| 00:00.06 | hno | already found it. |
| 00:01.04 | arokux | have you seen this guys? http://www.genesi-usa.com/products/smartbook |
| 00:01.42 | arokux | $199 |
| 00:07.46 | hno | arokux, I think that quote is malquoted. More likely meant NEON compared to VFP. Most A8 have pretty crap VFP. The A9 VFP is much improved. |
| 00:08.16 | arokux | hm.. |
| 00:13.00 | hno | Most descriptions of the A8 VFP-lite is "slow nonpipelined", "stalls" and so on.. |
| 00:13.56 | arokux | ah, ok, there there is indeed an error on this page |
| 00:15.39 | CIA-46 | rhombus-tech: jussi master * rb74a64530e9d /allwinner_a10/orders/jussi.mdwn: |
| 00:16.22 | arokux | hno, but look at the next page: http://wiki.debian.org/ArmHardFloatPort/VfpComparison |
| 00:16.27 | arokux | Cortex-A9 processors have far, far better FPU instruction timings meaning that the overhead is far, far higher. |
| 00:18.54 | hno | That's about softfp, which takes the penalty of transfers between integer and fp registers far more often. |
| 00:19.51 | hno | on A8 you don't care so much as the VFP instructions are so slow anyway, but on A9 the tansfer overhead becomes very noticeable because the VFP instructions as such are very fast. |
| 00:20.25 | arokux | oh, I see, thanks |
| 00:36.29 | toxicpsion | Anyone working on reversing the Livesuit/Phoenix Protocols? |
| 00:40.10 | hno | toxicpsion, need to get an A10 device first. |
| 00:40.22 | toxicpsion | have one.... Gemei G2. |
| 00:40.40 | toxicpsion | got Wireshark Dumps of my last flash too ;) |
| 00:41.44 | arokux | what are those protocols used for? |
| 00:41.48 | toxicpsion | but digging through 27m of packet captures sucks. |
| 00:42.06 | toxicpsion | arokux: flashing the device over USB |
| 00:42.14 | toxicpsion | 275M* |
| 00:42.57 | hno | should be pretty repetetive, if you eleminate the firmware data as such. |
| 00:44.03 | toxicpsion | it's just finding what is firmware, and what is control.... it uses 2 different Addresses, /w 2 Bulk - Vendor Specific Class Endpoints. |
| 00:47.20 | hno | should be possible to dump the rom and analyze stuff from that end as well. |
| 00:48.11 | toxicpsion | maybe... but flashing happens in Brom, or boot0... not totally sure where. |
| 00:48.29 | hno | BROM. |
| 00:48.52 | toxicpsion | kind of what i thought. |
| 00:48.52 | hno | the only ROM you have in the deviec most likely. |
| 00:49.05 | arokux | hno, have you order mele a1000 from tom also? |
| 00:49.10 | hno | yes |
| 00:50.50 | arokux | bastards.. to flash linux-based device they are going to make me using windows........ |
| 00:51.06 | hno | arokux, or use an SD card. |
| 00:51.48 | arokux | aha.. so not bastards, sorry |
| 00:54.06 | hno | Tom have already worked out how to make a correct u-boot spl for SD-card booting. |
| 00:54.30 | hno | NAND is not done yet. |
| 00:54.57 | arokux | mm.. so the official method is to use Livesuite? |
| 00:55.12 | hno | official tool is livesuite. |
| 00:55.24 | arokux | so bastards. |
| 00:55.43 | arokux | is it legal to reverseengineer the protocol? |
| 00:56.03 | hno | and official bootloader is eGON I think, all closed. |
| 00:56.27 | hno | arokux, why shouldn't it be legal? |
| 00:56.49 | arokux | I do not know, just asking. |
| 00:57.48 | hno | I doubt they have any trademark data within the protocol. |
| 01:00.11 | hno | reverse engineering the BROM is a little grey if the result is used to write another BROM, but not if the goal is to document how it works. |
| 01:01.21 | arokux | but the source code to write another BROM is documentation or? |
| 01:01.47 | hno | ? |
| 01:02.47 | arokux | once you know how it works you'll write a program to do it |
| 01:03.39 | arokux | then the source code is the documentation.. or sources are not considered as documentation? |
| 01:05.21 | hno | if you use a disassembler and translate the result to source is's still the original work under original copyright, with a bit of copyright owned by you as well. |
| 01:06.36 | arokux | mm.. and? |
| 01:06.50 | hno | but if you disassemble the code and document the effect of the code in terms of USB protocol and then uses that documentation to write another program doing the same then it's ok. |
| 01:07.20 | arokux | ok! thanks |
| 01:10.56 | hno | copyright do not allow you to reproduce material unless the copyright holder allows you to. Which means yoiu are perfectly allowed to do the first for your own use, but not to share that result with anyone else as it's not yours. |
| 01:12.43 | hno | in the second case there is no such restrictions, The protocol documentation is an independent work, and any program written from it is equally independent. |
| 01:16.00 | hno | toxicpsion, regarding livesuite. It's only the very first part that is interesting, where boot0 is uploaded to SRAM in FEL mode. |
| 01:17.43 | hno | if the actual BROM in the A10 is anything like the copy found in the SDK then it's bound to be very simple x-modem based transfer. |
