00:01.03 | penguin42 | mnemoc: the problem is without anything saying that you only find out when someone charges you later |
00:01.19 | mnemoc | true |
02:06.30 | *** join/#arm-netbook cromo (~cromo@90-255-251-94.net.stream.pl) |
02:15.53 | *** join/#arm-netbook hp_ (~kvirc@220-136-106-32.dynamic.hinet.net) |
02:53.07 | A0Sheds | Apple iPad3 screen massive power draw (8Watts?!) |
02:53.16 | A0Sheds | http://pixelqi.com/blog1/2012/04/21/next-generation-pixel-qi-screen-improves-over-ipad3-display/ |
02:55.27 | CIA-158 | rhombus-tech: Shawn master * r73b0a6a2efb7 /allwinner_a10/orders/178259201445.mdwn: |
08:20.41 | CIA-158 | rhombus-tech: Surendra master * r2d916913cc34 /allwinner_a10/orders/surenreddy.htm: |
08:45.48 | *** join/#arm-netbook eebrah (~Ibrahim_K@212.49.88.64) |
09:09.49 | *** join/#arm-netbook cromo (~cromo@90-255-251-94.net.stream.pl) |
09:31.06 | CIA-158 | rhombus-tech: lkcl master * rf9ad675ca5aa /allwinner_a10/a10_image.mdwn: |
09:35.00 | CIA-158 | rhombus-tech: lkcl master * r3ea55cca6f61 /allwinner_a10/a10_image.mdwn: |
09:36.08 | CIA-158 | rhombus-tech: lkcl master * r4f677ee658a0 /allwinner_a10/a10_image.mdwn: |
09:38.12 | *** join/#arm-netbook MaDMaLKaV (~IceChat9@188.86.253.69) |
10:27.30 | *** join/#arm-netbook gnurbs (~arian@77-58-105-100.dclient.hispeed.ch) |
10:49.18 | lkcl | A0Sheds: 8 watts???? holy cow |
11:08.55 | cromo | any idea why it only reports 313940 kB of memory? |
11:08.58 | cromo | should be 512 |
11:10.33 | specing | cromo: mine does that too |
11:10.58 | cromo | mnemoc: would you mind checking yours? I assume you run own kernel |
11:11.14 | cromo | u-boot says 512 on boot, so it's not hardware |
11:14.43 | specing | <PROTECTED> |
11:14.43 | specing | Mem: 322576 296676 25900 0 11452 |
11:14.43 | specing | -/+ buffers: 285224 37352 |
11:14.44 | specing | Swap: 0 0 0 |
11:18.04 | cromo | that's clearly almost 2/5 missing ;) |
11:18.13 | specing | well... |
11:18.18 | specing | 32 MB is for the GPU |
11:18.26 | specing | Where the rest went... no clue |
11:20.46 | specing | Also, my tablet is again not booting |
11:29.10 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
12:38.57 | *** join/#arm-netbook cromo (~cromo@90-255-251-94.net.stream.pl) |
12:57.32 | *** join/#arm-netbook eebrah (~Ibrahim_K@212.49.88.73) |
13:03.25 | zenitraM | kexec seems to work with 2.6.36... except kernel parameters. |
13:14.20 | cromo | anybody knows if there's a "build_sun4i" script available in linux-allwinner branch? |
13:39.29 | mnemoc | cromo: yes, same mem here, MemTotal: 316412 kB |
13:39.58 | cromo | any idea why? |
13:40.03 | cromo | should be more I think |
13:40.58 | mnemoc | [ 0.000000] Total Detected Memory: 512MB with 1 banks |
13:41.06 | mnemoc | [ 0.000000] Memory Reserved: |
13:41.06 | mnemoc | [ 0.000000] VE: 0x43000000, 0x04a00000 |
13:41.06 | mnemoc | [ 0.000000] FB: 0x5a000000, 0x02000000 |
13:41.06 | mnemoc | [ 0.000000] G2D: 0x58000000, 0x01000000 |
13:42.46 | cromo | hmm |
13:43.37 | zenitraM | i think i've got how to fix eth on the evb script.bin |
13:43.50 | zenitraM | the evb is missing a line: |
13:43.51 | cromo | zenitraM: with working X's? |
13:43.51 | zenitraM | emac_power = port:PH15<1><default><default><0> |
13:44.09 | zenitraM | i have to try it |
13:44.10 | cromo | if you could compile quickly the bin |
13:44.25 | cromo | I have a friend that will test it quickly to check both x and eth |
13:44.33 | zenitraM | i only need to find where did i put the script compiler |
13:44.35 | cromo | the same guy that initially reported X not working |
13:45.34 | zenitraM | the thing is, with the mele .bin |
13:45.41 | zenitraM | x works, but hangs after login |
13:46.46 | mnemoc | what driver you use on X? fb? |
13:47.03 | zenitraM | i think so. in fact there is no xorg.conf |
13:47.53 | mnemoc | cromo: by "linux-allwinner branch" you mean certain branch of this repo: https://github.com/amery/linux-allwinner ? |
13:48.06 | cromo | mnemoc: are the values after the comma? size? |
13:48.19 | zenitraM | cromo: https://www.dropbox.com/s/xptsdzelwtm821o/evb_eth.bin |
13:48.34 | mnemoc | cromo: yes, base comma size |
13:48.43 | zenitraM | i was trying to do the opposite, find out why the mele one doesn't work |
13:48.49 | cromo | cromo: yes |
13:48.52 | cromo | jeeez |
