≡
[00:00:14] <Niii> k :)
[00:00:17] <Marex> thanks
[00:15:22] <otavio> zumbi: Are you using what for it? Yocto? Ubuntu?
[00:18:47] <BusError_> Marex, sorry I pull the micro of the bench, into the 'nauhgty box for Tsv' :>
[00:20:31] <Marex> oh ok
[01:13:05] <xenoxaos> anyone have the chunk of the fex that deals with the 4.3"lcd? tsvestan was gonna post it somewhere...
[01:29:15] **************** Logging Ended ****************
[01:29:45] **************** Logging Ended ****************
[01:30:16] **************** Logging Ended ****************
[01:30:26] **************** Logging Started ****************
[02:11:21] <tinti> otavio: salvador?
[02:11:53] <otavio> tinti: yes
[02:34:01] <Marex> pamaury: I said nothing
[02:34:11] <Marex> pamaury: get yourself a remote machine with irssi :)
[02:34:21] <Marex> screen + irssi is your friend in times like this
[02:35:05] <pamaury> yeah
[04:12:41] <WarheadsSE> pushing packages for a13-olinuxino-* support tonight. rootfs tomorrow, fex & uboot to follow
[04:13:10] <WarheadsSE> I've had to chop up the a13-olinuxino-micro's FEX file to fix a few things, but I have it working on a unified kernel :)
[04:15:21] <Marex> Im making a debian mirror for when I'll be on HSPA/EDGE next week and will need to install my new rig
[04:15:29] <Marex> until I get the CATV guy to connect me
[04:40:45] <Marex> Niii: are you around ?
[04:40:53] <Marex> Niii: with micro ready ?
[04:47:56] <Marex> Niii: http://patchwork.ozlabs.org/patch/216413/ try this one and see if the memory is stable
[04:49:33] <Marex> BusError_: ^
[09:22:55] <Tsvetan> morning
[09:25:11] <gzamboni> morning Tsvetan
[09:30:20] <Niii> hi
[09:31:35] <Niii> Marex, mtest is running, I'm going to work, I'm back in ~1h
[09:31:46] <BusError_> Niii, cool
[09:32:07] <Niii> no error for moment, sounds good :>
[09:34:42] <Tsvetan> Niii: mx233 uboot?
[09:37:42] <HeHoPMaJIeH> Morning
[09:39:52] <BusError_> Niii, mine still reboots. but I suspect this one is definitely faulty now :/
[09:40:52] <BusError_> ACTION 's micro returns ti the naughty box
[09:41:07] <BusError_> wants to play with that A13 for a while anyway :-)
[09:58:22] <Niii> Tsvetan, yes uboot for mx233
[09:58:34] <Niii> ho no, some meme error :'(
[09:58:42] <Niii> mem*
[10:01:20] <Tsvetan> micro-mini-maxi?
[10:01:30] <Niii> Marex, http://pastebin.com/zgnk4n3d
[10:01:36] <Niii> micro
[10:02:35] <Tsvetan> only bit0 is read wrong
[10:02:39] <BusError_> Niii, you could try actually adding a bit of a voltage bump ...
[10:03:10] <BusError_> thats interesting... is the bit zero pcb trace an issue then?
[10:04:07] <Niii> Niii, you could try actually adding a bit of a voltage bump ... > maybe this night
[10:07:55] <Tsvetan> BusError_ we are going to investigate this
[11:21:41] <teatime07> question about olimexino-32u4: In the description of the Leonardo board on arduino.cc, it says "when the board first powers up, it will jump straight to the user sketch, if present, rather than initiating the bootloader". My olimexino doesn't work this way. What's the difference? A bit different bootloader?
[11:23:02] <Tsvetan> please send you question to support@olimex.com I doubt here we have lot of arduino users who can help
[11:28:44] <teatime07> OK, thanks
[12:01:08] <oliv3r> Tsvetan: Hey quick question. You state in your Readme no A20 boards until AW plays nicely with the GPL (and docs). While I strongly applaud this stance, was this brought clearly to AW's attention? And have they responded?
[12:02:30] <Tsvetan> yes, this is what I told allwinner but I think they do not care so much, their market is Android and they do not give s*it about Linux and GPL
[12:02:42] <Tsvetan> I mean Android is where their money come from
[12:02:51] <oliv3r> Tsvetan: well android is still linux and GPL
[12:03:07] <Tsvetan> until they do not see financial benefit they will not do anything
[12:03:23] <oliv3r> i know userland on android is quite different and if they'd ignore MALI and Cedar because of this, I could understand (somewhat) but the kernel is linux and thus bound to the GPL license
[12:03:38] <Tsvetan> oliv3r try to explain this to Chinese company
[12:03:43] <oliv3r> LOL i know :)
[12:04:22] <oliv3r> But I mean, legally, they are on shakey ground? As a mannnufacturer, e.g. 'user' of their hardware, I think at the least you have a right to demand source code from them thoughh
[12:04:45] <Tsvetan> oliv3r they are always kind and friendly when I ask them this or that they say yes, ok then do nothing
[12:05:11] <Tsvetan> oliv3r the money for the development came from the Chinese government
[12:05:12] <oliv3r> Tsvetan: Have you tried getting in touch with gpl-violations?
[12:05:23] <Tsvetan> I would love to see somebody to sue Allwinner in China and win
[12:05:54] <Tsvetan> and 99% of their sales are in China they do not have distributor outside China
[12:06:40] <oliv3r> I know legally its all very tricky
[12:07:00] <oliv3r> but from what I understand, I as a user of say, a olimex board, can demand the source from you, but not them
[12:07:16] <oliv3r> you however, can demand the source code from them, and thus sue them for not compling
[12:07:19] <Tsvetan> we use linux-sunxi
[12:07:32] <oliv3r> yes, but that doesn't mean you can demand it from them
[12:07:57] <Tsvetan> oliv3r the problem with Allwinner is
[12:08:05] <Tsvetan> 1. they do not understand the GPL problem
[12:08:08] <oliv3r> gpl-violations.org has a mailing list where this question surely could be asked :) I see similar things (mostly from users demanding code for hardware) pop by from time to time
[12:08:26] <oliv3r> maybe the EFF can even give some legal advise from that
[12:08:39] <Tsvetan> 2. they do not see what financial benefit they have if they do this, for them this is just extra work with nothing in return
[12:08:43] <oliv3r> but yes, chine just say 'screw your western laws!'