| 01:26.19 | arokux | I've just checked E-Mail, somebody have found the datasheet with confidantial across of every page |
| 01:26.36 | hno | yes |
| 01:26.54 | toxicpsion | mhm... just noticed that one too... just gives pinouts and clock refs. |
| 01:27.06 | hno | yes, it does not say much interesting. |
| 01:27.38 | arokux | what is left? |
| 01:27.50 | hno | left for what? |
| 01:28.19 | arokux | what would be interesting that is not known yet |
| 01:28.39 | hno | plenty. |
| 01:31.12 | hno | It's not much that is known really. |
| 01:33.20 | hno | <PROTECTED> |
| 01:33.25 | arokux | but there is the kernel sources? |
| 01:35.26 | *** join/#arm-netbook hipboi (~hipboi@221.4.213.95) |
| 01:36.32 | arokux | I assume this one is already known: http://www.cnx-software.com/2011/12/28/allwinner-a10a1x-processor-resources-development-board-and-sdk/ |
| 01:37.29 | arokux | ah, it even references our site.. |
| 01:39.25 | hno | Missing: GPU, 2D acceleration, Video CODEC (both in & out), etc, details all unknown. |
| 01:39.55 | arokux | so, surprisingly I've built my toolchain without any error. so crosstool-ng is highly recommended, if your distribution doesn't provide a crosscompiling packages. |
| 01:40.27 | arokux | hno, but Mele is already reverse engineered, there is still something unknown? |
| 01:41.04 | hno | "already".. they have started to scratch on it's surface only. |
| 01:41.50 | hno | assumin you meant ARM MALI GPU. |
| 01:41.59 | arokux | yes, mali. |
| 01:42.09 | arokux | :D they were saying it as though everything was done. |
| 01:42.15 | hno | Who? |
| 01:43.11 | arokux | http://www.phoronix.com/scan.php?page=news_item&px=MTA1NjY |
| 01:43.43 | arokux | ... that's providing a reverse-engineerED open-source ARM Mali driver. |
| 01:47.45 | hno | Read it again. No where does it say it's even close to ready. |
| 01:48.16 | A2Sheds | he's been able to draw lines with the open mali driver |
| 01:48.33 | hno | Yes. |
| 01:49.15 | hno | or even basic OpenGL ES objects. |
| 01:50.09 | A2Sheds | #lima for more info |
| 01:51.05 | arokux | I did not go to the end of the article.. ( |
| 01:51.53 | hno | Hope to get there in time. First focus is getting something that boots.. |
| 02:37.18 | *** join/#arm-netbook QingPei (~qingpei@inet-hqmc01-o.oracle.com) |
| 02:45.09 | arokux | http://www.cnx-software.com/2012/03/05/raspberry-pi-releases-arch-linux-sd-card-image/ |
| 03:26.52 | *** part/#arm-netbook QingPei (~qingpei@inet-hqmc01-o.oracle.com) |
| 03:31.02 | hno | wonders what is going on with the Fedora Raspberry pi image. Was supposed to be ready long ago. There is even a GUI installer for it. |
| 03:52.31 | CIA-46 | rhombus-tech: R_i master * r94b856544355 /allwinner_a10/orders/OurWorld-001.mdwn: as per request from luke (takes me a while!!) |
| 03:59.41 | CIA-46 | rhombus-tech: R_i master * r81db550fe9bd /allwinner_a10/orders/OurWorld-001.mdwn: |
| 03:59.42 | CIA-46 | rhombus-tech: my comments seem to have disappeared?? - assuming the board is working, ie |
| 03:59.42 | CIA-46 | rhombus-tech: final- when could 100,000 be ordered (per month) assuming $18.00 per unit, can |
| 03:59.42 | CIA-46 | rhombus-tech: the same supplier be used to get a 'mother' PCB assembled (with PS and adapters |
| 03:59.42 | CIA-46 | rhombus-tech: I want??)?? |
| 04:02.48 | *** join/#arm-netbook QingPei (~qingpei@inet-hqmc02-o.oracle.com) |
| 04:06.44 | *** join/#arm-netbook toxicpsion (~toxicpsio@out-ab-215.wireless.telus.com) |
| 04:31.46 | *** join/#arm-netbook toxicpsion (~toxicpsio@out-ab-209.wireless.telus.com) |
| 06:49.54 | CIA-46 | rhombus-tech: Lloyd master * re2b35fd79419 /allwinner_a10/orders/lrp.mdwn: |
| 08:25.39 | *** join/#arm-netbook arokux (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 09:03.42 | *** part/#arm-netbook QingPei (~qingpei@inet-hqmc02-o.oracle.com) |
| 10:45.30 | *** join/#arm-netbook arokux (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 10:53.01 | mnemoc | zenitraM: and it was sweden post after all |
| 10:53.10 | mnemoc | (not swiss) |
| 10:53.16 | zenitraM | sweden? |
| 10:53.29 | zenitraM | you got your tracking? |
| 10:53.36 | mnemoc | yes, RRxxxxxSE |
| 10:54.02 | zenitraM | did tom email it to you? he hasn't sent me anything yet |
| 10:54.29 | mnemoc | in the status of the aliexpress order |
| 10:54.34 | zenitraM | oh i'll check |
| 10:55.16 | zenitraM | mine's still on shipment required |
| 10:55.23 | zenitraM | maybe he hasn't added it yet |
| 10:58.59 | mnemoc | probably.... I doubt he went to shenzehn to ship only some boxes :) |
| 10:59.51 | mnemoc | I got two tracking numbers, one internal for espeedpost and the preasigned RRxxxxxSE |
| 11:00.41 | mnemoc | needs to buy a UPS for the adsl router... |
| 11:11.02 | zenitraM | oh well, let's hope it arrives fast :) |
| 11:12.07 | mnemoc | and untaxed :p |