13:48.54 | cromo | mnemoc: yes |
13:49.24 | mnemoc | cromo: the allwinner-*-android branches include the lichee scripts, the linux-* branches don't |
13:49.50 | mnemoc | cromo: but the linux-* branches only include core functionality not even the mmc or nand drivers |
13:49.53 | cromo | so you build uImage manually after compilation? |
13:50.01 | cromo | ah |
13:50.18 | mnemoc | i've never used those scripts |
13:50.31 | mnemoc | i use the normal `make` |
13:50.50 | cromo | and to build uImage? mkimage? |
13:50.53 | mnemoc | make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- CONFIG_DEBUG_SECTION_MISMATCH=y -j32 uImage modules |
13:50.59 | cromo | ah |
13:51.02 | cromo | I get it |
13:51.54 | mnemoc | and arch/arm/boot/uImage will be there at the end |
13:52.11 | cromo | I thought you need to use the mkimage tool to get the uImage generated |
13:52.17 | mnemoc | sure |
13:52.19 | cromo | since normally make builds bzImage only |
13:52.35 | mnemoc | the the uImage target calls uboot helpers for you |
13:52.48 | mnemoc | s/^the/but/ |
13:53.07 | cromo | nice |
13:54.20 | mnemoc | remove the CONFIG_DEBUG_SECTION_MISMATCH=y from the paste above |
13:54.30 | mnemoc | modules probably too |
13:55.19 | cromo | which branch do you recommend? linux-3.3-sun4i? |
13:55.28 | cromo | is it stable at all? |
13:55.57 | mnemoc | only if you are happy with an initrd based system |
13:56.18 | mnemoc | I work on allwinner-v3.0-android, trying to fix drivers |
13:57.07 | cromo | would rather want kernel only |
13:57.21 | mnemoc | hno used linux-3.3-sun4i to fix some core support, but doesn't include any driver |
13:57.29 | mnemoc | no mmc, no nand, no fb, .... |
13:57.57 | cromo | if there's no mmc, rootfs won't boot, so... |
13:58.12 | mnemoc | initramfs will work |
13:58.18 | mnemoc | but nothing facier |
13:58.21 | mnemoc | fancier* |
13:58.33 | cromo | hmm |
13:58.51 | mnemoc | allwinner-v3.0-android has working mmc, not nand is not working |
13:58.59 | mnemoc | sata seems to work, but haven't tested it |
13:59.04 | cromo | USB? |
13:59.14 | mnemoc | works fine |
13:59.15 | cromo | I want to connect USB cable modem |
13:59.19 | mnemoc | also usb-storage |
13:59.24 | mnemoc | might work |
13:59.28 | cromo | will try it then |
13:59.32 | mnemoc | but forget about X for now |
14:01.07 | cromo | does it compile with 4.6? |
14:01.11 | cromo | gcc that is |
14:02.05 | mnemoc | yes, I use linaro's gcc |
14:03.30 | mnemoc | cromo: If you are going to play with that branch, this is the diff of my .config http://dpaste.com/735562/ |
14:04.08 | cromo | what does it change compared to the config file in branch? in summary |
14:04.32 | mnemoc | finishes compiling :p |
14:04.42 | cromo | haha |
14:04.53 | cromo | CONFIG_VIDEO_SUN4I_CEDAR |
14:04.56 | cromo | interesting |
14:05.08 | mnemoc | some drivers are still broken, others still crash your kernel so I made them m temporarily |
14:07.10 | mnemoc | I wanted to fix the nand driver, but switched to the disp driver as fb seems to have more popular interest ;-) |
14:10.41 | *** join/#arm-netbook mikey_w (~mike@c-71-63-115-202.hsd1.va.comcast.net) |
14:11.00 | *** join/#arm-netbook mike_ (~mike@c-71-63-115-202.hsd1.va.comcast.net) |
14:11.47 | zenitraM | mmm the evb one has (obviously) spdif disabled |
14:12.13 | zenitraM | aand the system hangs on the mele exactly at the point where the ubuntu startup chime is played |
14:12.21 | mnemoc | meh... that's the only sound driver I already ported :< |
14:12.41 | zenitraM | this is with their 3.0.8+.. |
14:13.04 | zenitraM | i suspect spdif is what is making it really crash |
14:13.18 | zenitraM | let's try to disable alsa.. |
14:14.28 | cromo | I need to do some more reading on this |
14:14.44 | cromo | why can't we really use the one that is used for android? |
14:15.07 | mnemoc | we can |
14:15.07 | cromo | or in other words, why does it cause problems with vanilla linux kernel? |
14:15.12 | zenitraM | cromo: it crashes after ubuntu X login |
14:15.52 | mnemoc | cromo: they only implemented and fixed the parts needed to run android |
14:16.11 | mnemoc | cromo: linux, Xm and full fb was never of interest for allwinner |
14:16.34 | mnemoc | cromo: the drivers miss several not-optional callbacks |
14:16.47 | mnemoc | cromo: because android doesn't call them |