[12:09:26] <Tsvetan> oliv3r I'm not GPL expert
[12:09:51] <Tsvetan> but if you write letter with what exactly Allwinner do wrong
[12:10:01] <Tsvetan> with simple sentences so they understand
[12:10:07] <Tsvetan> I will forward it
[12:10:15] <Tsvetan> but please try to explain it simple
[12:10:26] <Tsvetan> as they never admit if they do not understand something
[12:10:36] <oliv3r> I think the best initial approach is gpl-violations. I'll write a message to the list, and ask them what the best course of action is?
[12:10:39] <Tsvetan> i.e. if they do not understand what you ask them they will answer some BS
[12:10:59] <oliv3r> may I add your address to the mail?
[12:11:03] <Tsvetan> if you tell them -please change the headers of this and that file to include this text, they will do it
[12:11:37] <oliv3r> As far as I know, right now they are required to release sources (for the A10 since that's the chip you use) for the latest android version they 'support'
[12:11:38] <Tsvetan> if they do not understand what you want from them will reply something just to be kind
[12:11:56] <Tsvetan> they do
[12:12:14] <Tsvetan> these "sources" can compile working android image
[12:12:17] <oliv3r> I belive their code drops is extremly old and doesn't properly support all harware?
[12:12:35] <oliv3r> I think they have since released newer android releases, without source however, and thus breaking the law
[12:12:49] <Tsvetan> but are full with binary blobs which are for the processor registers and initializations
[12:13:03] <oliv3r> Also illegal if i'm not mistaken
[12:13:11] <oliv3r> But i'm far far from a GPL expert :p
[12:13:28] <oliv3r> What e-mail address may I list as yours?
[12:13:39] <Tsvetan> usunov at olimex
[12:13:43] <oliv3r> rgr
[12:14:54] <Tsvetan> again I think the major problem is that they do not understand what they violate nor see what money they will get more if they put efforts in this direction, anything which do not have financial benefit is waste of time for them
[12:17:31] <Tsvetan> with A20 problem is that there will be no info for registers, datasheet, sources, so Linux support will be not possible
[12:17:58] <Tsvetan> and releasing board with only runs android limits it's use
[12:18:06] <Tsvetan> same is for A31
[12:18:36] <oliv3r> If only they understood. At linux-sunxi mainlining even has started. Their time-to-market could have been somewhat if not greatly reduced if that would have been done much sooner and easier
[12:18:54] <Tsvetan> this is why we now move our focus on AM335X
[12:19:30] <oliv3r> I saw, but that unfortunatly does not have a VPU :( which is something that made the A10 quite interesting
[12:19:54] <oliv3r> Tsvetan: unless of course the datasheet for the A20 also gets leaked ;)
[12:20:01] <Tsvetan> VPU without sources and datasheet is like you have it not
[12:20:13] <Tsvetan> oliv3r to leak it should first exist
[12:20:15] <Tsvetan> :)))
[12:20:17] <oliv3r> Tsvetan: you mention binary blobs with register information. Ignoring the Cedar VPU and the Mali GPU for a moment, which parts are you referring to?
[12:20:24] <oliv3r> Tsvetan: O.o
[12:20:45] <oliv3r> Tsvetan: over at #arm-netbook we speculated that the A20 is probably extremly similar to the A10
[12:20:54] <Tsvetan> oliv3r I'm not linux expert but the A13-SDK is in the wiki you can download it and view
[12:21:17] <Tsvetan> right with 2 cores and intern databus you have no info about :)
[12:21:24] <oliv3r> I'd just be curious to specific files to look for
[12:21:51] <Tsvetan> https://www.olimex.com/wiki/A13-OLinuXino#A13-SDK_Android_4.0.3_ICS
[12:21:59] <oliv3r> Even so, if they release android images/sdk, they are legally required to supply sources, so in_theory_ it should not be a problem
[12:22:45] <Tsvetan> the SDK have sources which you compile
[12:42:09] <oliv3r> Tsvetan: have they released an updated SDK at all?
[12:42:31] <oliv3r> I mean, all these manufactures are releasing android 4.1 and 4.2 images, so an SDK must exist?
[13:02:32] <Tsvetan> oliv3r they promised to send me the new SDK few weeks ago but still nothing received
[13:02:47] <Tsvetan> brb go to lunch
[13:04:33] <oliv3r> oh! food!
[13:06:51] <focus> Tsvetan: have you thought about iMX6 series - in particular the quad core iMX6Q?
[13:35:48] <Tsvetan> focus: yes we know imx6, the prices though make the BOM around $100 alone, so bit above low cost boards we target
[13:36:37] <BusError_> migth come down in price tho. there's already a few i.mx6 devices on dx etc
[13:36:48] <BusError_> $100 for an android 'stick' with i.mx
[13:36:49] <Tsvetan> as imx6 will need at least 2GB RAM and Gigabit Ethernet, PCIe etc features
[13:37:05] <Tsvetan> BusErros: are these 4core
[13:37:17] <BusError_> yeah
[13:37:34] <Tsvetan> Gigabit, PCIe, 2GB RAM?
[13:38:14] <BusError_> nope, no pcie, or Gb... just an android stick
[13:38:30] <Tsvetan> 2GB ram?
[13:38:43] <focus> Tsvetan: the price come down recently - $20 for quad core, $14 for dual core, $8 for single core. Not all features need be implemented - just bring to connector, let the developer tinker - keep costs down.
[13:39:20] <BusError_> 1GB, BT, Wifi, 8GB flash etc. pretty standard android stick really. lacks all the other interesting goodies
[13:39:40] <BusError_> but.. it means it's trickling down in the channel, so the prices will come down...
[13:40:37] <Tsvetan> when Allwinner wrote me about A31 they said that 4 core device without at least 2GB RAM works not efficient
[13:41:33] <BusError_> for android maybe? for linx I dont' see why, I used quad cores with less memory for ages in servers etc
[13:42:01] <Tsvetan> focus where you see $20?
