**** BEGIN LOGGING AT Mon Jun 11 03:00:03 2018
Jun 11 03:37:30 <Anu> Hi, which Beagleboard has KVM enabled ,where guest VMs could be created ?
Jun 11 03:40:35 <Anu> If someone has a KVM enabled beagleboard, please post the output of the command "kvm-ok"
Jun 11 05:29:17 <Keshav_> hi
Jun 11 05:32:55 <Anu> Hi, Please suggest which Beagleboard should I use to utilize KVM so as to create guest VMs.
Jun 11 06:27:30 <Anu> Hi, Please suggest which Beagleboard should I use to utilize KVM so as to create guest VMs.
Jun 11 06:37:17 <LetoThe2nd> Anu: x15.
Jun 11 06:38:19 <Anu> Thanks LetoThe2nd , may I please get the output of "kvm-ok" on x15, if somebody has it ?
Jun 11 06:38:43 <Anu> from the command prompt of x15
Jun 11 06:49:09 <Anu> may I please get the output of "kvm-ok" from the command prompt of beagleboard x15, if somebody has it ?
Jun 11 06:49:43 <LetoThe2nd> Anu: please, stop repeating yourself over and over again. if nobody answers, then nobody can help you at the moment.
Jun 11 06:53:50 <Anu> Sorry for spamming, I needed some data before going for a purchase
Jun 11 06:56:07 <zmatt> -bash: kvm-ok: command not found
Jun 11 06:57:39 <zmatt> :P
Jun 11 06:58:15 <Anu> Thanks zmatt,  does "dmesg | grep 'kvm' " return  anything ?
Jun 11 06:59:02 <zmatt> regardless of the existence of that command or lack thereof (it doesn't exist on any machine I tested, including two servers), I know I'm not running a kvm-capable kernel, it's not an lpae kernel
Jun 11 06:59:15 <zmatt> I still have trying one on my to-do list
Jun 11 06:59:21 <zmatt> I know the hardware is vm-capable
Jun 11 07:00:43 <Anu> These are couple of ways I know to ensure that KVM is enabled and initialized by the kernel,  I will be glad to know if there are other ways by which we can create VMs efficiently
Jun 11 07:01:23 <zmatt> TI seems to be fond of something called "Jailhouse": http://www.ti.com/lit/wp/spry317/spry317.pdf
Jun 11 07:01:30 <zmatt> never heard of it myself before
Jun 11 07:03:28 <zmatt> I've booted a kernel in HYP mode on the closely related omap5 before, it shouldn't be hard to replicate that on the x15
Jun 11 07:05:15 <Anu> Good to know that zmatt, how would you do that ? booting a kernel in Hyp mode for ARM based processor ?
Jun 11 07:07:21 <Anu> I have worked on an ARM processor where KVM is not initialized, used virt-manager GUI (that calls QEMU commands) , but it only allows to open an exiting image file, the options of creating a new VM are disabled for the lack of KVM
Jun 11 07:07:57 <zmatt> you need an LPAE kernel with virtualization enabled, and an u-boot capable of bringing the cpu into HYP mode
Jun 11 07:08:55 <zmatt> ah
Jun 11 07:09:00 <zmatt> root@bbx15:~# dmesg | grep -i hyp
Jun 11 07:09:00 <zmatt> [    0.068526] CPU: All CPU(s) started in HYP mode.
Jun 11 07:09:34 <zmatt> so it seems the default u-boot for bbx15 should already be fine, so with the right kernel it should support kvm
Jun 11 07:11:24 <Anu> Hmm , thank you for sharing that output
Jun 11 07:11:45 <zmatt> maybe I'll give it a try sometime this week
Jun 11 07:11:54 <zmatt> I think the big concern would be ram
Jun 11 07:15:20 <zmatt> although I guess it depends on how much crap is running on the host and in the VM(s).  a debian system without crap installed doesn't seem to require all that much memory: https://pastebin.com/raw/4q48A4DA
Jun 11 09:37:46 <TonyTheLion> can you assign an irq to a gpio pin?
