| 00:30.48 | ZaEarl | naobsd, I'm trying to get your recovery working on my tablet. https://github.com/naobsd/cm_bootable_recovery Vol+ doesn't work. keytest shows it sends keycode 115. What file do I edit to make that change? |
| 00:31.12 | Turl | oh, he has a repo :P |
| 00:32.30 | Turl | doesn't look like it's changed majorly though ZaEarl, wonder why does mine not work |
| 00:41.50 | CIA-91 | rhombus-tech: lkcl master * r9db59aa0a6c3 /mailing_list_policy.mdwn: add mailing list policy |
| 00:45.06 | naobsd | ZaEarl: at first, that repo is just a placeholder to cherry-pick my changes. currently it should be synced with latest CWM but sometimes it may not. when updating branch is rebased so I don't recommend to clone it. |
| 00:46.39 | naobsd | ZaEarl: and I put recovery_keys.c for custom key map and add reference to BoardConfig.mk under device tree like as device/allwinner/a10/. |
| 00:47.16 | naobsd | well, "add reference to it into BoardConfig.mk", sorry |
| 00:52.12 | naobsd | ZaEarl: btw it's CWM by koush. my change is very small. no big difference |
| 00:56.11 | Turl | naobsd: yep we got it working now, thanks :) |
| 00:58.00 | naobsd | Turl: well, is "he has a repo", about me? |
| 00:58.34 | Turl | naobsd: yeah, I didn't know you had a repo :) |
| 01:00.55 | *** join/#arm-netbook sysfwlab (~sysfwlab@eta91-1-82-227-11-158.fbx.proxad.net) |
| 01:01.47 | naobsd | from the beginning, I'm offering everything to build same image. links to my repos are provided on my site/XDA thread |
| 01:03.13 | Turl | naobsd: the problem is, I wasn't looking at your thread, ZaEarl handed me your image as a working one to have a look |
| 01:03.38 | sysfwlab | Hello everybody |
| 01:03.56 | naobsd | Quarx's repo should be better than mine. https://github.com/Quarx2k |
| 01:04.44 | naobsd | lkcl: may I ask about ML policy here? |
| 01:05.36 | Turl | naobsd: yep I'm working with him, we're setting up a common android repo for allwinner devices, to inherit from |
| 01:06.42 | sysfwlab | Sorry to disture, hope i can ask this, I want to know if it's common to clock an UART at 24 MHz (a 8250/16550 clone) ? |
| 01:13.30 | naobsd | lkck: sorry, don't mind |
| 01:13.35 | naobsd | oops |
| 01:13.45 | naobsd | lkcl: sorry, don't mind |
| 01:14.39 | naobsd | I'm new to ML and here. I'm not sure about policies. |
| 01:14.59 | naobsd | I'm not sure what should I do before say something on ML/here. |
| 01:15.34 | hno | naobsd, just say. |
| 01:15.44 | naobsd | in general, I like wiki, but sometime I may not have enough time to edit wiki |
| 01:17.55 | naobsd | I think this IRC channel is same as ML. someone may post useful information here, but it will be lost very easily |
| 01:23.38 | naobsd | now, question about Quarx's is arrived from ML. (I have no answer but for example) if I'm thinking to answer it, I need to start browser to edit wiki, add information to wiki, then I send e-mail "look here". |
| 01:24.19 | naobsd | I'm sure wiki page will be helpful for everyone include me. I just say "look here" when someone asked same question in future. |
| 01:25.33 | naobsd | but I may not have enough time to edit wiki, then I may think "I don't answer this time..." |
| 01:26.07 | naobsd | is it impossible to ask someone to make wiki page after posting information to ML? |
| 01:28.26 | naobsd | lately, I'm busy :( I had no time to make any update to my A10 project for several weeks. I made answer for some question on my XDA thread. I felt my answer should be summarized like as FAQ, but I had no time. no one said "I'll make FAQ page from your posts". |
| 01:29.08 | naobsd | I'm almost alone on my project but I can see a lot of people here/ML |
| 01:29.55 | naobsd | I shouldn't expect someone who can edit wiki page? |
| 01:31.21 | naobsd | of course I follow policy regardless of what I'm feeling |
| 01:32.55 | naobsd | sorry for stupid message :( |
| 01:34.28 | naobsd | again: I like wiki and similar system to centralize information. I said "please read my page" a lot on my XDA thread :) |
| 01:46.31 | naobsd | really sorry. I should not say anything like that. I'm new. I have no contribution to here yet. |
| 03:28.51 | mikey_w | I don't see any images for android that boots from an SD card like Ubuntu does? |
| 03:29.54 | lundman | android_4.0.img |
| 03:30.07 | lundman | although it boots SD to reprogram nvram |
| 03:34.49 | mikey_w | Where is the image? |
| 03:35.16 | mikey_w | I would prefer leaving the nand alone. |
| 03:35.31 | mikey_w | Pull the card and hae a fallback position. |
| 03:35.35 | mikey_w | have |
| 03:35.53 | lundman | there is that |
| 03:36.19 | mikey_w | Have many cards with different OSes. |
| 03:36.32 | lundman | the beta 4.0 image is fairly easy to find, but you have to use a soft to prep the sd card |
| 03:36.56 | lundman | that would be best yeah, I think booting sd card isnt particularly hard |
| 03:37.18 | mikey_w | The prep software only runs in windoze. |
| 03:37.40 | lundman | yep |
| 03:37.53 | lundman | ran in VM windows though :) |
| 03:38.26 | lundman | http://lundman.net/wiki/index.php/MeLe_A2000 |
| 03:38.39 | lundman | will point you there.. dunno if the channel has a wiki |
| 03:39.36 | mikey_w | I prefer DD. |
| 03:39.43 | mikey_w | No magic. |
| 03:41.40 | lundman | dd dont work so well with different sized cards, well, long term |
| 03:42.21 | lundman | I like my ZFS setup on cubox, since you can create snapshots and fs very easy, so at uboot I can pick the OS on just one storage |
| 03:43.40 | mikey_w | I dd to a large card the do gparted to resize the partitions to full up the card. |
| 03:48.21 | mikey_w | I want a nice standard uboot in the MBR, vfat partition with the bin flies various uimages for the kernels and various root files systems for the various OSes. |
| 03:48.59 | mikey_w | None of the overwriting of the internal nand. |
| 03:49.29 | mikey_w | Always done the same way. |
| 03:56.18 | mikey_w | The bin files not being human readable is also a unix abortion. |
| 03:57.35 | mikey_w | Later, good night. |
| 04:48.28 | *** join/#arm-netbook A2Sheds (~ly@unaffiliated/l84supper) |
| 05:34.02 | *** join/#arm-netbook marcan (marcan@marcansoft.com) |
| 06:08.22 | *** join/#arm-netbook serishema (~serishema@203.160.121.206) |
| 06:25.14 | *** join/#arm-netbook gimli (~gimli@178.73.212.231) |
| 06:25.14 | *** join/#arm-netbook gimli (~gimli@xbmc/staff/gimli) |
| 06:31.17 | *** join/#arm-netbook rell01 (~chatzilla@p4FE562F6.dip0.t-ipconnect.de) |
| 06:53.50 | *** join/#arm-netbook techman2 (~Glen@121.209.128.126) |
| 08:00.03 | *** join/#arm-netbook frantisheq (~frantishe@kapusany.3log.sk) |
| 08:01.15 | CIA-91 | rhombus-tech: nil master * rc576e696e586 /mailing_list_policy/discussion.mdwn: |
| 08:18.34 | hno | lundman, this channel do have a wiki. http://rhombus-tech.net/ |
| 08:21.51 | *** join/#arm-netbook arete (~arete@2-226-90-244.ip180.fastwebnet.it) |
| 08:29.03 | *** join/#arm-netbook arete_ (~arete@net-188-152-188-139.cust.dsl.teletu.it) |
| 08:42.17 | mnemoc | meh... I wrote address instead of added in that reply :< |
| 08:54.20 | lundman | hno: ah sounded like someones company :() |
| 08:55.46 | mnemoc | lundman: it is, lkcl's |
| 08:56.12 | zenitraM | mnemoc: mali drm module works, but seems like it's not enough for X11 |
| 08:56.14 | zenitraM | http://pastie.org/3934516 |
| 08:56.40 | mnemoc | zenitraM: I'm making a branch with what got mailed. wip/allwinner-v3.0-android-v2/mali |
| 08:57.38 | zenitraM | nice. i'm going to look on the igloo tree to see if there are GPL sources for the b2r2_blt missing device |
| 09:01.24 | mnemoc | wip/allwinner-v3.0-android-v2/mali pushed |
| 09:04.11 | mnemoc | zenitraM: can you file a ticket with that problem? |
| 09:06.52 | lundman | whats the deal? we are working for them? just doing hobby dev for them? they happy with hacking? |
| 09:08.15 | mnemoc | "they" started the project which broght us here, and that wiki ended up collecting all a10-related info |