[13:42:03] <Tsvetan> http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=i.MX6Q&tab=Buy_Parametric_Tab&nodeId=018rH3ZrDRB24A&pspll=1&fromSearch=false
[13:42:45] <Tsvetan> the real 4 core chip with media bus is $39 / 1000
[13:44:11] <focus> These are UK prices - usually convert $ to UK pound and then add a margin - http://uk.farnell.com/freescale-semiconductor/mcimx6q5eym10ac/mpu-i-mx6-quad-1ghz-624fcbga/dp/2253171
[13:45:04] <focus> Roughly comes out to same price as dollar price. So for 300 pieces imx6q is 20.56 UK pounds - which means they got it for about $20.
[13:46:13] <Tsvetan> I see now Freescale set their prices for 1000 pcs but when they had these budgetary prices for 100 pcs usually I got from our distributor price for 1000 pcs higher than the web price Freescale quote for 100 pcs, so these Farnell assumptions means nothing
[13:46:56] <Tsvetan> probably Freescale is trying to sell directly to the bigger customers with this policy
[13:47:16] <Tsvetan> but I guess the prices written on their web is what their distributors get
[13:49:17] <Tsvetan> I got e-mail asking me if Gentoo can run on A1x, is this possible?
[13:49:49] <xenoxaos> it can...
[13:50:05] <Tsvetan> http://linux-sunxi.org/Gentoo is empty ..
[13:50:11] <xenoxaos> but why would you want to compile EVERYTHING on a little arm board?
[13:50:50] <Tsvetan> have no idea
[13:51:02] <focus> Tsvetan: I go by the Farnel price because I buy in China a while ago, these assumptions are valid. The price on chip company web sites usually much higher than what can be bought from 'professional' buying routes.
[13:52:08] <Tsvetan> focus: in china you can get good prices with unknown quality only on some items
[13:52:56] <Tsvetan> there are many recycled/refurbished chips with unknown specs, for production it's non sense to use chips which came not from the official distributor
[13:53:01] <Tsvetan> at least for the processors
[13:53:17] <focus> Tsvetan: need to do business face to face in CN and build relations so as not to be sold rubbish
[13:53:42] <Tsvetan> right, but sometimes the seller don't know what he sells you
[13:53:51] <focus> Shenzhen is good and so is Guanzhou (more expensive for components)
[13:54:41] <focus> Tsvetan: decent sellers will tell you x is good quality y is lower price (hint - y is also lower quality) :)
[13:55:16] <Tsvetan> I buy components for 20+ years :)
[13:55:24] <focus> In CN we got burned from new suppliers, but never from old friends.
[13:56:26] <focus> Do you buy face to face, or get your buyers to buy?
[13:56:45] <Tsvetan> I have few agents
[13:56:54] <focus> That be your problem
[14:00:40] <Tsvetan> I'm sure that they will not let me down, but sometimes they are let down by the ppl they deal with, so we have to do 100% test on all items on arrival, which adds cost, which should be considered
[14:04:16] <focus> Tsvetan: I meant to say the buying agents are OK, but you need to get out there, meet the suppliers in person, buy components and test to destruction and select only the best and keep going back to the good guys by showing up at their booths. Don't need to buy anything other than samples when you meet face to face. Meeting face to face adds magic to the transaction.
[14:05:25] <focus> I turn up roughly every month and the more I turn up, the higher the quality and level of service.
[14:05:37] <focus> And better prices too.
[14:05:57] <Tsvetan> c'mon do you really think you talk to the producer when you go to their booth?
[14:06:56] <Tsvetan> I got at least 50 SPAM e-mails from PCB factories with address - blah blah building, 42 floor, office 4202
[14:07:04] <Tsvetan> every day
[14:07:34] <Tsvetan> every chinese who know few english is factory and produce everything :)
[14:07:58] <focus> That is a loaded question - but I will say, you get better service as you establish your relations. Go to the markets where they have a booth or a shop - not to their office unless its a lot of business.
[14:07:59] <Tsvetan> DX in action
[14:09:00] <focus> So anyway, I take it the imx6 series are not on your radar at the moment?
[14:10:43] <Tsvetan> I see no good prices, so for the moment we will pass on imx6
[14:21:57] <Tsvetan> xenoxaos: btw A1x is not so bad for compiling on board, IIRC there were gentoo for SAM9 200Mhz ARM9
[14:22:14] <xenoxaos> lol
[14:23:14] **************** Logging Ended ****************
[14:23:45] **************** Logging Ended ****************
[14:24:16] **************** Logging Ended ****************
[14:24:26] **************** Logging Started ****************
[14:30:02] <Marex> Niii: that sucks
[14:30:45] <Marex> Niii: can you revert only the part of the patch that sets the voltage to 0x10 instead of 0x12 (it's there, twice) but leave the EMI reset in, then retry ?
[15:40:01] <Niii> Marex, I make a mistake, I forget reset function ... I can't upgrade uboot remotely on micro, I'll test in ~3h
[15:59:38] <WarheadsSE> lol, gentoo on the a1x, native.
[16:32:43] <vinifm> WarheadsSE: i am so sorry. I thought a patch was just diff -up :)
[16:33:49] <WarheadsSE> :)
[16:34:07] <WarheadsSE> I need to send the a13-micro fex up
[16:34:24] <WarheadsSE> the a10s is doable @ staging/sunxi-3.X
[16:34:42] <vinifm> i saw: http://lxr.free-electrons.com/source/Documentation/SubmittingPatches
[16:35:39] <WarheadsSE> we submit patches to github.com/linux-sunxi/linux-sunxi they then submit sets upstream
[16:35:45] <vinifm> is somewhat complicated
[16:36:03] <WarheadsSE> well its just a "diff -up" if you half-ass it :P
[16:36:29] <WarheadsSE> If you want it right, and supporting more than just you, it is more intricate :)
[16:38:50] <Tsvetan> any idea for anti spam bot question? how many legs A13 have is way too hard for some ppl, I got two angry e-mails today
[16:40:49] <xenoxaos> hmmm
[16:41:48] <xenoxaos> if they get that angry, you don't want them around anyway
[16:41:49] <xenoxaos> :P
[16:44:47] <Tsvetan> one say I did this on purpose so ppl do not ask support questions on the forum :)))
[16:45:08] <xenoxaos> lol
[16:45:54] <xenoxaos> you did...didn't you?
[16:46:26] <Tsvetan> the other said - I will not search this info and waste my time!
