**** BEGIN LOGGING AT Thu Mar 30 03:00:02 2017 Mar 30 04:00:02 odd Mar 30 04:00:11 oh, too impatient Mar 30 06:15:29 rishabh_bb: stick around next time! Mar 30 07:57:04 gm Mar 30 07:58:41 zmatt: looking into the cke settings now Mar 30 07:59:51 zmatt: I can't see where linux is doing that. IIRC, I should gate and ground cke until I do cancelation of self refresh, no? Mar 30 08:00:26 zmatt: otherwise it might happen that restoring the EMIF context, something is waking up the SDRAM before clocks are back up Mar 30 08:04:33 hi Mar 30 08:04:38 sup guest Mar 30 08:13:28 Hy guys, my name is Marco and I'm a student. I would like to install embedded linux on the beaglebone bone but I don't know how. Could you help me Mar 30 08:20:11 sooo Mar 30 08:20:13 it comes pre installed Mar 30 08:20:48 and if you want to install a newer version, the instructions for using an SD card are pretty well documents on the werld wibe weeb Mar 30 08:38:47 hi I am trying to send remote At command to xbee from beaglebone but it is transmitted as normal tx frame Mar 30 08:39:13 I am not getting hoe to send remote AT command Mar 30 08:54:57 any idea about it??? Mar 30 08:58:30 r u scoping Mar 30 08:58:38 is it UART Mar 30 08:59:16 ssj Mar 30 09:00:32 if you're scoping and you can see that the BBB is not sending out uart you probably need to turn the pin from GPIO mode to UART mode using the program pin-config or config-pin i always forget what order Mar 30 09:00:43 but if you type in the program it'll give you a big old help menu that shows you how to switch pin modes Mar 30 09:01:21 and that being said it helps to have the latest image and `git pull` RCN's updates to the kernel and whatever else he's got Mar 30 09:01:37 you can find all of his updatable repositories by searching for the .git folders, tbh Mar 30 09:01:42 `find...` Mar 30 09:01:49 no actually I am able to use uart for sending data in normal 64 bit addr mode Mar 30 09:02:10 i dont know what normal 64 bit addr mode is in uart sorry Mar 30 09:02:12 but not remote AT command Mar 30 09:02:38 are you using ttyS{0,1,2,3,4}? Mar 30 09:02:44 sorry,it is related with xbee frame options Mar 30 09:02:56 I can read/write on uart Mar 30 09:03:00 but is this an xbee problem or a bbb problem? Mar 30 09:04:21 xbee ,i think :( Mar 30 09:04:33 soo #xbee Mar 30 09:04:46 ok Mar 30 09:39:50 thinkfat_: well dgerlach implied that that's not needed because emif initializes the sdram config register from the value configured (by u-boot) in the control module Mar 30 09:41:13 it wouldn't hurt to do it anyway Mar 30 09:41:58 does the a8 or the m3 reenable emif in prcm? Mar 30 10:03:38 zmatt: the A8 saves and restores the config and also flips the switch on the module. The M3 only does some additional powersaving stuff Mar 30 10:04:14 zmatt: M3 also handles the PLLs, so it can be assumed that the DDR PLL is locked stably before the A8 is woken up. Mar 30 10:08:27 zmatt: earlier sources show that the A8 was originally responsible for all that, but recently it seems that more and more of the functionality was moved to the M3 Mar 30 11:14:27 supposedly there's this single cycle FGPIO address range but when i "str r0, =0xFPGIO" it scopes the same as the regular GPIO through the AHB-Lite bus Mar 30 11:14:30 :-/ Mar 30 11:14:35 str is usually 2 cycles Mar 30 11:16:54 dgerlach: hm, the am33xx_do_wfi() from your github repo completely removes the "wfi_flags" stuff, but it this now also used in the cpu idle path? That would cause a lot of additional latency, no? Mar 30 11:17:54 https://github.com/dgerlach/linux-pm/blob/upstream/v4.11/amx3-suspend-dev/arch/arm/mach-omap2/sleep33xx.S Mar 30 11:27:39 is there a version of https://github.com/beagleboard/am335x_pru_package for