14:17.11 | cromo | I see |
14:17.14 | mnemoc | but linux over the same android kernel does use those callbacks, and crash |
14:17.56 | zenitraM | their 3.0.8+ seems to have them implemented though... but no sources |
14:17.59 | cromo | A10 is great chip |
14:18.05 | cromo | too bad it's not so popular |
14:18.28 | cromo | seems perfect for set-top-boxes |
14:18.30 | cromo | running linux |
14:19.54 | cromo | so basically only their 3.0.8+ runs X fine? |
14:21.21 | zenitraM | so far, it seems so |
14:21.21 | mnemoc | not even |
14:32.48 | *** join/#arm-netbook QingPei (~qingpei@114.241.176.76) |
14:39.05 | cromo | mnemoc: USB serial is not enabled in your config, do you use serial<>USB converter for serial console debugging? |
14:39.57 | mnemoc | cromo: the dongle is connected via usb to another machine |
14:40.45 | mnemoc | cromo: but if you want to connect a dongle to the mele you can enable it on your builds too |
14:41.13 | cromo | I need to get some more sleep |
14:41.16 | cromo | :o |
14:41.24 | mnemoc | :) |
14:41.31 | cromo | of course it is connected to another machine |
14:41.37 | cromo | sometimes I amuse myself |
14:41.45 | mnemoc | ^^ |
14:49.13 | zenitraM | i'm trying to isolate what on the mele .bin is making things go kaboom |
14:49.19 | zenitraM | so far i've got this: http://pastie.org/3833914 |
14:49.32 | zenitraM | mele_spdif crashes, evb doesn't |
14:49.40 | mnemoc | zenitraM: please, diff -u |
14:50.10 | zenitraM | sorry :) http://pastie.org/3833920 |
14:51.07 | mnemoc | changes on `uart_used = 0` sections shouldn't matter |
14:51.36 | zenitraM | yup.. but there isn't much left |
14:51.59 | mnemoc | emac_power = port:PH15<1><default><default><0> seems important |
14:52.45 | mnemoc | and +usb_host_init_state = 1 |
14:52.57 | cromo | <PROTECTED> |
14:53.07 | zenitraM | emac_power is *not* on the evb one |
14:53.12 | cromo | it should be OK to enable it by default in repository |
14:53.24 | mnemoc | zenitraM: uhm |
14:53.27 | zenitraM | that would explain ethernet not working |
14:53.58 | mnemoc | cromo: do you have github account? |
14:54.05 | cromo | not yet |
14:54.28 | cromo | all the programming I do I do only for my closed-source employer ;) |
14:54.33 | zenitraM | will try removing that one. maybe the reason X11 crashes is... well, ethernet working and network-manager doing something nasty |
14:55.26 | mnemoc | cromo: time to make a change on your life |
14:56.04 | cromo | I think about it each day, trust me ;) |
14:56.25 | mnemoc | cromo: doing closed programming for your money provider doesn't mean you can't make your own code open |
14:56.36 | cromo | time's the issue |
14:57.42 | cromo | anyways, I don't get why do I get "Restart config..." each time I make |
14:57.59 | cromo | I did make menuconfig and saved, so I assume no more configuration should be needed |
14:58.25 | mnemoc | you can edit .config manually |
14:58.34 | mnemoc | and consistency needs to be reviewed, each time |
14:58.53 | cromo | I'm confused |
14:59.19 | cromo | haven't compiled kernel for years now probably, but AFAIR it was enough to do make menuconfig and make |
14:59.39 | mnemoc | cromo: first, this isn't not native |
15:00.18 | cromo | it asks me some arm questions I don't know and don't want to ask |
15:00.27 | mnemoc | so every time you say `make` think: make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- |
15:00.31 | cromo | sure |
15:00.38 | cromo | I do that |
15:00.51 | mnemoc | the first thing is a defconfig, sun4i_defconfig in our case |
15:01.09 | mnemoc | then, every time you edit .config, oldconfig |
15:01.38 | mnemoc | and then, uImage and modules |
15:01.59 | mnemoc | with a good -j obviusly |
15:02.24 | mnemoc | I use 2 times the number of cores of the machine |
15:03.00 | cromo | I still don't understand why am I being asked the configuration questions when doing make |
15:03.17 | mnemoc | that happens when your .config is incomplete |
15:03.38 | cromo | AFIR running make menuconfig and saving should deal with it |
15:04.28 | mnemoc | haven't used menuconfig is years |
15:04.36 | mnemoc | in* |
15:04.46 | mnemoc | I suppose it should |
15:04.51 | mnemoc | oldconfig is faster |
15:04.53 | cromo | yeah |
15:04.57 | cromo | I went back to README |
15:05.05 | zenitraM | okay, even more isolation. http://pastie.org/3834005 |