Jun 11 09:39:11 <tbr> you can get interrupts from pins, yes
Jun 11 09:43:08 <TonyTheLion> yes but trying to enable wake from sleep from pins
Jun 11 09:44:47 <TonyTheLion> apparently you're supposed to be able to set the gpio irq as a wakeup irq
Jun 11 09:45:39 <TonyTheLion> per the next to last post on this thread https://e2e.ti.com/support/arm/sitara_arm/f/791/t/626971
Jun 11 09:47:54 <zmatt> the question isn't "can" but "how" ;P
Jun 11 09:48:45 <TonyTheLion> ye that is the question
Jun 11 09:49:03 <zmatt> the easy way to get wakeups from a gpio is by using the gpio-keys driver
Jun 11 09:49:17 <tbr> ah, wakeup, different can of worms. No clue
Jun 11 09:49:29 <zmatt> https://github.com/mvduin/overlay-utils/blob/master/gpio-wakeup-key.dtsi
Jun 11 09:49:39 <TonyTheLion> ye thats the easy way, but that isn't useful for my purposes
Jun 11 09:49:45 <zmatt> because?
Jun 11 09:50:15 <TonyTheLion> because its a key, has debouncing and can't use it for automated interrupt wakeups close together that are machine generated
Jun 11 09:50:16 <zmatt> if your purpose is specific, you also need to ask a more specific question
Jun 11 09:50:24 <zmatt> debouncing can be disabled
Jun 11 09:50:46 <zmatt> you can configure gpio-keys to use an interrupt as source instead of a gpio as source
Jun 11 09:51:14 <zmatt> (the gpio controllers are also irq controllers)
Jun 11 09:51:56 <TonyTheLion> ah I see
Jun 11 09:52:16 <TonyTheLion> is there some kind of docs on this?
Jun 11 09:52:21 <zmatt> https://www.kernel.org/doc/Documentation/devicetree/bindings/input/gpio-keys.txt
Jun 11 09:53:39 <zmatt> and then use an interrupt spec like https://github.com/mvduin/py-uio/blob/master/dts/gpio-irq.dtsi#L27-L28
Jun 11 12:26:23 <fred__tv> Hi A friend has a bbwhite with ancient kernel (ubuntu 12.04)  , he needs to connect a 8relays board to gpios, relays go active at 0 input state,
Jun 11 12:27:15 <fred__tv> there is no problem to configure gpios active_low/active_high, the problem is gpio state change at boot for a fraction of second
Jun 11 12:27:57 <fred__tv> since relays drive further hi-power step relays , this is a problem. How can it be avoided ?
Jun 11 12:29:14 <zmatt> does this happen at early boot or when he initializes the gpios?
Jun 11 12:29:22 <zmatt> actually, it's probably the latter
Jun 11 12:29:56 <zmatt> most likely he's simply initially setting the outputs to low
Jun 11 12:30:03 <fred__tv> honestly I don't know, but now he is explaining me:
Jun 11 12:30:22 <zmatt> when switching direction from input to output, you need to specify the initial level
Jun 11 12:30:43 <zmatt> i.e. you write "high" or "low" to the "direction" property
Jun 11 12:31:06 <zmatt> writing "out" is unfortunately also accepted as a synonym for "low"
Jun 11 12:33:53 <fred__tv> now it's working with active state high, gpio=1 relay=1, he has to change relay board that works with negate state, the problem is when beagle starts the initial transition (0 to 1) causes a quick change in relays state
Jun 11 12:34:19 <fred__tv> despite an external pullup resistor is present
Jun 11 12:34:34 <zmatt> yeah he's just initializing the output to low and then setting it high
Jun 11 12:34:40 <zmatt> he should be initializing the output to high
Jun 11 12:35:03 <fred__tv> can it be done ?
Jun 11 12:35:13 <zmatt> I just explained how
Jun 11 12:35:19 <fred__tv> sry
Jun 11 12:35:56 <zmatt> you write "high" to the direction attribute of the gpio to set it to output-high
Jun 11 12:37:00 <fred__tv> from what I remember this was already done, I suspect the spike occours before gpio initialization
Jun 11 12:38:05 <zmatt> no, all gpios are high-impedance (with 100KΩ ±50% pull-up or -down) until they are explicitly configured as output by userspace (or by the device tree)
Jun 11 12:38:10 <fred__tv> like gpio starts to 0 then it will bi initialized at 1 in "direction"
Jun 11 12:38:30 <zmatt> no
Jun 11 12:39:25 <fred__tv> so that is good , as they would have no influence in connected circuit
Jun 11 12:39:38 <zmatt> correct
Jun 11 12:40:19 <zmatt> so the problem is simply in how they are being initialized, i.e. he's unintentionally initializing them to output-low
Jun 11 12:40:57 <fred__tv> we are connecting to that board now....
