncertain who will exit reset when
May 17 14:49:04 <Jacobitz> can you summarize this in a bit more detail please? I don't really understand the root cause. what is the pmic?
May 17 14:49:52 <zmatt> power management IC
May 17 14:52:01 <zmatt> here's a plot of power-up right now:  https://goo.gl/photos/AKupVAFdx286oTrP6    yellow=current, blue=1.8V, cyan=3.3V, green="PGOOD"/power-on-reset, brown=reset
May 17 14:52:50 <zmatt> the whole power-up sequence basically happens all the way on the left side (where the current spikes are as supplies get turned on and capacitors need to charge)
May 17 14:53:24 <zmatt> then a bit of quiet as the PMIC lets things settle before releasing reset
May 17 14:54:31 <zmatt> that happens in the center, and you see the reset line has such a big cap that its journey upward takes ages
May 17 14:56:29 <zmatt> every major division is 5 ms if I remember correctly
May 17 14:56:54 <zmatt> the ethernet phy wants 25 ms between 3.3V supply stabilized and reset-release
May 17 14:59:10 <zmatt> is there 25 ms between 3.3V stable and reset-high in this picture?  I have no idea :)  depends on its input trigger level
May 17 15:00:29 <Jacobitz> So it's sometimes the right amount of time and sometimes not....
May 17 15:00:35 <Jacobitz> Well that's a bummer
May 17 15:00:41 <zmatt> it usually seems to be
May 17 15:01:19 <zmatt> do you have stuff connected to the expansion header pins?
May 17 15:02:28 <Jacobitz> Yes. In fact I'm powering another board from this power suppy. I have some external circuitry to make sure I'm only powering that external board after the reset pin goes high. This adds a bit of capacitance to the reset line
May 17 15:03:03 <zmatt> :)
May 17 15:04:56 <Jacobitz> Wouldn't more capacitance on the reset line help this situation because it would keep the reset line lower longer? Maybe long enough for the phy chip to be happy?
May 17 15:07:37 <zmatt> apparently not...
May 17 15:07:56 <Jacobitz> hmmmm
May 17 15:08:20 <zmatt> so, maybe my impression that the reset is too short is incorrect, although behaviourally it clearly doesn't get reset right
May 17 15:08:36 <Jacobitz> yeah, it's right in a gray area
May 17 15:14:24 <zmatt> there might be a lesson somewhere in there about low-cost ethernet phys ;)
May 17 15:14:35 <thinkfat> :-)
May 17 15:14:37 <zmatt> although for most people it works fine apparently
May 17 15:14:51 <zmatt> but it seems to fuss easily
May 17 15:14:57 <thinkfat> I have the problem just sometimes on the BBB. Not frequently
May 17 15:15:21 <zmatt> yeah it's annoying to attack since it's so rare
May 17 15:18:55 <Jacobitz> When you said "the ethernet phy wants 25 ms between 3.3V supply stabilized and reset-release" I was assuming you meant at a minimum. But you're saying it wants 25ms at a maxium?
May 17 15:29:03 <zmatt> no, minimum
May 17 15:30:00 <zmatt> I don't really have an explanation of why increasing the cap is causing issues, except the possibility that the cpu might get released before the phy and that might perhaps cause trouble
May 17 15:30:11 <Jacobitz> Oh. Hmmmm.
May 17 15:30:19 <Jacobitz> Well thanks for your help
May 17 15:30:32 <zmatt> or the phy just has issues with the reset rising so slowly
May 17 15:30:42 <zmatt> or something else I haven't thought of yet :)
May 17 22:00:55 <ayjay_t> whats the proper value for phys_offset
May 17 22:01:01 <ayjay_t> what does it even mean "main memory"
May 17 22:02:33 <zmatt> ayjay_t: ?