| 09:13.35 | zenitraM | ticket filed |
| 09:17.22 | mnemoc | \o/ |
| 09:19.59 | zenitraM | mnemoc: stupid git question: how do i toggle to the mali branch? |
| 09:21.53 | zenitraM | <PROTECTED> |
| 09:21.54 | zenitraM | error: unable to resolve reference refs/heads/wip/allwinner-v3.0-android-v2/mali: No es un directorio |
| 09:22.07 | mnemoc | first fetch it |
| 09:22.13 | mnemoc | `git fetch origin` |
| 09:22.37 | mnemoc | then branch it, `git checkout -b wip/allwinner-v3.0-android-v2/mali origin/wip/allwinner-v3.0-android-v2/mali` |
| 09:23.27 | mnemoc | (git is really ugly without some basic aliases like `git co`) |
| 09:23.54 | zenitraM | still getting the same error |
| 09:24.26 | mnemoc | paste me the whole thing |
| 09:25.01 | mnemoc | also `git remote -v` and `git branch -r` |
| 09:25.34 | zenitraM | http://pastie.org/3934594 |
| 09:26.06 | mnemoc | may you have a local branch just called "wip" > |
| 09:26.07 | mnemoc | ? |
| 09:26.27 | zenitraM | okay, solved it |
| 09:26.33 | mnemoc | what was it? |
| 09:26.57 | *** join/#arm-netbook arete_ (~arete@net-188-152-188-139.cust.dsl.teletu.it) |
| 09:27.03 | zenitraM | .git/logs/refs/heads/wip/allwinner-v3.0-android-v2 was a file |
| 09:27.09 | zenitraM | rm'ed it and made it a directory |
| 09:27.29 | mnemoc | .oO |
| 09:28.03 | mnemoc | btw, you may want to `git remote prune origin` to delete references to already gone branches |
| 09:28.46 | zenitraM | done, bye utf-8 branch |
| 09:29.18 | mnemoc | :) |
| 09:35.04 | mnemoc | zenitraM: don't we need something like modules/mali/DX910-SW-99002-r2p2-03rel0/driver too? |
| 09:35.36 | zenitraM | i'm not really sure |
| 09:36.13 | zenitraM | at least Xorg hasn't said anything about the absence of it yet |
| 09:36.36 | mnemoc | but it said something about a missing dev |
| 09:36.50 | mnemoc | libblt_hwCould not open device /dev/b2r2_blt |
| 09:36.51 | mnemoc | (EE) MALI(0): XVScreenInit failed |
| 09:36.59 | *** join/#arm-netbook naobsd (7c294571@gateway/web/freenode/ip.124.41.69.113) |
| 09:37.33 | naobsd | does anyone has connection to Quarx? I want to transfer my XDA thread to him, but I can't get any reply from him |
| 09:38.27 | zenitraM | but that device isn't anywhere on those mali sources |
| 09:38.54 | zenitraM | the only place i've seen it is on the snowball kernel tree |
| 09:39.55 | zenitraM | mmm but it's not on the x11 module from mali |
| 09:40.53 | mnemoc | we are surely missing another driver |
| 09:41.59 | zenitraM | not really sure, seems b2r2_blt is needed by the xserver-xorg-video-mali package on snowball, but that package is different from the opensource drivers published by arm |
| 09:42.30 | mnemoc | so, let's get the open source drivers working first :) |
| 09:42.50 | zenitraM | i'm on it :) |
| 09:43.26 | zenitraM | too lazy to crosscompile xorg so recompiling kernel with sata enabled so i don't burn my sd |
| 09:48.44 | naobsd | I have interest to small box with sata like as CuBox :) |
| 09:50.17 | lundman | i have one |
| 09:52.27 | frantisheq | anybody with goodix_touch_3F.ko for steev's 3.0.8 based novo 7 elf ubuntu? |
| 09:57.34 | frantisheq | http://goo.gl/9RFUx |
| 10:02.53 | steev_ | bah |
| 10:03.31 | steev_ | isn't that steev |
| 10:06.16 | mnemoc | steev_: should get a more unique nick :) |
| 10:06.36 | steev_ | pfft, i've been using this nick online since 1998 |
| 10:29.22 | *** join/#arm-netbook mikki (~quassel@140-104-22-46.fiber.hringdu.is) |
| 10:30.49 | *** join/#arm-netbook svarthol (~quassel@140-104-22-46.fiber.hringdu.is) |
| 10:45.45 | *** join/#arm-netbook destinal (destinal@cryptkcoding.com) |
| 10:46.16 | *** join/#arm-netbook dyoung-away (dyoung@cryptkcoding.com) |
| 10:46.28 | *** join/#arm-netbook dyoung (dyoung@nasadmin/dyoung) |
| 10:47.47 | *** join/#arm-netbook ka6sox (ka6sox@nasadmin/ka6sox) |
| 10:55.15 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 10:58.37 | CIA-91 | rhombus-tech: Alen master * r2aeffd8fd440 /allwinner_a10/orders/fasat2112.mdwn: |
| 10:58.57 | CIA-91 | rhombus-tech: lkcl master * r692cd8ca0739 /allwinner_a10/hacking_the_mele_a1000.mdwn: add server bugreport link |
| 10:59.20 | CIA-91 | rhombus-tech: lkcl master * rf42596cf61d9 /allwinner_a10/hacking_the_mele_a1000.mdwn: add toc |
| 10:59.41 | CIA-91 | rhombus-tech: lkcl master * redfa9286e0bf /allwinner_a10/hacking_the_mele_a1000.mdwn: add toc |
| 11:00.28 | CIA-91 | rhombus-tech: lkcl master * r8091b9a5fcd4 /allwinner_a10/hacking_the_mele_a1000.mdwn: add toc |
| 11:05.26 | *** join/#arm-netbook arokux_h (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 11:07.11 | CIA-91 | rhombus-tech: lkcl master * r79e304c54156 /mailing_list_policy/discussion.mdwn: add response to mailing list discussion question |
| 11:38.35 | *** join/#arm-netbook nALLITeT (nALLITeT@ppp-109-104-167-185.wildpark.net) |
| 12:01.14 | mnemoc | anyone here building the kernel with gcc 4.7? |
| 12:09.29 | specing | I could do it, but no... |
| 12:25.02 | mnemoc | It's about https://github.com/amery/linux-allwinner/issues/13 ... I removed the `inline` part from the .h and still compiles fine with 4.6, not sure if that makes 4.7 happy... I would prefer to start moving code around at this stage |
| 12:26.28 | mnemoc | I use linaro's toolchain, but they don't publish a 4.7-based one yet on their PPA |
| 12:27.00 | mnemoc | and I don't feel like building my own atm :p |
| 12:29.01 | specing | 's gcc: gcc (GCC) 4.7.0 20120107 (experimental) |
| 12:32.02 | mnemoc | specing: wanna try http://dpaste.com/750194/ ? |
| 12:45.34 | specing | Its going to take a while |
| 12:45.47 | specing | I only have 2.6.36 sources from idk where |
| 12:45.55 | mnemoc | :) |
| 12:46.11 | mnemoc | just add the remote to https://github.com/amery/linux-allwinner/ |
| 12:46.21 | mnemoc | it should save you quite a lot of download |
| 12:47.08 | specing | I only have |
| 12:47.10 | specing | * remote origin Fetch URL: git://github.com/allwinner/linux-2.6.36.git |
| 12:47.57 | mnemoc | git remote rename origin allwinner-github |
| 12:48.26 | mnemoc | git remote add origin git://github.com/amery/linux-allwinner.git |
| 12:48.30 | mnemoc | git fetch origin |
| 12:48.46 | mnemoc | :) |
| 12:49.49 | specing | git-fu... |
| 13:00.15 | mnemoc | that's a quite simple move... |
| 13:20.40 | *** join/#arm-netbook frantisheq (~androirc@kapusany.3log.sk) |
| 13:20.47 | specing | mnemoc: Ah, I have x86 4.7... |
| 13:20.55 | specing | totaly forgot |
| 13:42.46 | nALLITeT | hi everyone. can be EOMA-68 used without motherboard? |
| 13:43.40 | specing | EOMA-68 card is a motherboard? |
| 13:44.25 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 13:45.30 | nALLITeT | thats cpu card, right? |
| 13:45.44 | specing | yes |
| 13:45.50 | cromo | mnemoc: I forked your v2 branch for my custom kernel for mele, but some of the commits may be of your interest - https://github.com/wrobelda/linux-allwinner/commits/mele-a1000 |
| 13:47.02 | Turl | cromo: I see you didn't enable dvfs, is there any specific cause? |
| 13:48.18 | cromo | not really, I saw that but had no time to investigate further. it's not enabled by default if you turn the CPU FREQ on, so I left it as it |
| 13:52.20 | cromo | I enabled it now,will test if it works |
| 14:08.03 | mnemoc | ls -al |
| 14:08.05 | mnemoc | err |
| 14:10.54 | *** join/#arm-netbook gimli (~gimli@xbmc/staff/gimli) |
| 14:10.57 | gimli | hello |
| 14:11.14 | gimli | short question. whats the right mali config file for compiling ? |
| 14:11.41 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 14:12.02 | gimli | https://github.com/amery/linux-allwinner/tree/allwinner-v3.0-android-v2/modules/mali/DX910-SW-99002-r2p3-01rel0/src/devicedrv/mali try to figure out which is the right one for the mele a1000 |
| 14:14.49 | Turl | gimli: I suppose it's arch-ca8-virtex820-m400-1 |
| 14:15.11 | gimli | thx |
| 14:18.06 | *** join/#arm-netbook miles2 (~miles@cpc3-haye16-2-0-cust391.haye.cable.virginmedia.com) |