Jun 11 12:41:01 <fred__tv> let's see
Jun 11 12:54:47 <fred__tv> perhaps mistake is to set gpio pullup after they have been declared as outputs...
Jun 11 12:55:11 <zmatt> pull-up/down is irrelevant for outputs
Jun 11 12:55:28 <zmatt> like I said, internal pull-up/down is 100 KΩ (±50%)
Jun 11 12:55:55 <fred__tv> so high ??
Jun 11 12:56:02 <fred__tv> good to know..
Jun 11 12:56:46 <fred__tv> I'm burning a 12.04 console here , just to reproduce the fault
Jun 11 12:56:59 <zmatt> OS version is not relevant, what matters is the software
Jun 11 12:59:13 <fred__tv> yes , I mean the same console image used there
Jun 11 12:59:28 <zmatt> as in, the application software
Jun 11 12:59:49 <zmatt> whatever is initializing the gpios
Jun 11 13:02:44 <zmatt> without knowing what software he's using I can't really be any more specific than that
Jun 11 13:03:58 <zmatt> all I can say, based on what I've heard so far, it very much sounds like the gpios are initially driven low simply because his software asked the kernel to drive the gpios initially low
Jun 11 13:07:17 <fred__tv> so internal resistor has no matter for outputs ?
Jun 11 13:08:06 <zmatt> internal pulls have no influence since he has external pull-ups (which are presumably stronger than 100K)
Jun 11 13:08:31 <fred__tv> probably some 4,7/10k
Jun 11 13:08:47 <zmatt> exactly, so internal pulls are irrelevant
Jun 11 13:09:39 <fred__tv> but in case of no external resistors they are relevant also for output configuration, isn't it ?
Jun 11 13:11:44 <zmatt> they could be yes, but if the initial level is this important you shouldn't rely on those very weak internal pulls anyway
Jun 11 13:44:38 <fred__tv> zmatt: the problem is hw-related
Jun 11 13:45:14 <fred__tv> or better, the problem appears when I power the beagle
Jun 11 13:45:35 <fred__tv> well before boot
Jun 11 13:46:43 <fred__tv> when I insert jack relay is activated then deactivated immediately
Jun 11 13:47:09 <fred__tv> as if gpio is 0 (or near gnd)
Jun 11 13:49:07 <fred__tv> for a moment
Jun 11 13:49:39 <zmatt> ehm, of course is it, initially the power is off
Jun 11 13:49:54 <zmatt> hence all pins are effectively ground
Jun 11 13:50:07 <zmatt> so this sounds like a power sequencing issue
Jun 11 13:50:19 <zmatt> the relays shouldn't be powered until the 3.3v supply is up
Jun 11 13:51:11 <zmatt> the external pull-ups to 3.3v are obviously also effectively pull-downs until the 3.3v supply is up
Jun 11 13:56:12 <fred__tv> the relays board is powered by beagle pin VDD_5V , so at the same time
Jun 11 13:56:35 <zmatt> that's not "at the same time", that's considerably earlier
Jun 11 13:57:51 <zmatt> vdd_5v is directly tied to the 5v input, it is always on even if the board is powered off (by software or if the pmic detects a fault)
Jun 11 13:57:52 <fred__tv> yes, they are powered up at same time (beagle and board) gpios come later...
Jun 11 13:58:26 <fred__tv> so 3.3V come later
Jun 11 13:58:39 <zmatt> I think 50-100ms later or so?
Jun 11 13:58:46 <zmatt> can't remember exactly
Jun 11 14:00:01 <fred__tv> so I'm screwed....
