**** BEGIN LOGGING AT Sun Jan 10 02:59:59 2016 Jan 10 03:05:27 I imagine starting up those RF systems was ... interesting. Jan 10 03:06:56 I don't think there's much "starting up" beyond making sure noone is in a hazardous place before you enable the transmitter Jan 10 03:07:08 they're not engines Jan 10 03:08:26 * GenTooMan grins "yes those 100kw tubes could go to 30mhz 70mhz if you downed the power a bit. Yeah the filament current was 200amps..." Jan 10 03:10:04 I think they probably had to "warm up" before you could 'strike' them .. and such. Jan 10 03:13:07 well the shear scale of those things is just amazing. Jan 10 03:13:30 my Dad was workin gon MW transmitters until a couple of years ago Jan 10 03:13:38 as in mega-watt Jan 10 03:13:47 you just have to take extra precautions lol Jan 10 03:14:18 don't stand in front of the transmission dish for one. Jan 10 03:14:33 that's not so bad when its 00s of metres up :D Jan 10 03:16:16 veremit so on megawatt transmitters are they using klystron tubes? Jan 10 03:17:08 yea someties Jan 10 03:19:43 so mostly still vacuum tube type technology (just on a big scale). Not surprising however because it's easier to make large tubes than transistors. Jan 10 03:26:24 yes for the analogue stuff .. don't think there's a lot of 'high power' digital here .. Jan 10 03:33:07 the array of which I pasted images at least did FSK... they actually had some electronically controlled thingy that could slightly influence the tuning coil in tandem with the frequency shifts to ensure it would be correctly tuned at both frequencies Jan 10 03:33:34 zmatt.. thats pretty mental! Jan 10 03:34:23 yeah pretty nifty, especially considering it's a pretty old array Jan 10 03:35:16 "saturable core reactor" Jan 10 03:37:09 http://www.navy-radio.com/commsta/cutler.htm is the main page btw Jan 10 03:37:38 hola alguien esta despierto Jan 10 04:10:41 if data cache is disabled, what happens if a write to ddr3 sdram is too quick? Jan 10 04:10:55 do you get an exc_dabort? Jan 10 04:11:49 i should say, what happens if sequential writes to ddr3 sdram is too quick? Jan 10 04:11:54 are too quick.. Jan 10 04:11:58 (sheesh) Jan 10 04:14:22 this would be in a simple, tight assembly loop doing something like "str r2, [r0, r1]; add r1, #-1; bne loop" Jan 10 04:21:51 yates: cpu stalls Jan 10 04:24:14 that loop will however cause a data abort if the memory type is device or strongly ordered (including the case of the MMU being disabled) since 3 out of 4 writes are misaligned Jan 10 04:24:55 oh, yeah. sloppy. Jan 10 04:25:04 also, I hope you meant "adds" instead of "add" Jan 10 04:25:12 yes. Jan 10 04:25:18 gotta have my flags Jan 10 04:26:05 which register controls the memory type? Jan 10 04:26:31 or is it a coprocessor? Jan 10 04:26:55 is that in the am335x trm? or the arm architecture reference manual? or where? Jan 10 04:27:17 your section translation table, which is a 16 KB table (one word per 1 MB of memory), the base address of which is configured into a cp15 register Jan 10 04:27:30 cool. Jan 10 04:27:43 I have an example of making such a table somewhere... Jan 10 04:27:58 https://community.arm.com/docs/DOC-10098 Jan 10 04:28:23 danka. Jan 10 04:29:22 so i found the gel script that does ddr3 tests in ccs and ran it. Jan 10 04:29:28 everything passed. Jan 10 04:30:02 does that make this a pointless exercise? Jan 10 04:30:18 means you can also run the proggie to determine optimal leveling Jan 10 04:30:56 huh? the "proggie"? you mean the gel file? Jan 10 04:31:11 http://processors.wiki.ti.com/index.php/AM335x_DDR_PHY_register_configuration_for_DDR3_using_Software_Leveling Jan 10 04:33:48 so why are you interested in doing an assembly version? Jan 10 04:34:07 of the initialization/test, that is? Jan 10 04:34:23 I'm not, I use C++ Jan 10 04:35:03 you mean a baremetal c++ program? Jan 10 04:35:23 this is the only asm file in my own baremetal codebase: file:///home/dev/dd/barebone/start.S.html Jan 10 04:35:28 the rest is C++ Jan 10 04:35:38 ehh Jan 10 04:35:50 http://gerbil.xs4all.nl/start.S.html Jan 10 04:35:55 sorry bout that :P Jan 10 04:36:29 what i'm trying to get at is why you are bothering with any baremetal program at all? why not just "trust" the linux / u-boot to initialize things? Jan 10 04:37:20 ok, so maybe you patch u-boot to use your leveling parameters, but still... Jan 10 04:38:00 I use my basemetal codebase currently mostly for small tests of how some peripheral or other aspect of the SoC behaves, which may be difficult to do when an OS is running Jan 10 04:38:23 also, to back up a bit, you feel the gel tests thoroughly exercise the ddr3? Jan 10 04:38:39 also, I reuse the same headers for doing I/O from userspace in linux Jan 10 04:38:48 i see. Jan 10 04:39:35 and once all the pieces are there I intend to replace u-boot by my own bootloader, but it doesn't have high priority Jan 10 04:40:28 do you feel the gel tests thoroughly exercise the ddr3? Jan 10 04:40:45 no idea, I've never looked at them, but it sounds unlikely Jan 10 04:41:04 since unavoidably they will consist of single transactions with huge delays between them Jan 10 04:41:49 i see. ok, thanks man. Jan 10 04:41:58 sure helps to have someone to bounce these questions off of. Jan 10 04:42:58 you're welcome Jan 10 04:44:47 note that with MMU disabled you'll unavoidably get single transactions, and even if enabled it would be difficult to flood the memory controller if L2 cache is disabled, though it may be possible using Neon stores Jan 10 04:47:36 so is that going to be a good test? what's the alternative? enabling data L2 cache? Jan 10 04:47:45 i mean L2 cache? Jan 10 04:48:30 the goal of the test is to verify the layout is good. Jan 10 04:51:19 I'm not sure whether stress-testing is necessary, although it could be helpful. but you can just use the leveling program to get the optimal timing parameters for your board (as explained on the wiki page) and fill those into u-boot Jan 10 04:51:51 you can then run a memory test from u-boot (which does MMU/cache initialization for you) Jan 10 04:53:39 to genuinely confirm adequate signal integrity you'd also need to test in all process/voltage/temperature corners Jan 10 04:54:35 (which seems annoying as hell to set up to me) Jan 10 04:55:35 how do you run the memory test from u-boot? is it a flag you pass in Uenv.txt or somesuch? Jan 10 04:55:59 i.e., as part of booting linux? Jan 10 04:56:13 so if the board meets the layout guidelines specified in the datasheet and there's no obvious problem, I'd assume all is fine Jan 10 04:56:17 eh Jan 10 04:56:24 I think u-boot has commands for it Jan 10 04:56:36 (which you can enter in its shell) Jan 10 04:57:05 ah. Jan 10 09:30:48 Hey! everyone. I'm a EE undergraduate looking forward to join contribute to beagleboard during the upcoming GSoC. Where should I start? Thanks Jan 10 09:32:00 http://elinux.org/BeagleBoard/GSoC Jan 10 09:33:48 Thank you! anything else I need to know? Also what kind of skill set is expected from the applicant? Jan 10 09:34:13 not sure besides what's on there Jan 10 09:35:26 hey guys, I have connected my beaglebone black to my pc and use minicom for the stuff I want do in my bbBlack. an Ubuntu is installed in BBB. My problem is that minicom shows nothing to me. Is the problem the image I inserted in BBBlack SD or is it minicom problem. the only cable I use is the USB which power BBB and connect it to PC. Jan 10 09:35:46 torpico .. be sure to turn OFF flow control :) Jan 10 09:35:51 always a first gotcha Jan 10 10:36:23 veremit: how? may you explain? and is the USB TTL Serial Cable necessary? Jan 10 10:40:12 usb-ttl will give you a console no-matter-what if the beagle is powered Jan 10 10:40:40 usb-mini-B client will give you mass-storage/disk and networking and console when its connected Jan 10 11:16:29 veremit: so TTL is not a necessary