15:05.07 | cromo | quite a lot has changed |
15:05.33 | mnemoc | oldconfig is older than menuconfig... |
15:05.55 | cromo | but silentoldconfig |
15:06.04 | cromo | allyesconfig, etc. |
15:06.08 | cromo | don't remember that |
15:06.14 | mnemoc | those are for distros |
15:06.22 | mnemoc | to include in their build systems |
15:06.27 | mnemoc | not for human usage |
15:06.40 | cromo | see |
15:07.03 | mnemoc | zenitraM: do you have the .fex posted by Tom with all documentation (in chinese) etc? |
15:07.36 | mnemoc | zenitraM: still no ethernet on Tom's .fex+fixes ? |
15:07.54 | zenitraM | haven't tested yet.. but i think there is something else wrong |
15:08.09 | mnemoc | don't be greedy, test one change at the time |
15:08.17 | zenitraM | <PROTECTED> |
15:08.17 | zenitraM | 9c184caa59eb3deaa275a721cb0b16b5 /home/zenitram/Descargas/sys_config1.evb.bin |
15:08.17 | zenitraM | c79d30d3dac04220ac51b85268d8b4f2 sys_config1.evb.bin |
15:08.43 | zenitraM | the .bin tom posted (same as on evb.bin) |
15:08.50 | zenitraM | doesn't match the .fex compiled to .bin |
15:09.09 | mnemoc | uhm |
15:09.14 | zenitraM | and |
15:09.26 | zenitraM | it also doesn't match evb.bin decompiled to fex and compiled back again |
15:10.07 | zenitraM | which makes me think they're using another version of the script tool which adds something different at the overhead at the end |
15:10.55 | mnemoc | tried with the compiler for windows? |
15:11.05 | mnemoc | there are some versioning fields in the header |
15:11.13 | zenitraM | no windows at hand |
15:11.23 | zenitraM | going to try binary-diffing both .bin |
15:12.19 | mnemoc | it is possible that the windows compiler translates some of the "default" to a known default value, a 0 or a 1 for example |
15:12.28 | mnemoc | while the dumber compiler passes the default as is |
15:15.36 | mikey_w | I'e dd(ed) the 2 GB image to a 4GB sd card several times. The first partition on the card will mount but not the rootfs partition? |
15:16.03 | mnemoc | fsck it first |
15:16.10 | mikey_w | Also there is no u-boot in the first partition. Is the image a joke or what? |
15:16.11 | mnemoc | and check you have ext4 support |
15:16.38 | mnemoc | u-boot in injected in the first M, before the partitions |
15:16.54 | mnemoc | MBRish |
15:17.11 | mikey_w | What the heck does that mean, either it's a good image or it's not. |
15:17.42 | mnemoc | if works for others, it's a good image |
15:18.02 | vgrade | mikey_w, works here |
15:18.06 | mikey_w | I am confused or spoiled by my beagleboard's image files. |
15:18.27 | mnemoc | 2GB image.... iirc the image is 4GB *compressed* as 2GB |
15:18.55 | mikey_w | What is equivalent to MLO and u-boot.bin etc.? |
15:19.11 | mnemoc | mikey_w: not visible |
15:19.37 | mikey_w | What is the exact boot process from power on to linux running? |
15:19.43 | mnemoc | mikey_w: for A10 devices that doesn't go on a partition |
15:19.52 | mnemoc | mikey_w: it's documented in the wiki |
15:20.05 | mikey_w | URL please? |
15:20.06 | vgrade | there is a uboot binary written after the partition table on the sdcard |
15:20.19 | vgrade | before the first fat part |
15:20.37 | mikey_w | Actually that's rater cool. |
15:20.44 | mnemoc | mikey_w: http://rhombus-tech.net//allwinner_a10/a10_mmc_boot/ |
15:20.45 | mikey_w | rather |
15:21.13 | mikey_w | Makes sense now. Thank you. |
15:21.19 | mnemoc | sudo dd if=u-boot-spl-mmc.bin of=/dev/sdb bs=1024 seek=8 |
15:21.23 | mnemoc | sudo dd if=u-boot.bin of=/dev/sdb bs=1024 seek=32 |
15:22.10 | mnemoc | and http://rhombus-tech.net//allwinner_a10/a10_boot_process/ |
15:25.51 | mikey_w | So why doesn't the image just image the mbr so u-boot is automatically dd(ed)? |
15:26.11 | mikey_w | Or is it? |
15:26.30 | zenitraM | it is |
15:27.19 | mikey_w | So the dd of u-boot.bin to the sd card as a separate entity is not actually needed, it's already there? |
15:27.37 | mikey_w | Just invisible. |
15:27.39 | vgrade | correct |
15:28.19 | mnemoc | scratches his head..... didn't I said that 10m ago? :( |
15:28.36 | mikey_w | I just ordered my A1000 so I have a bit of time to wait before I can try all of this. Just trying to line up my ducks. |
15:28.51 | mikey_w | TU |
15:33.34 | mnemoc | yw |