Jun 11 14:00:41 <fred__tv> or I should add a relay to power up relay board from 3v3
Jun 11 14:01:57 <zmatt> the power to the relay should be gated yes
Jun 11 14:02:53 <zmatt> *relay board
Jun 11 14:03:05 <fred__tv> with just a led on VDD_3V3 you can note the delay from VDD_5V
Jun 11 14:05:10 <zmatt> this is what power-up/down sequencing looks like (20ms per major gridline): https://photos.app.goo.gl/ydDCofjjDUQTVW8N8
Jun 11 14:05:47 <zmatt> this is from powered shutdown state.  when starting from no power at all, the delay is considerably longer
Jun 11 14:12:58 <zmatt> this is the sequence from no power: https://liktaanjeneus.nl/2015-04-24-053316.png (pink = vdd_5v (input power), green = sys_5v, yellow = vdd_3v3b)
Jun 11 14:17:02 <fred__tv> so mystery revealed.....
Jun 11 14:28:56 <sateesh> hai
Jun 11 14:29:10 <sateesh> am new to this
Jun 11 14:29:31 <sateesh> i need help i bought rev c bagele bone black
Jun 11 14:32:18 <veremitz> a bagel-bone, now there's an idea .. a torus-shaped board ..
Jun 11 14:34:21 <sateesh[m]> hai frnds am new to this
Jun 11 14:34:39 <sateesh[m]> i bought beagle bone black rev c board i want to work with that board
Jun 11 14:35:17 <sateesh[m]> i want to work with debian os and in linux so what i do
Jun 11 14:37:21 <zmatt> ehm
Jun 11 14:38:31 <zmatt> my recommendation is to do what you want with the board
Jun 11 14:38:55 * LetoThe2nd recommends Schnitzel.
Jun 11 14:38:55 <zmatt> since you want to work with debian linux on that board, go do that
Jun 11 14:38:58 <sateesh[m]> hello is any body there
Jun 11 14:39:29 <sateesh[m]> k my need i want to write character driver
Jun 11 14:40:39 <sateesh[m]> so previously i  placed debian os 7.5 but in that kernel is 3.8.13-bone50 but i need linux-headres i didn't find any where so what i do
Jun 11 14:40:47 <sateesh[m]> yes
Jun 11 14:43:30 <sateesh[m]> give me any suggetions which one image i used?
Jun 11 14:47:09 <zmatt> why not just use the latest?
Jun 11 14:47:26 <zmatt> debian 7.5 and kernel 3.8 are both very very old
Jun 11 14:47:57 <zmatt> just grab the latest image from https://beagleboard.org/latest-images
Jun 11 14:49:02 <sateesh[m]> i need linux-headers also so which image i choose
Jun 11 14:49:30 <zmatt> every kernel package has a corresponding headers package that you can install, if it's not already installed by default (I think it actually is)
Jun 11 14:50:22 <sateesh[m]> bone-debian-9.3-lxqt-armhf-2018-01-28-4gb.img.xz
Jun 11 14:50:26 <sateesh[m]> is it k brother?
Jun 11 14:50:51 <zmatt> do you really need an lxqt image?
Jun 11 14:50:59 <zmatt> for most purposes the iot image is preferred
Jun 11 14:51:23 <zmatt> but if you really need a graphical desktop on the beaglebone itself, then yes that image is fine too
Jun 11 14:52:01 <sateesh[m]> or else which one i choose
Jun 11 14:52:06 <zmatt> the first one
Jun 11 14:52:52 <sateesh[m]> first one i installed but it is not working with linux terminal
Jun 11 14:53:01 <zmatt> ?
Jun 11 14:54:09 <sateesh[m]> Stretch IoT (without graphical desktop) for BeagleBone and PocketBeagle via microSD card
Jun 11 14:54:09 <sateesh[m]> Debian 9.3 2018-03-05 4GB SD IoT
Jun 11 14:54:28 <zmatt> yes you don't need to copy paste that, we can all read what the first link is
Jun 11 14:54:40 <zmatt> what do you mean by "not working with linux terminal"
Jun 11 14:55:38 <sateesh[m]> by using bash  terminal if i give ifconfig
Jun 11 14:56:08 <sateesh[m]> but it won't recognize my bbb
Jun 11 15:04:59 <sateesh[m]> hello is any body there?can u understand my problem
Jun 11 15:06:03 <sateesh[m]> hai jenna
Jun 11 15:06:20 <sateesh[m]> help me my problem
Jun 11 15:07:27 <LetoThe2nd> sateesh[m]: please stop randomly spamming people, ok?