| 14:19.41 | A2Sheds | specing, eoma68 is a computer on module, some may be used stand alone with just a suitable power supply (USB-OTG for instance) |
| 14:20.36 | A2Sheds | <PROTECTED> |
| 14:21.05 | nALLITeT | some? why other can't be used stand alone? |
| 14:22.36 | A2Sheds | nALLITeT : you just need to add IO connectors, any expansion IO you might want (LVDS for example) and power to the cards |
| 14:23.11 | mnemoc | cromo: what about making a mele_a1000_defconfig ? |
| 14:23.32 | cromo | mnemoc: makes sense |
| 14:24.32 | zenitraM | i'd call it mele_aseries_defconfig |
| 14:24.45 | nALLITeT | A2Sheds: thanks |
| 14:25.11 | mnemoc | zenitraM: fine too |
| 14:25.19 | cromo | will do it later today |
| 14:25.29 | mnemoc | zenitraM: btw, the leaked buildroot tree has some interesting mali stuff |
| 14:27.09 | zenitraM | where? |
| 14:27.32 | mnemoc | package/mali-3d* |
| 14:28.32 | zenitraM | yep, i've seen it, it's pretty much the headers and the binary libraries |
| 14:28.57 | mnemoc | nothing extra in the history? |
| 14:31.38 | zenitraM | i'd say no |
| 14:31.40 | hno | nALLITeT, EOMS68 A10 card can be used with only USB cable. |
| 14:32.23 | zenitraM | well, except three different versions of the libraries .tar's |
| 14:32.50 | Turl | is there any allwinner-based raspeberry-pi like device on a similar price range? |
| 14:33.29 | mnemoc | Turl: http://olimex.wordpress.com/ |
| 14:33.47 | mnemoc | Turl: I mean, http://olimex.wordpress.com/2012/05/14/a13-olinuxino-1ghz-cortex-a8-development-started/ |
| 14:34.06 | nALLITeT | hno: i know, asked just to be sure |
| 14:35.39 | gimli | mnemoc: indeed ;) |
| 14:35.57 | Turl | mnemoc: doesn't look like it's out yet |
| 14:35.58 | zenitraM | okay, i might have bad news |
| 14:36.05 | zenitraM | the mali x11 driver seems to require libraries from the mali ddk |
| 14:36.23 | mnemoc | Turl: right, not out yet |
| 14:36.30 | nikk|zzz | hi ppl |
| 14:37.20 | Turl | mnemoc: I was hoping to get a device to offload file serving & mp3 playback from my router, and add transmission to it while moving the stuff |
| 14:37.47 | mnemoc | Turl: at the moment the best you can buy is a mele |
| 14:37.55 | Turl | mnemoc: raspberry pi would be ok for that, but given their supply I'll probably get mine on 2013 :P |
| 14:38.02 | Turl | mnemoc: but mele is like 5x the price right? |
| 14:38.18 | gimli | zenitraM: who needs x11 :D |
| 14:38.41 | nikk | tried to add v3 tree to CM9 build, got that error: http://pastebin.com/20FMkQ0f |
| 14:38.53 | mnemoc | Turl: ~$85 with shipping is a great price for what it is |
| 14:38.58 | nikk | v2 android tree, sorry |
| 14:39.27 | mnemoc | Turl: the rpi is cheaper, sure, but junk compared to the mele |
| 14:40.23 | Turl | mnemoc: yeah, I'm not saying it's not worth what it costs |
| 14:40.27 | mnemoc | Turl: and even more if you want to run a torrents downloader there, usb can't compete with sata |
| 14:40.44 | Turl | but unfortunately I don't have any money-growing trees yet :P |
| 14:41.06 | gimli | mnemoc: depends what internet connection you have. if you realy need sata |
| 14:41.22 | Turl | my router USB does ~3MB/s |
| 14:41.29 | Turl | I expect at least that |
| 14:41.53 | gimli | mnemoc: it also depends what you are going to with the device, if it fits your needs or not. |
| 14:41.53 | mnemoc | bt is very io expensive |
| 14:41.55 | nikk | 3mb/s is possible on 300mhz mips ) |
| 14:42.05 | Turl | I have 6Mb/s internet, so it's not even near maxing out |
| 14:42.41 | Turl | nikk: cpu model: MIPS 24Kc V7.4 |
| 14:42.41 | Turl | BogoMIPS: 265.42 |
| 14:42.42 | Turl | :) |
| 14:42.47 | gimli | it is true that the PI is not the fastest one, but the video playback qualitity is extraordinary |
| 14:43.08 | mnemoc | :) |
| 14:43.43 | Turl | too bad it sucks at floating point |
| 14:43.48 | Turl | I cannot encode mp3 on it |
| 14:44.02 | Turl | :P |
| 14:44.14 | gimli | Turl: half the truth |
| 14:44.15 | zenitraM | well, the a10 with cedar on android can play 30mb/s 1080p MKVs without a hassle :P |
| 14:44.39 | Turl | gimli: why half? |
| 14:44.41 | gimli | zenitraM: and. whats so great in that ? |
| 14:44.46 | nikk | zenitraM: i'll try this ))) they proove 2160p ) |
| 14:45.04 | gimli | Turl: when they would sell the codec licences then you could do the encoding on the VideoCore |
| 14:45.07 | Turl | nikk: HDMI output is rated @ 1080p though |
| 14:45.19 | zenitraM | i mean that they are both equal at the playback quality i guess |
| 14:45.28 | Turl | gimli: I doubt my MIPS router has a VideoCore |
| 14:45.49 | gimli | oh. though you talk about the PI. sorry |
| 14:46.05 | Turl | the pi has decent floating point I suppose |
| 14:46.20 | mnemoc | nikk: very weird error |
| 14:46.52 | A2Sheds | any ARM board with cortex-a8 >1GHz, >512M-DDR3 for under $80 is a win |
| 14:47.12 | nikk | mnemoc yup. CM9 tried compiling only zImage for android by itself and crashes |
| 14:47.34 | mnemoc | nikk: fix CM9 :) |
| 14:47.49 | Turl | don't blame CM9 :) |
| 14:48.21 | mnemoc | make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- -j32 zImage |
| 14:48.22 | mnemoc | <PROTECTED> |
| 14:49.07 | A2Sheds | all of the $200 ARM modules, dongles, usb sticks .....whatevers have sold well, I'm not sure what they were thinking when they making them |
| 14:49.08 | mnemoc | the standard build works fine, so the problem is in something CM9 is wrongly assuming |
| 14:49.13 | Turl | mnemoc: that doesn't sound like android's prebuilt toolchain does it? |
| 14:49.35 | mnemoc | Turl: nope, it's linaro's toolchain |
| 14:49.48 | naobsd | -j32 sounds great ;) |
| 14:49.53 | nikk | mnemoc: i tried. there are only kernel.mk: http://pastebin.com/FLGEMV4Q |
| 14:50.24 | Turl | yeah, that's all the magic there is |
| 14:50.45 | nikk | heh, i'll try with arm-linux-gnueabi-, hope it helps |
| 14:51.14 | Turl | is the crane defconfig any good? |
| 14:51.19 | Turl | last I tried, it didn't build :P |
| 14:51.51 | mnemoc | I don't really use that one, but I fixed an error building android gadget some days ago |
| 14:52.05 | mnemoc | not sure if that's the only problem when using that .config |
| 14:52.25 | Turl | nikk: looks like you're on mac too? |
| 14:52.34 | nikk | Turl yup |
| 14:52.39 | mnemoc | I'll run a build of sun4i_crane now, 2m |
| 14:52.56 | Turl | CM9 often breaks build on mac, as nobody (except koush) builds on it :P |
| 14:54.46 | nikk | Turl I'll check this ;-) But, as 2.3 and 4.0.3 builds fine, I hope CM9 not a bitch ;-) |
| 14:55.45 | nikk | but i needed to test 4 versions of GCC and find correct one ) |
| 14:55.56 | Turl | historically android is harder to build on mac than linux |
| 14:56.08 | mnemoc | Turl: uImage on sun4i_crane_defconfig finished fine |
| 14:56.25 | mnemoc | 2m14.459s 3.80 MB |
| 14:56.27 | Turl | qemu breaks, and you need to set up your toolchains and use a volume |
| 14:56.34 | Turl | mnemoc: :) |
| 14:57.21 | mnemoc | modules finished too, no errors |
| 14:59.14 | nikk | so, the problem with toolchain |
| 14:59.23 | nikk | which CM9 uses? |
| 14:59.53 | mnemoc | i don't know, i'm not familiar with aosp's or cm build systems |
| 15:00.11 | mnemoc | but "works here" (tm) |
| 15:00.15 | Turl | let me setup inline kernel building on my tree |
| 15:00.32 | mnemoc | try using that toolchain but doing the build yourself |
| 15:01.25 | nikk | Turl i disabled qemu, all other stuff looks fine enough... hope i'll build it soon, at least with precopiled kernel ) |
| 15:10.38 | *** join/#arm-netbook nikk (~somebody@nikk.starlink.ru) |
| 15:18.37 | naobsd | looking lichee-3.0.8-sun4i branch... |
| 15:22.29 | nikk | naobsd it's superseeded by 3.0-v2 |
| 15:23.08 | *** join/#arm-netbook frafra (~frafra@host233-205-dynamic.41-79-r.retail.telecomitalia.it) |
| 15:23.13 | frafra | hi all |
| 15:23.14 | naobsd | I see. I read wiki. but I'm curious what's changed |
| 15:23.53 | mnemoc | git log origin/lichee-3.0.8-sun4i.. |
| 15:24.13 | frafra | when allwinner a10 soc will be released? |