| 11:14.11 | zenitraM | that's not what worries me the most |
| 11:14.30 | zenitraM | after having received huge boxes with 200$ in… t-shirts with lights |
| 11:15.09 | zenitraM | although.. isn't .se on the schengen space? |
| 11:16.34 | zenitraM | because if it gets stopped, it gets *on sweden* |
| 11:23.39 | mnemoc | yup, that is my fear |
| 11:26.04 | zenitraM | well let's pray |
| 11:26.10 | zenitraM | does the espeedpost tracking say.. anything? |
| 11:26.23 | mnemoc | yes, 3 entries :) |
| 11:26.45 | zenitraM | and where is it? |
| 11:27.32 | mnemoc | received by espeedpost yesterday before midnight and handed over to sweden post today at 16:25 ... assuming HK timezone |
| 11:28.27 | mnemoc | feels strange to see places where these things work on sunday until (at least) midnight |
| 11:28.55 | zenitraM | mmm have you tried the RR on sweden post? |
| 11:29.13 | mnemoc | yes, no match yet |
| 11:29.23 | mnemoc | it's only preassigned |
| 11:29.36 | zenitraM | handed to sweden post that fast… either they have handed it on China, or they're too fricking fast |
| 11:29.41 | mnemoc | I believe they are sent on bulk to .se |
| 11:30.17 | mnemoc | sweden post has to have a bulk shipment office in hongkong too |
| 11:31.50 | zenitraM | or they shipped to sweden using a *third* shipping agency |
| 11:32.07 | mnemoc | also possible |
| 11:32.56 | mnemoc | hno is from .se, he may know |
| 11:40.51 | mnemoc | btw, about ticket #3, I couldn't find any change in allwinner's tree not present in mine |
| 11:46.40 | CIA-46 | rhombus-tech: Luca master * r549e7e722ae6 /allwinner_a10/orders/http:__47____47__rhombus-tech.net__47__allwinner__95__a10__47__orders__47__.mdwn: |
| 11:48.43 | *** join/#arm-netbook gordan (~gordan@178.110.7.61) |
| 11:56.22 | *** join/#arm-netbook arokux (~arokux@merkur178.inf.uni-konstanz.de) |
| 12:01.25 | *** join/#arm-netbook zenitraM (~zenitraM@2607:f358:1:fed5:11:0:1018:14) |
| 12:08.18 | CIA-46 | rhombus-tech: Roman master * r1ec5deec320c /evaluated_cpus.mdwn: |
| 12:10.26 | CIA-46 | rhombus-tech: Roman master * r2bb3663e0bec /rock-chips_rk30xx.mdwn: Added Rockchip RK30xx SoC. |
| 12:11.32 | CIA-46 | rhombus-tech: Roman master * r5d8b65576425 /rock-chips_rk30xx.mdwn: |
| 12:21.08 | CIA-46 | rhombus-tech: lkcl master * raeec1da70cfb / (9 files in 3 dirs): tidyup |
| 12:22.50 | CIA-46 | rhombus-tech: lkcl master * re28f0025d151 /evaluated_cpus.mdwn: |
| 12:25.13 | *** join/#arm-netbook hipboi (~hipboi@123.89.249.238) |
| 12:28.37 | hno | mnemoc, I don't know what shipping Tom used. Also no shipment details on my order. |
| 12:29.56 | mnemoc | hipboi: ----^ :) |
| 12:30.23 | hipboi | hno: sorry |
| 12:30.42 | hipboi | Will send the mail right now |
| 12:31.49 | hipboi | European counties are the Sweden post air mail |
| 12:33.47 | mnemoc | my wonder is where will them pass the customs, in sweden (by swedish rules) or in the destination country, by local rules (spanish in my case) |
| 12:34.41 | mnemoc | hno is .se, so for him both are the same :p |
| 12:36.23 | zenitraM | i guess we'll have to wait |
| 12:36.55 | mnemoc | hno: how are cheap airmail packages sent from china normally taxed there? |
| 12:37.05 | zenitraM | (if they are) |
| 12:37.08 | gordan | Heh... My ISP is going to hate me. Since I mentioned that RedSleeve (RHEL6 port for ARM) is beta-ready on Saturday, I've had 40GB of downloads of it. :) |
| 12:37.52 | gordan | thinks that maybe he should have made a tar ball of the lot and set up a torrent tracker for it instead... |
| 12:38.41 | gordan | I'm actually quite amazed the DSL connection it lives on still feels perfectly normal despite the fact that it has been completely maxed on for 2 days now. |
| 12:40.09 | arokux | can I somehow read the logs of this channel of today morning? |
| 12:40.50 | zenitraM | !topic |
| 12:41.04 | arokux | of _today_ topic. |
| 12:41.16 | arokux | s/topic/morning |
| 12:41.53 | zenitraM | oh true, didn't notice the bot doesn't log real-time |
| 12:42.14 | mnemoc | slaps ibot |
| 12:44.14 | mnemoc | arokux: ssh+screen+irssi, the happy path to always-connected |
| 12:44.31 | zenitraM | or znc + whatever client you want |
| 12:47.12 | arokux | hipboi, does it mean there is already Linux 3.0.8 ported to A10 somewhere? http://www.wits-tech.com/pages/board.jsp |
| 12:47.55 | zenitraM | there is, as my tablet has a ICS firmware with a 3.0.8 kernel |
| 12:47.56 | hipboi | yes |
| 12:48.11 | hipboi | Our internal git tree |
| 12:49.33 | arokux | any plans to put it public? :) |
| 12:50.47 | mnemoc | hipboi: btw, my tree is not older that the "official-allwiner-2.6.36" on github... but I suppose the official-allwiner-2.6.36 is outdated :-/ |
| 12:55.35 | *** join/#arm-netbook kpotato (~lambda@wifi-81-11.uc3m.es) |
| 12:56.05 | hipboi | arokux: maybe I can persuade my manager to put it public |
| 12:56.44 | zenitraM | given that is GPL, he *has* to :P |