May 17 22:06:29 <ayjay_t> one of my friends is using CLFS and RCN's kernel to build linux from scratch on the beage
May 17 22:06:51 <ayjay_t> why wont they just join irc
May 17 22:07:01 <ayjay_t> they're really smart but they get so scared of people being mean to them on irc
May 17 22:12:08 <zmatt> ayjay_t: and where did your question above come from?  phys_offset ? what?
May 17 22:12:21 <zmatt> main memory = ram
May 17 22:13:51 <ayjay_t> yeah sorry zmatt, my friend who just joined, shpew12, was asking me this question
May 17 22:13:54 <ayjay_t> and i was feeling the same way you are
May 17 22:14:24 <zmatt> hi friend-of-ayjay-who-just-joined :)
May 17 22:15:01 <shpew12> hi! I was using rcn's stable-kernel repo, when I was compiling the kernel it prompted me for PHYS_OFFSET. I think that's what yall are talking about
May 17 22:15:44 <zmatt> no such var exists in my .config
May 17 22:16:43 <shpew12> it has a (NEW) next to it, maybe that's why?
May 17 22:17:17 <zmatt> 'new" just means "added since your previous config" or something like that
May 17 22:17:34 <ayjay_t> so are you guys using different versions?
May 17 22:17:41 <zmatt> what config did you start out with?
May 17 22:18:04 <ayjay_t> the beag has 64KB of internal RAM right, and then 512MB of external RAM?
May 17 22:18:14 <ayjay_t> or those treated the same by the kernel
May 17 22:18:34 <zmatt> shpew12: you need to select ARCH_MULTIPLATFORM
May 17 22:18:39 <zmatt> which is however the default
May 17 22:18:57 <zmatt> wait
May 17 22:18:57 <zmatt> or not
May 17 22:19:16 <zmatt> yeah
May 17 22:19:20 <zmatt> you do
May 17 22:19:25 <zmatt> then PHYS_OFFSET will vanish
May 17 22:19:53 <zmatt> but, may I suggest that if you're using rcn's kernel, you might also want to use his config? :)
May 17 22:20:24 <ayjay_t> LOL
May 17 22:20:24 <zmatt> at least to have some defaults before you tweak things to taste
May 17 22:21:31 <shpew12> That makes sense. I was looking for it before, this is the first time I compiled a kernel, so when I got the menu I wasn't actually sure if I was using the config file or not.
May 17 22:22:12 <zmatt> btw, are you compiling natively or cross-compiling?
May 17 22:22:22 <shpew12> cross-compiling
May 17 22:22:31 <zmatt> ok good, just checking :)
May 17 22:22:41 <zmatt> using rcn's scripts would be cheating I guess
May 17 22:23:12 <zmatt> which kernel are you trying to compile specifically?
May 17 22:23:43 <shpew12> This one is 3.7.10.
May 17 22:23:48 <zmatt> 4.9.28-bone4 ? 4.10.16-bone1 ?
May 17 22:23:49 <zmatt> what?
May 17 22:23:50 <ayjay_t> O_O
May 17 22:23:53 <ayjay_t> LOL
May 17 22:24:01 <ayjay_t> i'm literally laughing out loud
May 17 22:24:27 <shpew12> err... I'm guessing that means old
May 17 22:24:49 * ayjay_t facepalms
May 17 22:24:52 <zmatt> not so much *old* ...
May 17 22:24:55 <zmatt> more like antique
May 17 22:24:57 <ayjay_t> so i think 4.9 is where the patched the dirty cow
May 17 22:25:01 <ayjay_t> correct me if i'm wrong
May 17 22:25:14 <zmatt> parse error
May 17 22:25:43 <ayjay_t> ooh 4.4.26, 4.7.9, 4.83 it was patched
May 17 22:25:49 <zmatt> what  was patched?