15:33.50 | mikey_w | when I attempt to fsck I get "fsck.ext4: No such file or directory while trying to open /dev/sdb2 |
15:33.51 | mikey_w | Possibly non-existent device?". |
15:34.21 | mikey_w | brb |
15:34.34 | mnemoc | what `fdisk -l /dev/sdb` tells you? |
15:38.26 | mikey_w | My bad it's sdc. |
15:38.37 | mnemoc | :) |
15:42.54 | mikey_w | even after running fsck the partition will not mount? |
15:43.07 | mnemoc | check dmesg |
15:44.17 | mikey_w | EXT4-fs (sdc2): bad geometry: block count 974336 exceeds size of device (962304 blocks) |
15:45.19 | mnemoc | correct the partition table and fsck.ext4 again |
15:45.30 | mikey_w | ok |
15:45.36 | mnemoc | .... distributing raw images is a baaaaaaaad idea |
15:47.48 | mikey_w | No joy will start over and hope for a better result. |
15:48.39 | mnemoc | it's done in the name of user friendliness, but ends up been far more complex than constructing your own image |
15:49.15 | mikey_w | I actually dd(ed) the 2nd partition and copied it to another card after imaging it so that should work. It mounts. |
15:51.15 | mikey_w | The fs look good and the first partion contains: evb.bin LOST.DIR mele.bin sys_config1.mele_mod.bin uImage. |
15:51.39 | mikey_w | Have to wait for my unit to try it. |
15:52.20 | mikey_w | This whole exercise has been instructional. |
15:52.37 | mnemoc | how is your kernel-fu? |
15:52.52 | mikey_w | fu? |
15:53.19 | mnemoc | do you know C and some kernel programming? |
15:53.38 | mnemoc | there are still a bunch of drivers in the 3.0 tree that need to be ported |
15:53.47 | mnemoc | and no need of a device for that |
15:54.00 | mikey_w | yes started with first slackware release where a real man made his own drivers. |
15:54.50 | mnemoc | https://github.com/amery/linux-allwinner :) |
15:55.04 | mikey_w | ARM compiling is limited but I hae done some ARM assembly stuff. |
15:55.22 | mnemoc | allwinner-v3.0-android branch :) |
15:55.22 | mikey_w | have |
15:55.48 | mikey_w | Something to wile away the time.:) |
15:56.03 | mikey_w | lol |
15:56.10 | mnemoc | dive in ;-) |
15:56.24 | mikey_w | Maybe. |
15:56.36 | cromo | mnemoc: which drivers are these? |
15:57.20 | mnemoc | hdmi and i2s sound, and v4l2 |
15:57.35 | mikey_w | So now I hae a race going on as to whether my A1000, my raspberry pi on my openpandora paid for in April, 2009 will arrive first. |
15:58.18 | mnemoc | cromo: I ported the spdif sound driver, should be good a example.... but haven't tested in in runtime |
15:58.33 | mnemoc | cromo: I provably screwed up the references in the codec link |
15:58.39 | mnemoc | probably* |
15:58.58 | mikey_w | later |
15:59.38 | mnemoc | disp (fb) drivers builds but it's awfully incomplete, and nand builds but doesn't find sh*t |
16:00.39 | mnemoc | I'm trying to understand the disp driver to extend it... all the rest needs volunteers |
16:03.52 | *** part/#arm-netbook QingPei (~qingpei@114.241.176.76) |
16:21.17 | CIA-158 | rhombus-tech: Cristian master * radbf15710739 /allwinner_a10/orders/.mdwn: |
16:21.26 | cromo | mnemoc: my confusion about all the oldconfig/menuconfig/config was because I was not using make ARCH=arm menuconfig/oldconfig/config, so it would assume my system's default x86_64 |
16:21.42 | mnemoc | right |
16:22.06 | mnemoc | that's why the very first step is man ARCH=arm sun4i_defconfig |
16:22.12 | mnemoc | s/man/make/ |
16:22.34 | cromo | hah, funny bot |
16:22.38 | cromo | never seen such |
16:22.49 | cromo | although I don't spend too much time on IRC |
16:24.07 | cromo | anyone knows if mac802 driver for the realtek chip in mele is usable? |
16:24.13 | cromo | and if it does support AP mode? |
16:24.17 | CIA-158 | rhombus-tech: Cristian master * r3d1736dc8ed7 /allwinner_a10/orders/ZeroOneCris1991.mdwn: |
16:26.28 | mnemoc | cromo: doesn't see it, only the customized driver (drivers/net/wireless/rtl8192cu/8192cu.ko) from allwinner's 2.6.36 tree detects the chip in my 3.0.28+ installation |
16:27.10 | mnemoc | drivers/net/wireless/rtlwifi/rtl8192cu/rtl8192cu.ko doesn't even try to initialize anything |
16:41.51 | cromo | hmm |
16:42.12 | cromo | how come? it's not even in staging, so it should work I think |
16:42.30 | mnemoc | you are very welcomed to find why the stock driver doesn't see the chip in the mele :) |
16:42.38 | cromo | I'll try |
16:44.43 | cromo | but now I can't even finish compiling the kernel |