stuff. But the usb cable (which gives power too) do not show me any console when the bbb is powered on. Jan 10 11:16:43 do you see stuff on the ttl cable? Jan 10 11:16:57 veremit: I do not have that cable Jan 10 11:18:01 veremit: my host OS (my PC) is fedora. and it show no sign of bbb connected Jan 10 11:19:44 torpico .. I would suspect you don't have usb-gadget drivers available Jan 10 11:19:53 do you see anything in 'dmesg' when the beagle connects? Jan 10 11:20:11 veremit: NO Jan 10 11:20:43 hmm thats not good .. and ou see flashing LEDs on the beagle? Jan 10 11:21:07 USR0 and USR1 are light on Jan 10 11:21:22 yours is abeaglebone-black yes? Jan 10 11:21:32 they are on without blinking Jan 10 11:21:38 yes bbBlack Jan 10 11:22:02 you shuold have a heartbeat patter on the end led .. power led steady (by the dc jack) and flickering on one other Jan 10 11:23:12 most likely stuck in u-boot or something Jan 10 11:23:13 I think perhaps you should download a fresh image to a micro-SD card if that pattern doesn't come up Jan 10 11:23:25 torpico: do you hold down S2 while you power up the board? Jan 10 11:24:06 the boot button you mean? Jan 10 11:24:12 yes Jan 10 11:24:17 yes Jan 10 11:26:39 I changes the sd card now.USR0 is blinking, USR2 is blinking too with a different pace. USR1 blink very very slowly. Jan 10 11:27:01 but again grep tty shows nothing Jan 10 11:27:02 sounds like something is booting Jan 10 11:27:20 it'll take a good 2/4 minutes to show Anything at all Jan 10 11:27:38 because you need to wait for ttyACM0 to come up Jan 10 11:27:48 so you won't grep anythin cos the node won't exist :) Jan 10 11:28:37 good. then I will see that on minicom, yes? Jan 10 11:29:10 or it will give me a terminal by itself? Jan 10 11:29:35 no Jan 10 11:29:53 you will have to watch for the node appearing , and then connect to it via minicom if you wish Jan 10 11:30:15 you can probably use a command like 'watch -n5 ls /dev/ttyACM*' Jan 10 11:36:12 veremit: thanks for guiding me through connecting minicom, Now can I ask you about Ubuntu images? Jan 10 11:36:46 I'm not very familiar with ubuntu .. I'm usin debian in all my 'arms' Jan 10 11:36:52 or gentoo for my desktop Jan 10 11:38:06 ok, I ask maybe you could help me. in http://elinux.org/Beagleboard:Ubuntu_On_BeagleBone_Black they said download the image from here: http://elinux.org/BeagleBoardUbuntu then burne that image, put in in bbb, press the boot button. power bbb, now the ubuntu is installed (via sd card) but Jan 10 11:42:27 in the other page which is http://elinux.org/BeagleBoardUbuntu it is very complicated. Which image I should download? I downloaded the prebuild image for bbb and burned it to my SD card .I did none of the commands they wrote there. should I do the structures they wrote in that page Jan 10 11:43:08 oh and I should say again grep tty shows nothing Jan 10 11:44:33 torpico: I can only apologise the documentation is at best inconsistent, and often out-of-date Jan 10 11:48:21 torpico: let me see what I can digest .. Jan 10 11:50:22 torpico: I would start with .. http://elinux.org/BeagleBoardUbuntu#eMMC:_BeagleBone_Black.2FGreen Jan 10 11:51:44 Notes: Jan 10 11:51:44 Jan 10 11:51:44 If only two LED's stay lit and nothing happens, the board has crashed due to lack of power. Retry with a 5Volt DC power supply connected. Jan 10 11:52:00 This script will only take about 5-6 Minutes after power on. Jan 10 12:02:23 veremit: excuse me if I ask foolish questions, should I download prebuild image and then burn it on SDCard or should I only run the script? Jan 10 12:05:41 download the image .. dont' worry about the script :) that's just RN gettin keen lol Jan 10 12:50:27 veremit: I burned the image in 2 different sdCards, but every time I boot it up it and 4 LED's blink a constant pattern which means the eMMC write has failed. where I made mistake?!!! Jan 10 12:50:48 did you check the downloaded image? Jan 