| 00:00.33 | naobsd | then I want to try mnemoc's v2 too |
| 00:27.25 | hno | added a u-boot wiki page in github. |
| 01:40.39 | *** join/#arm-netbook ZaEarl (~malmrose@75.144.245.226) |
| 01:46.54 | *** join/#arm-netbook nikk (~somebody@nikk.starlink.ru) |
| 02:17.58 | *** join/#arm-netbook hipboi_ (~hipboi@123.89.120.50) |
| 02:33.04 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 03:04.43 | CIA-91 | rhombus-tech: David master * r9da1e240a936 /allwinner_a10/hacking_the_mele_a1000.mdwn: |
| 03:32.38 | *** join/#arm-netbook t0dbld (~t0dbld@unaffiliated/t0dbld) |
| 03:40.11 | *** join/#arm-netbook DonkeyHotei (~danielg4@50-0-149-248.dsl.dynamic.sonic.net) |
| 04:05.13 | *** join/#arm-netbook mikey_w (~mike@c-71-63-115-202.hsd1.va.comcast.net) |
| 04:15.58 | *** join/#arm-netbook ZaEarl (~malmrose@75.144.245.226) |
| 05:20.26 | CIA-91 | rhombus-tech: nil master * r467256fe6953 /allwinner_a10/hacking_the_mele_a1000.mdwn: mkimage requirement for building uimage |
| 06:36.16 | *** join/#arm-netbook RITRedbeard (~RITReadbe@c-68-37-165-37.hsd1.nj.comcast.net) |
| 06:36.36 | RITRedbeard | Anyone have the MK802? |
| 07:23.23 | *** join/#arm-netbook gimli (~gimli@178.73.212.231) |
| 07:23.31 | *** join/#arm-netbook gimli (~gimli@xbmc/staff/gimli) |
| 07:42.48 | *** join/#arm-netbook kaspter (~kaspter@116.237.61.223) |
| 07:46.21 | hp_ | RITRedbeard: i orderd one |
| 07:46.58 | RITRedbeard | Can you run bencharks for compile, opengl es stuff and battery life when you get it? |
| 07:53.23 | RITRedbeard | I think this is going to be my new hobby |
| 08:01.59 | RITRedbeard | hp_, any idea when they might be back in stock? |
| 08:02.31 | hp_ | RITRedbeard: there is also this other link.. rm showed me yesterday |
| 08:03.10 | RITRedbeard | PM? |
| 08:03.59 | hp_ | RITRedbeard: no i was just looking it up in my history on chrom |
| 08:04.00 | hp_ | http://www.asiapads.com/product_info.php?products_id=2246 |
| 08:04.08 | hp_ | but its also soldout there |
| 08:06.07 | RITRedbeard | is this a new device? has anyone had success slapping linux on there? |
| 08:09.51 | CIA-91 | rhombus-tech: Felix master * r0e85aef1f42d /allwinner_a10/orders/fixmeyer__64__gmail.com.mdwn: |
| 08:12.36 | nikk | fuf, successfully compiled CM9 and kernel 3.0.31 (from V2 branch), but tablet won't boot on it ( ... |
| 08:13.40 | nikk | it even doesn't leave uBoot image screen |
| 08:14.33 | RITRedbeard | It might seem crazy but I want open-ish source hardware to build ARM ultraportable with |
| 08:26.55 | RITRedbeard | Not free as in beer but free as in speech |
| 08:27.11 | RITRedbeard | I think Raspberry Pi Foundation and their fans really miss the point |
| 08:27.12 | naobsd | nikk: how about replacing only kernel on working system |
| 08:27.46 | nikk | naobsd: i'm in opposite situation. with packed stock kernel it works |
| 08:27.50 | DonkeyHotei | RITRedbeard: how so? they say the EOMA68 card misses the point |
| 08:28.12 | RITRedbeard | because for those people free as in beer is just a vehicle, a means to an end |
| 08:28.13 | nikk | naobsd: just pot it there in recovery and cat ramdisk to nandc |
| 08:28.18 | RITRedbeard | they don't care about free as in speech |
| 08:28.24 | nikk | and it boots |
| 08:29.03 | DonkeyHotei | RITRedbeard: do they have a binary-only video driver? |
| 08:29.14 | RITRedbeard | yep |
| 08:29.26 | DonkeyHotei | non-mali? |
| 08:29.49 | nikk | naobsd: can't find right way to make ramdisk for CM9. even after unpack / repack working one it doesn't boot. dunno why... |
| 08:29.55 | RITRedbeard | it's proprietary... it's Broadcom based |
| 08:29.58 | RITRedbeard | Powercore IV |
| 08:32.03 | naobsd | nikk: which tool and command line did you use to unpack/repack? |
| 08:33.28 | naobsd | --cmdline "console=ttyS0,115200 rw init=/init loglevel=8" --base 0x40000000 --pagesize 2048 |
| 08:33.57 | nikk | naobsd some perl scripts and mkbootfs mkbootimg... |
| 08:34.05 | naobsd | this is what I'm using. it should be same as stock ICS rom for A10 tablets |
| 08:34.25 | *** join/#arm-netbook lerc_ (~quassel@121-74-254-104.telstraclear.net) |
| 08:34.34 | naobsd | ah, I didn't use any scripts in allwinner source so I don't know about them |
| 08:34.51 | naobsd | I'm using mkbootimg in CM9 |
| 08:35.16 | naobsd | unpackbootimg in CM9 for unpacking |
| 08:35.50 | nikk | does CM9 need tweaking for that args ^^ --cmdline etc? |
| 08:36.50 | naobsd | nikk: I don't know these args are really used on A10 device(bootloader). I just use same values as stock rom |
| 08:37.20 | naobsd | I don't have any reason to change them |
| 08:38.16 | naobsd | nikk: please try: 1. get boot.img from stock rom(dump from working device or unimg.exe to unpack stock rom) 2. unpack boot.img by unpackbootimg in CM9 3. repack just unpacked files with same args |
| 08:38.19 | nikk | naobsd I desided to test 3.0.31 with not CM9 but with 4.0.3, as it provides working boot.img on build for me before (with self-compiled 3.0.8 kernel) for test. If it also will not boot, so the kernel is the issue. |
| 08:39.44 | naobsd | nikk: my suggestion is to use confirmed procedure. any way is ok if it's confirmed :) |
| 08:42.45 | naobsd | ah, you are sure that the procedure to build rom from allwinner 4.0.3 source is ok, but |
| 08:42.53 | nikk | naobsd: unpacked with -p 2048. BOARD_PAGE_SIZE 00000800 |
| 08:42.59 | nikk | in output |
| 08:43.29 | naobsd | but you can't use same procedure to build CM9 |
| 08:43.36 | naobsd | hmm |
| 08:44.51 | nikk | ls BOOT |
| 08:44.51 | nikk | boot.img-base boot.img-cmdline boot.img-pagesize boot.img-ramdisk.gz boot.img-zImage |
| 08:45.01 | nikk | how to pack it back? |
| 08:46.21 | naobsd | mkbootimg --kernel boot.img-zImage --ramdisk boot.img-ramdisk.gz --cmdline `cat boot.img-cmdline` --base `cat boot.img-base` --pagesize `cat boot.img-pagesize` --output boot-new.img |
| 08:47.03 | naobsd | I want to see cmdline/base/pagesize |
| 08:47.22 | nikk | BOARD_KERNEL_CMDLINE console=tty0,115200 rw init=/init loglevel=3 |
| 08:47.22 | nikk | BOARD_KERNEL_BASE 40000000 |
| 08:47.22 | nikk | BOARD_PAGE_SIZE 00000800 |
| 08:48.07 | nikk | but it's not stock boot.img it's from quark2k git (made for CM9) |
| 08:48.11 | nikk | which works |
| 08:50.51 | naobsd | oh |
| 08:51.06 | naobsd | BOARD_ prefix is really exist in boot.img-* files? |
| 08:51.13 | naobsd | which tool did you use to unpack? |
| 08:52.05 | naobsd | anyway, specify --cmdline "console=tty0,115200 rw init=/init loglevel=3" --base 0x40000000 --pagesize 2048 |
| 08:53.02 | naobsd | 0x prefix is important for base and pagesize need to be decimal |
| 08:53.31 | naobsd | (if these values are really used by bootloader, it's important) |
| 08:53.51 | nikk | unpacked with which unpackbootimg |
| 08:54.06 | naobsd | oops |
| 08:54.28 | nikk | unpacked with which unpackbootimg |
| 08:54.28 | nikk | it's in /Volumes/CM9/out/host/darwin-x86/bin/unpackbootimg |
| 08:54.47 | naobsd | ah |
| 08:55.07 | naobsd | I guess you pasted printed line to console |
| 08:55.29 | naobsd | boot.img-* files should have just values |
| 08:55.51 | naobsd | paste |
| 08:56.31 | naobsd | sorry, I'm busy |
| 08:57.34 | nikk | np |