16:45.18 | cromo | drivers/net/sun4i/sun4i_wemac.c:1876:6: error: ‘struct net_device’ has no member named ‘poll_controller’ |
16:45.29 | cromo | gcc 4.5 |
16:45.49 | mnemoc | o_O |
16:46.45 | mnemoc | gcc version has nothing to do on that |
16:46.59 | cromo | yeah, gcc 4.6 failed too |
16:47.06 | mnemoc | what remote/branch and how exactly did you configure it? |
16:47.20 | cromo | just a sec |
16:47.25 | cromo | I am trying with default .config |
16:47.41 | cromo | (I did some tweaks that shouldn't really matter) |
16:47.48 | cromo | (but still) |
16:48.25 | mnemoc | http://dpaste.com/735637/ |
16:48.56 | cromo | hah |
16:48.59 | cromo | same error |
16:49.12 | mnemoc | make distclean |
16:49.22 | mnemoc | make ... sun4i_defconfig |
16:49.27 | cromo | did |
16:49.30 | mnemoc | make ... uImage |
16:49.35 | mnemoc | no tweakings |
16:49.37 | cromo | I didn't do distclean though |
16:50.48 | penguin42 | remember to always include the ARCH=arm on every make; otherwise stuff gets rather confused |
16:51.00 | cromo | sure, I know |
16:51.13 | cromo | I am a programmer myself, too, just didn't touch kernel for a while ;) |
16:51.26 | cromo | and C in general |
16:51.32 | mnemoc | cromo: paste the whole thing, starting when you call sun4i_defconfig until you call uImage |
16:51.47 | cromo | just a sec, it's compiling now |
16:51.57 | cromo | could be that I didn't do distclean after some other tries |
16:52.24 | cromo | sound/soc/sun4i/hdmiaudio/sun4i-hdmiaudio.c:653:2: error: too few arguments to function ‘snd_soc_register_dai’ |
16:52.26 | penguin42 | yeh the kernel make system takes a bit of getting used to - like most large projects they have their own quirks |
16:52.31 | cromo | same thing |
16:52.44 | penguin42 | oh, well that just looks like a plane old bug/mismatch |
16:52.48 | cromo | make ARCH=arm sun4i_defconfig |
16:53.03 | cromo | make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- -j32 uImage modules V=99 |
16:53.36 | cromo | git branch: allwinner-v3.0-android |
16:53.45 | penguin42 | cromo: I find it useful to drop the -j32 when you're trying to figure out what's gone wrong - that way the error will always be the last thing |
16:53.58 | cromo | thx |
16:54.07 | cromo | I was wondering why does it continue ;) |
16:54.34 | penguin42 | -j32 is run 32 compiles in parallel - you might want to back it off if your build machine isn't as chunky as mnemoc's |
16:54.35 | mnemoc | you had another 31 workers finishing their taks :) |
16:54.42 | cromo | yeah |
16:55.22 | mnemoc | in general pick a -j between the numbers of cores and the double of it |
16:55.46 | cromo | build.sh script sets it to no of cores in system |
16:55.48 | penguin42 | wants mnemoc's 16 core box |
16:55.48 | mnemoc | but when hunting an error, drop -j |
16:55.58 | specing | Only 16? |
16:56.18 | cromo | core 2 duo here only :/ |
16:56.20 | specing | Wait till you get one of those 100-core CPUs from tilera |
16:56.22 | cromo | but waiting for new intel |
16:56.38 | mnemoc | http://www.hetzner.de/en/hosting/produktmatrix/rootserver-produktmatrix-ex has some nice deals :) |
16:57.59 | cromo | http://pastebin.com/GNRrY8z9 |
16:58.01 | cromo | same error |
16:58.08 | cromo | just cleaner output |
16:58.46 | mnemoc | not same error |
16:58.54 | mnemoc | that driver is known to be broken |
16:59.05 | cromo | indeed, not same error |
16:59.20 | cromo | I am hangovered today, things are a little blurry |
16:59.23 | mnemoc | apply this http://dpaste.com/735644/ |
16:59.31 | mnemoc | to get a safer defconfig |
16:59.48 | cromo | I did when you said, now I remember I did git clean |
16:59.59 | cromo | so it got reverted |
17:00.16 | mnemoc | save the .diff :) |
17:00.21 | mnemoc | or... commit it |
17:00.27 | mnemoc | but don't push |
17:00.42 | mnemoc | because it's a temporal hack |
17:00.50 | mnemoc | to have less noisy builds |
17:03.13 | cromo | had it saved, too, but again, git clean removed it anyway ;) |
17:13.31 | cromo | any reason for these USB mass storage readers to be configured as compiled into kernel? |
17:15.03 | mnemoc | I enabled none myself... so they come from the original defconfig from allwinner |
17:15.04 | cromo | it's probably from allwinners |