10 12:51:06 yes. that was ok Jan 10 12:59:00 are you sure that this means flashing has failed? Jan 10 12:59:18 tbr: documentation seems to suggest that Jan 10 12:59:23 and without a debug cable .. we're lost. Jan 10 12:59:46 tbr: Inthe page I sent it has written: If the 4 LED's blink a constant pattern, the eMMC write has failed. First REMOVE ALL capes, then retry again. Jan 10 12:59:49 and the image doesn't load a g_serial module? Jan 10 13:00:09 the ubuntu apparently has the usb ethernet .. Jan 10 13:00:24 torpico: which revision is your BBB? Jan 10 13:00:25 I would imagine if they're RN's images, they have the full composite gadget Jan 10 13:21:41 tbr: Rev B6 Jan 10 13:24:41 veremit: veremit: I would imagine if they're RN's images, they have the full composite gadget what it means : p Jan 10 13:25:03 torpico .. that was mroe intended for tbr Jan 10 13:30:48 I'm going to start having a look at the TDA19988 framer, and I'm in doubt as to whether or not I can simply copy parts of the code from the Linux driver and reuse it in u-boot? Jan 10 13:31:03 eliasbakken: yeah, why not?! Jan 10 13:31:12 Is it OK as long as the copyright holder is intact? Jan 10 13:31:37 oo now we're gettin into details lol Jan 10 13:31:48 I haven't a clue Jan 10 13:32:24 I saw that there also exist a Minix driver, and they mentioned that it was OK to study the Linux driver, but not copy code from it. Jan 10 13:33:22 This is a bit beyond my open source knowledge, but I do not want to step on Rob Claks toes, since he is the author of the Linux driver. Jan 10 13:34:21 It's probably OK since he works for TI and it is in the interest of TI to have this driver ported to u-boot. Jan 10 13:47:58 you may find you can't port the whole driver because of other dependencies Jan 10 13:52:06 Yeah, just looking for the minimal viable solution. No EDID for now. Jan 10 13:52:18 Or audio or anything... Jan 10 13:59:25 framebuffer/uboot imageload fails? Jan 10 14:03:42 torpico: then make sure the images are suitable for a 2GB emmc! Jan 10 14:04:22 eliasbakken: what license is minix under? Jan 10 14:04:37 eliasbakken: both u-boot and linux are under gplv2 IIRC Jan 10 14:06:02 tbr: I'm not sure about Minix, but it seems it is not compatibel with Linux: https://github.com/tcort/minix-i2c/tree/master/drivers/tda19988 Jan 10 14:06:50 As is always the case, Jan 10 14:06:50 great care has to be taken to only study the code to learn the Jan 10 14:06:51 functionality of the chip and not reproduce/copy the code. Jan 10 14:07:02 eliasbakken: https://github.com/tcort/minix-i2c/blob/master/LICENSE Jan 10 14:07:12 it's one of the BSD licenses Jan 10 14:07:44 and it's correct, you can't move GPL code into BSD code Jan 10 14:08:57 OK, but it's not a problem, I'll just use the Linux driver as a reference. Jan 10 14:18:03 eliasbakken: moving from the linux kernel to u-boot should be ok Jan 10 14:26:31 hello, I have a problem. My board is the BeagleBone A6a. Yesterday I had loaded on the SD an debian image (found on the site). Everything worked well. Today I uploaded the same image on the sd, but the board has stopped working. The power LED is always on, while the user LEDs do not light . What can I do? Jan 10 14:27:13 lucaItaly: how do you power the board? Jan 10 14:29:53 I tried both usb computer and charger 5v 2A Jan 10 14:31:52 veremit: tbr it seems I flashed the ubuntu image into the bbb correctly. I removed the sdcard and once again powered on the board without sdcard. USR1 and USR2 are blinking but dmesg | grep tty shows the same thng it alwayes show mw. there is no sign of BBB flash connection to my PC. Jan 10 14:32:50 torpico: hook it up to ethernet and check your dhcp server if it gets an IPv4 address Jan 10 14:33:01 or just check the ipv6 neighbors Jan 10 14:33:37 tbr: what? how? what is the meaning of the things you said :p Jan 10 14:34:15 no need to insult me about your own lack of knowledge... Jan 10 14:35:36 