May 17 22:25:51 <ayjay_t> the dirty cow
May 17 22:25:59 <ayjay_t> a priveledge escalation exploit
May 17 22:26:14 <ayjay_t> it was patched in Oct 2016
May 17 22:26:25 <zmatt> lol I managed to miss that one somehow
May 17 22:26:47 <ayjay_t> reddit.com/r/linuxmasterrace
May 17 22:26:48 <zmatt> presumably that will have been backported to all stable releases probably?
May 17 22:27:24 <ayjay_t> i'm actually not 100% on that
May 17 22:27:33 <ayjay_t> the wikipedia is like "you better be using a new version of linux!"
May 17 22:28:06 <ayjay_t> btw, shpew12 isn't using glibc
May 17 22:28:10 <ayjay_t> he's using musl
May 17 22:29:23 <zmatt> isn't that still lacking some functionality systemd needs?
May 17 22:29:29 <zmatt> or has that been fixed?
May 17 22:31:14 <shpew12> I got the impression that it was useable because they use it in clfs
May 17 22:31:35 <zmatt> trying to remember what the c stands for
May 17 22:31:54 <zmatt> ah
May 17 22:31:55 <zmatt> cross
May 17 22:33:02 <zmatt> eh
May 17 22:33:03 <ayjay_t> i bet lfs doesn't even do systemd
May 17 22:33:10 <zmatt> the entire clfs userspace consists of busybox
May 17 22:33:10 <ayjay_t> i bet lfs still uses openrc or something
May 17 22:33:13 <ayjay_t> LOL
May 17 22:33:36 <shpew12> yah it's true.
May 17 22:33:43 <zmatt> shpew12: this is more something for really limited devices
May 17 22:33:53 <shpew12> that was the plan
May 17 22:34:24 <zmatt> as in, devices where you typically don't spend much time actually logged in and doing stuff
May 17 22:34:31 <zmatt> like a router
May 17 22:35:10 <zmatt> this is not going to yield anything resembling a comfortable environment to do stuff in
May 17 22:35:20 <ayjay_t> i wonder what it will be like though
May 17 22:35:22 <ayjay_t> i'm curious
May 17 22:35:36 <zmatt> ayjay_t: just drop into initramfs :P
May 17 22:35:52 <ayjay_t> oh thats actually a good idea
May 17 22:35:56 <zmatt> although that might be more luxuriously equipped than clfs
May 17 22:35:57 <ayjay_t> but if it's like puppy linux thats a full OS
May 17 22:36:27 <ayjay_t> https://busybox.net/live_bbox/live_bbox.html
May 17 22:36:30 <ayjay_t> :-p
May 17 22:36:39 <ayjay_t> so basically thats what we can expect
May 17 22:36:43 <zmatt> shpew12: it all depends of course on what your goals are
May 17 22:37:03 <ayjay_t> woooowwww thats a javascript pc emulator
May 17 22:37:08 <ayjay_t> someone has WAY too much time on their hands
May 17 22:37:27 <ayjay_t> i think we're going to try and run electron as process #1
May 17 22:37:41 <ayjay_t> electron as in node.js+chromium combined into one thing
May 17 22:38:51 <ayjay_t> but don't tell anyone i said that
May 17 22:38:56 <ayjay_t> <_< >_>
May 17 22:39:13 <zmatt> ayjay_t: oh yeah, that thing, never seen it before?
May 17 22:39:25 <ayjay_t> the busy box emulator?