| 15:24.38 | naobsd | well... I'm curious about difference bitween leaked source and v2 |
| 15:24.54 | mnemoc | frafra: like a year ago |
| 15:25.20 | frafra | I mean the rhombustech one |
| 15:25.22 | miles2 | Hey guys, I'm hoping to cook up a version of Ubuntu to run on my Ainol Elf (A10 Allwinner). I've been googling around but I wondered if anyone had any tips (I've played with kerels before but far from an expert) |
| 15:26.49 | mnemoc | naobsd: v2 includes the latest fixes in 3.0.y and android's kernel-common, proper earlyprintk and serial console, sun3i support and some other fixes |
| 15:27.35 | nikk | and it includes all 3.0.8 allwinners additions as i understand ) |
| 15:27.38 | mnemoc | frafra: the eoma68 module based on the allwinner's a10 soc from rhombustech doesn't have a date yet |
| 15:27.38 | naobsd | mnemoc: thank you. sun3i support? where it comes from? |
| 15:28.05 | frafra | thanks mnemoc |
| 15:28.06 | mnemoc | naobsd: fron the 2.6.36 release, but there is still stuff to move and fix |
| 15:28.26 | naobsd | mnemoc: ah, I see. thanks again. |
| 15:28.28 | mnemoc | nikk: yes, it's starting point was lichee-3.0.8-sun4i |
| 15:29.21 | mnemoc | naobsd: also 3.0-v2 has sun5i (A12/A13) support which was removed during the history of the 3.0.8 branch |
| 15:30.22 | nikk | mnemoc: another question, why it adds "+" sign to kernel version string? it's not in kernel options. |
| 15:30.46 | mnemoc | nikk: it happens when building from git |
| 15:30.58 | nikk | i see |
| 15:31.03 | mnemoc | nikk: to mark it's not a pure build from a tag |
| 15:33.28 | Turl | mnemoc: can you do a build using O= ? |
| 15:33.57 | Turl | that looks like the problem in here, the compilation chokes because the directories don't get created somehow |
| 15:36.53 | mnemoc | trying |
| 15:38.59 | mnemoc | confirmed, with O=foo I get odd errors too |
| 15:39.18 | Turl | mnemoc: I think it's because those dirs have no Makefile in them |
| 15:39.22 | Turl | let me try something |
| 15:39.40 | mnemoc | Turl: file a ticket first please :) |
| 15:39.47 | Turl | mnemoc: ok |
| 15:40.59 | nikk | heh, so my problem not the toolchain? make -C kernel/allwinner/novo7a O=/Volumes/CM9/out/target/product/novo7a/obj/KERNEL_OBJ <- that's what not good? |
| 15:41.27 | Turl | indeed, toolchain is not your issue nikk |
| 15:42.34 | nikk | what i need to parch Turl? -) |
| 15:43.06 | Turl | nikk: I'm investigating that :) |
| 15:46.05 | nikk | naobsd: about cm9 quack2k additions, i need to replace system/core/include/system/window.h with one from 4.0.3 source to build with cedarx |
| 15:50.26 | *** join/#arm-netbook frantisheq (~androirc@kapusany.3log.sk) |
| 15:55.23 | naobsd | nikk: I don't try quarx's CM9 yet |
| 15:59.22 | Turl | mnemoc: I found out what's wrong |
| 15:59.43 | Turl | mnemoc: which branch should I submit a pull request to? v2? |
| 16:01.12 | nikk | hm, i remember, when i tried debug wints a81g (omap4 platform), there was cdc_acm enabled uBoot for it, and I was able access uboot via USB from terminal. Is this possible to add that functionality to a10 uboot too? |
| 16:03.43 | nikk | Turl diff url? |
| 16:04.45 | Turl | nikk: I'm still building all the stuff to make sure there's no more build breakage, I'll get you a link in a minute |
| 16:13.28 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 16:15.51 | mnemoc | Turl: yes, v2, but a clean merge please |
| 16:16.36 | mnemoc | nikk: hno is the one to answer that |
| 16:18.40 | mnemoc | nikk: i know it was on his todo, but I don't know what's the state |
| 16:19.08 | nikk | mnemoc this useful, if you want to access uBoot for changing boot params, or for loading from MMC, uBoot console goes not to internal uart but to usb via cdc_acm |
| 16:21.12 | nikk | but you lost it on kernel loading |
| 16:22.51 | *** join/#arm-netbook QingPei (~qingpei@123.118.104.193) |
| 16:27.06 | traeak | mnemoc: i see a few branches created |
| 16:28.55 | mnemoc | see github's wiki for the explaination of the branches |
| 16:30.49 | *** part/#arm-netbook QingPei (~qingpei@123.118.104.193) |
| 16:35.33 | Gumboot | gimli: Have you actually heard the VideoCore ported MP3 encoder? |
| 16:36.55 | gimli | iirc there is an openmac videocore encoder. sadly not free available |
| 16:37.16 | Gumboot | "openmac"? |
| 16:37.56 | nikk | it's always fun to talk with iirc ) |
| 16:58.25 | mnemoc | Turl: so...? :) |
| 16:59.43 | traeak | uhoh are they upset with us? |
| 17:00.22 | nikk | we all connected ))) |
| 17:00.57 | Turl | mnemoc: https://github.com/amery/linux-allwinner/pull/22 :) |
| 17:01.10 | Turl | give it a try just to make sure I didn't break anything |
| 17:03.33 | nikk | van't wait it will be merged =) |
| 17:03.40 | nikk | can't ) |
| 17:04.20 | *** join/#arm-netbook nikk (~somebody@nikk.starlink.ru) |
| 17:04.49 | Turl | mnemoc: github created a new issue together with the pull request :P |
| 17:05.19 | mnemoc | yeah, just noticed, quite stupid |
| 17:05.57 | mnemoc | doesn't the "pull request" thing let you pick an existing issue? |
| 17:06.15 | Turl | I didn't see any obvious way to do so |
| 17:06.21 | mnemoc | :< |
| 17:06.24 | Turl | and mentioning issues doesn't seem to do the trick either |
| 17:07.37 | nikk | Turl I don't know about Android.mk trick. Just deleted those modules ) |
| 17:08.21 | mnemoc | Turl: #21 in the description seems to bind them |
| 17:08.51 | mnemoc | https://github.com/amery/linux-allwinner/issues/21 now links to #22 |
| 17:09.53 | mnemoc | Turl: I'm confused by the pointer to dram/ |
| 17:09.54 | Turl | it still created a new one though |
| 17:10.15 | mnemoc | Turl: the original Makefile doesn't include dram/'s .o files |
| 17:10.36 | Turl | mnemoc: it does, '        dram/dram.o dram/dram_init.o \' |
| 17:10.59 | Turl | github people on irc are super friendly and just told me jfgi :) |
| 17:11.26 | mnemoc | lovely :) |
| 17:11.40 | mnemoc | Turl: but that goes in standby-y, not obj-y |
| 17:12.45 | mnemoc | Turl: ok, no problem, dram/ in standby-y fails badly |
| 17:16.06 | Turl | yeah that's why I used obj-y :) |
| 17:17.07 | nikk | Turl do you port CM9 to A10? |
| 17:18.09 | Turl | nikk: I'm working with quarx to do so |
| 17:18.18 | nikk | wow |
| 17:18.31 | nikk | I mady my fork from his repo |
| 17:18.42 | Turl | but I still didn't get my tablet, hopefully on monday :) |
| 17:18.55 | nikk | i see. which one? |
| 17:19.02 | nikk | I like hyundai A7HD |
| 17:19.06 | Turl | nikk: https://zareason.com/shop/zatab.html |
| 17:19.13 | nikk | it's with IPS and so cheep |
| 17:19.53 | nikk | in case 7" with 1024x600 (1:1) and 150$... |
| 17:20.33 | nikk | iurl CM9 still buggy |
| 17:21.54 | traeak | 14 hour battery |
| 17:22.18 | traeak | at bottom says 349USD |
| 17:22.46 | traeak | THAT is a ripoff |
| 17:23.05 | mnemoc | it's not a large massive gpl-violating tablet |
| 17:23.21 | traeak | so nook color/amazon kindle do violate gpl ? |
| 17:23.22 | mnemoc | it's an open device built in lower quantities |
| 17:23.24 | nikk | 3G doesn't work and its ztab, h7hd more cheap http://www.hyundai-digital.cn/ShowProduct.asp?ID=145 |
| 17:23.37 | Turl | 9.7" IPS and 8Ah battery |
| 17:23.41 | traeak | grr i have to hurry today |
| 17:23.43 | traeak | and fix this bug |
| 17:23.44 | Turl | besides I didn't pay 349US :PP |
| 17:23.46 | mnemoc | traeak: amazon looses money with the kindle |
| 17:23.59 | mnemoc | loses* |
| 17:23.59 | traeak | just got a call...a guy i know is in a coma from an accident....brain injury |
| 17:24.01 | traeak | grrr |
| 17:24.10 | mnemoc | :( |
| 17:24.22 | nikk | :( |
| 17:24.35 | traeak | haven't seen him in a few years though |
| 17:24.52 | nikk | hope hi'll survive |
| 17:24.54 | traeak | still doesn't matter, he lives a couple hours away (pita generally) |
| 17:25.01 | Turl | :( |
| 17:25.01 | traeak | can't say, he's been under 3 days apparently |
| 17:25.11 | *** join/#arm-netbook steev_ (u1383@gentoo/developer/steev) |
