**** BEGIN LOGGING AT Wed Jul 29 02:59:58 2015 Jul 29 03:24:19 what the hell.... http://pastebin.com/4LxKrFqv Jul 29 03:25:24 just pulled my bb black out of storage. plugged it into my windows 8 laptop and no usb devices shows up at all. I see the pwr is lit and a couple of the usb indicators flash. Should I at least be able to see a device without drivers in device manager ? Jul 29 03:26:32 I don't recall what state I left it in. Is a sane software linux build on the device required for it to appear at all ? Jul 29 03:27:15 rcn-ee: thoughts? I've had this before, but back then I assumed I botched the kernel config. This time the only changes since my system still booted are: corrected the eeprom size in DT to match the actual size, and added an ordering dependency between two services Jul 29 03:27:22 i.e. nothing panicworthy Jul 29 03:27:46 hmm, who pulled in pvrsrvkm... I never asked for that Jul 29 03:32:01 uhhhhh, that virtual address looks like it's an access at some offset from r3 Jul 29 03:32:19 which contains 0f0e0d0c Jul 29 03:38:59 k, pvrsrvkm is not to blame Jul 29 03:39:52 lol Jul 29 03:41:38 (booting without ethernet worked, allowing me to get rid of pvrsrvkm just in case) Jul 29 03:42:07 but alas, even without it... insert ethernet, link up, boom Jul 29 03:42:20 wait, this means I can inspect journal at the same time Jul 29 03:57:51 ok, now I got a different value in r3, it's definitely trying to access r3 + 1440 with r3 being what looks like something scribbled all over the place Jul 29 04:05:43 reverting kernel build works, time to bisect Jul 29 04:10:19 hrmpz, it also depends on DT Jul 29 04:25:08 ok Jul 29 04:25:50 so now I got two DTs, one consistently crashes in the ethernet driver after link up, the other works just fine Jul 29 04:25:57 --- 1/broken.dts Jul 29 04:25:57 +++ 2/working.dts Jul 29 04:26:02 sgx@56000000 { Jul 29 04:26:04 + status = "disabled"; Jul 29 04:26:21 no other differences Jul 29 04:26:37 (and no video actually being used in any way or fashion) Jul 29 04:26:51 *graphics Jul 29 04:37:31 yup, sgx (either the peripheral itself or its driver) appears to be scribbling over ethernet data structures, nice Jul 29 05:07:17 oh how nice... adding a status = "okay"; property makes the crash disappear too Jul 29 05:08:25 as does basically any other change to the DT Jul 29 05:08:42 (though --sort does not sufficiently perturb it) Jul 29 05:19:26 oh no don't you fucking DARE suddenly vanish on your own Jul 29 05:19:46 I've seen you, you evil memory corruption bug, don't go fucking play hide and seek now Jul 29 05:30:15 * myself hands zmatt a compass, and a rucksack, and a stick that's pointed at one end Jul 29 05:31:52 s/at/on/ Jul 29 05:35:04 to no avail... the DT which previously consistently caused the crash now consistently works without any problem -.- Jul 29 05:35:47 i.e. some timing changed, and now whoever is scribbling is doing it somewhere else where no crash immediately occurs Jul 29 05:36:11 * zmatt needs to set up a honeypot or something Jul 29 05:36:21 a big chunk of memory with a watchpoint on it Jul 29 05:37:36 not that I have watchpoints available on this beaglebone, but... detail Jul 29 05:37:47 hmm, the L3 firewalls Jul 29 05:38:45 its ok, i know how it feels talking to voices nobody else can hear! Jul 29 05:38:55 * zmatt slaps LetoThe2nd Jul 29 05:39:38 LetoThe2nd: I'm not the first one to have mysterious panics with no readily identifiable cause Jul 29 05:40:30 * LetoThe2nd munches on zmatt Jul 29 05:42:49 found the line where it crashes Jul 29 05:43:28 in cpsw_tx_interrupt() in drivers/net/ethernet/ti/cpsw.c Jul 29 05:44:25 the "priv" returned by cpsw_get_slave_priv() is the garbage pointer Jul 29 05:44:33 ->txch causes the crash Jul 29 05:45:35 ehh Jul 29 05:46:24 * zmatt glares at the cpsw code Jul 29 05:50:44 "git blame" is a very well-named command in this situation Jul 29 06:11:24 * adibis wonders what you guys put the BBB through to hit these random bugs... Jul 29 06:34:42 adibis: my transgression was setting slaves = <1>; in the ethernet subsystem (and correspondingly skipping configuration of the second MII interface) Jul 29 06:34:59 given that this interface isn't used on the BBB and gives kernel warnings on boot Jul 29 06:36:05 but the bright minds that wrote the CPSW driver first allocates an array (of per-interface data) based on that property Jul 29 06:36:54 but then