May 17 22:39:30 <ayjay_t> i mean, now i have a good idea what busy box is
May 17 22:39:57 <ayjay_t> i'm guessing you can't run xserver :-)
May 17 22:40:03 <zmatt> sure you can
May 17 22:40:13 <ayjay_t> oh god zmatt now that i'm thinking about it, running electron as pid 1 is going to be so so painful
May 17 22:40:16 <zmatt> if you compile and install it along with all its dependencies
May 17 22:40:23 <zmatt> yes, please don't do that
May 17 22:40:58 <ayjay_t> were doing awful things zmatt
May 17 22:41:00 <ayjay_t> awful awful things
May 17 22:41:02 <ayjay_t> we're heretics
May 17 22:41:09 <zmatt> please also try to ignore that this exists -> https://node-os.com/
May 17 22:41:15 <ayjay_t> HAHAHA
May 17 22:41:47 <shpew12> O_O
May 17 22:42:15 <ayjay_t> i think my first ever tweet was "javascript is spredding like a disease"
May 17 22:42:33 <zmatt> and with "never seen it before?" I meant that x86 emu in javascript
May 17 22:42:39 <zmatt> there's also http://copy.sh/v86/
May 17 22:42:47 <ayjay_t> no i've never seen a x86 emu js
May 17 22:43:02 <ayjay_t> LOL WUT
May 17 22:43:09 <ayjay_t> boothcess
May 17 22:43:11 <ayjay_t> bootchess*
May 17 22:43:15 <ayjay_t> i would love to install that into a vending machine
May 17 22:43:25 <ayjay_t> the one with the credit card acceptor and touch screen
May 17 22:43:48 <zmatt> https://vimeo.com/46465140
May 17 22:45:24 <zmatt> :)
May 17 22:45:48 <ayjay_t> thats so silly
May 17 22:45:59 <ayjay_t> i mean, honestly though, i want to run electron as pid 1
May 17 22:46:10 <ayjay_t> but ill admit, that i don't know enough about any of these frameworks to know if its possible
May 17 22:46:14 <zmatt> you can't
May 17 22:46:36 <ayjay_t> i mean i realize that i can't natively, i would need to hack it
May 17 22:46:39 <ayjay_t> but thats the plan
May 17 22:47:35 <zmatt> https://benja.io/
May 17 22:48:21 <ayjay_t> WHY ARE YOU LIKE THIS
May 17 22:48:25 <ayjay_t> is this all just in your memory?
May 17 22:48:42 <zmatt> thank god not, hopefully I'll get it out again
May 17 22:48:52 <zmatt> I just googled
May 17 22:48:59 <zmatt> also, eh?  "an Electron instance that runs on the lightweight, Wayland driven, Weston compositor, falling back to a bare minimum accelerated Xorg."
May 17 22:49:16 <ayjay_t> shpew12 LOOK AT THAT LINK
May 17 22:49:23 <zmatt> DON'T DO IT
May 17 22:49:49 <zmatt> ayjay_t: afaict Electron doesn't even run on wayland yet
May 17 22:50:01 <zmatt> well, other than via xwayland presumably
May 17 22:50:05 <zmatt> it needs x11
May 17 22:50:23 <ayjay_t> zmatt thats literally our business model
May 17 22:50:25 <ayjay_t> and they're already doing it
May 17 22:51:25 <zmatt> required 8GB of space
May 17 22:51:27 <ayjay_t> i guess its good to have some competition.
May 17 22:51:28 <ayjay_t> STOP
May 17 22:51:33 <zmatt> *requires
May 17 22:51:38 <ayjay_t> whhhhhhhat
May 17 22:51:51 <zmatt> to burn their iso
May 17 22:51:54 <zmatt> "You will need a Micro SD card, possibly Class 10, and with at least 8GB of space."
May 17 22:52:18 <ayjay_t> yeah i get it i just, i mean, i don't know what this guy is doing
May 17 22:53:29 <ayjay_t> we're pursuing a similiar angle but my vision is slightly different
May 17 22:53:43 <zmatt> I mean, it's just an Arch linux system with electron on autostart
May 17 22:53:51 <zmatt> that's what it sounds like anyway
May 17 22:53:57 <ayjay_t> yeah, thats the sense i'm getting
May 17 22:54:05 <ayjay_t> we're putting way more effort into the workflow for a developer
May 17 22:54:16 <ayjay_t> and we're putting way more effort to makeing it easy to use advanced features
May 17 22:55:07 <zmatt> I also found that electron has no concrete plans yet to begin work towards native wayland support, so yeah it'll be using xwayland
May 17 22:55:21 <ayjay_t> tbh, i mean it's just using chromium's compositor right?