tbr: oh, don't miss understand me, I did not mean to insult you, :P I am really newbie and also a little fool ;p Jan 10 14:35:51 *misunderstand Jan 10 14:36:18 then stop fucking using a smilie with a stuck out tongue all the fucking time Jan 10 14:36:28 that's fucking insulting Jan 10 14:37:06 tbr: ok, don't get angry. Jan 10 14:39:21 dmesg | grep tty should work whenever we connect a USB device, right? when I connect my own USB it again shows me this: Jan 10 14:39:24 dmesg | grep tty Jan 10 14:39:25 [ 0.000000] console [tty0] enabled Jan 10 14:39:25 [ 0.644207] 00:05: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A Jan 10 14:39:55 exactly like the time no USB is connected. Jan 10 14:43:19 have you checked HDMI output? there might be a fbconsole Jan 10 14:44:14 tbr: it seems for HDMI connection I should have a micro HDMI port which I do not have unfortunately Jan 10 14:44:54 well, do you have a LAN? aka a router with Ethernet ports? Jan 10 14:46:15 yes Jan 10 14:47:59 so? what should I do with it? Jan 10 14:52:08 hook it up to the BBB and boot the BBB Jan 10 14:52:37 then check in your router if the BBB has requested an IP address. If your router has that option Jan 10 15:11:25 On BBB, is the bootloader in the eMMC read first? So that adding an SD card will not override the bootloader? Jan 10 15:13:22 Hold on, reading the boot chart now... Jan 10 15:19:38 eliasbakken: default boot order is eMMC and then some other things like UART and USB. if you hold S2 then it's microSD ... UART.. USB Jan 10 15:19:55 for details see TRM/SRM Jan 10 15:20:15 Yeah, i read the fucking manual, and figured it out. Jan 10 15:20:31 :) Jan 10 15:21:16 older eMMC factory images had U-Boot configured in a way that it would load from eMMC, but then if there was an microSD it would continue booting from that Jan 10 15:21:35 this started failing quickly, as it was not an intended way of working anyway Jan 10 15:22:05 so if a BBB stopps booting when a microSD is inserted → update U-boot, or the whole eMMC image Jan 10 16:07:55 tbr: thanks for your helps. have as nice day Jan 10 17:04:26 tbr: eMMC,μSD,uart,usb vs μSD,spi,usb,uart Jan 10 17:04:39 sorry, Jan 10 17:04:52 spi,μSD,usb,uart Jan 10 17:08:26 zmatt: right, just couldn't be arsed to look it up at that time. ;-) Jan 10 17:11:22 hehe, yeah I'm close to having repeated it enough to just know it... it's mostly the usb/uart order I forget, although I know they're swapped for the two Jan 10 17:11:53 :D Jan 10 17:38:48 tbr Is this a problem with the Uboot shiped with Rev B boards? Jan 10 17:39:21 I think all boards? the fix is pretty recent Jan 10 17:40:07 installing just a new u-boot isn't sufficient though, you also need to put an uuid= line in your /boot/uEnv.txt Jan 10 17:40:47 yes it's a u-boot/uEnv combination Jan 10 17:41:39 Ok I have 3 rev B boards so I guess I will need to fix all 3 as I intended to use test kernels in the SD slot and "normal" boot in the eMMC. Jan 10 17:42:44 GenTooMan: you can ofc abuse this behaviour Jan 10 18:14:48 Hello, I am trying to get a TPLink TL-WN725N V1 to work on a beaglebone Rev A6A. Currently I have a 3.8.13-bone41 Kernel running. My first approach was to use a precompiled 8188eu.ko + firmware. I got the driver loaded, but no interface showed up. So I tried to compile the driver myself. Same result. Then I wanted to update to a new kernel and I used a mainline kernel with the bb.org_defconfig as it contains the driverin the staging f Jan 10 18:16:15 but the kernel does not load, It stops at starting kernel .... So I tried the kernel from robert c martin (bb-kernel) with the same result. Any suggestion what would be the best approach to get the WN725n wifi running on the beaglebone ? Jan 10 18:17:41 well not from robert c martin but from robert nelson :) Jan 10 18:30:31 gardi: are you sure that device has a 8188eu chipset? Jan 10 18:34:56 I used it also in a RPI and there I also used the