| 17:25.32 | traeak | (had to come in today to prepare for demo) |
| 17:27.59 | gimli | hehe. xbmc gui comes up on the a1000. quite nice :D |
| 17:28.46 | mnemoc | would prefer cedarx support in gstreamer |
| 17:28.49 | mnemoc | hides |
| 17:29.43 | gimli | haha |
| 17:30.09 | gimli | my plan is to write native xbmc support for it ;) |
| 17:30.17 | traeak | xbmc support for gstreamer? |
| 17:30.23 | gimli | nope |
| 17:30.50 | gimli | gstreamer is it's own pain in the ass |
| 17:32.00 | nikk | mnemoc i made commit in GitHub app, is it post it to public? |
| 17:32.23 | mnemoc | nikk: in your fork, yes |
| 17:35.19 | *** join/#arm-netbook nikk (~somebody@nikk.starlink.ru) |
| 17:35.22 | nikk | for linux-allwinner, but i'm not forked? it, just cloned https://github.com/amery/linux-allwinner.git |
| 17:36.04 | mnemoc | then the commit is local |
| 17:36.14 | nikk | i see, with r.o. access no posts. ok |
| 17:36.47 | mnemoc | you publish your local commits making a "push" |
| 17:37.00 | nikk | nevr used github before. i'll learn it ) |
| 17:37.01 | mnemoc | but that requires write access |
| 17:37.35 | mnemoc | please note github is only one of many source code hosting services for `git` |
| 17:38.07 | nikk | heah, with friendly interface ) |
| 17:38.23 | mnemoc | :) |
| 17:38.55 | nikk | on mac it so simple, no options at allll )))) |
| 17:39.37 | mnemoc | as expectable for that nazi system |
| 17:39.58 | traeak | what's a mac ? |
| 17:40.11 | nikk | yup |
| 17:40.25 | Turl | mnemoc: why did github do a commit merge? it should've been a simple fastforward :/ |
| 17:40.50 | mnemoc | that only happens if the branch is rebased |
| 17:41.04 | Turl | I did rebase it |
| 17:41.15 | mnemoc | nope, you merged |
| 17:41.15 | nikk | http://dl.dropbox.com/u/2258618/2012-05-19%20%D0%B2%2021.40.17.png |
| 17:41.50 | mnemoc | nikk: I have to admit it looks nice |
| 17:42.11 | nikk | yup |
| 17:42.17 | Turl | mnemoc: I rebased :P https://github.com/allwinner-dev-team/linux-allwinner/commits/allwinner-3.0 |
| 17:42.57 | mnemoc | Turl: odd... next time I'll better cherry-pick |
| 17:43.51 | nikk | i mean this: http://dl.dropbox.com/u/2258618/2012-05-19%20%D0%B2%2021.43.04.png |
| 17:44.07 | *** join/#arm-netbook steev_ (u1383@gentoo/developer/steev) |
| 17:44.56 | mnemoc | nikk: than one is local, or you push it somewhere or generate a patch and mail it |
| 17:45.13 | nikk | i cee |
| 17:48.04 | Turl | mnemoc: some warnings fixed https://github.com/allwinner-dev-team/linux-allwinner/commit/b2fd4b2c892f1ece4d8b6147650bbf11b7d802d0 |
| 17:48.14 | Turl | hopefully I didn't break anything :) |
| 17:51.44 | mnemoc | I had that one in v1... |
| 17:53.02 | nikk | do i need git checkout after repo sync? |
| 17:53.52 | mnemoc | nikk: are you using android's `repo` ? |
| 17:53.58 | Turl | mnemoc: yours looks better, cherrypick it :) |
| 17:54.10 | nikk | yup, for android |
| 17:54.25 | nikk | for kernel i use GitHub app |
| 17:54.29 | Turl | nikk: no, repo checks out automatically |
| 17:57.08 | A2Sheds | if luke screws up twice, who kicks him off the ML? :) |
| 17:57.28 | mnemoc | prefers IRC anyway |
| 17:58.10 | A2Sheds | what if I unsubscribe from the ML can I still post to the wiki? |
| 17:58.26 | mnemoc | sure |
| 17:58.31 | Turl | ML? I'm missing all the fun |
| 17:58.32 | Turl | :P |
| 17:58.46 | A2Sheds | http://rhombus-tech.net/mailing_list_policy/ |
| 17:59.42 | mnemoc | things are more.... complicated... when there are economic interest involed in free software communities |
| 18:02.55 | traeak | my problem is already fixed. i've tried to subscribe 3 times to the list and never go on |
| 18:02.59 | traeak | so that solved my problem |
| 18:03.06 | mnemoc | doh |
| 18:04.33 | nikk | heh, after git checkout it deleted all modules/* directory? |
| 18:04.50 | nikk | for android 3.0 v2 branch |
| 18:05.13 | mnemoc | nikk: `git checkout` alters what you have expanded on the filesystem to certain point in the history |
| 18:05.39 | mnemoc | if you lost the modules/* dir you switched back to the mainline tree |
| 18:06.09 | nikk | ah i'm not tell what branch i want to checkout yeah? |
| 18:06.23 | nikk | but i think v2 was default in my case |
| 18:06.26 | mnemoc | i don't know your interface |
| 18:06.41 | nikk | shell ) but github.app also |
| 18:06.43 | traeak | mnemoc: btw, do you post binary snapshots of your builds and modules directories ? |
| 18:07.11 | mnemoc | traeak: s/do/can/ ? |
| 18:09.32 | traeak | reason i ask is because someone could do the following: grab a uboot.img (without uImage), grab a distro snapshopt, grab a tarball with uImage and lib/modules/<version>. then install uboot image, install distro snapshot, overlay uImage and modules to get a working sd card going |
| 18:10.06 | mnemoc | traeak: I'm not in the business of supporting people who can't use a compiler on their own |
| 18:11.13 | traeak | mnemoc: understood, but it would just be a starting place for some people :-p |
| 18:12.30 | nikk | hm, can't understand: |
| 18:12.31 | nikk | Pro:linux-allwinner nikk$ pwd |
| 18:12.32 | nikk | Pro:linux-allwinner nikk$ git branch |
| 18:12.32 | nikk | * allwinner-v3.0-android-v2 |
| 18:12.32 | nikk | Pro:linux-allwinner nikk$ git checkout |
| 18:12.32 | nikk | Mdrivers/i2c/busses/i2c-sunxi.c |
| 18:12.32 | nikk | Dmodules/example/Makefile |
| 18:12.33 | nikk | Dmodules/example/example.c |
| 18:12.40 | nikk | Dmodules/mali/DX910-SW-99002-r2p1-05rel1/src/devicedrv/mali/Makefile |
| 18:12.42 | nikk | Dmodules/mali/DX910-SW-99002-r2p1-05rel1/src/devicedrv/mali/Makefile.common |
| 18:12.43 | nikk | Dmodules/mali/DX910-SW-99002-r2p1-05rel1/src/devicedrv/mali/arch-ca8-virtex820-m400-1/config.h |
| 18:12.44 | nikk | Dmodules/mali/DX910-SW-99002-r2p1-05rel1/src/devicedrv/mali/arch-pb-virtex5-m300/config.h |
| 18:12.54 | nikk | Dmodules/mali/DX910-SW-99002-r2p1-05rel1/src/devicedrv/mali/arch-pb-virtex5-m400-1-direct/config.h |
| 18:13.28 | mnemoc | traeak: I'm all for helping those who make the distro images (.deb for example), and eventually I plan to do a .deb (for linaro's ubuntu) myself.... but the tree is not ready for that level of users yet |
| 18:13.35 | mnemoc | nikk: git reset --hard |
| 18:14.28 | traeak | mnemoc: i agree it's too cutting edge for that just now |
| 18:14.39 | nikk | why it stalls? i deleted that part before for cm9 build (without Android.mk) |
| 18:15.13 | mnemoc | nikk: git needs clean dirs to work |
| 18:15.30 | nikk | Pro:linux-allwinner nikk$ git reset --hard |
| 18:15.30 | nikk | Checking out files: 100% (2619/2619), done. |
| 18:15.30 | nikk | HEAD is now at 182c204 usb: gadget: android: fix merge mistake |
| 18:15.30 | nikk | Pro:linux-allwinner nikk$ ls modules |
| 18:15.30 | nikk | example mali wifi |
| 18:15.58 | nikk | not it's back. clean dirs? what it means? |
| 18:16.08 | mnemoc | nikk: don't paste on the channel :( |
| 18:16.18 | nikk | ok |
| 18:16.19 | mnemoc | nikk: clean = empty `git status` |
| 18:18.25 | mnemoc | nikk: if `git status` shows you checked you don't want to keep, `git co -- path/to/the/file/to/reset` |
| 18:18.26 | nikk | i see, but i don't see Turl patches, head at 182c204 |
| 18:18.45 | mnemoc | nikk: as long as your dir is dirty you can't update it |
| 18:19.46 | hno | zenitraM, yes the mali X11 drivers require DDK libraries. |
| 18:22.51 | zenitraM | :( |
| 18:24.42 | nikk | hno does it possible add cdc_acm support to uboot? |
| 18:25.01 | nikk | it can be useful have usb uboot console |
| 18:26.04 | nikk | but it needs alter uboot configuration to choose console port as cdc_acm |
| 18:29.53 | ssvb | hi, could anybody do me a favour: compile and run the following simple memory benchmark program on Allwinner A10 hardware? https://github.com/ssvb/ssvb-membench |
| 18:30.45 | ssvb | for example the results from IGEPv2 board, TI DM3730 (Cortex-A8 1GHz) look like this - http://people.freedesktop.org/~siamashka/files/20120425/igepv2-dm3730-cortex-a8-1ghz.txt |