PRUSSv2? My BBG only has the v2 kernel modules Mar 30 11:31:35 there seems to be a mismatch between the supplied dtbo files, which include uio_pruss_enable-00A0 and the kernel modules which don't Mar 30 12:23:39 hi there, can I somehow load a DTB fragment/overlay on vanilla/mainline kernel? (I'm on 4.9.18) Mar 30 12:24:23 I'm playing with building my own images via buildroot, and stuff works, but I'm having troubles enabling additional UARTs and I2C buses Mar 30 12:43:14 jkt: just customize your dts for your application Mar 30 12:44:14 that's how it's normally done, not with overlays Mar 30 12:44:59 Simon-: https://github.com/beagleboard/am335x_pru_package is for prussv2 Mar 30 12:45:13 prussv2 refers to the pruss version that's on the am335x, not a kernel driver Mar 30 12:45:50 and afaik both uio_pruss and remoteproc_uio are still supported Mar 30 12:46:16 ayjay_t: on what mcu? Mar 30 13:00:06 zmatt, sorry I could not respond to you yesterday on abe dpll stuck in transition Mar 30 13:00:32 zmatt, what I can find in my mailbox is as follows: if abe dpll on omap5 was'nt locked at boot, it fails to transition.. Mar 30 13:00:54 zmatt, there has been hacks around this in the past.. http://review.omapzoom.org/#/c/31450/ etc.. Mar 30 13:03:12 zmatt, OMAP5_DPLL_ABE_DEFFREQ in drivers/clk/ti/clk-54xx.c in kernel attempted to be locked in omap5xxx_dt_clk_init i think attempts to do this.. Mar 30 13:05:27 zmatt, i think i810 erratum explains.. I think.. Mar 30 13:07:32 zmatt, hope that helps.... /me drops off Mar 30 13:14:05 zmatt: yeah, I was naively hoping that simple #include "BB-UART1.dts" in the relevant .dts file would do the trick, but not so much :) Mar 30 13:14:39 zmatt: also, I think I would actually make use of the dynamic things, like having the same DTS for two variants of the add-on HW that we use, etc Mar 30 14:38:54 NishanthMenon: ah, i810 very much sounds like this problem indeed Mar 30 14:43:14 so maybe mainline linux lacks the appropriate workarounds in case of the abe pll, I'll nee to check Mar 30 15:26:12 I am searching for a small debian image for my beagleBone Green. Mar 30 15:43:14 on https://beagleboard.org/latest-images is a typo in "Debian 7.11 2015-06-15 4GB SD LXDE". Should be 2016 instead of 2015. Mar 30 17:05:43 txt-file: the console images are quite small Mar 30 17:05:57 http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_console Mar 30 17:06:32 zmatt: will look into them Mar 30 17:17:38 thinkfat: hi, sorry, yes the wfi_flags are not present in that repo Mar 30 17:17:47 thinkfat: just getting deepsleep0 merged first Mar 30 17:17:52 will add cpuidle later Mar 30 18:46:41 Hi Mar 30 18:46:46 Hello Mar 30 18:47:49 . Mar 30 18:52:55 t.me/beagleirc bridge for this IRC created in telegram. Currently running in my local machine.. Will try to put in some free servers. The telegram users can use this. Mar 30 18:56:33 please don't, bridges are really annoying Mar 30 20:08:43 haha Mar 30 20:09:11 jesh_ do you know what the eternal september is? Mar 30 20:10:18 zmatt i'm working on an nxp m0+ <_< >_> but they're m0+ std features Mar 30 20:21:58 well, the single-cycle I/O port is optional for the m0+, and iirc its location is implementation-defined Mar 30 20:23:15 well, thats true, but it is implemented Mar 30 20:23:19 and i did find it Mar 30 20:24:16 how are you testing it exactly? Mar 30 20:24:35 generating a square wave with gpios and measuring the frequency Mar 30 20:24:50 using both the regular gpio and fgpio on different compiles Mar 30 20:25:00 check the disassembly Mar 30 20:25:11 i wrote the assembly Mar 30 20:25:24 ah since you said "on different compiles" Mar 30 20:25:31 yeahh sorry, whats that stage called then Mar 30 20:25:35 from assembly