| 09:01.24 | nikk | mkbootimg --kernel boot.img-zImage --ramdisk boot.img-ramdisk.gz --cmdline "console=tty0,115200 rw init=/init loglevel=3" --base 0x40000000 --pagesize 2048 -o new-boot.img |
| 09:01.48 | nikk | new-boot.img goes to /dev/block/nandc and tablet doesn't boot |
| 09:02.12 | nikk | goes to recovery to restore boot to working one |
| 09:02.32 | *** join/#arm-netbook arokux_h (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 09:04.47 | CIA-91 | rhombus-tech: ribuck master * r7410a00544aa /allwinner_a10/orders/RichieB.mdwn: Would like more information about suitable header boards. Will Phombus Release their own header board ? |
| 09:07.19 | CIA-91 | rhombus-tech: ribuck master * ree72e0eadcbe /allwinner_a10/orders/RichieB.mdwn: |
| 09:07.37 | nikk | with original one (boot.img) it boots fine again... can't understand what's wrong.... |
| 09:09.30 | *** join/#arm-netbook kaspter (~kaspter@116.237.61.223) |
| 09:10.41 | *** join/#arm-netbook fjfalcon (~fjfalcon@fjfalcon.ru) |
| 09:11.04 | CIA-91 | rhombus-tech: lkcl master * r59bb65ebb352 /allwinner_f20.mdwn: add allwinner f20 page |
| 09:11.04 | CIA-91 | rhombus-tech: lkcl master * r80b83411eadd /: Merge branch 'master' of ssh://git.hands.com/rhombus |
| 09:14.28 | CIA-91 | rhombus-tech: ribuck master * ra4a740442d4f /allwinner_a10/orders/RichieB.mdwn: |
| 09:15.31 | CIA-91 | rhombus-tech: ribuck master * rc33c08fde452 /allwinner_a10/orders/RichieB.mdwn: |
| 09:21.47 | naobsd | minor correction: boot.img-cmdline doesn't have 0x prefix. `cat boot.img-cmdline` was wrong ;) |
| 09:23.14 | naobsd | nikk: original boot.img and repacked new-boot.img with original files must be identical. please compare |
| 09:23.27 | nikk | nope, they not... |
| 09:23.42 | nikk | even 2 generated not identical thou... |
| 09:23.50 | naobsd | oops |
| 09:24.06 | nikk | in size - yes, in contents - no |
| 09:25.43 | naobsd | unpacked files are same if you unpacked twice? |
| 09:26.05 | nikk | i inspected it in araxis merge ;-) nope. in size all 3 equals |
| 09:26.07 | nikk | ls -la ../device/allwinner/novo7a/boot.img |
| 09:26.07 | nikk | -rw-r--r-- 1 nikk staff 8718336 18 май 10:30 ../device/allwinner/novo7a/boot.img |
| 09:26.26 | nikk | ls -la |
| 09:26.27 | nikk | total 51112 |
| 09:26.27 | nikk | drwxr-xr-x 2 nikk staff 306 20 май 12:58 . |
| 09:26.27 | nikk | drwxr-xr-x 30 nikk staff 1258 20 май 12:40 .. |
| 09:26.27 | nikk | -rw-r--r-- 1 nikk staff 8718336 20 май 12:52 boot-new.img |
| 09:26.27 | nikk | -rw-r--r-- 1 nikk staff 9 20 май 13:13 boot.img-base |
| 09:26.27 | nikk | -rw-r--r-- 1 nikk staff 45 20 май 13:13 boot.img-cmdline |
| 09:26.36 | nikk | -rw-r--r-- 1 nikk staff 9 20 май 13:13 boot.img-pagesize |
| 09:26.37 | nikk | -rw-r--r-- 1 nikk staff 618169 20 май 13:13 boot.img-ramdisk.gz |
| 09:26.38 | nikk | -rw-r--r-- 1 nikk staff 8095748 20 май 13:13 boot.img-zImage |
| 09:26.39 | nikk | -rw-r--r-- 1 nikk staff 8718336 20 май 13:16 new-boot.img |
| 09:27.45 | nikk | original - boots, new 2 - no ;-) |
| 09:28.37 | nikk | mayby it's my darwin not good compiled mkbootimg? |
| 09:28.49 | naobsd | maybe |
| 09:29.09 | naobsd | hmm |
| 09:29.27 | nikk | i dunno which tool used in 3.0.8 source, but boot from there was working one... |
| 09:29.28 | naobsd | do you have another mkbootimg which is used to build allwinner images? |
| 09:30.34 | nikk | aand now it compiles 3.0.8 with 3.0.31 kernel... even with other compiler (gcc-4.2) |
| 09:31.38 | nikk | cm9 uses llvm-gcc-4.2 |
| 09:31.39 | naobsd | I don't have working/usable Mac OS X machine |
| 09:32.28 | nikk | i have ubuntu also, but it's partition too small ) |
| 09:32.30 | naobsd | how about installing ubuntu to VM ;) |
| 09:32.35 | naobsd | mmm |
| 09:32.49 | nikk | 30GB only there ) |
| 09:33.23 | naobsd | ah, only unpackbootimg and mkbootimg are necessary for now :) |
| 09:33.57 | nikk | yup |
| 09:34.25 | nikk | i'll check it there. i bet it's compiler ))) |
| 09:36.09 | nikk | there are only 13GB left on my 1TB drive *-) |
| 09:36.13 | naobsd | I hate non-portable code too ;) |
| 09:36.17 | naobsd | lol |
| 09:53.13 | CIA-91 | rhombus-tech: Alejandro master * r8fa8afdc82fc /allwinner_a10/kernel_compile.mdwn: Updated to mnemoc's kernel tree. |
| 10:13.09 | *** join/#arm-netbook hipboi_ (~hipboi@222.50.105.122) |
| 10:55.38 | gimli | https://github.com/amery/allwinner-a10-video is realy usefull for video decoding :D |
| 11:19.24 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 11:26.20 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 11:36.19 | hno | gimli, you got it working? |
| 11:42.07 | gimli | i work on it ;) |
| 11:48.41 | Gumboot | RITRedbeard: I think you mean "Videocore". |
| 11:48.54 | Gumboot | And why does everybody always talk about "the point" like they're the only one who knows what it is. |
| 11:49.06 | Gumboot | ? |
| 11:58.36 | *** join/#arm-netbook Charlie (~quassel@64.31.59.70) |
| 12:00.43 | *** join/#arm-netbook arian_ (~arian@77-58-105-100.dclient.hispeed.ch) |
| 12:21.05 | *** join/#arm-netbook arokux_h (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 12:22.40 | hno | Gumboot, because everyone talks about their view. But I have to agree that R-Pi do miss the point, but A10 is pretty much in a similar boat. |
| 12:23.03 | hno | but better in all other aspects. |
| 12:29.46 | penguin42 | remember that a point is very local and very narrow |
| 13:33.28 | Gumboot | What point has been missed? |
| 14:28.44 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 14:39.51 | CIA-91 | rhombus-tech: Battle master * r01f691644435 /allwinner_a10/orders/battle.wesnoth.mdwn: |
| 14:57.41 | A2Sheds | http://www.cnx-software.com/2012/04/16/sunlike-umpc-1021-85-usd-allwinner-a10-based-10-2-android-4-0-netbook/ |
| 14:58.06 | A2Sheds | they used the same junk enclosure as some other awful arm netbooks |
| 14:58.34 | A2Sheds | same junk case as this http://www.aliexpress.com/product-fm/533639478-New-arrival-10-inch-EPC-laptop-Android-2-3-OS-10-Netbook-Rockchip-CPU-10-2inch-wholesalers.html |
| 14:58.52 | A2Sheds | for $5 more they could have a great case |
| 14:59.59 | mnemoc | cheap-thinking :( |
| 15:00.21 | A2Sheds | the touchpad is just a form in the top of the aluminum shell |
| 15:00.32 | mnemoc | o_O |
| 15:00.55 | A2Sheds | and it almost works |
| 15:01.36 | A2Sheds | "almost works" has become the new "works" |
| 15:06.11 | A2Sheds | http://www.google-android-tablet-pcs.com/diversed-amlogic-a9-android-netbook-released.html |
| 15:08.02 | A2Sheds | http://www.laptopspec.net/2011/01/shenzhen-pc101v-android-super-slim-netbook/ |
| 15:12.13 | nikk|zzz | naobsd hm, it definitly was mkbootimg ... |
| 15:12.42 | nikk|zzz | the same unpacked files on linux gust produced working boot image |
| 15:13.13 | nikk|zzz | 8-))) |
| 15:14.37 | nikk|zzz | wonder how many other binaries do their stuff wrong |
| 15:15.42 | Byan | so, with this allwinner devices running ubuntu, how do we play video? |
| 15:16.04 | Byan | do we have an mplayer patch that supports acceleration or anything? |
| 15:25.21 | *** join/#arm-netbook Hexxeh (u1532@gateway/web/irccloud.com/x-avdfalrljewzrtds) |
| 15:25.57 | nikk | naobsd: the strange thing is, mkbootimg, compiled in osx with 4.0.3 android source works fine too. the only buggy one in CM9 |