| 12:57.03 | hipboi | mnemoc: yes, we do not update the kernel source once we started porting |
| 12:57.36 | hipboi | When we need newer kernel, we just re port it |
| 12:57.58 | hipboi | based on a newer version kernel |
| 12:59.02 | mnemoc | there is a 3.3 android kernel tree already.... not sure if they aim to use it in ICS or in the next release |
| 13:08.58 | arokux | when allwinner normally publishes there kernels? does it make sense to port ourselves if they publish theirs soon? |
| 13:10.01 | mnemoc | arokux: they first leaked as part of ainol's SDK.... then hipboi got their 2.6.36 tree disclosed in github |
| 13:10.45 | arokux | mnemoc, so it happens randomly? |
| 13:11.28 | mnemoc | not sure if "randomly" is the right term |
| 13:11.38 | mnemoc | but there is no schedule |
| 13:24.22 | zenitraM | hipboi: my melee's are coming using DHL? |
| 13:25.32 | hipboi | Yes |
| 13:25.50 | hipboi | Check my mail |
| 13:26.26 | zenitraM | well thanks :) i hope the package isn't stopped by customs |
| 13:26.51 | mnemoc | uh |
| 13:27.00 | hno | mnemoc, the ones I have bougt have been delivered without customs processing. |
| 13:27.20 | mnemoc | hno: neat |
| 13:27.29 | hipboi | Because two boxes is more than 2kg weight |
| 13:27.43 | arokux | mnemoc, so why not ask about the schedule before porting ourselves? |
| 13:28.06 | mnemoc | arokux: because there is no schedule |
| 13:28.29 | arokux | :( |
| 13:29.03 | CIA-46 | rhombus-tech: Oniz master * r3bc9c4417b4a /allwinner_a10/orders/myreg_at_live_dot_com.mdwn: |
| 13:29.58 | zenitraM | hipboi: did you put any price on the customs declaration? |
| 13:30.35 | mnemoc | I've got DHL charging me $80 for a $50 package once... |
| 13:30.54 | mnemoc | hates couriers with all his guts |
| 13:31.33 | hipboi | There is a discount |
| 13:31.55 | hipboi | About 200rmb to Spain of 2kg |
| 13:32.30 | mnemoc | let see who's faster, .se post or dhl :p |
| 13:33.08 | zenitraM | hipboi: i mean: did you have to fill any paper (for the custom) detailing the price of what's inside? |
| 13:33.31 | hipboi | Yes |
| 13:33.37 | zenitraM | what price did you put? |
| 13:33.46 | hipboi | 130 used |
| 13:33.51 | hipboi | Usd |
| 13:34.01 | hipboi | For yours |
| 13:34.05 | zenitraM | uh-oh |
| 13:34.07 | hno | mnemoc, the fee is in two parts. A fixed handlign fee of quite significant amount, and customs/taxes. |
| 13:34.27 | hipboi | Others are 65 usd |
| 13:34.56 | zenitraM | hipboi: you probably shouldn't have put the real price :( |
| 13:35.19 | zenitraM | because they stop the parcels at our countries and bill us import taxes |
| 13:35.37 | mnemoc | to spain is far better to ship multiple under-2kg packages than using fancy couriers |
| 13:36.10 | zenitraM | but well let's hope they don't charge us too much (or charge us at all) |
| 13:36.28 | mnemoc | as hno said the problem is not really the tax. but the fee they charge for the kindness of forcing the customs to tax you |
| 13:36.30 | hipboi | I have packed the two boxes together |
| 13:37.13 | hipboi | It's the next day I know it's over weight that can not shipped by Sweden post |
| 13:37.52 | hipboi | At that time the boxes are already in espeed post |
| 13:38.06 | zenitraM | don't worry, we'll just pay the customs |
| 13:38.14 | zenitraM | it's not your fault, we should have told you |
| 13:38.27 | zenitraM | to put a lower price for customs |
| 13:38.41 | hno | Yes, for letters there is a weight limit of 2kg. parcels can be heavier but costs more. |
| 13:46.21 | mnemoc | hipboi: btw, if the 3.0 git is complicated, a fat git diff common/android-3.0..HEAD would be welcomed too :p |
| 13:46.43 | *** join/#arm-netbook QingPei1 (~qingpei@124.64.127.218) |
| 13:47.28 | hipboi | mnemoc: ok |
| 14:06.45 | *** join/#arm-netbook hipboi (~hipboi@123.89.249.238) |
| 14:24.35 | *** join/#arm-netbook kadaki01 (~lambda@wifi-81-11.uc3m.es) |
| 14:38.46 | *** join/#arm-netbook arokux (~arokux@merkur178.inf.uni-konstanz.de) |
| 14:45.15 | CIA-46 | rhombus-tech: Peter master * r284bcd4c9061 /allwinner_a10/orders/oosmdizajn.mdwn: |
| 14:46.44 | arokux | we have so many orders already.. |
| 14:48.31 | mnemoc | and still no connectors |
| 14:48.49 | arokux | what is the last news? |
| 14:49.28 | mnemoc | A2Sheds: ---^ ? |
| 14:50.35 | hno | How many do we have at alpha? And is there any decision on how to handle the alpha? |
| 14:51.45 | arokux | as of 2012Mar01 the alpha units have reached 141 |
| 14:53.50 | CIA-46 | rhombus-tech: Kristoffer master * r0158cc694fb8 /allwinner_a10/orders/ke84.mdwn: |
| 14:54.18 | hno | given the number of unknowns I would be very surprised if the alpha works out of the box. But hopefully at least ddr & usb works without too much rework. |
| 14:58.31 | mnemoc | that's why I only counted myself with one alpha board |
| 14:59.02 | mnemoc | i don't want to end up with a dozen of defective cards :-/ |