in the irq handler just blindly uses indices 0 and 1 Jul 29 06:40:56 asshats Jul 29 06:48:33 but if it's not used - why would it generate irq? Jul 29 07:08:40 it doesn't generate one, they just blindly do something like: process( slave[0] ); process( slave[1] ); return IRQ_HANDLED; Jul 29 07:09:07 I don't even know yet why or how the distinction is made (unless running in dual mac mode) Jul 29 07:10:13 normally you don't have to give a shit from which interface the packet came Jul 29 07:10:17 zmatt: don't know, if it is related, but have you looked at these patch series: https://www.mail-archive.com/netdev@vger.kernel.org/msg71824.html Jul 29 07:11:09 yegorich: heh, yes the cpsw driver is really badly written Jul 29 07:11:37 it doesn't need to be patched, it needs to be tossed and rewritten. this time also separating the host ethernet interface from the switch code Jul 29 07:11:48 especially since the switch can remain operational across reboots Jul 29 07:13:46 zmatt: do you have patches in the queue? Jul 29 07:14:32 I've written my own driver code once for a baremetal application... as for the mess in linux and u-boot I don't even know where to begin Jul 29 07:15:16 I tried to slightly fix the u-boot version once, but I got stuck since the managed to get the switch code and host ethernet code intensely interwoven Jul 29 07:15:43 zmatt: OK Jul 29 07:15:56 *they managed Jul 29 07:16:46 my implementation just initialized the switch during early boot if necessary (without resetting the PHYs unless they seemed to be wonky) Jul 29 07:16:51 zmatt: have you already contacted Mugunthan V N? Jul 29 07:17:15 and much later the host ethernet / cpdma code can initialize if the host wants to do networking Jul 29 07:17:26 it can also reinitialize without affecting the switch Jul 29 07:18:01 I only just discovered this Jul 29 07:18:35 and I'm still staring at the cpsw code Jul 29 07:18:45 zmatt: :-) Jul 29 07:29:26 I just started reading about writing device drivers (ldd3) - interesting book. Jul 29 07:30:06 yegorich: it also annoys the hell out of me that the PHY on power up negotiates link up, then gets reset, negotiates link up, gets reset again, negotiates link up... :P Jul 29 07:31:45 and while that's just an annoyance on the BBB that delays final link up (it's currently a bottleneck in startup time) and risks confusing the link partner, resetting the PHYs will be absolutely unacceptable for the x15 Jul 29 07:32:00 since you'll cut two network segments down the middle Jul 29 07:33:43 zmatt: then it is really important, that you report all this stuff on the netdev mailing list. At least for future reference. Jul 29 07:39:08 I'm not currently on the netdev list, I'm not familiar with linux kernel internals, and writing such emails always costs me a _lot_ of time, time I don't really have Jul 29 07:40:31 and I've noticed in the past that merely pointing out glaring flaws isn't always well-received Jul 29 07:40:45 your 09:13 < yegorich> zmatt: do you have patches in the queue? is a nice example of that Jul 29 07:43:06 I was just hunting down a crash, I fortunately found it Jul 29 07:44:05 I'm gonna stare a bit more at the cpsw code mostly to see what the best approach is for configuring DT when one port is unused Jul 29 07:46:38 other than that, I mostly try very hard to ignore mediocre code as long as it works sufficiently well Jul 29 07:47:37 if I'm gonna try my hand at fixing a driver it'll be omap2-mcspi Jul 29 07:49:54 (or I just add it to the ever-growing list of peripherals which I simply use directly, bypassing the kernel) Jul 29 10:19:51 Dear All, do you know where can buy the ViFFF-024 camera board for BeagleBoard XM? I cannot from it from the supplier. Jul 29 13:42:35 Hello, I kindly need an information related to BeagleBone Black: what is the maximum level of humidity at which it can be installed? Jul 29 13:46:46 isaacdevice: I can't speak for the project, but almost every electronic device specifies 10-90%, non-condensing. Jul 29 13:48:01 isaacdevice: if you need a confirmed spec, ask the manufacturer of your concrete implementation. Jul 29 13:48:57 Generally if you're asking this question it means you're about to do something stupid, and I hope you understand that a development board isn't intended or warranted for use in a finished product. Get your own designers to do that. (And your own designers will make a board that