[16:47:04] <Tsvetan> I told them the answer
[16:47:28] <Tsvetan> the problem is that some bots easily answer questions like: "how many is three plus five?"
[16:47:48] <Tsvetan> and I got hundreds bots registered for 24 hours until I change the question
[16:47:56] <Tsvetan> so I have to be careful what is the question
[16:48:09] <Tsvetan> in other hand some questions are hard for some humans to answer
[16:48:34] <Tsvetan> is there some practice about this? I know you are more deep in the anti spam stuff than me
[16:56:44] <xenoxaos> actually....we just started using a different antispam package
[16:59:00] <xenoxaos> http://www.stopforumspam.com/
[16:59:17] <xenoxaos> we've seen quite a bit less spam since that
[17:00:56] <Tsvetan> wow nice :)
[17:02:00] <xenoxaos> looking at the admin log....it looks like we've only banned one acct in the past week
[17:02:21] <xenoxaos> vs what used to be a few a day
[17:03:38] <Tsvetan> looking our stats the peak was November 2012 with 1700 attempts from bots
[17:05:05] <Tsvetan> after this we introduced the tech question and the attempts drop significant, actually we have few bots registered MANUALLY per month which I delete
[17:05:21] <xenoxaos> btw, back to lcds....with the script.bin that works on the 4.3"....should that power up and at least show something on the 7"?
[17:05:39] <Tsvetan> wait I will upload it on the wiki in few minutes
[17:05:47] <Tsvetan> yesterday was out of office
[17:05:50] <xenoxaos> sure
[17:06:01] <xenoxaos> but should the 7" at least turn on?
[17:09:33] <xenoxaos> brb
[17:10:47] <Tsvetan> https://www.olimex.com/wiki/A13-OLinuXino#Using_an_operating_system_-_Nand_flash
[17:11:05] <Tsvetan> 7" will turn on if you turn it on :)
[17:20:57] <xenoxaos> but you have to tell it to turn on...because the 4" comes on no problem, the 7" does nothing
[18:01:16] <NicoM> Hi everyone, did someone ever experiment segfault while trying to use alsa (like speaker-test or aplay) with the A13 Micro or A13 ? (this is a custom build using buildroot, I tried with udev and static /dev result is the same)
[18:01:55] <NicoM> (mplayer and speaker-test just work fine on a Debian pre-built image)
[18:03:14] <WarheadsSE> hmm, haven't tried that on mine yet
[18:03:40] <NicoM> sun5i audio codec seems to load fine at kernel boot
[18:10:33] <NicoM> well I guess if it works on the debian rootfs, it should be an issue with my buildroot fs, any idea ?
[19:01:01] <Marex> Niii: no prob
[19:02:00] <Niii> Marex, just tested with reset function and 0x12, got lot of errors
[19:02:23] <Niii> original works for moment, it's running without error
[19:02:47] <Marex> Niii: we probably have further issues
[19:03:16] <Niii> original patch*
[19:04:23] <Marex> Niii: I suspect some further errors in the mem init
[19:06:58] <Niii> http://pastebin.com/3mRdsi2N :/
[19:13:00] <Marex> Niii: what, further errors ?
[19:13:15] <Marex> Niii: my current ssh client (tf700 BTEP) doesn't support console copy-paste
[19:15:25] <Niii> git uboot = lots of errors, with 0x12 and reset func = less, with 0x10 and reset func = really less error, but still here
[19:16:36] <Marex> ok ... so we have multiple bug
[19:16:37] <Marex> s
[19:19:43] <jinzo> Tsvetan, I know it's a long shot, but are you thinking of making any non-graphics-video-focused SoC boards? (for example with AR9883 SoC - or something there that is focused/used in cheap routers - ofcourse running openwrt)
[19:24:14] <Marex> woo, got my 4-channel 200MHz 2G/s scope today :)
[19:29:43] <mnemoc> jinzo: AM3352
[19:31:13] <jinzo> mnemoc, I only glanced at that one - but tought it's in line with Allwinners offerings
[19:31:33] <jinzo> I'll check it a bit more, sorry for not doing that before.
[19:31:35] <mnemoc> Marex: but just 4 channels? :p j/k
[19:32:32] <jinzo> ah it does have VGA/LCD connectors, I'm looking something in a more "half that price, with WiFi" :P
[19:32:42] <jinzo> (ofcourse, half or even less performance too.)
[19:44:44] <LFSveteran> hmm, no brd file of the olinuxino maxi?
[19:45:32] <xenoxaos> https://github.com/OLIMEX/OLINUXINO/blob/master/HARDWARE/iMX233-OLinuXino-Maxi/iMX233-OLINUXINO-MAXI_Rev_B1.brd
[19:46:15] <LFSveteran> yeah, but that doesn't show the board layout
[19:46:35] <LFSveteran> only some drawing of the case
[23:41:34] <WarheadsSE> Tsvetan: I've put out 2 kernels & a rootfs for the a13 & a10s
[23:41:58] <WarheadsSE> I have an updated fex file I need to send you at some point that properly turns off the axp features on the -micro
[23:42:32] <mnemoc> WarheadsSE: found the cause of the disp_alloc() crash?
[23:43:18] <WarheadsSE> mnemoc: yeah, remember, I accidentally left an reserve in the kernel config :P
[23:43:59] <WarheadsSE> so i have built up linux-sun5i (sunxi-3.0) and linux-sun5i-staging (stage/sunxi-3.0)
[23:44:38] <mnemoc> WarheadsSE: \o/
[23:46:25] <WarheadsSE> so now all linux-sunXi are @ 3.0.57+
[23:47:02] <BusError_> anything more recent on A13's ? 3.0.x is ancient...
[23:53:35] <mnemoc> 3.4
[23:54:08] <mnemoc> people says it actually runs faster...
[23:54:29] <mnemoc> but usb gadget is broken
[23:57:29] <WarheadsSE> I'd bet that 3.4 does run faster.
[23:57:48] <WarheadsSE> when I am told (& convinced) that 3.4 is stable, i'll bump them.
[23:58:09] <WarheadsSE> until then, I know most stuff is working in 3.0.x, and stably so.