| 15:01.49 | mnemoc | software defects or missing supports I can easily deal with... but hardware is a differnt story |
| 15:05.13 | A2Sheds | if they assembled the boards with high priced connectors everybody would be complaining about how much the boards cost |
| 15:06.02 | mnemoc | which is the maximal final price considering high priced connectors? |
| 15:06.37 | A2Sheds | mnemoc: why don't you think about it? |
| 15:07.03 | mnemoc | you already know sample prices, i don't |
| 15:07.54 | mnemoc | i don't have any reference |
| 15:08.30 | A2Sheds | heh backseat driver |
| 15:08.36 | mnemoc | but still find the "advertised" price tag very unrealistic |
| 15:09.34 | A2Sheds | "are we there yet" ? |
| 15:10.01 | mnemoc | i only would like to know an honest price range |
| 15:10.56 | mnemoc | considering the "high prices" you already have and the ideals you are aiming |
| 15:11.37 | mnemoc | but not "BOM", but the prices one would have to pay |
| 15:11.39 | A2Sheds | complaining and discouraging other devs... |
| 15:11.49 | mnemoc | I'm not complaining |
| 15:12.40 | mnemoc | I would perfectly use the eoma68 cards at work for $100 a piece |
| 15:13.08 | mnemoc | maybe even $150 |
| 15:16.00 | mnemoc | but I can't go to my boss and "sell" him a solution based on $15 modules if they end up costing $150 |
| 15:20.37 | mnemoc | but also, I don't need them to cost $15, nor $50 |
| 15:23.12 | mnemoc | and asking for prices or times when not been interested/involved in hardware production doesn't mean "backseat driver" it only means one needs references to set prices and times of whatever one is interested in building upon that hardware |
| 15:24.22 | hno | mnemoc, so sell the idea to him at $100 or whatever high figure you deem realistic. The lower range is only when this hits mass production. |
| 15:24.37 | hno | If the price ends up at $15 he will only be happier. |
| 15:24.52 | mnemoc | is $100 a safe upper limit? |
| 15:28.38 | zenitraM | runs make |
| 15:33.29 | zenitraM | oh, now this is fun. |
| 15:33.49 | zenitraM | my tablet screen decided to keep pressing itself at one point randomly |
| 15:39.30 | mnemoc | zenitraM: found a way to get the kernel verbose yet? |
| 15:39.59 | zenitraM | mnemoc: going to try just now |
| 15:41.38 | zenitraM | mnemoc: doesn't work :( |
| 15:41.52 | zenitraM | sun4i#setenv bootargs console=ttyS0 loglevel=8 earlyprintk=ttyS0 root=/dev/nandd |
| 15:41.53 | zenitraM | sun4i#fatload nand 7 50000000 uImage |
| 15:41.53 | zenitraM | reading uImage |
| 15:41.53 | zenitraM | 3875728 bytes read |
| 15:41.53 | zenitraM | sun4i#bootm 50000000 |
| 15:42.03 | zenitraM | and then it gets stuck at "Starting kernel ..." |
| 15:42.39 | mnemoc | narf |
| 15:42.46 | zenitraM | gonna try something else |
| 15:43.04 | zenitraM | try to make a bImage and boot it directly from boot1 instead of u-boot |
| 15:43.46 | mnemoc | remember our kernel expects to find the boot params at 0x40000100 |
| 15:43.58 | zenitraM | build_sun4i.sh: ${OBJCOPY} -R .note.gnu.build-id -S -O binary vmlinux output/bImage |
| 15:44.06 | zenitraM | the thing is |
| 15:44.16 | zenitraM | u-boot was chain loaded by boot1 |
| 15:44.28 | zenitraM | and put the params on its place |
| 15:44.38 | zenitraM | so unless u-boot has overwritten that address.. they should still be there |
| 15:53.09 | zenitraM | and now, for more fun: weird partitioning |
| 15:53.28 | zenitraM | according to df: |
| 15:53.37 | zenitraM | <PROTECTED> |
| 15:54.07 | zenitraM | that is the boot partition… *but* on NAND the partition is smaller |
| 15:54.11 | zenitraM | <6>[ 618.170000] attempt to access beyond end of device |
| 15:54.11 | zenitraM | <6>[ 618.180000] nanda: rw=1, want=32849, limit=32768 |
| 15:57.47 | mnemoc | 32768*512=16M |
| 15:58.15 | zenitraM | so, not enough space to make both bImages fit |
| 15:59.17 | zenitraM | great, now I can't copy anything without going outside the limits and I have erased the original bImage. |
| 15:59.20 | zenitraM | FEL to the rescue! |
| 15:59.29 | mnemoc | I don't know where that limit comes from, but it's clearly not about the partition size |
| 16:00.24 | zenitraM | from u-boot: |
| 16:00.25 | zenitraM | -total partitions:8- |
| 16:00.25 | zenitraM | -name- -start- -size- |
| 16:00.26 | zenitraM | BOOTFS : 100000 1000000 |
| 16:01.35 | *** part/#arm-netbook QingPei1 (~qingpei@124.64.127.218) |
| 16:01.40 | zenitraM | the FAT header (or the nand driver? where is the partition info stored?) reports the partition to be much larger than it really is |
| 16:06.33 | arokux | what happened with A2Sheds? |
| 16:24.40 | zenitraM | booting from boot1 results in.. the kernel also silent :( |
| 16:25.39 | mnemoc | :( |
| 16:31.13 | zenitraM | so it is either not outputting anything or.. outputting to the real uart and ignoring script.bin |
| 16:36.34 | mnemoc | from hno I understood early printk will use the address returned by addruart |