| 18:31.25 | ssvb | I'm just trying to decide whether Allwinner A10 is interesting for me or not :) |
| 18:31.44 | zenitraM | ssvb: downloading.. |
| 18:32.03 | hno | nikk, I am aiming to add USB serial gadget to A10 u-boot. But life have interrupted a bit. |
| 18:32.32 | nikk | i see. it will be cool ;-) |
| 18:33.53 | Turl | ssvb: if nobody does by monday let me know |
| 18:34.01 | Turl | I'll prolly have my hardware by then :) |
| 18:34.19 | zenitraM | ssvb, Turl: i'll have it in a few secs :) |
| 18:34.42 | ssvb | zenitraM, Turl: thanks |
| 18:34.58 | traeak | ssvb: and you'll get a sample from me too |
| 18:35.02 | traeak | be interesting to see the variance |
| 18:35.17 | zenitraM | http://pastie.org/3936586 |
| 18:35.17 | traeak | i guess sample of 2 doesn't really give that |
| 18:35.20 | zenitraM | this is on armhf |
| 18:35.32 | hno | nikk, there is absolutely nothing that stops it from being added, just time to do it. |
| 18:36.53 | traeak | http://pastie.org/3936602 |
| 18:37.05 | traeak | zenitraM: not huge differences, i'm using gcc 4.7 here |
| 18:39.18 | zenitraM | mmm interestingly enough, the snowball version of the x11 mali module doesn't use the ddk |
| 18:42.19 | ssvb | zenitraM, traeak: thanks a lot, looks like copy (memcpy) is a bit better than TI DM3730, fill (memset) is somewhat slower, latency is about the same |
| 18:42.34 | ssvb | was it running at 1GHz or 1.2GHz? |
| 18:43.08 | traeak | how expensive is the dm3730 and what gpu does it use? |
| 18:43.17 | zenitraM | 1GHz afaik |
| 18:43.31 | traeak | crap how do you tell the clock rate? |
| 18:45.49 | ssvb | traeak: don't know the DM3730 SoC price, but the board itself is beagleboard-xm clone (SGX530 GPU 200MHz), which means that it is more expensive than Allwinner A10 based hardware :) |
| 18:46.26 | ssvb | traeak: does cpufreq-info from cpufrequtils work? |
| 18:47.05 | traeak | ssvb: missing driver |
| 18:47.30 | traeak | zenitraM: you're running gcc 4.6 right ? |
| 18:47.53 | zenitraM | traeak: yup |
| 18:48.03 | hno | traeak, cat /proc/cpuinfo don't work? |
| 18:48.43 | traeak | hno: no frequency info, just the other information |
| 18:49.16 | traeak | http://pastie.org/3936647 |
| 18:50.13 | traeak | zenitraM: i've got an omap3 arch linux image on there, that has gcc-4.7 with other stuff as well |
| 18:50.42 | ssvb | looks like 1GHz (for Cortex-A8 the bogomips rating is roughly equal to clock frequency) |
| 18:51.47 | Turl | ssvb: I'm running the benchmark on a Qualcomm MSM7201A, just for fun :) |
| 18:53.01 | hno | nikk, there is a patch to add usb serial console to the kernel as well. |
| 18:53.14 | ssvb | traeak, zenitraM: btw, do you have monitor connected and active? Screen refresh can steal quite a bit of memory bandwidth |
| 18:55.11 | ssvb | the compiler version should have no effect on the performance of assembly code, and NEON copy/fill is done via assembly |
| 18:55.23 | zenitraM | huh, i can also run it on the 1.2GHz TI OMAP 4460 on my phone :P |
| 18:55.26 | zenitraM | i don't think so |
| 18:56.06 | Turl | ssvb: http://paste.debian.net/170165/ |
| 18:56.26 | Turl | zenitraM: got a galaxy nexus? :P |
| 18:56.53 | zenitraM | Turl: yup :P |
| 18:57.14 | Turl | zenitraM: mine probably beats yours ;) j/k :) |
| 18:58.09 | Turl | ssvb: why is ARM fill (STM with 8 registers) so much faster? |
| 18:59.23 | penguin42 | Turl: It's one instruction, probably manages to kick an entire cacheline in one go |
| 18:59.52 | penguin42 | Turl: Can also be very CPU specific - they're saying on the A15 cores that LDRD/STRD is much better than ldm/stm |
| 19:00.04 | ssvb | Turl: for less advanced processors, the way how to do memory accesses matters a lot |
| 19:00.14 | penguin42 | also specific to size of memory area and which cache it hits (and some core options that can sometimes be configured) |
| 19:01.39 | ssvb | Turl: in the case of ARM11, you have no write-allocate cache policy and the memory writes go through write combining buffer |
| 19:02.29 | penguin42 | ssvb: you might want to look at https://wiki.linaro.org/WorkingGroups/ToolChain/Benchmarks/InitialMemset some benchmarks I did a while ago with different memset implementations |
| 19:03.04 | ssvb | Turl: it is important to ensure that optimal burst writes are used, and nudging the code a bit may hit the sweet spot |
| 19:03.16 | mnemoc | ssvb: another, http://dpaste.com/750278/ ... mele a1000 |
| 19:04.26 | traeak | ssvb: monitor is connected but i don't have x11 or anything loaded...no video curently |
| 19:04.54 | *** join/#arm-netbook sysfwlab (~sysfwlab@eta91-1-82-227-11-158.fbx.proxad.net) |
| 19:05.19 | sysfwlab | hello everybody |
| 19:05.46 | ssvb | penguin42: thanks for the link, looking at it now |
| 19:07.57 | *** join/#arm-netbook j1nx (55df3ea3@gateway/web/freenode/ip.85.223.62.163) |
| 19:09.41 | j1nx | gimli: which ARM version are you using? r2p4 or r2p3? |
| 19:10.25 | j1nx | ARM=MALI drivers (sorry for the confusion) |
| 19:10.27 | penguin42 | ssvb: I was primarily optimising for A9 though, it's interesting to note what's different on A8 |
| 19:11.17 | ssvb | penguin42: I'm mostly interested in large memory blocks, way bigger than L2 cache (for software rendered graphics), this corresponds to the numbers on the right part of the graphs from your chart |
| 19:12.13 | penguin42 | ssvb: Right note how on A8 there isn't much difference for the things I tried at the top end - but Neon stinks on A9 for that |
| 19:13.08 | ssvb | penguin42: for this workload and large blocks, it does not matter what memset implementation is used on Cortex-A8 and Cortex-A9 (unless it is totally braindead :) ), but ARM9 and ARM11 are different |
| 19:13.36 | penguin42 | ssvb: I've not looked at optimising for anything that old |
| 19:14.10 | gimli | j1nx: r2p3 dont have r2p4 |
| 19:14.22 | penguin42 | ssvb: It's pretty hard to get stuff that works well on everything |
| 19:16.31 | *** join/#arm-netbook zewelor (~x@89-78-93-36.dynamic.chello.pl) |
| 19:16.58 | nikk | guys, how can i add this commit with git? https://github.com/allwinner-dev-team/linux-allwinner/commit/f8959c5aad791230b6485ecb9d540e28391dfe28 |
| 19:17.39 | *** join/#arm-netbook Lerc (~Lerc@121-74-254-104.telstraclear.net) |
| 19:18.00 | j1nx | gimli: OK, I will work on a patch to update the kernel drivers and GLES libraries to r2p4. Just for the sake of have the latest, because I don't think there are many differences. |
| 19:18.13 | ssvb | penguin42: yes, memcpy from glibc is pretty bad for Cortex-A8, memset from glibc is bad for ARM9 and ARM11, Cortex-A9 is mostly fine, and I have no idea about Cortex-A15 yet |
| 19:18.37 | ssvb | penguin42: but MIPS appears to be in a much worse shape :) |
| 19:19.01 | j1nx | gimli: what are you working on at the moment? |
| 19:19.47 | mnemoc | nikk: git pull --rebase |
| 19:19.48 | Turl | ssvb: did you have a look at the memcpy/set from bionic? |
| 19:20.22 | nikk | Pro:linux-allwinner nikk$ git pull --rebase |
| 19:20.22 | nikk | Current branch allwinner-v3.0-android-v2 is up to date. |
| 19:20.30 | gimli | j1nx: xbmc on the a1000 |
| 19:20.35 | mnemoc | nikk: then you have that commit |
| 19:20.59 | nikk | hm, ok, but i don't see it in history... |
| 19:21.00 | ssvb | Turl: it's easy to tune memcpy/memset for each particular hardware, the big problem with glibc is their requirement for copyright reassignment |
| 19:21.34 | gimli | j1nx: basicly gui is up |
| 19:22.15 | nikk | mnemoc sorry, it's there |
| 19:22.49 | j1nx | gimli: Is there any thing I an help with? I will go on a business travel to Abu Dhabi tomorrow (5 weeks), so don't have access to my A1000, but do have access to my server and use that for all my cross compiling stuff. |
| 19:24.17 | j1nx | gimli: I guess nothing is done on the hardware decoding? Good to hear the GUI is up, how about the resourses consumed? |