Jun 11 15:08:12 <LetoThe2nd> sateesh[m]: zmatt already tried to help you, and unless you can provide accurate information on what "is not working with linux terminal" means, there is no point in continueing.
Jun 11 15:08:37 <sateesh[m]> k
Jun 11 15:10:20 <sateesh[m]> no i already told if use bash terminal from ubuntu 18.04 if i give from terminal "ifconfig" it doesn't recognize my board
Jun 11 15:11:10 <LetoThe2nd> why should "ifconfig" on the host tell you anything about your board?
Jun 11 15:12:52 <sateesh[m]> yes i want to work with board without putty i can give 1.ifcongif  2.step sudo ifconfig  up that name 3.sudo debian@912.168.7.2
Jun 11 15:13:13 <sateesh[m]> or tell tell me any other procedure i will fallow those steps
Jun 11 15:14:13 <sateesh[m]> debian@192.168.7.2
Jun 11 15:14:30 <LetoThe2nd> that sounds totally nonsensical, to be honest.
Jun 11 15:15:22 <LetoThe2nd> why not make it work with putty first?
Jun 11 15:15:31 <LetoThe2nd> how are you connected, anyways?
Jun 11 15:15:41 <sateesh[m]> bro help me what u tell finally
Jun 11 15:15:48 <LetoThe2nd> i am not your bro.,
Jun 11 15:16:23 <sateesh[m]> in putty what i do
Jun 11 15:16:30 <sateesh[m]> tell me steps bro i will fallow those steps
Jun 11 15:16:49 <LetoThe2nd> *sigh*
Jun 11 15:16:54 <sateesh[m]> sry sir
Jun 11 15:17:48 <sateesh[m]> ok tell me procedure
Jun 11 15:18:01 <LetoThe2nd> anything should be explained here https://beagleboard.org/getting-started
Jun 11 16:04:59 <sateesh[m]> k .how to provide network beagle bone black
Jun 11 17:06:29 <zmatt> hah, wow this bbb's ethernet phy is really borked... a customer had managed to connect it to a phone line instead of an ethernet jack
Jun 11 17:07:50 <zmatt> although it still *kinda* works? at least usually link goes up correctly at boot, but if you plug it in after boot then you need to manually set the link down and up again to kick the link detection
Jun 11 17:08:52 <zmatt> i.e. make a service that does this in the background every few seconds if no ip is configured on the link -- ta-da, beaglebone "fixed" ! :D
Jun 11 17:10:03 <sateeshmamidala1> hello
Jun 11 17:10:57 <sateeshmamidala1> Booting off SD card how much time it takes bbb
Jun 11 17:11:17 <TonyTheLion> can you run a normal torvals/linux kernel on a beagle?
Jun 11 17:11:49 <TonyTheLion> or do I have to use beaglebone/linux kernels?
Jun 11 17:11:59 <zmatt> sateeshmamidala1: that's extremely heavily dependent on the system you boot, e.g. all the services and such that are enabled
Jun 11 17:12:14 <zmatt> sateeshmamidala1: it can range from mere seconds to many minutes
Jun 11 17:12:26 <zmatt> TonyTheLion: mainline kernel should boot just fine
Jun 11 17:12:37 <TonyTheLion> oh nice
Jun 11 17:12:58 <TonyTheLion> does the beaglebone one have specific beagle changes?
Jun 11 17:13:54 <zmatt> I wouldn't want to use a mainline kernel since I depend on the patches in rcn's kernels (along with some of my own), but it should boot
Jun 11 17:14:04 <sateeshmamidala1> but am booting on my board using sd card it takes more time nearly 20 minitutes is it correct?
Jun 11 17:14:43 <TonyTheLion> sdcard boot doesn't take 20 mins here
Jun 11 17:14:52 <zmatt> sateeshmamidala1: are you asking me how long it takes for your system to boot? do you not have a stopwatch?