[23:58:28] <WarheadsSE> and anyone can always feel free to adjust my PKGBUILDs to make their own
[23:58:58] <mnemoc> :)
IRC #olimex 2013-01-29
[00:00:09] <Marex> Niii: then one in all the uboot patches ;-)[00:00:14] <Niii> k :)
[00:00:17] <Marex> thanks
[00:15:22] <otavio> zumbi: Are you using what for it? Yocto? Ubuntu?
[00:18:47] <BusError_> Marex, sorry I pull the micro of the bench, into the 'nauhgty box for Tsv' :>
[00:20:31] <Marex> oh ok
[01:13:05] <xenoxaos> anyone have the chunk of the fex that deals with the 4.3"lcd? tsvestan was gonna post it somewhere...
[01:29:15] **************** Logging Ended ****************
[01:29:45] **************** Logging Ended ****************
[01:30:16] **************** Logging Ended ****************
[01:30:26] **************** Logging Started ****************
[02:11:21] <tinti> otavio: salvador?
[02:11:53] <otavio> tinti: yes
[02:34:01] <Marex> pamaury: I said nothing
[02:34:11] <Marex> pamaury: get yourself a remote machine with irssi :)
[02:34:21] <Marex> screen + irssi is your friend in times like this
[02:35:05] <pamaury> yeah
[04:12:41] <WarheadsSE> pushing packages for a13-olinuxino-* support tonight. rootfs tomorrow, fex & uboot to follow
[04:13:10] <WarheadsSE> I've had to chop up the a13-olinuxino-micro's FEX file to fix a few things, but I have it working on a unified kernel :)
[04:15:21] <Marex> Im making a debian mirror for when I'll be on HSPA/EDGE next week and will need to install my new rig
[04:15:29] <Marex> until I get the CATV guy to connect me
[04:40:45] <Marex> Niii: are you around ?
[04:40:53] <Marex> Niii: with micro ready ?
[04:47:56] <Marex> Niii: http://patchwork.ozlabs.org/patch/216413/ try this one and see if the memory is stable
[04:49:33] <Marex> BusError_: ^
[09:22:55] <Tsvetan> morning
[09:25:11] <gzamboni> morning Tsvetan
[09:30:20] <Niii> hi
[09:31:35] <Niii> Marex, mtest is running, I'm going to work, I'm back in ~1h
[09:31:46] <BusError_> Niii, cool
[09:32:07] <Niii> no error for moment, sounds good :>
[09:34:42] <Tsvetan> Niii: mx233 uboot?
[09:37:42] <HeHoPMaJIeH> Morning
[09:39:52] <BusError_> Niii, mine still reboots. but I suspect this one is definitely faulty now :/
[09:40:52] <BusError_> ACTION 's micro returns ti the naughty box
[09:41:07] <BusError_> wants to play with that A13 for a while anyway :-)
[09:58:22] <Niii> Tsvetan, yes uboot for mx233
[09:58:34] <Niii> ho no, some meme error :'(
[09:58:42] <Niii> mem*
[10:01:20] <Tsvetan> micro-mini-maxi?
[10:01:30] <Niii> Marex, http://pastebin.com/zgnk4n3d
[10:01:36] <Niii> micro
[10:02:35] <Tsvetan> only bit0 is read wrong
[10:02:39] <BusError_> Niii, you could try actually adding a bit of a voltage bump ...
[10:03:10] <BusError_> thats interesting... is the bit zero pcb trace an issue then?
[10:04:07] <Niii> Niii, you could try actually adding a bit of a voltage bump ... > maybe this night
[10:07:55] <Tsvetan> BusError_ we are going to investigate this
[11:21:41] <teatime07> question about olimexino-32u4: In the description of the Leonardo board on arduino.cc, it says "when the board first powers up, it will jump straight to the user sketch, if present, rather than initiating the bootloader". My olimexino doesn't work this way. What's the difference? A bit different bootloader?
[11:23:02] <Tsvetan> please send you question to support@olimex.com I doubt here we have lot of arduino users who can help
[11:28:44] <teatime07> OK, thanks
[12:01:08] <oliv3r> Tsvetan: Hey quick question. You state in your Readme no A20 boards until AW plays nicely with the GPL (and docs). While I strongly applaud this stance, was this brought clearly to AW's attention? And have they responded?
[12:02:30] <Tsvetan> yes, this is what I told allwinner but I think they do not care so much, their market is Android and they do not give s*it about Linux and GPL
[12:02:42] <Tsvetan> I mean Android is where their money come from
[12:02:51] <oliv3r> Tsvetan: well android is still linux and GPL
[12:03:07] <Tsvetan> until they do not see financial benefit they will not do anything
[12:03:23] <oliv3r> i know userland on android is quite different and if they'd ignore MALI and Cedar because of this, I could understand (somewhat) but the kernel is linux and thus bound to the GPL license
[12:03:38] <Tsvetan> oliv3r try to explain this to Chinese company
[12:03:43] <oliv3r> LOL i know :)
[12:04:22] <oliv3r> But I mean, legally, they are on shakey ground? As a mannnufacturer, e.g. 'user' of their hardware, I think at the least you have a right to demand source code from them thoughh
[12:04:45] <Tsvetan> oliv3r they are always kind and friendly when I ask them this or that they say yes, ok then do nothing
[12:05:11] <Tsvetan> oliv3r the money for the development came from the Chinese government
[12:05:12] <oliv3r> Tsvetan: Have you tried getting in touch with gpl-violations?
[12:05:23] <Tsvetan> I would love to see somebody to sue Allwinner in China and win
[12:05:54] <Tsvetan> and 99% of their sales are in China they do not have distributor outside China
[12:06:40] <oliv3r> I know legally its all very tricky
[12:07:00] <oliv3r> but from what I understand, I as a user of say, a olimex board, can demand the source from you, but not them
[12:07:16] <oliv3r> you however, can demand the source code from them, and thus sue them for not compling
[12:07:19] <Tsvetan> we use linux-sunxi
[12:07:32] <oliv3r> yes, but that doesn't mean you can demand it from them
[12:07:57] <Tsvetan> oliv3r the problem with Allwinner is
[12:08:05] <Tsvetan> 1. they do not understand the GPL problem
[12:08:08] <oliv3r> gpl-violations.org has a mailing list where this question surely could be asked :) I see similar things (mostly from users demanding code for hardware) pop by from time to time
[12:08:26] <oliv3r> maybe the EFF can even give some legal advise from that
[12:08:39] <Tsvetan> 2. they do not see what financial benefit they have if they do this, for them this is just extra work with nothing in return
[12:08:43] <oliv3r> but yes, chine just say 'screw your western laws!'