doesn't have 92 pins floating out in space to gather moisture, etc) Jul 29 13:50:30 well, not necessary stupid. but in fact there are at least two manufacturers of the BBB, and i don't know if and under which terms they are willing to explicitly specify maximum environmental levels. best bet is still looking into the SRM. Jul 29 13:51:07 http://elinux.org/Beagleboard:BeagleBoneBlack#Terms_of_Use "We do not encourage the use of the board that we manufacture under the BeagleBoard.org logo in commercial products." Jul 29 13:52:49 I simply need that information just because I must include it into an user manual Jul 29 13:53:34 I never meant to try some kind of experiment Jul 29 13:55:11 yeah, so as i said: look into the SRM Jul 29 13:56:39 I already tried to find this information all around the universe, but unsuccessfully Jul 29 13:58:16 hello Jul 29 13:58:22 I'm falling back on "if you have to tell someone this information, it's because the device is being used outside its design intent". If we knew what product this was going into I'd call my bookie right now... Jul 29 13:58:47 My beaglebone driver software for windows 64 bit is failing to install every time. Jul 29 13:58:49 isaacdevice: that maybe means that nobody is willing to guarantee any ratings at the moment. Jul 29 13:58:58 I dont know what to do now Jul 29 14:00:00 alok_ windows 8 ? Jul 29 14:00:12 LetoThe2nd: OK, clear, thank you! Jul 29 14:00:37 yes, by the way, just upgraded to windows 10. Still facing the same problem Jul 29 14:01:02 for windows 8 you have to temporarily turn off driver signing verification. one sec I'll find a link Jul 29 14:01:08 or, you pay extra for the guarantee, because that comes with additional support. The cheap-as-possible (cheaper than the individual components except in huge quantities) board is made for hobbyists who can do their own support Jul 29 14:01:32 ok... Thanks Jul 29 14:01:51 https://learn.sparkfun.com/tutorials/disabling-driver-signature-on-windows-8/disabling-signed-driver-enforcement-on-windows-8 Jul 29 14:02:13 Thank you so much sabo, I will try it now. Jul 29 14:02:36 Do that to install them. You don't need to keep it enabled after the install.(Next reboot puts windows back in normal mode and the drivers will still work) Jul 29 14:02:38 np Jul 29 14:03:21 Thanks Sabo!! Jul 29 14:05:07 myself: you may have bought your kindness at a very special supermarket... if you are not able to answer a simple customer question, you shouldn't have been so talkative. Thank you anyway! Jul 29 14:05:35 isaacdevice: And if you wanted commercial help, you shoul pay for it. Jul 29 14:06:46 So you are just able to talk about money... i thought it was a technical information, not a discount request! Jul 29 14:06:56 isaacdevice: YOu get what you get when you ignore the warnings from the manufacturer of the product. Jul 29 14:07:09 myself: isaacdevice agmlego: lets just end the topic here, ok? Jul 29 14:07:42 isaacdevice: Everything you have said so far indicates commercial use of the device, which is explicitly not kosher. Then, you ask about things that only matter to commercial uses, in a free volunteer support channel. Jul 29 14:07:49 isaacdevice: Excuse us for taking umbrage at that. Jul 29 14:08:12 LetoThe2nd: Yup, I am done. isaacdevice is now yet another entry in the waste-of-time blacklist. Jul 29 14:12:31 mostly just a bunch of end-users here helping eachother Jul 29 16:38:14 hello i ve been trying to configure my debian system with a static ip using ifconfig and route commands but the system couldn't keep the static ip that i configured. My question is how can i configure correctly the static ip for debian or for another ubuntu distro? Jul 29 16:38:43 im using the beaglebone black with debian 2014 on it Jul 29 16:38:56 i hoppe you'll help me pleas Jul 29 16:40:21 sometimes setting things up in /etc/network/interfaces helps. Jul 29 16:42:28 when i modified interfaces file works fine but after reboot sometimes get the dynamic ip and sometimes the static ip Jul 29 17:28:53 Hi, anyone there use beagleboard for antminer? Jul 29 17:29:58 I try install the antminer system using HDD Raw Copy Tool, and my computer turn off witouth finish Jul 29 18:48:33 AYUDA! Jul 29 18:48:36 HELP! Jul 29 18:52:50 BALTA00: Silence in this channel means one of three things: 1) you failed to ask at a time people were watching, and you need to be more patient 2) you failed to ask a question or 3) you failed to add sufficient information about the issue to catch someone's interest and attention. Jul 29 18:53:02 Randomly shouting will not help fix any of these three. Jul 29 18:55:36 not familiar with antminer, which computer randomly turned off? Jul 29 19:02:19 or, why is antminer even relevant here? Sounds like your PC doing the image-writing had a problem, perhaps a bad driver, perhaps a bad card, and didn't handle the error well. Try again with a different card. Jul 29 19:02:54 or the write went to the wrong place ~.