rtl8188eu driver. Jan 10 18:37:04 did you check dmesg output? did lsusb show it? Jan 10 18:38:23 Yes lsusb showed it and also dmesg showed that the driver was loaded (but it did not say anything about the firmware nor that it failed nor that it loaded it) Jan 10 18:39:34 sounds like something is wrong then Jan 10 18:40:16 have you tried one of the official debian images with a mainline kernel prebuilt? Jan 10 18:40:27 does the stick need firmware? Jan 10 18:41:46 I did not try an official debian image as I have got some work on the sd card and no other card available. Yes, the stick needs firmware Jan 10 18:51:14 root@beaglebone:~# lsusb Bus 001 Device 002: ID 0bda:8179 Realtek Semiconductor Corp. Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Jan 10 18:53:38 [ 3.314979] usb 1-1: default language 0x0409 [ 3.315179] usb 1-1: udev 2, busnum 1, minor = 1 [ 3.315202] usb 1-1: New USB device found, idVendor=0bda, idProduct=8179 [ 3.322332] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 [ 3.329866] usb 1-1: Product: 802.11n NIC [ 3.334109] usb 1-1: Manufacturer: Realtek [ 3.338442] usb 1-1: SerialNumber: 00E04C0001 [ 3.350325] usb 1-1: usb_probe_device Jan 10 18:54:32 ation #1 chosen from 1 choice [ 3.350596] usb 1-1: adding 1-1:1.0 (config #1, interface 0) [ 3.351409] hub 1-0:1.0: state 7 ports 1 chg 0000 evt 0002 [ 3.351459] hub 1-0:1.0: port 1 enable change, status 00000503 [ 3.436729] udevd[91]: starting version 175 [ 5.213978] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null) [ 6.829182] udevd[316]: starting version 175 [ 8.595919] r8188eu 1-1:1. Jan 10 18:55:06 6011] r8188eu 1-1:1.0: usb_probe_interface - got id [ 8.604628] Chip Version Info: CHIP_8188E_Normal_Chip_TSMC_D_CUT_1T1R_RomVer(0) [ 8.790201] EEPROM ID = 0x8129 [ 8.824554] usbcore: registered new interface driver r8188eu Jan 10 18:55:32 well it is quite hard to post logs on irc? Jan 10 18:56:51 This is the output of the 3.8.13-bone41 kernel and the problem is that I do not have an interface in ifconfig (no wlan0) Jan 10 18:57:04 just lo and usb0 Jan 10 19:23:14 Ok, so I've implemented a minimal TDA19988 driver in u-boot. It appears as though there is data coming out of the HDMI plug, which is good, but no test-pattern on the screen yet. Jan 10 19:29:46 does the bbb require some closed source software/binary blobs to normaly run? Jan 10 19:31:25 Looking at the pixel clock, it appears as though the clock coming from the BBB is too low. I must have some problem with the LCD controller perhaps... Jan 10 19:31:57 floss_lover: Only for the SGX engine. Jan 10 19:32:30 eliasbakken: and this stands for what usage? GPU? Jan 10 19:32:41 Yes Jan 10 19:33:17 eliasbakken: so i dont need it when not using hdmi or when vesa is fine? Jan 10 19:34:17 only for using the 3D engine. Jan 10 19:34:24 2D is fine. Jan 10 19:36:23 is this like the binary blob on amd/radeon cards when not installing the firmware files? its just slow, thats all? Jan 10 19:38:58 other question: when i install a 2gb image on the 4gb memory, it boots just fine, right? Jan 10 19:46:07 It should. Jan 10 19:47:06 Not sure what Radeon does when the firmware is missing. Jan 10 19:47:40 eliasbakken: it write in the dmesg that it cant find and the firmware and just works fine but not that fast Jan 10 19:48:47 I don't think the SGX complains about that during boot. It's perhaps the PRU? Jan 10 19:49:00 What is the message? Jan 10 19:49:22 did i understand right, that i can just dd any bbb image that is been for a sd-card and that is smaller then 4gb into the internal eMMC and it just boots fine? Jan 10 19:50:43 It should just use the first 2GB worth of sectors, as far as I know. Jan 10 19:52:25 why are there so many people warning about writing something to eMMC and so on? It does not really matter because a bbb is not brickable thanks to the button to boot from sd-card. or am i missing something? Jan 10 19:54:37 