May 17 22:55:22 <ayjay_t> called blink.
May 17 22:55:30 <ayjay_t> blink is actually my favorite part of that framework
May 17 22:55:46 <ayjay_t> they abstracted the html renderer from the graphics rendered really well
May 17 22:56:46 <zmatt> we briefly used electron for something
May 17 22:56:57 <zmatt> ditched it because it was way too slow and bulky
May 17 22:57:47 <ayjay_t> you guys make sound equipment right?
May 17 22:58:22 <zmatt> this wasn't for an audio project
May 17 22:58:36 <ayjay_t> ive always been confused by the nature of your work
May 17 22:58:45 <zmatt> don't worry, so are we
May 17 22:58:52 <ayjay_t> because whenever i got a link LOL to your github and looked up the company...
May 17 22:59:01 <ayjay_t> the dutchanddutch ones, i was like... "speakers?"
May 17 22:59:09 <zmatt> yeah :D
May 17 22:59:19 <ayjay_t> and i just don't know what kind of speakers you have
May 17 22:59:24 <ayjay_t> that you have to modify the kernel to run them
May 17 22:59:31 <ayjay_t> or that you hvae to do any of the things you're good at
May 17 22:59:56 <ayjay_t> maybe you're a spy
May 17 23:00:01 <ayjay_t> like a secret agent engineer
May 17 23:00:05 <zmatt> http://dutchdutch.com/#scroll-to
May 17 23:00:06 <zmatt> these
May 17 23:00:48 <ayjay_t> ....
May 17 23:00:53 <ayjay_t> so is there a feedback loop?
May 17 23:01:07 <ayjay_t> nvm
May 17 23:01:09 <ayjay_t> i'm not even gonna try
May 17 23:01:15 <ayjay_t> it still just looks like you're making speakers
May 17 23:01:52 <ayjay_t> don't take that the wrong way
May 17 23:01:58 <zmatt> well, the bbb is right now mostly for getting the dsp programmed, the web interface, that sort of stuff
May 17 23:02:33 <ayjay_t> are you guys using the new one with wifi?
May 17 23:02:35 <zmatt> although there are audio links between bbb and dsp, so you could run mpd on it ;)
May 17 23:03:14 <zmatt> no we're not using wifi yet
May 17 23:03:28 <zmatt> I think we're looking into it, but not really high priority right now I think
May 17 23:03:45 <zmatt> the shower otoh does have wifi (that's also where we briefly used electron)
May 17 23:04:21 <ayjay_t> shower otoh?
May 17 23:04:32 <zmatt> on the other hand
May 17 23:04:36 <ayjay_t> oh
May 17 23:04:38 <ayjay_t> you made a shower?
May 17 23:04:41 <ayjay_t> WHAT IS GOING ON
May 17 23:05:07 <ayjay_t> so anyway, i just set up an atmel (microchip) atwinc3400 and atwinc1500
May 17 23:05:14 <ayjay_t> both precertified for FCC
May 17 23:05:20 <ayjay_t> about 40mbit/s UDP
May 17 23:05:35 <ayjay_t> 1x1
May 17 23:11:52 <ayjay_t> my goal was to take a 1/4 inch male to female and basically just wedge the wifi module inbetween there
May 17 23:12:04 <ayjay_t> and just stream the data to a central router
May 17 23:13:15 <zmatt> ehm, I have no idea what topic we just shifted to?
May 17 23:13:22 <zmatt> also, I kinda still have some stuff to do
May 17 23:14:52 <ayjay_t> ha yeah good luck
May 17 23:14:55 <ayjay_t> thanks for your help earlier
**** ENDING LOGGING AT Thu May 18 03:00:02 2017