~ Jul 29 19:05:22 ahh, could be. I've blown up enough things with hasty dd commandlines... Jul 29 19:15:38 command lines are generally not a good place for haste Jul 29 19:17:14 and that's why I'm so glad the Chrysocome port of 'dd' has a little idiotproofing -- rename it to dd-removable.exe and it restricts itself to only being able to write to removable devices. Jul 29 19:18:00 that, like reversible power connectors, has saved me from a number of mishaps since it came into my toolbox :) Jul 29 19:22:10 I liked a patch I saw for ddrescue, which prompted you for a confirmation if the target was mounted. Jul 29 19:41:56 hah Jul 29 19:42:49 I had the weird problem of the framebuffer being shifted by one pixel: the rightmost pixel would end up at the start of the next line Jul 29 19:43:18 manually flipped the pixel clock polarity -- problem gone Jul 29 19:43:22 Heh. Jul 29 19:45:11 so... it would seem that although both the am335x and the HDMI frame support configuring the pixel clock polarity, they haven't been configured right? Jul 29 19:46:28 (either that or there's something silly going on like something configuring the pixel clock polarity between the am335x and the hdmi framer based on EDID data or whatever) Jul 29 19:48:38 also, looks like we should be able to support 1920x1080 .. perhaps at a reduced frame rate Jul 29 19:49:57 depends a bit on how tolerant the screen is I guess Jul 29 20:02:38 What can i do if I delete the beagleboard partition? Jul 29 20:08:28 beagleboard or beaglebone? Jul 29 20:08:52 (in all cases, fix it, but the details differ and I have no experience with beagleboards) Jul 29 20:11:46 if you only delete the partition, then recreate it in the exact same place, the filesystem may actually be intact Jul 29 20:27:48 beagleboard Jul 29 20:30:47 When I try conect it to my computer for USB, I think it is restarted, because it sounds as if you connect and disconnect Jul 29 20:32:17 I use "HDD Raw Copy Tool" for try edit the partition, and my computer turn off (I don't know how), and don't finish the partition creation. Jul 29 22:42:37 Recently, i've been using debian distribution and i 've had several problems with the configuration for wired network settings. I tried to configure my host with a static ip and afterwards configure a dhcp ip through bash script. Besides, configure the static ip for eth0 interface(set the appropriate parameters such as ip, net mask, gateway, dns-nameservers in the file /etc/network/interfaces) as follows : auto eth0 iface e Jul 29 22:46:28 really bad that beagle cannot handle more than 1,8v in analogic Jul 29 22:46:46 will the new beagle have higher voltage input on analogic? Jul 29 23:04:43 why's that? Jul 29 23:04:56 you got a high voltage signal that is impedence sensitive? Jul 29 23:14:13 You could use ye olde voltage divider as ds2 implied. Jul 29 23:15:03 However you may want to buffer the output with something with a very low offset like OPA333 or LMP2102. Jul 29 23:15:34 if it is a large signal, why bother with low offset? Jul 29 23:15:52 lm324 should be fine for a large enough signal Jul 29 23:17:28 Well depends before or after the voltage divider I suppose. The input impedance to the A2D is an issue too. An LM324 is just a quad LM741/ua741 without the null offset circuit. Jul 29 23:18:30 The OPA333 is a choperstablized device so it has a low offset and almost no drift. The LM2102 is faster but with just a low offset and a higher drift. All depends I guess. Jul 29 23:18:51 not quite Jul 29 23:18:58 LM324 is happy with unipolar supplies Jul 29 23:19:09 LM741 does not like it unless you play games with dividers, etc Jul 29 23:19:23 but all this is moot w/o knowing more about the signal Jul 29 23:22:30 indeed lacking information is a cruel reality in this case :D **** ENDING LOGGING AT Thu Jul 30 02:59:58 2015