17:54.22 | cromo | drivers/net/sun4i/sun4i_wemac.c:1876:6: error: ‘struct net_device’ has no member named ‘poll_controller’ |
17:57.12 | mnemoc | very odd |
17:57.21 | cromo | will try again |
17:57.25 | mnemoc | when what config setting enables poll_controller |
17:57.30 | mnemoc | s/when/check/ |
17:57.32 | cromo | maybe I mised something |
17:57.58 | mnemoc | look for the definition of the struct |
17:58.19 | mnemoc | the field has to be in an #ifdef #endif block |
17:58.33 | mnemoc | and for some reason that's false for you |
18:15.12 | cromo | it was because I enabled netconsole |
18:27.34 | *** join/#arm-netbook vgrade2 (~martinbro@cpc2-nrte22-2-0-cust128.8-4.cable.virginmedia.com) |
18:34.42 | CIA-158 | rhombus-tech: nibble master * rd99af5834295 /allwinner_a10/a10_image.mdwn: Replaced option "skip" for "seek" in dd uboot command |
18:35.30 | mnemoc | cromo: it would be good to add the corresponding #ifdef / #endif in drivers/net/sun4i/sun4i_wemac.c to support the case when someone enables netconsole and insist he hasn't changed the defconfig :p |
18:38.04 | *** join/#arm-netbook cromo (~cromo@90-255-251-94.net.stream.pl) |
18:52.19 | mikey_w | What does the the mele A1000 use instead of MLO during boot if anything? |
18:53.02 | mnemoc | 17:30:05 < mnemoc> and http://rhombus-tech.net//allwinner_a10/a10_boot_process/ |
19:39.30 | vgrade2 | ok, \o/, X up on stock allwinner 2.6.36-andriod |
19:39.50 | vgrade2 | well stock plus a hack to fix the framebuffer driver |
19:42.15 | mnemoc | vgrade2: would you share that hack with me? |
19:43.13 | vgrade2 | absolutly |
19:43.54 | vgrade2 | usb host is fkd though so no keyboard or mouse, but progress |
19:48.49 | vgrade2 | sec for patch |
19:52.44 | mnemoc | vgrade2: do you have github account? |
19:53.06 | vgrade2 | http://pastie.org/3835332 |
19:53.16 | vgrade2 | yes |
19:53.26 | vgrade2 | sec for user |
19:53.33 | mnemoc | ouch |
19:53.49 | vgrade2 | martinbrook |
19:54.00 | vgrade2 | told you it was a hack |
19:54.07 | mnemoc | somehow that hack looks far uglier than.... expected |
19:55.46 | *** join/#arm-netbook ARE-TEE (~leebradbu@cpc20-stok15-2-0-cust539.1-4.cable.virginmedia.com) |
19:56.46 | mnemoc | vgrade2: I'm trying to fix fbcon support currently... do you have a log of the crash you were having with X that motivated this nice patch? :) |
19:57.13 | vgrade2 | same as zenitraM from yesterday |
19:57.16 | vgrade2 | sec |
19:57.46 | vgrade2 | http://pastie.org/3830099 |
19:58.52 | vgrade2 | fbcon is currently good enough to boot to ubuntu login manager |
19:59.23 | mnemoc | o_O |
19:59.25 | zenitraM | fbcon = framebuffer console != framebuffer |
19:59.41 | zenitraM | FBIOPUTCMAP = put character map? |
19:59.56 | vgrade2 | colourmap |
20:00.11 | mnemoc | how comes you can get fbcon while it relies in .fb_imageblit which is not implemented at all? |
20:01.16 | vgrade2 | sry, disp driver is good enough to boot to ubuntu login manager |
20:01.25 | mnemoc | still your crash is within DE_BE_Set_SystemPalette so it will also strike me soon |
20:01.47 | vgrade2 | I'm guessing thats never called from andriod |
20:02.01 | mnemoc | right |
20:03.05 | mnemoc | DE_BE_Set_SystemPalette+0x28 is drivers/video/sun4i/disp/de_bsp/de/ebios/de_be.c:347 |
20:03.15 | mnemoc | all very nasty in there |
20:03.23 | mnemoc | also full of blobs |
20:05.28 | mnemoc | vgrade2: have you noticed any side effect of your hack yet? |
20:05.58 | vgrade2 | nope colours look ok on login screen |
20:06.24 | vgrade2 | need to get usb host working so I can get a mouse and keyboard |
20:06.35 | *** join/#arm-netbook cromo (~cromo@90-255-251-94.net.stream.pl) |
20:06.43 | vgrade2 | has there been any work done on usb host ? |
20:06.50 | mnemoc | odd, my 3.0 detects my usb keyboard just fine |
20:07.22 | mnemoc | but haven't had the chance to type on it yet as I don't have a console or X yet |
20:07.55 | mnemoc | vgrade2: using a hub or something fancy? |
20:08.44 | vgrade2 | nope, |
20:09.10 | mnemoc | and still no reaction on dmesh when you plug it? |
20:10.33 | vgrade2 | I have hub 1-0:1.0: cannot reset port 1 (err = -32) |
20:10.37 | vgrade2 | in boot log |
20:10.43 | vgrade2 | let me try again |
20:10.59 | mnemoc | Apr 22 20:10:43 localhost kernel: [ 695.320000] usb 1-1.2: new low speed USB device number 4 using sw-ehci |