| 15:57.32 | *** join/#arm-netbook Brandon15811 (~Brandon15@ip72-207-255-84.br.br.cox.net) |
| 16:18.38 | hno | mnemoc, have uploaded tools with sun5i restored to my github. |
| 16:20.27 | hno | Byan, none yet. |
| 16:21.50 | Byan | so no video decoding in gnu/linux? |
| 16:21.55 | Byan | fffff |
| 16:21.58 | Byan | oops |
| 16:22.24 | hno | Gumboot, <RITRedbeard> Not free as in beer but free as in speech |
| 16:22.59 | hno | Byan, sure ther is, just nothing using it. |
| 16:23.40 | hno | s/er/ere/ |
| 16:24.28 | mnemoc | hno: only for reference or to be turned into a BSP? |
| 16:25.34 | hno | mnemoc, for now it's their tools tree only with sun5i undeleted. |
| 16:25.54 | hno | do not really plan on using it, but need some info from there. |
| 16:26.57 | hno | or.. will probably use it a bit to collect what is needed on the differences from sun4i. |
| 16:27.52 | mnemoc | thanks, I was unsure about the legal implications of adding those in sunxi-tools |
| 16:27.56 | nikk | mnemoc: hm, 3.0.31+ boots ;-) but I was have trouble booting it with framebuffer console enabled (screen goes black to pixe white on display init) |
| 16:28.17 | mnemoc | nikk: fbcon is not supported (yet) |
| 16:28.27 | nikk | i see ;-) |
| 16:28.51 | mnemoc | the fb driver misses several needed callbacks and tables for that |
| 16:29.12 | mnemoc | they only implemented what was needed to run android |
| 16:30.05 | nikk | thats why it wont boot initialy ;-) reverted to stock kernel options and now i'm on 3.0.31+ |
| 16:30.15 | nikk | can't wait to test it in CM9 ;-) |
| 16:30.37 | nikk | it compiled also ) |
| 16:31.56 | nikk | hm |
| 16:32.21 | nikk | seems like some problems with additional wifi modules (nano_wifi) |
| 16:32.23 | nikk | D/WifiHW ( 162): begin insmod nano wifi fw. |
| 16:32.23 | nikk | I/WindowManager( 162): Lock screen displayed! |
| 16:32.23 | nikk | D/WifiHW ( 162): faied to read proc/net/wireless |
| 16:32.40 | nikk | and it won't come up :( |
| 16:34.04 | A2Sheds | http://www.ibtimes.com/articles/343028/20120519/arm-intel-battle-heats-up.htm |
| 16:34.51 | nikk | seems like driver loaded correctly, but wi-fi won't come up. nothing changed except kernel, with 3.0.8 it was working fine... http://pastebin.com/iW4guv88 |
| 16:36.54 | nikk | http://pastebin.com/8zair2CE |
| 16:37.04 | zenitraM | [ 316.120000] [nano] read mib file: /system/vendor/modules/mib.bin failed |
| 16:37.13 | nikk | it was olways there |
| 16:37.24 | nikk | in 3.0.8 it's the same output |
| 16:37.31 | nikk | and it doesn't matter |
| 16:40.06 | nikk | thought he need to by USB wi-fi and forget about that buggy noano-wifi |
| 16:42.56 | mnemoc | nikk: patches to turn modules/wifi/nano into a proper (and updated) drivers/wireless/nano are welcomed :) |
| 16:44.15 | nikk | mnemoc i have only for sun4i-ir.ko yet ;-) |
| 16:44.36 | mnemoc | uh |
| 16:45.44 | zenitraM | iirc nano required two modules to be loaded in a specific order and one of them with a parameter pointing to something |
| 16:46.00 | zenitraM | this is, on 2.6.36 and with a ubuntu rootfs |
| 16:46.06 | CIA-91 | rhombus-tech: sysfw master * r1df939e9c7ae /allwinner_f20.mdwn: Reformating and adding information, i need to adjust image link |
| 16:46.34 | nikk | but probably it's not widespread ;-) it doublepressed all buttons of my Creative SB Live IR remote, so I found a bug and fixed it (on 2.6.36) |
| 16:47.06 | hno | mnemoc, my future plans fot A10 do not actually involve Allwinner A10 tools at all. here is no need for them unless wanting to flash Allwinner android images. |
| 16:47.07 | penguin42 | what's the current guess on availability of the A10 EOMA-68 hardware? |
| 16:47.08 | nikk | zenitraM yeah, i check it now, thx |
| 16:47.10 | CIA-91 | rhombus-tech: sysfw master * r370aa88eb15d /allwinner_f20.mdwn: |
| 16:47.18 | hno | penguin42, when it's done. |
| 16:47.40 | penguin42 | hno: As I said, current guess - not wanting a promise |
| 16:47.44 | mnemoc | hno: I was actually only interested in the .fex files |
| 16:47.51 | hno | mnemoc, same here. |
| 16:48.13 | mnemoc | they have lots of knowledge on them |
| 16:48.37 | hno | worth noting is that those (SD + NAND) are also included in firmware files when there is one released. |
| 16:49.28 | hno | I think I only need 1-1 & 1-2 fex files, but not sure yet. |
| 16:49.44 | zenitraM | i wonder why there are separate "linux" and "crane" .fex |
| 16:50.11 | mnemoc | crane is more.... android |
| 16:50.26 | zenitraM | yes, but why also a linux one |
| 16:50.27 | hno | there is many different kinds of fex files in the tools. |
| 16:50.34 | nikk | zenitraM yahoooo! just put missed x_max.axf to moduled (why it's not in firmware) and voila: http://pastebin.com/XTeUCvAm |
| 16:50.59 | nikk | wi-fi is on |
| 16:51.46 | mnemoc | nikk: as luke would say.... please write it on the wiki :) |
| 16:52.14 | hno | there is many fex files of different kinds. Wonder what .fex actually stands for. |
| 16:53.16 | nikk | mnemoc how kernel detect firmware dir in android version? i just discovered in 2.3.4 in was /system/etc/firmware (put my DVB-T firmware to many locations till finded correct one), and i'm not sure about ICS, cause other files (for touchscreen) goes to /vendor/firmware/ |
| 16:53.18 | mnemoc | I notice they use .fex for two different kind of ini-like files (fine), but also for fs images :< |
| 16:54.21 | mnemoc | nikk: it's not the kernel who loads the firmwares, it's the hotplug tool |
| 16:54.25 | mnemoc | in userspace |
| 16:54.53 | mnemoc | kernel requests, udev (e.g.) loads it |
| 16:54.55 | nikk | didn't know android have one ;-) |
| 16:55.02 | hno | a13 linux & crane sys_config.fex barely differs at all. |
| 16:55.12 | hno | > pll4_freq = 960 |
| 16:55.12 | hno | > pll6_freq = 960 |
| 16:55.21 | hno | those two are only in crane. |
| 16:55.23 | mnemoc | nikk: maybe their init does the whole thing? |
| 16:56.00 | hno | for USB stuff they need some module loading I think. |
| 16:56.22 | nikk | mnemoc probably so, it also have ueventd and config script for it |
| 16:56.22 | zenitraM | the init.sun4i.rc does insmod some modules, including lcd, hdmi, etc |
| 16:56.24 | mnemoc | implementations differ, but they still need a userpsace helper |
| 16:56.30 | hno | but standard hardware is loaded by modprobe in system startup. |
| 16:56.39 | penguin42 | I thought there was a kernel firmware loader these days |
| 16:56.42 | zenitraM | no, they couldn't have compiled the kernel with them built-in... |
| 16:57.29 | nikk | goest to CM9 build |
| 16:58.48 | nikk | 4.0.3 seems work fine with 3.0.31, all modules loaded |
| 16:59.00 | hno | err.. pll4 & 6 are only configured in the linux sys_config.fex, not crane. And pll4 is apparently the MALI clock source. |
| 16:59.06 | hno | odd. |
| 16:59.11 | nikk | time to test 4.0.4 CM edition |
| 17:01.25 | mnemoc | penguin42: right, drivers/base/Kconfig:72:config FW_LOADER |
| 17:02.01 | mnemoc | and EXTRA_FIRMWARE_DIR |
| 17:02.30 | mnemoc | I had forgotten that |
| 17:02.44 | nikk | in case android 2.3.4 it's /system/etc/firmware (missed in Ainol's FM) |
| 17:03.10 | nikk | you need to create that dir and put firmware there |
| 17:03.22 | nikk | found it when loaded DVB-T stick |
| 17:05.58 | mnemoc | wiki wiki :) |