http://ahsoftware.de/Beaglebone_Black_Boot_explained.svg Jan 10 19:58:13 i have some questions about the BeagleBoard-X15. Does it have PWM? How many? WHen will it be released? Jan 10 19:58:47 eliasbakken: thanks. i already had this in my head. So when i boot from external and then dd /dev/zero to the eMMC i can just copy u-boot to sd-card and boot again without any problems :) Jan 10 20:32:09 Are there some eMMC wearing out problems on the bbb? Jan 10 20:39:33 haven't heard of any Jan 10 20:40:00 ofc if you write continuously to it and stress it, it will break eventually Jan 10 20:51:19 tbr: ok. as i tought. thanks for the information. i dont have to worry then when Jan 10 21:02:50 Ok, so I'm back to that pixel clock... From the TRF, the LCD_PCLK is derived from LCD_CLK which is 200MHz? Jan 10 21:03:19 So If I want a 25 MHz clock, i should use a divider of 8. Jan 10 21:04:18 Ah... Hold on, the MAX frequency of that clock is 200MHz... Jan 10 21:42:02 Ok, I have a big fat 25 MHz pixel clock on the BBB now, but the recovered clock on the screen is still not right... I think there must be some more settings on the TDA19988 missing, right? Jan 10 23:40:09 you want a boot splash via HDMI? Jan 10 23:42:54 eliasbakken: I thought you were just targeting an lcd cape? hdmi is a whoooole different story Jan 10 23:43:01 eliasbakken: also that boot chart is both confusing and wrong Jan 10 23:43:27 Well, I'm pretty close I think... Jan 10 23:44:21 eliasbakken: hdmi requires significant interaction with the nearly-undocumented tda19988 Jan 10 23:44:53 among other things to fix the sync signals of the lcdc into something that's actually VESA-compliant Jan 10 23:45:19 Hm... Jan 10 23:45:27 That might be what I'm seeing. Jan 10 23:45:29 plus you'd technically need to read EDID information to configure your timings Jan 10 23:45:44 though I guess you can try a standard "safe" mode and hope for the best Jan 10 23:45:46 Yeah, that's not really an issue. Jan 10 23:46:14 I only need this working for one particular display! Jan 10 23:47:15 but not the lcd cape you showed in previous pics? Jan 10 23:47:47 (which would not use the hdmi framer at all) Jan 11 00:35:05 bonescript vs octalbonescript, opinions? Jan 11 00:37:00 * zmatt prefers /dev/mem Jan 11 00:50:18 you use /dev/mem with the frame buffer? Jan 11 00:50:43 I was replying semi-jokingly to ayjay Jan 11 00:51:41 (in the sense that I do use /dev/mem or uio for things like GPIO, ADC, PWM) Jan 11 01:22:18 The ADC is 12bits correct? Jan 11 01:30:01 10 iirc Jan 11 01:30:08 but I may be wrong Jan 11 01:30:41 beaglebone? 12 I think Jan 11 01:31:27 ah it is 12-bit Jan 11 01:31:29 my bad Jan 11 01:34:47 though I think you run into the noise floor before that gets relevant (maybe driving a low-impedance signal into the adc will help with that, but then you'll have fun avoiding the powerup erratum) Jan 11 01:42:03 or rather, it might be 10-bits with 2 bits of noise :) Jan 11 01:42:31 I've heard someone talk about ARM Cortex-M MCUs from STM32 et al being like that... 9 bits of signal and 3 bits of noise Jan 11 01:42:49 same guy said the MSP430 has a more legit ADC Jan 11 01:42:54 but I'm no expert on those things... Jan 11 01:43:08 I don't hold the stm32 in particularly high esteem myself Jan 11 01:43:11 but they're cheap :P Jan 11 01:43:26 yeah I've never touched them, only done stuff with the TI Tiva-C so far ARM-wise Jan 11 01:43:32 rather expensive but eh Jan 11 01:43:46 I'm not making 100 of some gadget either Jan 11 01:44:11 yeah the Tiva Snowflake looked interesting, but never played with one Jan 11 01:46:05 ah yeah I have one of the early boards Jan 11 01:46:16 that thing's too freaking big for me to figure out wtf to do with it Jan 11 01:46:30 blizzard is closer to my use-case right now Jan 11 01:47:07 I heard the Tiva Snowflake (TM4C129) series was originally invented with the use-case of "big server baseboard management controller" in mind Jan 11 