to byteecode Mar 30 20:27:00 also assembly I think, or assembling maybe? I don't often need to refer to it really :P Mar 30 20:27:38 but what do you mean by "different" then? I mean, there's only one way to assemble something Mar 30 20:28:41 oh i mean, i'll change the address from 0x400ff000 which si the normal gpio to 0xf8000000 which is the fpgio Mar 30 20:28:46 and program it Mar 30 20:28:47 with each Mar 30 20:32:25 hm, if you put a few stores in a row to toggle a gpio you ought to be able to see the difference as far as I understand Mar 30 20:32:36 yeah thats what i've been trying Mar 30 20:32:54 ill give it anotehr shot and then request support if it doesn't work Mar 30 20:33:25 i chose a uC a little too small for this project and i dont want to redesign hardware so i'm trying to make this guy work Mar 30 20:33:35 assembly is fun tho Mar 30 20:36:57 Hi, is it possible to add x15 support in Openwrt/LEDE project? Mar 30 20:38:56 milani_: sure it is possible. We "just" need somebody who does the work. ;-) Mar 30 20:39:23 +1 to that ^ :] Mar 30 20:39:36 As lede uses kernel 4.4, I guess it is not hard for you guys ;) Mar 30 20:40:07 meh … there is not even support for beagleBone Mar 30 20:41:20 ayjay_t: it helps that the m0+ instruction set is really simple and tidy Mar 30 20:42:12 yeah i'm totally into it Mar 30 20:42:20 txt-file: Under TI OMAP3/AM33xx, I can see EBV Beagleboard profile. Mar 30 20:42:45 omap3/beagleboard is not am335x/beaglebone Mar 30 20:43:49 my information: https://wiki.openwrt.org/toh/texas.instruments/beagleblack Mar 30 20:44:48 and https://lede-project.org/toh/views/toh_extended_all?dataflt[Model*~]=beagle Mar 30 20:44:57 am335x is not an omap, although it is a relative thereof... the wiki page is wrong about needing a separate boot partition Mar 30 20:46:06 it's a wiki … you are allowed to change it Mar 30 20:46:16 a prerequisite would be me caring Mar 30 20:46:27 true Mar 30 20:47:12 you are allowed to change it even if you do not care ;-) Mar 30 20:48:08 it might be interesting to note btw that the ethernet port itself is gigabit, it's just the phy that limiting it to 100mbit Mar 30 20:49:21 Is it possible to change it? Mar 30 20:49:35 at least one person has managed that, hehe Mar 30 20:49:59 and there's a poorly available derivative, the BeagleBone Enhanced, that has a gigabit phy Mar 30 20:51:55 (the second ethernet port (also gbit) of the am335x is not accessible, it conflicts with other stuff) Mar 30 21:01:51 I only wonder why they used mini-USB … otherwise it look reasonable Mar 30 21:02:29 a usb cable is included anyway Mar 30 21:13:09 txt-file: I'm not a pro in CPUs or things related to it, is this the same instruction set as AM5728? Mar 30 21:13:29 https://lede-project.org/docs/instructionset/arm_cortex-a15_neon-vfpv4 Mar 30 21:16:11 AM5728 is Cortex-A15 Mar 30 21:24:25 but I am unable to find information if AM5728 has NEON and/or VFPv4 Mar 30 21:28:24 txt-file: wanted to know if we can reuse the already included instruction set for x15. I am reading https://lede-project.org/docs/guide-developer/adding_new_device Mar 30 21:29:10 txt-file: it has: http://www.ti.com/product/AM5728 Mar 30 21:30:13 "Programmability is provided by dual-core ARM Cortex-A15 RISC CPUs with Neon™ extension" Mar 30 21:30:56 milani__, instruction set should not be the problem in LEDE Mar 30 21:31:12 worst case is that you run without NEON and VFP Mar 30 21:32:02 without neon is no problem but without vfp makes a big difference Mar 30 21:32:12 NEON and VFPv4 are a single unit Mar 30 21:33:14 txt-file: it has both Mar 30 21:33:31 i thought neon is one and vfp is another part of the cpu Mar 30 21:34:52 so adding x15 to lede should not be a big deal. should get a