| 17:10.13 | nikk | i promise i'll post some guide there (about building CM9), when it'll be done (leaked 4.0.3 not good enough ;-) |
| 17:14.30 | *** join/#arm-netbook rell01 (~chatzilla@p4FE56282.dip0.t-ipconnect.de) |
| 17:24.25 | nikk | mnemoc khm, can bImage be replaced with zImage in case of android boot from ramdisk? |
| 17:24.40 | nikk | why they use uncompressed one? |
| 17:27.02 | CIA-91 | rhombus-tech: hno master * ref1734d16e38 /allwinner_f20.mdwn: |
| 17:32.28 | mnemoc | nikk: I assume it's because decompression is expensive (time and battery).... |
| 17:32.53 | nikk | i see. but it'll work also, yeah? |
| 17:33.05 | mnemoc | don't know, try |
| 17:33.11 | mnemoc | but I feel it won't work |
| 17:34.00 | zenitraM | i think it didn't |
| 17:53.36 | nikk | zenitraM do you know right way to do boot.img by hand, if i have root folder with all contents and kernel? |
| 17:54.05 | zenitraM | it was with mkbootimg iirc |
| 17:54.26 | zenitraM | google "create boot.img android" and you'll find it, it's a "generic" procedure |
| 17:55.23 | nikk | in which format must be mkbootimg --ramdisk <filename> |
| 17:55.40 | nikk | all other parts I understand clearly |
| 17:57.11 | nikk | i have all files, but images, generated by CM9 won't boot, dunno why... some weird bug in my mkbootimg... |
| 18:07.50 | hno | nikk, kernel uimage is uncomressed because it containz an compressed bzimage. |
| 18:08.18 | hno | compressing twice only adds decompression time. |
| 18:09.26 | hno | ah, you are talking about Android ICS boot images. Ignore me then. |
| 18:11.23 | hno | anyone else measured power draw of the A10? |
| 18:17.31 | *** join/#arm-netbook A0Sheds (~ly@unaffiliated/l84supper) |
| 18:18.30 | mnemoc | hno: < 3W :D amazing |
| 18:40.21 | hno | mine measured with a multimeter. hope it's correct. |
| 18:41.08 | DonkeyHotei | to measure power draw with a multimeter, you need to break the circuit |
| 18:41.27 | hno | yes. |
| 18:42.21 | hno | it's powered by loose wires. |
| 18:42.33 | DonkeyHotei | ah |
| 18:43.12 | hno | didn't have a suitable AC outlet adapter, so cut the cable and fit it on the connector of another 5V supply. |
| 18:46.11 | DonkeyHotei | for some odd reason, i thought you meant power draw of just the chip itself rather than the whole device |
| 18:48.29 | hno | no, it's the whol device. But there isn't much else using noticeable amounts of power I think. |
| 18:49.58 | nikk | ( modified ramdisk with success. now tab loaded with 2.0.31+ (from V2). ump.ko and mali.ko modules loaded fine. but there is no display init in android. stacktraced: http://pastebin.com/sDK2v1vu |
| 18:50.13 | nikk | it worked with 3.0.8 thou |
| 18:52.19 | nikk | W/libEGL ( 78): 0: eglInitialize(0x1) failed (EGL_BAD_ALLOC) |
| 18:52.56 | zenitraM | nikk: did you modprobe disp, modprobe lcd? |
| 18:53.12 | nikk | they inside kernel as i understand |
| 18:53.28 | zenitraM | at least on the original kernel they are built as modules |
| 18:53.39 | zenitraM | and insmodded at boot |
| 18:53.59 | zenitraM | (this is, on 4.0 firmwares) |
| 18:54.06 | nikk | all compiled in. i checked this, it's config from 4.0.3 (v1.0) |
| 18:54.35 | nikk | ~ # lsmod |
| 18:54.35 | nikk | Module Size Used by Not tainted |
| 18:54.35 | nikk | goodix_touch 12178 0 |
| 18:54.35 | nikk | rtl8150 9190 0 |
| 18:54.35 | nikk | mcs7830 6605 0 |
| 18:54.35 | nikk | qf9700 8775 0 |
| 18:54.45 | nikk | asix 17998 0 |
| 18:54.46 | nikk | usbnet 23901 3 mcs7830,qf9700,asix |
| 18:54.47 | nikk | sun4i_csi0 25920 0 |
| 18:54.48 | nikk | gt2005 15207 1 |
| 18:54.49 | nikk | mma7660 6423 0 |
| 18:54.49 | nikk | videobuf_dma_contig 5744 1 sun4i_csi0 |
| 18:54.50 | nikk | videobuf_core 18944 2 sun4i_csi0,videobuf_dma_contig |
| 18:54.51 | nikk | mali 131360 0 |
| 18:55.01 | zenitraM | try yes >/dev/fb0 |
| 18:55.01 | nikk | ump 41936 1 mali |
| 18:55.02 | nikk | sun4i_vibrator 2278 0 |
| 18:55.04 | nikk | even ump and mali loaded |
| 18:55.11 | mnemoc | hno: http://www.aliexpress.com/product-fm/467538897-Hot-Sale-free-shipping-EU-Version-Power-Balance-Energy-Meter-Monitor-Electricity-Test-Equipment-wholesalers.html <--- these are great :) |
| 18:55.56 | zenitraM | ump and mali don't check for disp/lcd |
| 18:56.01 | nikk | zenitraM: echo "yes" > /dev/fb0 ? nothing happens |
| 18:56.19 | zenitraM | ls /dev/fb* |
| 18:56.20 | hno | mnemoc, yes, but more interested in the DC side of things not including overhead of a random chineese AC/DC converter. |
| 18:56.57 | nikk | # ls /dev/fb* |
| 18:57.20 | hno | plus that I have the 5V lead open on the desk.. |
| 18:57.23 | nikk | - /dev/fb0 exists |
| 18:57.24 | zenitraM | is the screen turned on? as in, the backlight |
| 18:57.25 | nikk | yes >/dev/fb0 |
| 18:57.25 | nikk | yes: No space left on device |
| 18:57.34 | nikk | zenitraM: yes |
| 18:57.34 | zenitraM | ls -l /dev/fb0 |
| 18:57.49 | nikk | zenitraM it tries OpenGL and fails |
| 18:58.06 | nikk | -rw-rw-rw- 1 root root 168898560 Jan 4 14:17 /dev/fb0 |
| 18:58.19 | nikk | backlight on |
| 18:58.22 | zenitraM | no, /dev/fb0 doesn't exist |
| 18:58.50 | zenitraM | in fact you just created a 160mb file of "y" instead of writing to the framebuffer |
| 18:59.10 | nikk | i see |
| 18:59.19 | zenitraM | although.. mmm check in /dev/graphics/ or something like that |
| 18:59.23 | zenitraM | cd /dev; find . -name fb0 |
| 18:59.30 | zenitraM | (and rm that fb0 file) |
| 19:00.03 | penguin42 | cat /proc/fb should show if the kernel knows of a fb shouldn't it? |
| 19:00.18 | *** part/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 19:00.22 | zenitraM | mm it should, try that too |
| 19:00.32 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 19:01.02 | nikk | rebooted: fb in dev /graphics/fb0 |
| 19:01.22 | zenitraM | yes >/dev/graphics/fb0 |
| 19:01.30 | zenitraM | it should turn your screen purple |
| 19:01.35 | nikk | - /proc/fb: 0 to 7 |
| 19:02.10 | nikk | yes, i see purple ;-) some kind |
| 19:02.20 | nikk | but now disappeared ;-) |
| 19:02.23 | zenitraM | mmm it could be |
| 19:02.33 | zenitraM | the libEGL_mali.so libraries not matching the mali kernel driver |
| 19:02.43 | zenitraM | s/driver/module version/ |
| 19:03.42 | nikk | hm. there are some variants, but i never defined which one to build |
| 19:04.21 | nikk | just made ./build -p su4i_crane with new kernel tree 3.0.31 or v2 |
| 19:05.10 | nikk | and all will be fine if i'll replace kernel and modules to ones from 3.0.8 |
| 19:05.22 | nikk | it was working before |
| 19:06.47 | nikk | the strange thing i can use the same 3.0.31 with Android 4.0.3, builded from source also. this trouble only in CM9 |
| 19:07.35 | nikk | What I was done: compiled 4.0.3 with new kernel and modules - that works fine (opengl etc) |
| 19:08.36 | nikk | After, I tried build CM9 with it - CM9 builds, but with non-bootable ramdisk. so, i repacked it by hand with new kernel and pushed new modules to FS. that give me the error. |
| 19:08.58 | zenitraM | the .so modules are the same on both? |
| 19:08.59 | nikk | If i repack again with stock CM9 kernel and modules - it'll boot |
| 19:09.11 | nikk | no, it's 2 different sources |
| 19:09.37 | zenitraM | try replacing the libEGL_mali.so and all mali files with the ones with 4.0.3 |