| 19:24.17 | ssvb | zenitraM: is your galaxy nexus rooted? by default it runs some weird frequency scaling governor which does not let it go up to 1.2GHz, and this messes up the benchmarks |
| 19:24.27 | zenitraM | ssvb: rooted and with.. not sure which kernel |
| 19:24.40 | Turl | ssvb: thermal protection kicks in |
| 19:24.43 | zenitraM | also, an android ROM whose logo is a pink unicorn |
| 19:25.10 | ZaEarl | aokp? |
| 19:25.28 | Turl | ZaEarl: how many roms with a pink unicorn are out there? :) |
| 19:25.40 | ZaEarl | not enough |
| 19:25.41 | zenitraM | yup :) |
| 19:26.16 | ssvb | zenitraM: so you should be able to set the 'performance' scaling governor during the test and get more meaningful numbers :) |
| 19:27.05 | gimli | j1nx: at the moment i prepare to work on the HW decoding. resources used it didn't cre so far. getting the development env up was important |
| 19:28.59 | MaDMaLKaV | gimli, if you have a repository we can follow to see what you are up to, it will be really nice |
| 19:29.07 | zewelor | did someone did network performance benchmark on a1000 / a2000 ? how fast is it ? |
| 19:29.35 | zewelor | im wondering if it will be able to play fullhd mkv from mounted nfs over lan |
| 19:30.45 | j1nx | gimli: I am interested to see how the different ARM devices perform compared to each other. RPi, A1X, Beagle, Panda, etc.. I think we are really on the edge of having ARM taking over the world. |
| 19:30.57 | *** join/#arm-netbook Lerc (~Lerc@121-74-254-104.telstraclear.net) |
| 19:31.01 | gimli | MaDMaLKaV: i got the a1000 yesterday. when i feel it's worth to put online i go for it. at the moment all is in the starting stage |
| 19:31.16 | MaDMaLKaV | aha |
| 19:31.49 | gimli | j1nx: arm will take over on the mediaplayer side, for sure ;) |
| 19:32.02 | gimli | MaDMaLKaV: sorry to have no better anwser for you at the moment |
| 19:32.09 | j1nx | gimli: Take your time! ussually it is better not to show it to fast as a lot of people will get there hopes up to fast |
| 19:32.32 | gimli | j1nx: i know. |
| 19:32.42 | j1nx | gimli: ARM is also very much of fun if we talk aboutr multi core servers ;) |
| 19:33.09 | gimli | thanks to google and android that arm got pushed |
| 19:34.52 | j1nx | Yes indeed, thanks to google but let not forget Steve Jobs. It is because of him Google dove into it. Without him we would still walk around with GSM phones in our pocket instead of continuously in our hand ;) |
| 19:34.56 | ZaEarl | if we just had solid x11 drivers for arm |
| 19:35.50 | nALLITeT | oh, tell me about first tablet pc |
| 19:36.09 | nALLITeT | answer to > Without him we would still walk around with GSM phones |
| 19:36.13 | j1nx | ZaEarl: have a look into the home folder of Tom his released ubuntu rootfs. Comapre those files against the content of ".bash_history" files of the users and you will get a long way |
| 19:36.22 | nALLITeT | and other "innovations" by apple |
| 19:36.26 | gimli | j1nx: at leats we have cool toys to play now ;) |
| 19:37.35 | ZaEarl | j1nx, we're still a long way from a full open source, 2d/3d/video accelerated driver stack. |
| 19:37.58 | j1nx | gimli: HW decode, how do you want to tackle that one? A decoder or like the RPi an alternative player? |
| 19:38.05 | zenitraM | ssvb, Turl: on the Galaxy Nexus: http://pastie.org/3936836 |
| 19:38.29 | Turl | ssvb: I linked it to bionic, I'll have newer numbers in a minute :) |
| 19:38.35 | zenitraM | that's compiled staticly with glibc |
| 19:38.44 | zenitraM | no bionic at hand |
| 19:38.47 | j1nx | ZaEarl: Yes, we have to hope for the Limadriver to be ussuable any time soon, but it is a bit quite lately |
| 19:39.15 | Turl | ssvb: http://paste.debian.net/170172/ |
| 19:39.29 | Turl | std memset > arm fill :) |
| 19:39.48 | Turl | std memcopy > arm copy |
| 19:39.51 | gimli | j1nx: that will be decided when i know how the handling will be |
| 19:40.59 | j1nx | gimli: OK, I wish you the best of luck with it. I will keep my eyes on the forum during my next 5 weeks of business trip. |
| 19:41.13 | zenitraM | also, glibc > bionic |
| 19:41.45 | Turl | zenitraM: eh, no, glibc was all slow |
| 19:42.16 | Turl | zenitraM: glibc http://paste.debian.net/170165/ |
| 19:42.33 | ssvb | Turl: yep, thanks, at least they are comparable :) I quickly added these ARM assembly tests recently specifically for Raspberry Pi without giving them any real tuning. Maybe bionic does more unrolling |
| 19:43.26 | penguin42 | Turl: newlib is also worth a try |
| 19:43.47 | zenitraM | ah, thought yours were on the nexus too and was comparing with mine. nvm |
| 19:44.59 | Turl | ssvb: btw http://paste.debian.net/170174/ |
| 19:45.19 | Turl | that's to build with bionic if you want it |
| 19:47.35 | ssvb | Turl: ARM11 performance is really sensitive to memory access pattern, fortunately this issue got mostly resolved in newer ARM chips :) Cortex-A9 even has automatic hardware prefetcher (admittedly somewhat less efficient than explicit software prefetch) |
| 19:48.02 | ssvb | Turl: you commented out the assembly blocks, did they fail to compile? |
| 19:48.07 | Turl | yeah |
| 19:49.01 | Turl | external/ssvb-membench/./main.c: In function 'random_read_test': |
| 19:49.02 | Turl | external/ssvb-membench/./main.c:162: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' |
| 19:49.02 | Turl | external/ssvb-membench/./main.c:162: error: 'asm' operand has impossible constraints |
| 19:49.55 | MaDMaLKaV | have to leave, seeya later |
| 19:49.57 | Turl | and after commenting that one |
| 19:49.58 | Turl | external/ssvb-membench/./main.c: In function 'random_dual_read_test': |
| 19:49.59 | Turl | external/ssvb-membench/./main.c:240: error: can't find a register in class 'LO_REGS' while reloading 'asm' |
| 19:49.59 | Turl | external/ssvb-membench/./main.c:240: error: 'asm' operand has impossible constraints |
| 19:50.13 | ssvb | Turl: I see, maybe it runs out of registers because of enabled frame pointers or something like this |
| 19:53.33 | ssvb | are you all using Mele A1000/2000? does it have serial console? |
| 19:55.00 | mnemoc | internal, but yes |
| 19:55.54 | ssvb | is it possible to buy a "development kit" including Mele A1000 and a serial cable? |
| 19:56.17 | mnemoc | ssvb: sure, from hipboi's shop on aliexpress |
| 19:58.37 | ssvb | mnemoc: yes, thanks, I have seen http://www.aliexpress.com/product-fm/532332455-Mele-A1000-TV-box-Allwinner-A10-hackable-device-wholesalers.html |
| 19:58.56 | ssvb | is "UART port connector" a standard IDC10 or something else? |
| 19:59.09 | mnemoc | no, ttl to usb |
| 19:59.38 | mnemoc | ehm |
| 20:00.00 | mnemoc | the uart port connector is 4 2.00mm pins |
| 20:00.44 | mnemoc | there you connect 3-4 jumper wires to a 3.3V ttl to usb adapter, which goes to your computer |
| 20:02.18 | ssvb | heh, a ready made cable would be the best, so that I don't need to shop for any adapters and/or do soldering |
| 20:02.23 | ssvb | is it included? |
| 20:02.44 | mnemoc | he sells one |
| 20:03.02 | CIA-91 | rhombus-tech: lkcl master * r733d77ca8e0c /allwinner_a10/hacking_the_mele_a1000.mdwn: tidyup |
| 20:03.07 | ssvb | ok, sounds good |
| 20:03.07 | j1nx | http://www.aliexpress.com/fm-store/511685/0-542914515/TTL-to-USB-debug-board.html |
| 20:03.38 | ssvb | j1nx: thanks! |
| 20:04.36 | j1nx | Is from the same "hipboi". just mention it while ordering both and he will put it nicely within the A1000 box |
| 20:05.03 | CIA-91 | rhombus-tech: lkcl master * r83ee4a71e03a /allwinner_a10/hacking_the_mele_a1000.mdwn: tidyup |
| 20:05.03 | j1nx | And is you ask him nicely, he will put the required 2.00 mm cable with it for free ;) |
| 20:05.26 | CIA-91 | rhombus-tech: lkcl master * r4c74822a23cf /allwinner_a10/hacking_the_mele_a1000.mdwn: tidyup |
| 20:06.06 | Turl | do you know if there's any chance hipboi can ship with china/hk post? |
| 20:06.07 | CIA-91 | rhombus-tech: lkcl master * r7aef9e0ddb71 /allwinner_a10/orders/fasat2112.mdwn: order processing cleanup |
| 20:12.11 | mnemoc | Turl: he usually does china post |