Jun 11 17:14:53 <TonyTheLion> doesn't take much longer than a normal boot
Jun 11 17:15:09 <zmatt> booting from sd card may take longer than booting from eMMC
Jun 11 17:15:09 <sateeshmamidala1> how much time take
Jun 11 17:15:30 <zmatt> the interface to the sd card has half the bandwidth of the interface to eMMC
Jun 11 17:15:34 <sateeshmamidala1> Debian 9.3 2018-03-05 4GB SD IoT am using this
Jun 11 17:15:37 <zmatt> so reading data is potentially much slower
Jun 11 17:15:55 <zmatt> yeah the iot image takes annoyingly long to boot because it has so much crap enabled
Jun 11 17:16:01 <TonyTheLion> for me it takes a bit longer, but not 20 mins longer
Jun 11 17:16:03 <zmatt> cape-universal also slows down boot a lot I think
Jun 11 17:16:14 <TonyTheLion> oh IoT
Jun 11 17:16:14 <sateeshmamidala1> my board is rev c
Jun 11 17:16:24 <zmatt> board rev is irrelevant
Jun 11 17:16:50 <sateeshmamidala1> means how much time its continuing
Jun 11 17:17:04 <zmatt> parse error
Jun 11 17:18:17 <sateeshmamidala1> what i do tell me am confusing no one help me right path
Jun 11 17:19:37 <sateeshmamidala1> so which one i install
Jun 11 17:20:17 <zmatt> sateeshmamidala1: because your question is unclear and everything that can be said based on your vague questions has already been said.  booting from sd card takes up to twice as long as booting from eMMC, depending on how much of the boot time is reading data from eMMC/sd.  the actual boot time very heavily depends on the system you're booting, e.g. which services are enabled and such.  yes the iot ...
Jun 11 17:20:23 <zmatt> ...image probably boots fairly slow because it has a lot of services and other stuff
Jun 11 17:22:01 <zmatt> if you want faster boot, you can either try to disable unnecessary stuff in the iot image, or start with a console image and just install the things you need
Jun 11 17:22:17 <zmatt> you can find the latest console image here: https://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Stretch_Snapshot_console
Jun 11 17:22:56 <zmatt> if you're booting from sd card, you may also want to check the performance of the sd card itself
Jun 11 17:23:07 <zmatt> a very slow sd card would obviously also make booting from it very slow
Jun 11 17:24:55 <sateeshmamidala1> oh now i understood clear.but my assumption in previous i was flashed Debian 7.5 emmc it takes less than 10 miniates .so now am doing Debian image 9.3 without graphic so it takes more time it is continuing so am thinking am doing wrong
Jun 11 17:24:56 <zmatt> you can try something like:   sudo dd if=/dev/mmcblk0 of=/dev/null bs=4M count=16 iflag=direct     to test the performance of your card
Jun 11 17:25:29 <zmatt> it is possible the latest image is slower than a very old one... an unfortunate side-effect of feature-creep
Jun 11 17:25:30 <tbr> does first boot still take longer?
Jun 11 17:26:11 <zmatt> also, cape-universal (which is enabled by default nowadays) causes significant slowdown I think
Jun 11 17:26:54 <sateeshmamidala1> is i am doing wrong ?or correct way
Jun 11 18:06:19 * tbr dances the exact steps
Jun 11 18:21:13 <nobe> hi
Jun 11 18:21:38 <nobe> anyone?
Jun 11 18:22:06 <vagrantc> nobe: do you have a more specific question?
Jun 11 18:22:30 <what> hi
Jun 11 18:22:54 <Guest3005> hey
Jun 11 18:23:01 <Guest3005> hello
Jun 11 18:25:25 <hahha> hi
Jun 11 18:37:42 <zmatt> ookay
Jun 12 00:50:36 <patrick_here> Hello ...is anyone here building their own Linux images for BBB (eg: debian ...etc) ?
Jun 12 01:33:00 <zmatt> patrick_here: it seems fairly plausible to assume some people do
Jun 12 01:33:32 <zmatt> (tip: if you have an actual question, ask that instead)
**** ENDING LOGGING AT Tue Jun 12 03:00:04 2018