| 19:09.39 | nikk | one from Qware (leaked), another - Quark2K git CM9 |
| 19:17.26 | nikk | zenitraM heh, it did the trick |
| 19:17.34 | nikk | zenitraM: u r my hero ;-) |
| 19:17.43 | nikk | I have working CM9 ;-) |
| 19:17.46 | nikk | yahoooo |
| 19:18.07 | nikk | "almost working" ;-) |
| 19:18.48 | zenitraM | yipee :D |
| 19:19.21 | nikk | replaced this ones: libMali libUMP and all in /egl |
| 19:19.36 | nikk | as it was in 4.0.3 source for ainol |
| 19:21.38 | *** join/#arm-netbook gimli (~gimli@xbmc/staff/gimli) |
| 19:22.20 | nikk | still need to build wi-fi modules, change config etc, setup accelerometer, but in general, it works ;-) |
| 19:23.00 | nikk | thanks guys! |
| 19:26.51 | zenitraM | you're welcome :) what device are you working on? |
| 19:27.04 | nikk | novo 7 advanced |
| 19:29.27 | nikk | now i'm think about hyundai A7HD ;-) |
| 19:29.34 | zenitraM | oh, nice :) i'll try it on mine once i finish exams then :P |
| 19:30.36 | nikk | zenitraM it's much to be done, i plan to add FM, DVB, GPS and USB audio support there ;-) |
| 19:30.50 | *** join/#arm-netbook t0dbld|work (~t0dbld@unaffiliated/t0dbld) |
| 19:31.07 | nikk | as it was done for 2.3 ) |
| 19:33.05 | nikk | zzzz |
| 19:36.12 | Gumboot | hno: Can you really miss something if you're not aiming at it? |
| 19:42.55 | hno | nikk, there is 4.0.3 sources for ainol? |
| 19:43.51 | nikk | hno: in qware sources, there is crane-aino (Novo7A) |
| 19:43.57 | nikk | in device section |
| 19:44.31 | nikk | with all the stuff (keys mapings camera configs etc) |
| 19:45.24 | nikk | but i bet it's little old (about v.0.7 4.0.3 release) |
| 19:46.16 | nikk | i taked wi-fi driver to CM9 from there |
| 19:46.38 | nikk | wi-fi.c i mean |
| 19:47.24 | nikk | CM9 much better ;-) |
| 19:48.13 | nikk | and it'll definitely work soon ;-) |
| 21:31.11 | *** join/#arm-netbook cromo (~cromo@139-255-251-94.net.stream.pl) |
| 21:46.27 | t0dbld|work | woo hoo ... gentoo with e16 working with external display ... now jsut to figure out this internal display mess |
| 21:46.48 | t0dbld|work | hno: mnemoc zenitraM any you guys around today ? |
| 21:47.18 | t0dbld|work | any of you guys ever get display other than hdmi or vga on your net books ? |
| 21:47.43 | mnemoc | ? |
| 21:47.54 | zenitraM | net books? |
| 21:48.05 | t0dbld|work | is that not what a10 device you guys have ? |
| 21:48.10 | t0dbld|work | sorry if im confused |
| 21:48.21 | mnemoc | I don't have any netbook |
| 21:48.34 | t0dbld|work | oh what a10 device do you have mnemoc ? |
| 21:48.52 | mnemoc | well... I do, on a box on the cellar |
| 21:48.58 | t0dbld|work | my bad for assuming |
| 21:49.25 | t0dbld|work | what devices are you guys working with than ? |
| 21:49.33 | mnemoc | a mele a1000 and a bunch of 3 tablets which I don't use.... my daughters do |
| 21:49.45 | mnemoc | s/bunch of// |
| 21:50.31 | t0dbld|work | gotcha well have you ever gotten one of the tablets displays to work ? |
| 21:51.08 | mnemoc | they all have the stock os, and yes, hdmi works fine |
| 21:51.26 | t0dbld|work | oh :-) |
| 21:51.30 | t0dbld|work | i meant in linux :-) |
| 21:51.31 | ZaEarl | the qware sources should be using lcd display |
| 21:51.51 | t0dbld|work | ZaEarl: got gentoo display working now on external |
| 21:51.57 | t0dbld|work | trying to get internals now working |
| 21:52.05 | ZaEarl | excellent |
| 21:52.15 | t0dbld|work | also got ubuntu 12.04 |
| 21:52.20 | t0dbld|work | if thtas more your flavor |
| 21:52.25 | t0dbld|work | and a debian build as well |
| 21:52.32 | mnemoc | nice |
| 21:52.53 | ZaEarl | I wanna see 12.04. can i dl it? |
| 21:52.58 | mnemoc | t0dbld|work: but the point is? some patches to share? |
| 21:53.10 | t0dbld|work | yeah i will once i get everything working :-) |
| 21:53.29 | t0dbld|work | mnemoc: the point in the end is duel booting a10 tablets |
| 21:53.37 | t0dbld|work | cm9 and your choice of linux |
| 21:53.45 | mnemoc | nice |
| 21:54.04 | t0dbld|work | Turl: i think already sent you some pull requests on kernel side |
| 21:54.11 | mnemoc | even if i"m not sure my daughters will like CM |
| 21:54.26 | t0dbld|work | well theortically you can use stock ICS as well |
| 21:54.30 | mnemoc | but it's good for geeks who likes tablets |
| 21:54.57 | zenitraM | well, it has a theme manager. you can put them a pink theme |
| 21:55.06 | t0dbld|work | :-) ^ |
| 21:55.09 | gimli | haha |
| 21:55.12 | mnemoc | t0dbld|work: no need to wait until "it's ready" to start sending patches ;-) |
| 21:55.45 | mnemoc | they are more of purple than pink ;-) |
| 21:55.59 | t0dbld|work | well the kernel stuff im sure well send pull requests as we go , like i said i think turl sent you at least one so far |
| 21:56.04 | ZaEarl | t0dbld|work, you booting from nand or sd? |
| 21:56.35 | t0dbld|work | ZaEarl: sd as of now wanted to get it working before i figured out how to use live suit to repartition the HD's on tablet |
| 21:56.50 | zenitraM | well, there are fugly cm9 themes to match anyone :) |
| 21:57.01 | t0dbld|work | going to have to remap the whole device controls at boot |
| 21:57.17 | t0dbld|work | like dolume down recovery for android , volume up for linux |
| 21:57.24 | t0dbld|work | nothing for cm |
| 21:57.34 | t0dbld|work | and than repartion every thing |
| 21:57.37 | zenitraM | i hope i can find some free time to try to make the multitouch module work properly for linux |
| 21:57.43 | t0dbld|work | leave android about 12gb |
| 21:57.53 | t0dbld|work | zenitraM: I really hope you can |
| 21:58.05 | t0dbld|work | that is a huge challenge on the herizon |
| 21:58.08 | zenitraM | i don't think you can do that... only recovery is available |
| 21:58.09 | ZaEarl | adb has a "boot bootloader". do these a10 devices have a bootloader to boot into? |
| 21:58.30 | zenitraM | well in fact the problem is, the multitouch module works but at least on my tablet it emited the touch events as numbers |
| 21:59.04 | zenitraM | if you put the multitouch module and run.. xevent? or some X11 event debugging tool you'll see it |
| 21:59.05 | t0dbld|work | ZaEarl: im planing on rewritting from the begining |
| 21:59.19 | t0dbld|work | we have the boot chain |
| 21:59.26 | ZaEarl | define beginning :) |
| 21:59.30 | t0dbld|work | haha |
| 21:59.34 | ZaEarl | brom, boot0, boot1, uboot, etc |
| 21:59.35 | zenitraM | all the data is there but reported as Major 1, Major 2.. up to 6. it's a matter of making it make sense |
| 21:59.51 | t0dbld|work | right and we can write to all but brom correct |
| 22:00.06 | zenitraM | you know what would be great? |
| 22:00.09 | zenitraM | check if kexec works |
| 22:00.16 | zenitraM | and then make a small linux image |
| 22:00.29 | zenitraM | and a simple touch-based boot manager |
| 22:00.48 | t0dbld|work | why bother with kexec if we can boot natively ? |
| 22:00.55 | ZaEarl | I'd guess uboot would be as far back as you'd want to go |
| 22:01.09 | t0dbld|work | well clearly we have working uboot , |
| 22:01.18 | mnemoc | ZaEarl: many devices don't even have uboot |
| 22:01.27 | zenitraM | we can only choose between a kernel, two at most if the device supports recovery |
| 22:01.49 | mnemoc | t0dbld|work: hno is going to add usb console support soon |