[12:09:26] <Tsvetan> oliv3r I'm not GPL expert
[12:09:51] <Tsvetan> but if you write letter with what exactly Allwinner do wrong
[12:10:01] <Tsvetan> with simple sentences so they understand
[12:10:07] <Tsvetan> I will forward it
[12:10:15] <Tsvetan> but please try to explain it simple
[12:10:26] <Tsvetan> as they never admit if they do not understand something
[12:10:36] <oliv3r> I think the best initial approach is gpl-violations. I'll write a message to the list, and ask them what the best course of action is?
[12:10:39] <Tsvetan> i.e. if they do not understand what you ask them they will answer some BS
[12:10:59] <oliv3r> may I add your address to the mail?
[12:11:03] <Tsvetan> if you tell them -please change the headers of this and that file to include this text, they will do it
[12:11:37] <oliv3r> As far as I know, right now they are required to release sources (for the A10 since that's the chip you use) for the latest android version they 'support'
[12:11:38] <Tsvetan> if they do not understand what you want from them will reply something just to be kind
[12:11:56] <Tsvetan> they do
[12:12:14] <Tsvetan> these "sources" can compile working android image
[12:12:17] <oliv3r> I belive their code drops is extremly old and doesn't properly support all harware?
[12:12:35] <oliv3r> I think they have since released newer android releases, without source however, and thus breaking the law
[12:12:49] <Tsvetan> but are full with binary blobs which are for the processor registers and initializations
[12:13:03] <oliv3r> Also illegal if i'm not mistaken
[12:13:11] <oliv3r> But i'm far far from a GPL expert :p
[12:13:28] <oliv3r> What e-mail address may I list as yours?
[12:13:39] <Tsvetan> usunov at olimex
[12:13:43] <oliv3r> rgr
[12:14:54] <Tsvetan> again I think the major problem is that they do not understand what they violate nor see what money they will get more if they put efforts in this direction, anything which do not have financial benefit is waste of time for them
[12:17:31] <Tsvetan> with A20 problem is that there will be no info for registers, datasheet, sources, so Linux support will be not possible
[12:17:58] <Tsvetan> and releasing board with only runs android limits it's use
[12:18:06] <Tsvetan> same is for A31
[12:18:36] <oliv3r> If only they understood. At linux-sunxi mainlining even has started. Their time-to-market could have been somewhat if not greatly reduced if that would have been done much sooner and easier
[12:18:54] <Tsvetan> this is why we now move our focus on AM335X
[12:19:30] <oliv3r> I saw, but that unfortunatly does not have a VPU :( which is something that made the A10 quite interesting
[12:19:54] <oliv3r> Tsvetan: unless of course the datasheet for the A20 also gets leaked ;)
[12:20:01] <Tsvetan> VPU without sources and datasheet is like you have it not
[12:20:13] <Tsvetan> oliv3r to leak it should first exist
[12:20:15] <Tsvetan> :)))
[12:20:17] <oliv3r> Tsvetan: you mention binary blobs with register information. Ignoring the Cedar VPU and the Mali GPU for a moment, which parts are you referring to?
[12:20:24] <oliv3r> Tsvetan: O.o
[12:20:45] <oliv3r> Tsvetan: over at #arm-netbook we speculated that the A20 is probably extremly similar to the A10
[12:20:54] <Tsvetan> oliv3r I'm not linux expert but the A13-SDK is in the wiki you can download it and view
[12:21:17] <Tsvetan> right with 2 cores and intern databus you have no info about :)
[12:21:24] <oliv3r> I'd just be curious to specific files to look for
[12:21:51] <Tsvetan> https://www.olimex.com/wiki/A13-OLinuXino#A13-SDK_Android_4.0.3_ICS
[12:21:59] <oliv3r> Even so, if they release android images/sdk, they are legally required to supply sources, so in_theory_ it should not be a problem
[12:22:45] <Tsvetan> the SDK have sources which you compile
[12:42:09] <oliv3r> Tsvetan: have they released an updated SDK at all?
[12:42:31] <oliv3r> I mean, all these manufactures are releasing android 4.1 and 4.2 images, so an SDK must exist?
[13:02:32] <Tsvetan> oliv3r they promised to send me the new SDK few weeks ago but still nothing received
[13:02:47] <Tsvetan> brb go to lunch
[13:04:33] <oliv3r> oh! food!
[13:06:51] <focus> Tsvetan: have you thought about iMX6 series - in particular the quad core iMX6Q?
[13:35:48] <Tsvetan> focus: yes we know imx6, the prices though make the BOM around $100 alone, so bit above low cost boards we target
[13:36:37] <BusError_> migth come down in price tho. there's already a few i.mx6 devices on dx etc
[13:36:48] <BusError_> $100 for an android 'stick' with i.mx
[13:36:49] <Tsvetan> as imx6 will need at least 2GB RAM and Gigabit Ethernet, PCIe etc features
[13:37:05] <Tsvetan> BusErros: are these 4core
[13:37:17] <BusError_> yeah
[13:37:34] <Tsvetan> Gigabit, PCIe, 2GB RAM?
[13:38:14] <BusError_> nope, no pcie, or Gb... just an android stick
[13:38:30] <Tsvetan> 2GB ram?
[13:38:43] <focus> Tsvetan: the price come down recently - $20 for quad core, $14 for dual core, $8 for single core. Not all features need be implemented - just bring to connector, let the developer tinker - keep costs down.
[13:39:20] <BusError_> 1GB, BT, Wifi, 8GB flash etc. pretty standard android stick really. lacks all the other interesting goodies
[13:39:40] <BusError_> but.. it means it's trickling down in the channel, so the prices will come down...
[13:40:37] <Tsvetan> when Allwinner wrote me about A31 they said that 4 core device without at least 2GB RAM works not efficient
[13:41:33] <BusError_> for android maybe? for linx I dont' see why, I used quad cores with less memory for ages in servers etc
[13:42:01] <Tsvetan> focus where you see $20?