01:47:07 also notable for actually documenting the crypto accelerators, several of which are essentially identical to the ones in omap4 and later SoCs such as the am335x (but undocumented in all of those) Jan 11 01:47:23 I can imagine them for that Jan 11 01:47:32 oh, huh, yeah. that's on the newer ones I take it? I think they just came out with the crypto version of their TM4C129E launchpad Jan 11 01:47:35 crypto accelerators, ethernet, adc for sensor stuff Jan 11 01:47:40 yep Jan 11 01:47:53 it's on the Snowflake Jan 11 01:48:07 a bazillion I2C buses probably for your multitude of SMBus enabled drives and other stuff Jan 11 01:48:34 well that is cool I guess, bout time someone documented it eh :P Jan 11 01:48:58 documentation is incomplete and not always right, but hey Jan 11 01:49:04 yeah Jan 11 01:49:44 I hope its i2c peripheral is then better than the one on their big SoCs Jan 11 01:50:05 I'm not sure, haven't played with it yet. I've heard some folks on 43oh gripe about it before Jan 11 01:50:28 the MSP430 USCI I2C implementation is a little squirrely but I managed to beat it into submission Jan 11 01:50:54 omap-i2c is buggy as shit Jan 11 01:51:19 fortunately it has good company... stm32's i2c is buggy as shit, the i2c peripheral on that broadcom thing on the rpi is buggy as shit :P Jan 11 01:52:00 fucking up i2c controllers seems to be a popular hobby of μC/SoC makers Jan 11 01:54:11 wtf 10 i2c peripherals? holy shit you weren't kidding Jan 11 01:54:59 yeah Jan 11 01:55:02 Tiva's like that Jan 11 01:55:09 sort've like Renesas with their MCUs Jan 11 01:55:17 having like 12 I2C and 6 SPI Jan 11 01:55:39 ...and most packages for their chips not having enough pins able to take the alternate function settings for them all Jan 11 01:55:51 and not having a switchable pin architecture like NXP's ARM MCUs Jan 11 01:56:28 ick, they lack good pinmuxing? Jan 11 01:56:39 that's one of the things I probably hate the most about stm32 Jan 11 01:57:01 yeah each peripheral has only 1 or 2 or maybe 3 pins where each of its functions can go Jan 11 01:57:07 usually 2 Jan 11 01:57:23 but some CAN peripherals on the Tiva blizzard (TM4C123) might only have 1 spot for its TX or RX Jan 11 01:57:41 well, actually this is on the PM package, 64-LQFP Jan 11 01:57:52 Tiva TM4C123 can go up to a BGA or LQFP-144 package I think Jan 11 01:58:04 so you might find more alternate locations for things with that Jan 11 01:58:27 C129 is 128-TQFP or 212-BGA Jan 11 01:58:41 yeah Jan 11 02:00:30 stm32 is particularly annoying since you can only move all peripheral pins together as a single block rather than having per-pin muxing, and the combining is done in some really weird way that often results in a conflict if two peripherals are enabled with an overlapping pin, even if you have no need for the corresponding signal on one of the two Jan 11 02:03:05 oh, now that's hideous Jan 11 02:03:09 special ed hideous Jan 11 02:03:17 never heard of that kind of treachery Jan 11 02:05:09 also, there's almost no runtime mapping possible of dma request to dma channel... want to use dma for UART3 Tx and SPI1 Rx at the same time? too bad, you can't, they use the same channel Jan 11 02:07:13 also, nearly all timers have different features Jan 11 02:07:42 all these things combined mean it's complete hell to figure out whether it's actually possible to use some combination of features Jan 11 02:13:13 * zmatt notes to all that a new revision of the am335x TRM (spruh73m) is available for download Jan 11 02:17:54 lol, nicely done TI Jan 11 02:19:11 "The reg_refresh_rate value at reset is config_refresh_def_val port value. [..] The user must tie off the config_refresh_def_val port with a correct value to meet ..." Jan 11 02:19:42 that's an internal tie-off port you asswipes, you copy-pasted this from your own internal chip integration documentation Jan 11 02:19:54 lol **** ENDING LOGGING AT Mon Jan 11 02:59:58 2016