| 22:02.00 | t0dbld|work | that would be huge |
| 22:02.14 | zenitraM | that would be *very* useful |
| 22:02.32 | zenitraM | what about usb console on linux? |
| 22:03.05 | t0dbld|work | first and for most i need to get familiar with livesuite figure out what it is capable of |
| 22:03.23 | mnemoc | zenitraM: isn't that just about .config dance around the usb gadget? |
| 22:03.37 | ZaEarl | i'd rather just use recovery than livesuit |
| 22:03.50 | zenitraM | mnemoc: mmm have to try then :) |
| 22:03.59 | t0dbld|work | well CWM iusnt going to repartion the internal sdcard |
| 22:04.14 | t0dbld|work | inwhich case duel booting would be out |
| 22:04.18 | zenitraM | livesuit is the only way to repartition for now |
| 22:04.23 | t0dbld|work | yeah |
| 22:04.24 | ZaEarl | ah |
| 22:04.36 | t0dbld|work | thats why i need to get familiar with it |
| 22:04.43 | t0dbld|work | i jsut wanted to get working images first |
| 22:05.07 | zenitraM | it's not about livesuit but more about creating the images |
| 22:05.08 | zenitraM | and in fact |
| 22:05.09 | zenitraM | http://www.flashmyandroid.com/forum/showthread.php?573-WIP-Novo-7-Elf-Dual-Boot-Elvish-0-3-Ubuntu-12-04-LTS-UPDATED-04-28-12 |
| 22:05.17 | mnemoc | sunxi-tools is begging for a toy to repartition the nand, patches welcomed |
| 22:06.13 | zenitraM | i'm not sure if that can even be done from userspace |
| 22:06.34 | zenitraM | at least a raw way to access the nand from the kernel would be needed (which, well, wouldn't be difficult either) |
| 22:06.37 | t0dbld|work | well tonight if i get some time ill play with livesuit |
| 22:06.59 | mnemoc | zenitraM: yes |
| 22:07.01 | t0dbld|work | i mean first step is understanding it |
| 22:07.11 | mnemoc | zenitraM: I forgot we don't have a /dev/nand |
| 22:07.18 | t0dbld|work | knowing what form its images are and how it gets created |
| 22:07.38 | mnemoc | t0dbld|work: unimg and filesumadd |
| 22:07.51 | t0dbld|work | i hate windows so i have not booted in and tried it |
| 22:07.53 | t0dbld|work | :-) |
| 22:08.03 | mnemoc | wine works :) |
| 22:08.08 | t0dbld|work | it didnt for me |
| 22:08.13 | t0dbld|work | i got sick of trying to get it too |
| 22:08.24 | mnemoc | for unimg I mean |
| 22:08.33 | ZaEarl | wine works for unimg.exe |
| 22:08.36 | mnemoc | for livesuit it won't because that one requires usb2 |
| 22:08.46 | t0dbld|work | i mean all my machines tripple boot |
| 22:08.55 | t0dbld|work | so ill jsut play with it nativley tonight |
| 22:09.12 | t0dbld|work | i tell ya what i almost bought a tegra3 to speed up these arm compiles lol !! |
| 22:09.27 | zenitraM | has just discovered the "t" key on github |
| 22:09.39 | t0dbld|work | a10 compiling power leave something to be desired |
| 22:09.56 | t0dbld|work | has still not changed all his git hub keys |
| 22:10.00 | ZaEarl | crosscompiling ftw |
| 22:10.03 | mnemoc | cpufreq on performance? |
| 22:10.20 | mnemoc | my mele lives at 60Mz most of the time... |
| 22:10.25 | t0dbld|work | ZaEarl: find me a way to cross compile arm gentoo packages and ill put my beast to work |
| 22:10.37 | t0dbld|work | i sure as hell am not doing kernels on the a10 lol |
| 22:11.24 | t0dbld|work | but "emerge gnome " is an all day event on the a10 :-P |
| 22:11.30 | ZaEarl | lol |
| 22:11.49 | t0dbld|work | and looks like im going to have to do that cause e16 is not ready on its own ill have to hybrid em |
| 22:12.03 | Brandon15811 | t0dbld|work, http://www.gentoo.org/doc/en/distcc.xml |
| 22:12.14 | t0dbld|work | Brandon15811: al ready read that |
| 22:12.14 | Brandon15811 | why not try distcc when you get home? |
| 22:12.43 | ZaEarl | distcc with a farm of mele's? |
| 22:12.48 | t0dbld|work | i need to get xlive up and going on gentoo so i can start coding from bed |
| 22:12.50 | Brandon15811 | hahaha |
| 22:12.51 | hno | t0dbld, uboot-mmc only uses BROM + u-boot. |
| 22:13.04 | hno | no boot0/boot1/... |
| 22:13.35 | t0dbld|work | hno: ok so why cant i repartion to have one side of internal sd be exatly that |
| 22:13.51 | hno | t0dbld|work, not sure what you talk about. |
| 22:13.52 | t0dbld|work | jsut like volume down takes me to boot recovery kernel |
| 22:14.07 | t0dbld|work | i want to add a third option is all |
| 22:14.32 | hno | volume key check is done by boot1 I think. |
| 22:14.42 | t0dbld|work | yeah illl remap that |
| 22:14.51 | t0dbld|work | but i want to shirnk android to 12gb |
| 22:15.04 | hno | is not familiar with the standard boot procedure or even less with Android |
| 22:15.07 | zenitraM | you can't add a third option |
| 22:15.07 | t0dbld|work | to allow for /boot and rfs for linux |
| 22:15.19 | t0dbld|work | why not ? |
| 22:15.28 | zenitraM | there are no sources for boot1 |
| 22:15.53 | hno | there is an os chooser somewhere in the Allwinner boot files. But unknown how to use it. |
| 22:16.18 | zenitraM | i think it's not even for the a10 |
| 22:16.23 | mnemoc | someone reads those .ini files in nanda |
| 22:16.27 | ZaEarl | like chaining grubs, could you chain uboots? |
| 22:16.46 | zenitraM | it was used on sochip mp4 players to choose what option (.axf) to "play" |
| 22:16.55 | zenitraM | ZaEarl: yes, i think so |
| 22:16.56 | hno | there is no need to chain grubs, and even less to chain uboots. But yes you can. |
| 22:17.22 | t0dbld|work | i dont understand why this is so different , i could duelboot my tf101 |
| 22:17.28 | t0dbld|work | using nvflash |
| 22:17.44 | t0dbld|work | my understanding is livesuit can write to everything but brom |
| 22:17.56 | mnemoc | good thing is you can't brick your A10 |
| 22:18.12 | mnemoc | uSD will always be able to rescue you |
| 22:18.15 | hno | It's not different. Just that no one here knows how to use the Allwinner bootloaders for multi-choice OS selection. |
| 22:18.20 | t0dbld|work | are livesuit and nvflash not equal than ? |
| 22:18.47 | t0dbld|work | ok but android uses full chain and linux we only need uboot right ? |
| 22:18.52 | mnemoc | livesuit actually executes some binaries that come inside the image |
| 22:19.06 | hno | livesuit is an odd beast. |
| 22:19.26 | hno | or is it allwinner firmware file that is the odd beast, not sure.. |
| 22:19.36 | mnemoc | both? |
| 22:19.42 | hno | it's odd in either case. |
| 22:20.01 | t0dbld|work | yeah i mean dont get me wrong im not expecting this to be easy |
| 22:20.05 | mnemoc | going to sleep, cu tomorrow |
| 22:20.07 | t0dbld|work | but i dont see why we cant |
| 22:20.23 | mnemoc | "can't" is probably too strong for this, right |
| 22:20.43 | zenitraM | easy: uboot can't read the keys afaik |
| 22:20.46 | t0dbld|work | jsut need more research and understanding |
| 22:20.56 | hno | t0dbld|work, sure you can. It's just that no one here knows how to do it with the standard Allwinner bootloader. |
| 22:20.59 | zenitraM | and the built-in bootloader just supports one recovery image |
| 22:21.04 | Turl | zenitraM: I was planning on getting kexec working :) |
| 22:21.12 | hno | and with u-boot you are missing many things. |
| 22:21.14 | mnemoc | let's kidnap hipboi |
| 22:21.20 | zenitraM | +1 |
| 22:21.26 | t0dbld|work | haha |
| 22:21.26 | hno | someone reported kexec working earlier. |
| 22:21.38 | t0dbld|work | hay didnt one of you say you were adding support for fastboot ? |