| 16:37.41 | mnemoc | therefor ignoring script.bin would be expectable |
| 17:09.22 | CIA-46 | rhombus-tech: joph master * r82f7119d5564 /allwinner_a10/orders/joph.mdwn: |
| 17:15.20 | CIA-46 | rhombus-tech: joph master * r575ba9d93ad4 /allwinner_a10/orders/joph.mdwn: |
| 17:17.10 | CIA-46 | rhombus-tech: joph master * r0ade78dd4741 /allwinner_a10/orders/joph.mdwn: |
| 17:23.11 | lkcl | arokux: he's waiting for the connectors, like we all are. |
| 17:24.32 | arokux | lkcl, he could have said it |
| 17:24.34 | lkcl | QingPei has found some, but there's no datasheets! *lol*. so we have to get some samples and actually measure them up. |
| 17:24.58 | lkcl | well he's got other things to do! he's running a business and has had to go out on some errands. patience! |
| 17:25.23 | lkcl | dings arokux round the ear :) |
| 17:25.24 | mnemoc | QingPei = seven? |
| 17:25.35 | arokux | lkcl, cann't you write an e-mail asking for datasheets? |
| 17:25.44 | lkcl | mnemoc: i honestly don't know. i believe it's a different person |
| 17:25.57 | mnemoc | :) |
| 17:25.57 | lkcl | arokux: yeess, we've done that... the factory said that there aren't any :) |
| 17:26.10 | lkcl | it's *cough* best not to ask why |
| 17:26.32 | arokux | lkcl, I wonder how this factory sells them :D |
| 17:26.32 | mnemoc | people really needs an special mindset to build things in china :p |
| 17:27.02 | lkcl | in very large boxes that fell off a back of a lorry, arokux. in some cases, fell of the back of a _lot_ of lorries. |
| 17:27.17 | lkcl | teehee |
| 17:28.00 | lkcl | yes, it's rather interesting to hear that it is normal for companies to main *four* sets of accountancy books. |
| 17:28.34 | lkcl | one for the government, one to show to investors, another to show only to the boss, one set to show to anyone else... |
| 17:28.44 | mnemoc | o_O |
| 17:33.21 | arokux | lkcl, if they do not have a datasheet, they could change connector properties anytime... |
| 17:33.44 | arokux | so how it'll help to measure some of them |
| 17:34.03 | mnemoc | you but enough to not need to buy again :p |
| 17:34.08 | mnemoc | buy* |
| 17:34.37 | arokux | i thought he buys some to measure, then he buys more |
| 17:34.59 | mnemoc | :) |
| 17:42.37 | CIA-46 | rhombus-tech: lkcl master * re31595fb3d2c /allwinner_a10/orders/ (9 files): order processing corrections |
| 17:43.14 | lkcl | arokux: it means that they're just a warehouse in a supply chain. which means that we can get the connectors at wholesale prices. |
| 17:43.23 | lkcl | this is gooood |
| 17:47.21 | arokux | I do not understand.. why allwinner engineer is helping to hack *allwinner* device..... http://rhombus-tech.net/a10_mmc_boot/ |
| 17:48.20 | mnemoc | that's hipboi (Tom) |
| 17:48.38 | arokux | isn't he working for allwinner? |
| 17:48.44 | mnemoc | yes |
| 17:49.05 | mnemoc | while they (allwinner) is product driven, they also want to get mainlined |
| 17:49.25 | mnemoc | so Tom is authorized to help us |
| 17:50.10 | lkcl | because they sell chips, arokux. they don't sell *information* about chips, they sell chips. |
| 17:50.37 | lkcl | this is what these fucking morons in marvell, nvidia and so on don't understand. |
| 17:51.17 | lkcl | you want to make it *easy* for people to buy your chips. |
| 17:51.18 | arokux | other companies later do somethings with chips and this is what tom trying to hack? |
| 17:51.34 | lkcl | if you can't fucking well get any information on how to program it, why the fuck would you want to buy it?? |
| 17:51.34 | mnemoc | allwinner is willing to provide everything they can |
| 17:51.45 | mnemoc | but not all "IP" on the A10 belongs to them |
| 17:52.17 | arokux | lkcl, they provide it, but not for us. |
| 17:53.19 | mnemoc | some large IC makers don't even want to have new customers |
| 17:53.19 | arokux | mnemoc, that's why there cannot be a public kernel tree? |
| 17:53.39 | arokux | s/there/they |
| 17:53.45 | mnemoc | they can, but he needs the be authorized |
| 17:53.54 | mnemoc | and those requests don't have high prio |
| 17:54.34 | mnemoc | they only care about getting the new products out ASAP |
| 17:54.57 | mnemoc | things as "doing it right" are not relevant |
| 17:54.59 | arokux | mnemoc, I thought putting a tree to github is a matter of minutes |
| 17:55.23 | zenitraM | the fun thing is, the kernel tree *has* to be released due to being GPL, but they don't even care about that |
| 17:55.28 | mnemoc | for hipboi, yes... but the people who has to authorize it doesn't waste time in doing so |
| 17:56.30 | arokux | wis-tech is selling dev. boards already. so legally the kernel should be there already or? |
| 17:56.36 | mnemoc | zenitraM: git repo is luxury.... the fat 2GB tarball over a cotton line is enough to "comply" when there is time to think about it |
| 17:57.18 | mnemoc | they do have an SDK for android 4.0 including a kernel source snapshot |