[13:42:03] <Tsvetan> http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=i.MX6Q&tab=Buy_Parametric_Tab&nodeId=018rH3ZrDRB24A&pspll=1&fromSearch=false
[13:42:45] <Tsvetan> the real 4 core chip with media bus is $39 / 1000
[13:44:11] <focus> These are UK prices - usually convert $ to UK pound and then add a margin - http://uk.farnell.com/freescale-semiconductor/mcimx6q5eym10ac/mpu-i-mx6-quad-1ghz-624fcbga/dp/2253171
[13:45:04] <focus> Roughly comes out to same price as dollar price. So for 300 pieces imx6q is 20.56 UK pounds - which means they got it for about $20.
[13:46:13] <Tsvetan> I see now Freescale set their prices for 1000 pcs but when they had these budgetary prices for 100 pcs usually I got from our distributor price for 1000 pcs higher than the web price Freescale quote for 100 pcs, so these Farnell assumptions means nothing
[13:46:56] <Tsvetan> probably Freescale is trying to sell directly to the bigger customers with this policy
[13:47:16] <Tsvetan> but I guess the prices written on their web is what their distributors get
[13:49:17] <Tsvetan> I got e-mail asking me if Gentoo can run on A1x, is this possible?
[13:49:49] <xenoxaos> it can...
[13:50:05] <Tsvetan> http://linux-sunxi.org/Gentoo is empty ..
[13:50:11] <xenoxaos> but why would you want to compile EVERYTHING on a little arm board?
[13:50:50] <Tsvetan> have no idea
[13:51:02] <focus> Tsvetan: I go by the Farnel price because I buy in China a while ago, these assumptions are valid. The price on chip company web sites usually much higher than what can be bought from 'professional' buying routes.
[13:52:08] <Tsvetan> focus: in china you can get good prices with unknown quality only on some items
[13:52:56] <Tsvetan> there are many recycled/refurbished chips with unknown specs, for production it's non sense to use chips which came not from the official distributor
[13:53:01] <Tsvetan> at least for the processors
[13:53:17] <focus> Tsvetan: need to do business face to face in CN and build relations so as not to be sold rubbish
[13:53:42] <Tsvetan> right, but sometimes the seller don't know what he sells you
[13:53:51] <focus> Shenzhen is good and so is Guanzhou (more expensive for components)
[13:54:41] <focus> Tsvetan: decent sellers will tell you x is good quality y is lower price (hint - y is also lower quality) :)
[13:55:16] <Tsvetan> I buy components for 20+ years :)
[13:55:24] <focus> In CN we got burned from new suppliers, but never from old friends.
[13:56:26] <focus> Do you buy face to face, or get your buyers to buy?
[13:56:45] <Tsvetan> I have few agents
[13:56:54] <focus> That be your problem
[14:00:40] <Tsvetan> I'm sure that they will not let me down, but sometimes they are let down by the ppl they deal with, so we have to do 100% test on all items on arrival, which adds cost, which should be considered
[14:04:16] <focus> Tsvetan: I meant to say the buying agents are OK, but you need to get out there, meet the suppliers in person, buy components and test to destruction and select only the best and keep going back to the good guys by showing up at their booths. Don't need to buy anything other than samples when you meet face to face. Meeting face to face adds magic to the transaction.
[14:05:25] <focus> I turn up roughly every month and the more I turn up, the higher the quality and level of service.
[14:05:37] <focus> And better prices too.
[14:05:57] <Tsvetan> c'mon do you really think you talk to the producer when you go to their booth?
[14:06:56] <Tsvetan> I got at least 50 SPAM e-mails from PCB factories with address - blah blah building, 42 floor, office 4202
[14:07:04] <Tsvetan> every day
[14:07:34] <Tsvetan> every chinese who know few english is factory and produce everything :)
[14:07:58] <focus> That is a loaded question - but I will say, you get better service as you establish your relations. Go to the markets where they have a booth or a shop - not to their office unless its a lot of business.
[14:07:59] <Tsvetan> DX in action
[14:09:00] <focus> So anyway, I take it the imx6 series are not on your radar at the moment?
[14:10:43] <Tsvetan> I see no good prices, so for the moment we will pass on imx6
[14:21:57] <Tsvetan> xenoxaos: btw A1x is not so bad for compiling on board, IIRC there were gentoo for SAM9 200Mhz ARM9
[14:22:14] <xenoxaos> lol
[14:23:14] **************** Logging Ended ****************
[14:23:45] **************** Logging Ended ****************
[14:24:16] **************** Logging Ended ****************
[14:24:26] **************** Logging Started ****************
[14:30:02] <Marex> Niii: that sucks
[14:30:45] <Marex> Niii: can you revert only the part of the patch that sets the voltage to 0x10 instead of 0x12 (it's there, twice) but leave the EMI reset in, then retry ?
[15:40:01] <Niii> Marex, I make a mistake, I forget reset function ... I can't upgrade uboot remotely on micro, I'll test in ~3h
[15:59:38] <WarheadsSE> lol, gentoo on the a1x, native.
[16:32:43] <vinifm> WarheadsSE: i am so sorry. I thought a patch was just diff -up :)
[16:33:49] <WarheadsSE> :)
[16:34:07] <WarheadsSE> I need to send the a13-micro fex up
[16:34:24] <WarheadsSE> the a10s is doable @ staging/sunxi-3.X
[16:34:42] <vinifm> i saw: http://lxr.free-electrons.com/source/Documentation/SubmittingPatches
[16:35:39] <WarheadsSE> we submit patches to github.com/linux-sunxi/linux-sunxi they then submit sets upstream
[16:35:45] <vinifm> is somewhat complicated
[16:36:03] <WarheadsSE> well its just a "diff -up" if you half-ass it :P
[16:36:29] <WarheadsSE> If you want it right, and supporting more than just you, it is more intricate :)
[16:38:50] <Tsvetan> any idea for anti spam bot question? how many legs A13 have is way too hard for some ppl, I got two angry e-mails today
[16:40:49] <xenoxaos> hmmm
[16:41:48] <xenoxaos> if they get that angry, you don't want them around anyway
[16:41:49] <xenoxaos> :P
[16:44:47] <Tsvetan> one say I did this on purpose so ppl do not ask support questions on the forum :)))
[16:45:08] <xenoxaos> lol
[16:45:54] <xenoxaos> you did...didn't you?
[16:46:26] <Tsvetan> the other said - I will not search this info and waste my time!