| 22:21.55 | t0dbld|work | cause with fastboot i wouldnt need a recovery any how :-) |
| 22:21.56 | Turl | hno: I'll confirm their findings then :P |
| 22:22.04 | hno | the u-boot version chainloaded from allwinner bootloader has fastboot. |
| 22:22.31 | t0dbld|work | ok so than flashing android updates no longer requires a recovery :- ) |
| 22:22.33 | zenitraM | i think the easiest option right now is booting a linux kernel with a simple fb-based menu on initrd |
| 22:23.05 | hno | zenitraM, agreed. Simple, and open. |
| 22:23.26 | t0dbld|work | well i will work on these images a little more and ill upload them |
| 22:23.32 | Turl | zenitraM: I talked with one of the defy guys |
| 22:23.33 | t0dbld|work | so we can all have working linux images |
| 22:23.42 | zenitraM | and kexec wherever you want to go |
| 22:23.45 | t0dbld|work | what does everyone prefer for linux distors |
| 22:23.51 | Turl | zenitraM: I was told bootmenu's UI library should be easy to use |
| 22:23.58 | Turl | and would allow us to do a UI on fb |
| 22:24.02 | Turl | and it has touch support :) |
| 22:24.22 | hno | is a Fedora guy. Some others like Ubunut, some Arch, some Gentoo, some Debian. |
| 22:24.35 | Turl | you know I'm a debian user :) |
| 22:24.42 | t0dbld|work | haha ok well ill suply all , i have not done an arch or fedora yet |
| 22:24.47 | t0dbld|work | but ill do them next |
| 22:24.59 | Turl | I don't wanna start a flamewar but I don't like rpms :p |
| 22:25.00 | t0dbld|work | i was thinking enlightenment for tablet window manager |
| 22:25.29 | t0dbld|work | but that is seemingly alot of work so i am open to alternatives in mean time |
| 22:25.48 | t0dbld|work | ubuntu will come standard cause i dont like it enough to spend my time there switching it |
| 22:25.50 | Turl | t0dbld|work: something with gnome3 and kde's plasma active thing for tablets :) |
| 22:25.52 | t0dbld|work | you can do on your own :-P |
| 22:26.08 | t0dbld|work | hmm i have not checked out the kde thing |
| 22:26.14 | zenitraM | if we only had hw accelerated x11.. |
| 22:26.14 | t0dbld|work | ill youtube it if its nice ill go with that |
| 22:26.24 | t0dbld|work | haha we need to steal a ddk |
| 22:26.31 | t0dbld|work | whos near china ? |
| 22:26.32 | ZaEarl | I wish the kubuntu active guys would get in touch. |
| 22:26.33 | t0dbld|work | :-) |
| 22:26.40 | Turl | let's take over allwinner factories :P |
| 22:27.05 | hno | allwinner have no factories. |
| 22:27.11 | *** join/#arm-netbook naobsd (7c294571@gateway/web/freenode/ip.124.41.69.113) |
| 22:27.17 | t0dbld|work | haha its jsut peoples basement |
| 22:27.27 | hno | it's fabless. |
| 22:27.56 | hno | like most others. |
| 22:28.13 | naobsd | disconnected for a while... |
| 22:28.43 | Turl | hno: fabless in china? |
| 22:29.06 | zenitraM | AllwinnerTech, 4th floor, Building B1, South Software Park, Zhuhai, China |
| 22:29.06 | naobsd | nikk: on android, absolute path is used for insmod (no modprobe) |
| 22:29.09 | hno | Turl, yes? |
| 22:29.29 | zenitraM | that's their address (tom told me to send my faulty mele there) |
| 22:29.55 | nikk | naobsd i see. or, it's devined in sourcecode, as i see for nanowifi |
| 22:29.55 | Turl | hno: I had never heard of fabless on china |
| 22:30.20 | nikk | naobsd now i'm remompiling it with correct options, hope it helps |
| 22:30.44 | naobsd | nikk: for x_mac.axf, nrx_config= can be used to specify path while loading nano_if.ko |
| 22:30.53 | t0dbld|work | we should find out cost of getting a license and ddk from allwinner |
| 22:31.00 | nikk | yup it's the option |
| 22:31.09 | nikk | in wifi.c |
| 22:31.12 | hno | t0dbld|work, license for what? |
| 22:31.37 | t0dbld|work | DDK |
| 22:31.45 | t0dbld|work | so we have access to full stack |
| 22:31.45 | nikk | naobsd problem is = android loads and unloads driver by jni, so it must be defined there too |
| 22:31.53 | t0dbld|work | not binary blobs |
| 22:32.01 | t0dbld|work | and maybe can get hw accel |
| 22:32.03 | hno | I don't think Allwinner can distribute it in source form. Only binary which they do. |
| 22:32.20 | naobsd | I found these and noted in my source. no one have interest and now obsolete by leaked source ;) |
| 22:32.27 | t0dbld|work | i was under impression that you did get source with ddk |
| 22:32.35 | t0dbld|work | from their dev site |
| 22:32.38 | t0dbld|work | but i could be wrong |
| 22:32.45 | t0dbld|work | its still worth asking about |
| 22:32.57 | zenitraM | the snowball x11-mali driver (which is gpl) doesn't need the ddk to compile |
| 22:32.59 | t0dbld|work | i mean if they want 30k were out anyhow :-) |
| 22:33.01 | hno | the BSP package from Allwinner only includes a binary DDK. |
| 22:33.18 | zenitraM | but it uses the 2d engine on that SoC |
| 22:33.50 | nikk | naobsd where that absolute path defied? for 2.3 it was /system/etc/firmware. i need to create it and put DVB firmware there, only in that case it loads |
| 22:35.54 | naobsd | nikk: sorry, it may depend on what you used. insmod doesn't know about /system/etc/firmware/ |
| 22:36.22 | hno | the 2d engine is not MALI. |
| 22:36.42 | lundman | wooo panic , solar eclipse |
| 22:36.42 | nikk | naobsd but, in kernel driver for DVB, there was no any defines of the path, so i experiment and find that one |
| 22:36.50 | naobsd | nikk: "it must be defined there too" sorry, what? |
| 22:37.11 | naobsd | nikk: hmm. I may miss something :( |
| 22:38.06 | naobsd | wifi.c has issue while loading 2 module |
| 22:38.16 | naobsd | it was fixed in CM9 while ago |
| 22:38.44 | naobsd | very beginning I defined necessary lines in BoardConfig.mk for 2 nano module |
| 22:38.56 | naobsd | but it didn't work by "issue" |
| 22:39.18 | naobsd | so I added "insmod ..." lines in init.rc for workaround |
| 22:39.30 | naobsd | and commented out lines in BoardConfig.mk |
| 22:39.36 | hno | Turl, Allwinners own words on who/what they are: "Allwinner Technology, established in 2007, is a leading fabless semiconductor company in China, supplying customers with cost-efficient SoC products as well as market-oriented total solutions." |
| 22:39.47 | naobsd | my CM9 doesn't load/unload on demand ;) |
| 22:39.48 | nikk | naobsd: i think it's my case. i defined there 4 lines for it, but wlan0 doesn't up |
| 22:40.10 | naobsd | from my log, you just forgot x_mac.axf :) |
| 22:40.55 | naobsd | current CM9 should not have that issue... well, I'm not sure Quarx's CM9, it may have another issue? not sure :( |
| 22:41.26 | hno | Turl, and there is quite many such companines in China now. |
| 22:41.49 | nikk | naobsd: yup, the same error was in 4.0.3 without x_max.axf. here all modules loaded, but the error exist... |
| 22:42.00 | naobsd | I'm busy to see eclipse, sorry ;) |
| 22:42.24 | naobsd | ah, 4.0.3 should have that issue... I'll check fix in CM9 later |
| 22:42.25 | nikk | x_mac.axf putted in all places, even to /vendor/firmware ))) |
| 22:42.43 | naobsd | did you specify nrx_config= ? |
| 22:43.02 | nikk | on driver load - no, in wi-fi.c - yes |
| 22:43.46 | nikk | now i'm backport code from 4.0.3 to CM9 for test |
| 22:43.58 | nikk | SW_BOARD_USR_WIFI := nanowifi |
| 22:44.08 | nikk | and all descendants |