| 17:57.23 | zenitraM | that's what i meant, sources (tree is a bonus) |
| 17:58.13 | mnemoc | if you send some mails you'll probably get a link to download it... with a week of delay between each mail gets answered |
| 17:58.39 | zenitraM | if they do, because i think there's no official 4.0 from them yet |
| 17:58.55 | zenitraM | all of the 4.0 firmwares floating out there for A10 devices are unstable betas |
| 17:59.26 | zenitraM | and devices that come with 4.0 preinstalled often also have a 2.3 firmware out there that ppl end up installing because all of the reboots of the 4.0 version |
| 17:59.40 | mnemoc | for the scripts and documentation they have a git repo server for their customers |
| 18:00.04 | mnemoc | with a link so bad that it only works from .cn |
| 18:00.36 | mnemoc | iirc pushing the 2.6.36 it's in github took hipboi almost a whole day |
| 18:01.10 | arokux | ?? they are an IT company! |
| 18:01.38 | zenitraM | well the slow link is not something to blame them for |
| 18:01.50 | zenitraM | all of .cn is slow by itself |
| 18:02.21 | mnemoc | also in most places the international and the national bandwidth are charged differently |
| 18:02.58 | mnemoc | they may have a good enough national connectivity... but international is not interesting |
| 18:03.29 | mnemoc | even 115.com ... the chinese "dropbox" has a pityful download speed from outside china |
| 18:04.33 | arokux | ok, it's all very funny. |
| 18:05.04 | mnemoc | it's a different reality |
| 18:05.37 | zenitraM | you can get lucky and find a 600kb/s link by clicking around on the page |
| 18:05.43 | arokux | I'm googling for "brom FEL" but cannot find anything about it. where can I read something about it? |
| 18:06.48 | zenitraM | i don't think there's anything about it anywhere |
| 18:07.16 | arokux | what are those things? |
| 18:07.23 | mnemoc | zenitraM: so start documenting what you already know about it :) |
| 18:09.06 | zenitraM | i guess i could, but there's not much I know |
| 18:09.55 | mnemoc | it's more than what I do :) |
| 18:10.07 | mnemoc | and it would give others a better starting point |
| 18:10.40 | zenitraM | all i know is already here: http://rhombus-tech.net/allwinner_a10/a10_boot_process/ |
| 18:12.55 | zenitraM | i might add a few things about boot1 through :) |
| 18:13.16 | mnemoc | go go go go |
| 18:14.06 | arokux | with Livesuite google also does not help much. |
| 18:14.19 | zenitraM | you will only be able to find links to download it |
| 18:14.36 | arokux | who developed it? |
| 18:14.45 | zenitraM | it's a quite closed program, you feed it a .bin firmware file, plug the tablet if it's on FEL, and it will flash it |
| 18:14.48 | zenitraM | Allwinner i think |
| 18:15.21 | arokux | so why tom is trying to hack there own protocol?! |
| 18:15.50 | zenitraM | i guess it is faster than convince all winner to release their own tools |
| 18:15.58 | arokux | :D |
| 18:16.02 | zenitraM | also, the tool is windows-only |
| 18:16.18 | arokux | well, protocol is the most important |
| 18:16.49 | zenitraM | the .bin is not a raw image but it is somewhat obfuscated |
| 18:16.56 | zenitraM | and the tools to obfuscate it are also windows-only |
| 18:17.08 | zenitraM | and we only have a tool that builds android gingerbread images |
| 18:17.21 | zenitraM | (well, anything with the partition structure of them) |
| 18:19.18 | arokux | zenitraM, could you upload wireshark logs? |
| 18:19.40 | zenitraM | i don't have any, but I guess I could capture them |
| 18:20.10 | arokux | somebody told yesterday it has some, I thought it was u |
| 18:21.23 | zenitraM | nope, it wasn't me, but I have to flash a firmware image now so i could capture |
| 18:21.29 | zenitraM | downloading wireshark... |
| 18:21.53 | arokux | u run windows in virtual machine? |
| 18:22.01 | zenitraM | yup |
| 18:22.07 | arokux | ok |
| 18:22.31 | zenitraM | and it's strange because the flashing process involves the usb device disconnecting and appearing again with another device ID |
| 18:22.46 | zenitraM | so i have to plug it again to the machine |
| 18:22.50 | mnemoc | uhm.... and you don't lose the link with the usb when powering off your tablet? |
| 18:23.13 | mnemoc | what VM you use? |
| 18:23.19 | zenitraM | Parallels (i'm on Mac) |
| 18:23.30 | zenitraM | powering off? what do you mean? |
| 18:23.56 | mnemoc | to get to FEL... for livesuite |
| 18:24.11 | zenitraM | the way to get to FEL for this tablet is |
| 18:24.16 | zenitraM | with the tablet turned off |
| 18:24.23 | zenitraM | hold Volume+ |
| 18:24.26 | zenitraM | plug the USB |
| 18:24.30 | zenitraM | and then hit the power key 5 times |
| 18:25.05 | arokux | :D |
| 18:25.05 | zenitraM | that or, in extreme cases where boot.axf is destroyed (happened a few times) using the recovery pinhole button |
| 18:25.06 | mnemoc | and then you somehow hand over the new device to the emulator? |