board and give it a try... unless I miss something ;) Mar 30 21:37:11 you might be able to find information about partition layout and such in buildroot Mar 30 21:37:32 https://git.busybox.net/buildroot/tree/configs/beagleboardx15_defconfig?h=2017.02.x Mar 30 21:37:52 https://git.busybox.net/buildroot/tree/board/beagleboardx15?h=2017.02.x Mar 30 21:42:22 txt-file: it is not on the cortex-a15 Mar 30 21:42:34 good to know Mar 30 21:43:34 "For area sensitive designs, the Cortex-A15 offers the option to remove the sizable Neon-VFP unit for embedded applications." from https://www.arm.com/products/processors/cortex-a/cortex-a15.php Mar 30 21:44:02 BTW: Is there a kvm driver for ARM, yet? Mar 30 21:45:01 That sentence made me wonder if beagleBoardX15 has both or not. Mar 30 21:46:10 the ti datasheet says it has neon so it has both Mar 30 21:46:16 https://en.wikipedia.org/wiki/ARM_Cortex-A15 talks about being neon and vfp being "onboard" Mar 30 21:46:16 [WIKIPEDIA] ARM Cortex-A15 | "The ARM Cortex-A15 MPCore is a 32-bit processor core licensed by ARM Holdings implementing the ARMv7-A architecture. It is a multicore processor with out-of-order superscalar pipeline running at up to 2.5 GHz...." Mar 30 21:47:32 Theo Debrouwere here? Mar 30 21:50:19 txt-file: kvm support for the cortex-a15 was added in 3.9, I never used it though Mar 30 21:50:26 zmatt: I downgraded to a different kernel that had the uio_pruss module and it's working now Mar 30 21:50:59 Now I test if 4.9.x is able to boot from emmc. Otherwise I need to user 4.4.ti Mar 30 21:51:17 jannau: awesome Mar 30 21:53:23 yeah … booting BeagleBoneGreen from emmc. Now I have a BBG running debian stretch. Mar 30 22:17:57 Simon-: uhh, you really don't need to downgrade, all the latest kernels support it Mar 30 22:21:54 ok, it seems you're right 4.4.55-ti-r94 only has pruss-remoteproc Mar 30 22:22:00 but 4.9.13-ti-r26 has uio-pruss and pruss-remoteproc Mar 30 22:22:26 and the latest -bone kernels (4.4.58-bone17, 4.9.19-bone4, 4.10.7-bone1) all have uio-pruss _only_ Mar 30 22:26:22 it seems it's only absent from very recent 4.4-ti kernels, strangely Mar 30 22:27:47 txt-file: cortex-a15 can be configured without vfp/neon, with vfp but without neon, or with vfp/neon. the bbx15 most certainly does have them Mar 30 22:28:08 zmatt: I used the upgrade script and it downgraded the kernel version Mar 30 22:28:20 to? Mar 30 22:28:44 zmatt, This is what I thought but didn't find any information which parts the AM5728 has. Mar 30 22:28:47 4.1.38-bone24 Mar 30 22:28:57 Simon-: that's really old Mar 30 22:29:19 Simon-: all -bone kernels have uio_pruss Mar 30 22:29:36 the previous one was a -ti kernel Mar 30 22:29:56 Simon-: like I said, the latest 4.4-ti kernels seem to lack uio_pruss for some reason Mar 30 22:30:03 while the 4.9-ti kernels do have it Mar 30 22:30:12 4.4.9-ti-r25 Mar 30 22:30:20 ok Mar 30 22:30:28 all -bone kernels have it too Mar 30 22:30:39 (I usually prefer those myself) Mar 30 22:32:47 --lts-kernel|--lts|--lts-4_1-kernel|--lts-4_1) kernel="LTS41" ;; Mar 30 22:32:49 txt-file: then you grab the manual ( http://www.ti.com/lit/ug/spruhz6h/spruhz6h.pdf ), go to chapter 4 ("Dual Cortex-A15"), and the first block diagram there already shows neon and vfp :) Mar 30 22:33:00 (from https://github.com/RobertCNelson/boot-scripts) Mar 30 22:33:32 Simon-: he probably forgot to update it I guess, I've never used an update-script myself Mar 30 22:33:47 4.1 hasn't been the default for quite a while I think Mar 30 22:37:09 zmatt: reminds me of hitchhikers guide to the galaxy → https://www.youtube.com/watch?v=L4q6qBuSjHk&feature=youtu.be&t=197 Mar 30 22:42:53 hehe Mar 31 02:53:47 Hello Everyone **** ENDING LOGGING AT Fri Mar 31 03:00:01 2017