| 22:44.09 | Turl | hno: nice to know :) |
| 22:47.49 | nikk | fuf |
| 22:47.58 | nikk | with manual loading it's up: |
| 22:47.59 | nikk | <7>[ 7946.740000] [nano] mac: 84:5d:d7:0b:96:cf |
| 22:47.59 | nikk | <7>[ 7947.000000] mmc1:0001:1: Clock switched to 45MHz |
| 22:47.59 | nikk | <7>[ 7947.530000] mmc1:0001:1, Registered interface wlan0 |
| 22:48.08 | nikk | <7>[ 7947.530000] [nano] NULL BSSID sent |
| 22:51.22 | Turl | BSSID sent? :p |
| 22:51.29 | nikk | naobsd, found also wpa_supplicant dependencies in 4.0.3: L_CFLAGS += -DNANO_SDIO_WIFI_USED for it. Do I need add this too? |
| 22:59.01 | naobsd | nikk: https://github.com/CyanogenMod/android_hardware_libhardware_legacy/commit/a5be083afc54933b0403dcd48e85d9af14f75f54 this is "fix" in CM9 |
| 23:00.21 | naobsd | I'm not sure it's necessary for allwinner 4.0.3. you can load nano_if.ko in init.rc or somewhere, then this issue should not happen |
| 23:00.45 | naobsd | nikk: I didn't try allwinner code yet. no idea for now |
| 23:03.22 | naobsd | I didn't try that fix in CM9 because I was lazy... I used same libhardware_legacy for other tablets, so I didn't specify wifi driver information in it. |
| 23:04.34 | nikk | naobsd: i see. by BoardConfig have TARGET_CUSTOM_WIFI := ../../device/allwinner/novo7a/wifi.c and there are braces in wifi_load_driver, i think it was the problem with missing nrx_config= parameter, and module can't find it automatically in any location (dumb) |
| 23:04.37 | naobsd | my init.sun4i.rc loads 8192cu.ko then nano*.ko. it works fine both NOVO7 Advanced(nano) and Adv2/Elf/Aurora/others(8192cu) |
| 23:05.02 | naobsd | and I can use external USB 8192cu wifi on my Adv1 w/o any modification :) |
| 23:05.38 | naobsd | nikk: I see. thank you for clarify |
| 23:06.12 | naobsd | I guess you can specify default nrx_config while building kernel (module) |
| 23:06.41 | naobsd | grr |
| 23:07.00 | nikk | cool. i compiled rtl8187 module (for 2.6.36) for a guy, who lost internal wifi and wanna to use usb one, and he report it worked without any mods on novo ) |
| 23:07.27 | naobsd | loading nano*ko on Adv2/Elf/Aurora must not good. boot loops. I need to remove nano*ko from /system, then it works w/o init.rc mod |
| 23:07.44 | naobsd | I was too lazy lol |
| 23:07.49 | nikk | ;-) |
| 23:08.16 | naobsd | my strategy was: add necesarry insmod lines for all tablets to init.sun4i.rc |
| 23:08.39 | naobsd | (cont) and remove unusable *.ko from /system for each target specific update.zip |
| 23:08.49 | nikk | naobsd my idea is compile alll in kernel, so it must support as many as possible. is it bad? |
| 23:09.11 | naobsd | nikk: if we can share kernel, it should be ok |
| 23:09.44 | nikk | naobsd now im on CM9 with 3.0.31+ |
| 23:09.47 | naobsd | I understand all A10 devices are very similar, hardware information is descripted in script.bin, so we can share kernel and modules |
| 23:09.48 | nikk | and it works ;-) |
| 23:09.57 | naobsd | described |
| 23:10.39 | naobsd | but it's possible to make local change for specifiy device so I can't judge it's really safe for "all A10 device" |
| 23:11.04 | naobsd | I know some ICS rom for NOVO7 series share same kernel |
| 23:11.14 | naobsd | maybe ver1.0 |
| 23:11.15 | nikk | yup. even better make external SD-card loaded build, so anyone can test without repartition nand |
| 23:11.49 | naobsd | yup |
| 23:12.01 | mikey_w | I vote for the external SD-card. |
| 23:12.04 | nikk | it's easyli can be made as A10 have mmc-load option and we have tools to do so. |
| 23:12.22 | nikk | as many do for mele and ubuntu |
| 23:12.47 | naobsd | ah |
| 23:13.03 | naobsd | I'm sure loading nano* is harmfull for Adv2/Elf/Aurora |
| 23:13.07 | nikk | think class 10 cards will be faster and bigger than nand ;-) |
| 23:13.21 | naobsd | your kernel which has nano may not work on them |
| 23:13.33 | naobsd | I don't know detail |
| 23:13.56 | nikk | naobsd nano can't be compiled internally, as i understand. but many others can |
| 23:13.57 | naobsd | I just asked some person to try 2 my CM9, one has nano* and another doesnt have nano |
| 23:14.03 | naobsd | I see |
| 23:14.27 | naobsd | s/some person/& who has Aurora/ |
| 23:14.31 | nikk | pl2303 dvb usb-audio etc |
| 23:14.52 | naobsd | oh where is sweet bot |
| 23:14.57 | nikk | why we need modules, if we can include it all |
| 23:15.40 | nikk | it'll be not much bigger, but mode universal at all |
| 23:16.17 | nikk | for example, why not to add all touchscreens in kernel rather than distribute separate patches for different hardware |
| 23:16.41 | nikk | script.bin will provide info which one must be used, no? |
| 23:18.59 | nikk | naobsd: CM9 still not doing right boot.img for me, but i found the way to make proper one by hand with you instructions! many thanks! |
| 23:20.02 | nikk | I'm just using another mkbotimg binary (steal it from my 4.0.3 source) ;-) |
| 23:44.20 | naobsd | distribute separate *patches* is feiyu's way, not official |
| 23:46.07 | naobsd | on A10 devices, device dependent unit (e.g. TP) can be specified in script.bin and kernel can judge each one is necessary or not |
| 23:46.32 | naobsd | but in general loading(probing) non-existing unit may be harmful |
| 23:47.16 | naobsd | official way is safe(load only necessary modules) |
| 23:48.22 | naobsd | but trying to include all modules as much as possible should be ok for us :) |
| 23:48.31 | naobsd | except nano of course! |
| 23:50.04 | naobsd | but as I said I'm not sure kernel can be shared. some tablet maker may put some module upside down, and it can't be tweaked by script.bin, then they need to make special module only for their device;) |
| 23:50.54 | naobsd | btw I like module |
| 23:51.33 | naobsd | even though today's devices have enough resources |
| 23:51.45 | nikk | many kernels have many supported devices/options included, so probably this is the BUG in nano, yup, module give options, and it cam be much mo on .system partition for enthusiasts ;-) |
| 23:52.45 | nikk | grrr. can be much more ;-) |
| 23:53.19 | nikk | apple bluetooth multitouch panel for example ;-) |
| 23:53.53 | nikk | i saw it inside menuconfig =) |
| 23:54.18 | nikk | it will be cool connect it to A10 TV box |
| 23:54.27 | nikk | and many DVB sticks also |
| 23:54.53 | naobsd | any app is available to see DVB outputs? |
| 23:55.02 | nikk | yup, in test phase |
| 23:55.17 | nikk | https://github.com/chrulri/droidtv |
| 23:55.31 | nikk | and it scans channels fine ;-) |
| 23:56.01 | nikk | but needs some tweaking (symlinks in /dev/) |
| 23:56.59 | naobsd | oh, good |
| 23:57.19 | *** join/#arm-netbook arokux_h (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 23:57.36 | nikk | yeah, it will be possible to see DVB TV in any place with only TAB and STICK |
| 23:57.54 | naobsd | I really like this kind of hack |
| 23:58.00 | nikk | or, on any A10 like mele |
| 23:58.13 | nikk | http? |
| 23:58.14 | naobsd | one of the reason to build Android myself |
| 23:59.17 | nikk | I made it work with external DAC even on 2.2 a year ago ;-) |
| 23:59.47 | naobsd | anything(up to my skil limitation) can be done on it :) |
| 23:59.49 | nikk | but only in stereo. as i saw in 4.0 source, it will support multichannel soon |