| 18:25.15 | arokux | and spit over the left shoulder |
| 18:25.23 | mnemoc | :) |
| 18:25.25 | zenitraM | parallels asks me where to use it, on mac or windows |
| 18:26.05 | mnemoc | qemu/kvm isn't that friendly :-/ |
| 18:26.21 | zenitraM | use virtualbox :) |
| 18:26.22 | arokux | virtualbox is good |
| 18:26.31 | zenitraM | that, or vmware |
| 18:26.55 | mnemoc | ponders which is the lesser evil |
| 18:27.02 | arokux | vb |
| 18:27.21 | zenitraM | given that it's now owned by oracle, i'd be doubtful too |
| 18:27.48 | arokux | well, the other is closed at all |
| 18:28.01 | zenitraM | mmm how do I capture USB with wireshark? |
| 18:28.07 | zenitraM | trying to find the option but can't see it |
| 18:28.20 | arokux | I could tell you for Linux |
| 18:28.40 | arokux | you should enable something in host |
| 18:28.54 | mnemoc | I don't have a tablet to test currently anyway :-/ |
| 18:29.06 | arokux | http://wiki.wireshark.org/CaptureSetup/USB |
| 18:30.05 | zenitraM | You cannot directly capture raw USB traffic on Windows with Wireshark/WinPcap. |
| 18:30.34 | arokux | wireshark should be in the host |
| 18:30.53 | zenitraM | wireshark for mac doesn't support usb sniffing either i think |
| 18:30.57 | zenitraM | let's check.. |
| 18:31.57 | arokux | ok, I'll be later. see u |
| 18:32.10 | zenitraM | ok, cya! |
| 18:33.03 | zenitraM | from another USB sniffer tool i found for windows: |
| 18:33.05 | zenitraM | ========================================================================== |
| 18:33.06 | zenitraM | FAQ: |
| 18:33.06 | zenitraM | ========================================================================== |
| 18:33.06 | zenitraM | Q: Is this a cool piece of software? |
| 18:33.06 | zenitraM | A: Yes. |
| 18:33.09 | zenitraM | that's all. |
| 18:33.28 | arokux | :D |
| 18:33.54 | zenitraM | doesn't seem to work on w7 :( |
| 18:50.20 | *** join/#arm-netbook gordan (~gordan@178.108.177.117) |
| 19:04.29 | *** join/#arm-netbook arokux_h (~arokux@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 19:42.22 | *** join/#arm-netbook gordan (~gordan@78-86-195-86.zone2.bethere.co.uk) |
| 21:11.20 | CIA-46 | rhombus-tech: Mike master * refa50f156e11 /allwinner_a10/orders/mikenboy84__64__yahoo.de.mdwn: |
| 21:44.43 | arokux_h | zenitraM, how is sniffing? |
| 22:00.20 | zenitraM | arokux_h: i only found two tools and none of them seemed to capture anything |
| 22:00.32 | arokux_h | ok |
| 22:00.46 | zenitraM | maybe they have some trouble running inside a win7 64bits virtual machine (all of the trouble factors are there...) |
| 22:06.40 | arokux_h | zenitraM, it'd be best to run sniffing software on the host |
| 22:07.11 | zenitraM | yup, but i found no usb sniffing tool for mac and the linux machines I have don't have enough RAM to properly run a virtual machine :( |
| 22:07.25 | arokux_h | ok |
| 22:07.38 | arokux_h | althought wait... |
| 22:08.03 | arokux_h | you could sell you mac and buy some ram and several another machines |
| 22:08.22 | arokux_h | just kidding :) |
| 22:08.38 | zenitraM | haha probably, but for now i'm just saving for an asus zenbook |
| 22:09.10 | zenitraM | after having to wait 3 months for the third warranty replacement motherboard for the macbook, i'm not falling on a macbook again |
| 22:09.13 | arokux_h | go for sony vaio Z |
| 22:10.33 | zenitraM | mm they also look great |
| 22:39.47 | CIA-46 | rhombus-tech: lkcl master * rffe9b0bc8e45 /allwinner_a10/orders/mikenboy84__64__yahoo.de.mdwn: |
| 22:41.03 | *** join/#arm-netbook roman__ (~roman@HSI-KBW-046-005-103-006.hsi8.kabel-badenwuerttemberg.de) |
| 22:52.54 | hno | is Tom working on FEL support? |
| 22:55.10 | CIA-46 | rhombus-tech: Pawe master * ra09ed23a3a29 /allwinner_a10/orders/jerzmanlask__64__gmail.com.mdwn: |
| 23:02.15 | hno | zenitraM, it's just boot0 that is loaded via BROM FEL. |
| 23:02.55 | zenitraM | you mean, when flashing? |
| 23:03.05 | hno | and the soft method (pressing volume and power buttons in magic sequence) is done by boot1 I think. |
| 23:03.06 | hno | yes. |
| 23:03.30 | hno | If you use the recovery button then the CPU immediately enters FEL on poweron. |
| 23:03.52 | hno | not even attempting to load anything from flash or sdcard. |
| 23:04.55 | hno | It looks like FEL is xmodem over some simple USB protocol. Probably one char at a time. |
| 23:11.20 | zenitraM | mmm that makes sense |
| 23:11.49 | zenitraM | well let's hope we come up with something to use it |
| 23:12.10 | zenitraM | i'm tired of powering up my win7 vm just to reflash the tablet |
| 23:38.32 | hno | would be nice to be able to load u-boot via fel. |
| 23:39.11 | hno | something I might look into when I get a device to play with. |
| 23:55.59 | hno | Not too worried about FEL as such. That's most likely something very simple. The tricky part is making a boot0 bootloader that's useful. |