[16:47:04] <Tsvetan> I told them the answer
[16:47:28] <Tsvetan> the problem is that some bots easily answer questions like: "how many is three plus five?"
[16:47:48] <Tsvetan> and I got hundreds bots registered for 24 hours until I change the question
[16:47:56] <Tsvetan> so I have to be careful what is the question
[16:48:09] <Tsvetan> in other hand some questions are hard for some humans to answer
[16:48:34] <Tsvetan> is there some practice about this? I know you are more deep in the anti spam stuff than me
[16:56:44] <xenoxaos> actually....we just started using a different antispam package
[16:59:00] <xenoxaos> http://www.stopforumspam.com/
[16:59:17] <xenoxaos> we've seen quite a bit less spam since that
[17:00:56] <Tsvetan> wow nice :)
[17:02:00] <xenoxaos> looking at the admin log....it looks like we've only banned one acct in the past week
[17:02:21] <xenoxaos> vs what used to be a few a day
[17:03:38] <Tsvetan> looking our stats the peak was November 2012 with 1700 attempts from bots
[17:05:05] <Tsvetan> after this we introduced the tech question and the attempts drop significant, actually we have few bots registered MANUALLY per month which I delete
[17:05:21] <xenoxaos> btw, back to lcds....with the script.bin that works on the 4.3"....should that power up and at least show something on the 7"?
[17:05:39] <Tsvetan> wait I will upload it on the wiki in few minutes
[17:05:47] <Tsvetan> yesterday was out of office
[17:05:50] <xenoxaos> sure
[17:06:01] <xenoxaos> but should the 7" at least turn on?
[17:09:33] <xenoxaos> brb
[17:10:47] <Tsvetan> https://www.olimex.com/wiki/A13-OLinuXino#Using_an_operating_system_-_Nand_flash
[17:11:05] <Tsvetan> 7" will turn on if you turn it on :)
[17:20:57] <xenoxaos> but you have to tell it to turn on...because the 4" comes on no problem, the 7" does nothing
[18:01:16] <NicoM> Hi everyone, did someone ever experiment segfault while trying to use alsa (like speaker-test or aplay) with the A13 Micro or A13 ? (this is a custom build using buildroot, I tried with udev and static /dev result is the same)
[18:01:55] <NicoM> (mplayer and speaker-test just work fine on a Debian pre-built image)
[18:03:14] <WarheadsSE> hmm, haven't tried that on mine yet
[18:03:40] <NicoM> sun5i audio codec seems to load fine at kernel boot
[18:10:33] <NicoM> well I guess if it works on the debian rootfs, it should be an issue with my buildroot fs, any idea ?
[19:01:01] <Marex> Niii: no prob
[19:02:00] <Niii> Marex, just tested with reset function and 0x12, got lot of errors
[19:02:23] <Niii> original works for moment, it's running without error
[19:02:47] <Marex> Niii: we probably have further issues
[19:03:16] <Niii> original patch*
[19:04:23] <Marex> Niii: I suspect some further errors in the mem init
[19:06:58] <Niii> http://pastebin.com/3mRdsi2N :/
[19:13:00] <Marex> Niii: what, further errors ?
[19:13:15] <Marex> Niii: my current ssh client (tf700 BTEP) doesn't support console copy-paste
[19:15:25] <Niii> git uboot = lots of errors, with 0x12 and reset func = less, with 0x10 and reset func = really less error, but still here
[19:16:36] <Marex> ok ... so we have multiple bug
[19:16:37] <Marex> s
[19:19:43] <jinzo> Tsvetan, I know it's a long shot, but are you thinking of making any non-graphics-video-focused SoC boards? (for example with AR9883 SoC - or something there that is focused/used in cheap routers - ofcourse running openwrt)
[19:24:14] <Marex> woo, got my 4-channel 200MHz 2G/s scope today :)
[19:29:43] <mnemoc> jinzo: AM3352
[19:31:13] <jinzo> mnemoc, I only glanced at that one - but tought it's in line with Allwinners offerings
[19:31:33] <jinzo> I'll check it a bit more, sorry for not doing that before.
[19:31:35] <mnemoc> Marex: but just 4 channels? :p j/k
[19:32:32] <jinzo> ah it does have VGA/LCD connectors, I'm looking something in a more "half that price, with WiFi" :P
[19:32:42] <jinzo> (ofcourse, half or even less performance too.)
[19:44:44] <LFSveteran> hmm, no brd file of the olinuxino maxi?
[19:45:32] <xenoxaos> https://github.com/OLIMEX/OLINUXINO/blob/master/HARDWARE/iMX233-OLinuXino-Maxi/iMX233-OLINUXINO-MAXI_Rev_B1.brd
[19:46:15] <LFSveteran> yeah, but that doesn't show the board layout
[19:46:35] <LFSveteran> only some drawing of the case
[23:41:34] <WarheadsSE> Tsvetan: I've put out 2 kernels & a rootfs for the a13 & a10s
[23:41:58] <WarheadsSE> I have an updated fex file I need to send you at some point that properly turns off the axp features on the -micro
[23:42:32] <mnemoc> WarheadsSE: found the cause of the disp_alloc() crash?
[23:43:18] <WarheadsSE> mnemoc: yeah, remember, I accidentally left an reserve in the kernel config :P
[23:43:59] <WarheadsSE> so i have built up linux-sun5i (sunxi-3.0) and linux-sun5i-staging (stage/sunxi-3.0)
[23:44:38] <mnemoc> WarheadsSE: \o/
[23:46:25] <WarheadsSE> so now all linux-sunXi are @ 3.0.57+
[23:47:02] <BusError_> anything more recent on A13's ? 3.0.x is ancient...
[23:53:35] <mnemoc> 3.4
[23:54:08] <mnemoc> people says it actually runs faster...
[23:54:29] <mnemoc> but usb gadget is broken
[23:57:29] <WarheadsSE> I'd bet that 3.4 does run faster.
[23:57:48] <WarheadsSE> when I am told (& convinced) that 3.4 is stable, i'll bump them.
[23:58:09] <WarheadsSE> until then, I know most stuff is working in 3.0.x, and stably so.
[23:58:28] <WarheadsSE> and anyone can always feel free to adjust my PKGBUILDs to make their own
[23:58:58] <mnemoc> :)