20:11.02 | mnemoc | Apr 22 20:10:44 localhost kernel: [ 695.860000] input: USB Keyboard as /devices/platform/sw-ehci.1/usb1/1-1/1-1.2/1-1.2:1.0/input/input6 |
20:11.06 | mnemoc | Apr 22 20:10:44 localhost kernel: [ 695.960000] generic-usb 0003:04D9:1603.0003: input: USB HID v1.10 Keyboard [ USB Keyboard] on usb-sw-ehci-1.2/input0 |
20:11.34 | zenitraM | vgrade2: are you using the evb script.bin? |
20:11.34 | mnemoc | and I haven't done anything special after moving the driver from 2.6.36 |
20:11.36 | zenitraM | evb.bin |
20:11.39 | zenitraM | (well the default one) |
20:11.40 | zenitraM | if so |
20:11.48 | zenitraM | plug the kb/mouse to another port |
20:11.55 | zenitraM | the evb has a port configured differently |
20:12.03 | mnemoc | uh |
20:12.29 | zenitraM | and the config for that port doesn't match the port, so it complains about it even if nothing is connected |
20:12.49 | mnemoc | zenitraM: got the .fex fixed yet? :) |
20:13.11 | zenitraM | not yet. all those cold-reboots somehow messed my root partition and now it doesn't boot anything |
20:13.23 | mnemoc | narf |
20:13.38 | zenitraM | but all that was left was... the gpio ports and not much more |
20:13.41 | vgrade2 | zenitraM, I'm still on the .bin which cam with the ubuntu iamge |
20:13.50 | zenitraM | vgrade2: that is the evb one |
20:13.52 | zenitraM | use another port |
20:14.09 | vgrade2 | or use the script.bin renamed to evb.bin |
20:14.42 | zenitraM | X will hang after login on that one |
20:15.48 | vgrade2 | zenitraM, nothing on any port |
20:15.48 | zenitraM | for still unknown reasons |
20:16.17 | vgrade2 | zenitraM, how did you get X up? |
20:16.22 | zenitraM | mmm then it might be a 2.6.36 issue |
20:16.32 | zenitraM | couldn't get it up on 2.6.36 |
20:17.00 | zenitraM | only on the 3.0.8+, and was investigating why it works on the evb .bin but not on the .bin used by the mele |
20:17.20 | vgrade2 | ok, |
20:17.56 | vgrade2 | I'll give the script.bin a go now I have X up on stock 2.6.36 |
20:18.23 | vgrade2 | mnemoc, do you have a .config for your 3.0 |
20:21.54 | CIA-158 | rhombus-tech: Andreas master * r647ee5561cd6 /allwinner_a10/orders/lonelycoder.mdwn: |
20:22.14 | mnemoc | vgrade2: apply http://dpaste.com/735714/ to turn sun4i_defconfig is one that finishes fine |
20:22.41 | mnemoc | it disables broken drivers |
20:23.05 | mnemoc | and turns problematic into modules for better debugging |
20:23.47 | mnemoc | vgrade2: it's the allwinner-3.0-android branch |
20:25.12 | cromo | mnemoc: I managed to compile and allwinner-3.0 branch |
20:25.14 | mnemoc | zenitraM: vgrade2 "fixed" 2.6.36 with http://pastie.org/3835332 |
20:25.28 | cromo | however it appears _much_ slower than allwinner's 3.0.8? |
20:25.31 | cromo | is this expected? |
20:25.47 | mnemoc | no idea, I've never used allwinner's 3.0.8 |
20:26.15 | cromo | logging in with ssh takes 2 secs now, it was almost immediate before |
20:26.21 | zenitraM | mnemoc: i know, and that doesn't seem like an ideal patch, but oh well |
20:26.33 | mnemoc | cromo: it can also be about the governor.... ondemand likes to set mine to 60Mhz |
20:28.12 | cromo | CONFIG_CPU_FREQ_GOV_ONDEMAND=y |
20:28.34 | mnemoc | try interactive |
20:29.04 | mnemoc | 60Mhz ain't sexy |
20:30.30 | cromo | interactive didn't help |
20:30.37 | cromo | echo 1008000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq did |
20:30.44 | cromo | so it's def governor's fault |
20:31.31 | cromo | seems fantasy did the trick |
20:31.46 | mnemoc | now that I'm clonning the git repo to the SD *cough* ondemand was kind enough to set it to 1008000 too |
20:33.29 | cromo | which of the governors should be able to scale to max frequency quickiest? |
20:33.43 | mnemoc | i'm analphabet there |
20:33.43 | cromo | AFIK interactive is the one |
20:34.22 | vgrade2 | zenitraM, mnemoc, I agree its fugly |
20:35.28 | mnemoc | vgrade2: if it doesn't have obvius side effects and lets people work on fixing other problems in the meantime, it's perfect :) |
20:36.15 | vgrade2 | I'll let you know if I see anything wrong once I get a mouse/kbd |
20:36.33 | vgrade2 | still no usb host with script.bin |
20:37.07 | vgrade2 | does the stock andriod have /proc/config.gz |
20:41.37 | mnemoc | vgrade2: no |
20:42.31 | mnemoc | meh, I run out of candles :( |
20:42.59 | vgrade2 | mnemoc, thanks |