| 20:12.28 | mnemoc | Turl: if ali doesn't offer you that option, contact him |
| 20:13.48 | j1nx | indeed, I am in holland without shipping. He advised to select germany and mention the real address in the remarks |
| 20:14.00 | mnemoc | unlike most ali sellers he actually understands and speaks english :p |
| 20:14.48 | j1nx | and ships a hell lot faster then the alternative DX |
| 20:14.57 | j1nx | received mine after 2 weeks already |
| 20:15.04 | mnemoc | ^^ |
| 20:16.44 | mnemoc | I actually had a surprise with DX yesterday. they shipped me some stuff only 2 days after buying it |
| 20:18.08 | j1nx | you are kidding right? |
| 20:18.22 | mnemoc | ssvb: http://www.aliexpress.com/fm-store/511685/0-566678042/Mele-A1000-TV-box-debug-board-pack.html <--- this product includes both, so you save shipping $ |
| 20:18.28 | mnemoc | j1nx: no, seriously |
| 20:18.47 | j1nx | must be a very much willing item then |
| 20:19.35 | mnemoc | http://dx.com/p/cartoon-figure-water-cup-with-cover-strap-800ml-57828 |
| 20:19.55 | mnemoc | i never expected them to actually have it in stock |
| 20:20.00 | ssvb | mnemoc: great, many thanks, that's exactly what I need |
| 20:21.13 | j1nx | Hah, never knew Tom is selling it as a "kit". Smrt move |
| 20:21.58 | Turl | mnemoc: on the past DX would ship me all my stuff same day or the next day |
| 20:22.14 | Turl | now it's been a month and they still haven't shipped 2/3 of a package :) |
| 20:28.01 | j1nx | Ah they are still so much more cheaper then buying it locally. Ussually worth the waiting ;) |
| 20:28.17 | Turl | yeah |
| 20:28.45 | Turl | I got an HDMI cable for 3.60US |
| 20:29.09 | Turl | they charge you like 25 for it on a brick and mortar store |
| 20:29.46 | j1nx | I got my RPi power supply for 3,20 But never ctually bought the Rpi :) |
| 20:30.05 | Turl | haha |
| 20:30.17 | Turl | doesn't the RPi run over USB power? |
| 20:30.20 | j1nx | So now I hve to SII chargers ;) |
| 20:30.45 | Turl | the one place I'll never buy again from is ebay |
| 20:31.08 | Turl | bought a battery for my gnex and a USB controlled relay |
| 20:31.12 | Turl | neither arrived |
| 20:31.37 | Turl | the sellers had a lot of sales, diff countries, 99.99%+ ratings |
| 20:31.48 | Turl | luckily I got my money back :P |
| 20:41.29 | rm | eBay sellers typically don't send with tracking |
| 20:41.34 | rm | especially cheap items |
| 20:41.48 | rm | on Aliexpress, I think anything you buy will be shipped tracked |
| 20:42.18 | zenitraM | it's because they don't get paid until you receive it |
| 20:42.52 | zenitraM | if they don't track it they risk a lot of people saying they didn't receive them and they wouldn't get their money, with less hassles than on paypal |
| 20:43.11 | zenitraM | hell i bought a 2$ flashlight on gaga deals on aliexpress and it got shipped with tracking |
| 20:47.23 | Turl | haha |
| 20:47.38 | Turl | DX adds free tracking over 15US |
| 20:47.59 | zenitraM | yep, but if they do partial shipment which they usually do |
| 20:48.05 | zenitraM | they will ship smaller items untracked |
| 20:48.52 | Turl | yeah they shipped my HDMI cable untracked ~.~ |
| 20:49.38 | zenitraM | although sometimes untracked items arrive even faster, at least here |
| 20:50.11 | Turl | I haven't noticed such a thing |
| 20:50.19 | Turl | but they often skip customs |
| 20:50.47 | Turl | I guess customs people figure you won't ship something valuable completely untracked and complain when it gets lost :) |
| 20:55.16 | zenitraM | heh, the mele was the only thing that was caught by customs for me |
| 20:55.21 | *** join/#arm-netbook frantisheq (~frantishe@kapusany.3log.sk) |
| 20:55.31 | zenitraM | (well, they were two, and also hipboi put their full price on the customs paper) |
| 21:05.34 | Turl | how much (%) did they charge you? |
| 21:09.31 | CIA-91 | rhombus-tech: Gregory master * r04940a95ce05 /allwinner_a10/orders/miodzicho.mdwn: |
| 21:09.39 | rm | "more than one of the same thing" can cause a customs problem, yes |
| 21:13.13 | Turl | rm: here they tax you 50% on anything >25US |
| 21:15.52 | nikk | heh, we have no tax for < EUR1000 ;-) |
| 21:25.30 | hno | mnemoc, have you found any clue on what those last 64MB of the first 512MB are used for? |
| 21:26.49 | hno | or why they are removed from the accounted memory rather than reserved? |
| 21:29.23 | *** join/#arm-netbook ZaEarl (~malmrose@173-11-125-162-SFBA.hfc.comcastbusiness.net) |
| 21:40.44 | *** join/#arm-netbook sysfwlab (~sysfwlab@eta91-1-82-227-11-158.fbx.proxad.net) |
| 21:41.19 | hno | Found it. It's MALI. |
| 22:14.52 | Turl | mnemoc: are you still around? |
| 22:15.49 | *** join/#arm-netbook mmf (b1136d34@gateway/web/freenode/ip.177.19.109.52) |
| 22:33.58 | naobsd | If my memory is correct, bionic in CM9 has NEON-optimized mem* functions (from other project) |
| 22:39.16 | naobsd | hmm, my memory was wrong a bit... |
| 22:39.22 | naobsd | https://github.com/CyanogenMod/android_bionic/commit/c5e5bb2e2a458a574134b43a0e11ca016a0a6572 |
| 22:39.29 | naobsd | https://github.com/CyanogenMod/android_bionic/commit/af7b7b32c900015de8e5815120c04c3f067e0052 |
| 22:54.08 | naobsd | I have devices which have Amlogic AML8726-M. it's bit expensive than Allwinner A10, but similar cheap chinese SoC. Cortex-A9 800/1GHz w/NEON, Mali-400, USB/SATA/HDMI/SD card/etc etc |
| 22:54.39 | naobsd | if you want benchmark, I'll try |
| 22:55.46 | naobsd | similar cheap chinese SoC = no document, no source except leadked one! |
| 22:56.30 | MaDMaLKaV | I know, i leaked it to you on XDA forum PM IIRC |
| 22:56.31 | MaDMaLKaV | XD |
| 22:57.01 | naobsd | oh |
| 22:57.25 | MaDMaLKaV | I got access to their customers repository in an error but they already cutted it off. it was good the months it lastest |
| 22:57.34 | MaDMaLKaV | s/lastest/lasted |
| 22:57.43 | naobsd | glad to meet you again :) |
| 22:57.52 | MaDMaLKaV | yw |
| 22:57.53 | MaDMaLKaV | :) |
| 22:58.27 | MaDMaLKaV | unfortunately I no longer have access to their repository :/ |
| 22:59.56 | naobsd | hmm |
| 23:00.30 | naobsd | sometimes I feel pain while playing with binary blobs |
| 23:00.41 | naobsd | but I'm ok it's just my hobby :) |
| 23:00.51 | MaDMaLKaV | I'm going to bed, nite all |
| 23:00.58 | naobsd | good night |
| 23:25.51 | hno | binary blobs are evil |
| 23:26.20 | hno | and we have at leats two major areas of binary blobs to live with. |
| 23:26.41 | hno | GPU and VPU. |
| 23:33.20 | naobsd | hmm... my XDA thread is really chaostic... I can't understand newly posted question is for what :( |
| 23:34.54 | naobsd | it may be my obsolete CM9 which will be removed... it may be Quarx's CM9 but I'm not sure Quarx has interest to my thread... it may be stupid question about another rom |
| 23:36.45 | naobsd | I have no idea what should I do ;) |
| 23:37.56 | naobsd | if Quarx has no interest to that thread, it should not be used to discuss his work :( |
| 23:41.45 | naobsd | s/to discuss/to ask/ |
| 23:42.10 | naobsd | oh, nice bot... |
| 23:45.47 | Turl | yeah :) I need to implement that on mine |
| 23:46.00 | Turl | wonder if.... |
| 23:46.07 | Turl | s//a/g |
| 23:46.22 | Turl | s/./a/g |
| 23:46.34 | Turl | :P it probably doesn't use a regex |
| 23:52.16 | naobsd | hehe |
| 23:52.53 | naobsd | Turl: can you talk with Quarx? I want to know he want to use my XDA thread or not |
| 23:54.08 | naobsd | if no, I'll recommend people to use another thread to ask his work |
| 23:54.41 | naobsd | probably slatedroid is better. I noticed his recent post |
| 23:55.16 | *** join/#arm-netbook arokux_h (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 23:55.25 | Turl | naobsd: he is offline now |
| 23:55.44 | naobsd | Turl: I see |
| 23:57.12 | naobsd | sigh. I can't get enough time to try his CM9 yet :( |
| 23:58.20 | naobsd | so far I only tried to compile leaked allwinner kernel source lol |
| 23:58.48 | naobsd | I want to check difference between AOSP and allwinner android source |
