hno changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See | | Logs at
vicenteH has quit [Ping timeout: 248 seconds]
derethor has joined #linux-sunxi
<servili007> libv: Yep, was my binary drivers, works beautifully after a local recompile, thanks
<jlj> is the edid hdmi code only in 3.4 or should it work in 3.0 too?
<jlj> so that the card autodetects the resolution at start up
auxym has joined #linux-sunxi
<auxym> bah. hours of messing around to figure that my problem seems to be buggy wifi drivers
<auxym> do any of you know if "8192cu" are drivers provided by realtek or community-developed?
egbert has quit [Disconnected by services]
egbert_ has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
\\Mr_C\\ has quit []
\\Mr_C\\ has joined #linux-sunxi
<Turl> jlj: should be on both
<Turl> auxym: there's some drivers on mainline, and then there's the ones provided by realtek
<Turl> I think 8192cu is the realtek one and rtl8192cu the mainline ones
hramrach__ has quit [Remote host closed the connection]
hramrach__ has joined #linux-sunxi
BJfreeman has quit [Quit: had a good time]
[7] has quit [Read error: Operation timed out]
TheSeven has joined #linux-sunxi
<auxym> yeah thats what I figured. Guess I'm out of luck for bug reports in that case?
<Turl> auxym: what bug are you experiencing?
servili007 has quit [Quit: Leaving]
<auxym> trying to connect to my apt block's wifi, which uses WPA2 with EAP-PEAP (mschapv2) auth. the exact same wpa_supplicant config that works on my suse box with ralink 2800 fails on my a10.
<auxym> the a10 (with rtl8188cus) does work with regular wpa2-psk and about anything else, just eap-peap that I can't get working, no matter what tweaks i do to wpa_supplicant.conf
<auxym> I'm about 99% percent sure its not a config issue but something with the driver (or h/w)
<auxym> tried with manual wpa_supplicant config but also wicd and lxde's network manager, all the same result
ibrah has quit [Ping timeout: 260 seconds]
<Turl> auxym: you could try with the other driver
<auxym> yeah, i did try a bit yesterday, albeit with an older kernel (3.0 something) it was way worst, i couldnt get anything to connect. googling around, a lot of people are reporting that rtl8192cu has a lot of issues. hence with it seems 8192cu is the default
<auxym> ill try getting the latest one from realtek and compiling it tomorrow, it was released nov '12.
jlj has quit [Ping timeout: 252 seconds]
derethor has quit [Ping timeout: 268 seconds]
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
Tsvetan has joined #linux-sunxi
rz2k has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest56091
_BJFreeman is now known as BJfreeman
Guest56091 has quit [Ping timeout: 252 seconds]
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest2866
_BJFreeman is now known as BJfreeman
wingrime has joined #linux-sunxi
Guest2866 has quit [Ping timeout: 252 seconds]
mripard_ has quit [Ping timeout: 248 seconds]
mripard has joined #linux-sunxi
<oliv3r> mornin'
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest2720
_BJFreeman is now known as BJfreeman
MadSpark has joined #linux-sunxi
Guest2720 has quit [Ping timeout: 252 seconds]
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest88069
_BJFreeman is now known as BJfreeman
Guest88069 has quit [Ping timeout: 252 seconds]
rellla has joined #linux-sunxi
<rellla> morning
<rellla> Turl: ping
BJfreeman has quit [Read error: Connection reset by peer]
<oliv3r> mornin
<oliv3r> turl is probably sleeping :)
Black_Horseman has quit [Remote host closed the connection]
<rellla> wiki spammers got more clever last days :p
<oliv3r> again?
<oliv3r> turl just changed the questions
rz2k has quit []
<hramrach__> mornin
theOzzieRat has quit []
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
eebrah|away is now known as eebrah
<oliv3r> hmm, taht's strange, my cubie1 doesn't have the boot magic csel value in dram controller
<oliv3r> ssvb: ping
<ssvb> oliv3r: pong
<oliv3r> hi
<oliv3r> so, what do I need to look for/at
<oliv3r> though I think it's all solved now, isn't it?
<oliv3r> i've put default boot0 values here:
<oliv3r> btw, boot0 is identical (for the same version) amongst all boards
<ssvb> I think the ball is now on Tom's and/or Allwinner's side
<oliv3r> luke was saying in the 'how easy it is to use allwinner SoC's' that they have 1 binary dump, and only change the fex file
<oliv3r> tom doesn't compile boot0 as it's fully closed (until recently of course)
<oliv3r> i do hope sincerly we get more dram info
<oliv3r> but i still doubt it's allwinner's own IP
<ssvb> it would be interesting if they shed any light on CAS6 vs. CAS9 configuration choice
Superpelican has joined #linux-sunxi
<oliv3r> i mean, look at things like tpr1, tpr2, tpr3; they simply copy the values from the config into the register
<ssvb> maybe CAS7 would be the best for this hardware
<oliv3r> and there's more options they just blalantly copy from the config -> registers
<ssvb> sure, I mean these settings affect memory latency
<oliv3r> i hate working with a black-box like that :)
<ssvb> probably Allwinner could review them and advise what is best for reliability & performance
<oliv3r> but i sincerly doubt we get any more info :p
<oliv3r> we have their 'latest changes' since that's the a20 memory settings
<oliv3r> and they haven't changed much really
<oliv3r> well in the code
<oliv3r> and i bet nobody dares touchding dram_para :)
heffer has joined #linux-sunxi
heffer has quit [Changing host]
heffer has joined #linux-sunxi
<oliv3r> gah, not even the timing refresh rates are the same, which is crazy concidering the ram and pcb layout is identical
<ssvb> to me it just looks like some random reshuffling of dramc setting is being done, and then they keep whatever does not obviously fail
* ssvb fails to see any intelligence in this process
vicenteH has joined #linux-sunxi
<oliv3r> i doubt there is
<oliv3r> sun4i sun5i sun7i in 1 dram.c is slowly becomming messy
paulk-desktop has joined #linux-sunxi
rz2k has joined #linux-sunxi
<oliv3r> sun4i doesn't have option c), sun5i wants this option done before option a), sun7i has this additional option
<oliv3r> if it where only performance tunings
<oliv3r> as for the pll5 vs pll6 options, those are 'tuning' settings, which should be in dram_para, with some safe default
<oliv3r> not hardcoded imo
<oliv3r> but that's all for later
<oliv3r> just like removing GPS clock in dram.c
rellla has quit [Remote host closed the connection]
rellla has joined #linux-sunxi
<hramrach__> ssvb: that's how chip design works as well. you test it until there are no obvious flaws
<hramrach__> so no wonder chip settings follow similar pattern
<hramrach__> hmm, I have this patch for sun[67]i nand partition support in kernel somewhere
<hramrach__> the top two commits here (or on ML)
<hramrach__> so if somebody feels bored you can look at it and suggest how to do it sanely
<hramrach__> basically I have 1 code which I #include with different header which makes it support two layouts
<hramrach__> OO code in C for you
<hramrach__> it's ugly but the alternative is to have the code twice which is also ugly
techn__ has joined #linux-sunxi
<hramrach__> but given that sun[45]i fits all four mbr copies into one page and sun[67]i requires get_part_sector dances to read just one mbr maybe it is worth splitting after all
techn_ has quit [Ping timeout: 246 seconds]
<hramrach__> github upload 15k/s /o\
<mnemoc> can someone make me a branch to merge into stage/3.4 ?
<mnemoc> (same for 3.0 is also welcomed)
<mnemoc> but can't review the ML atm
<hramrach__> mnemoc: tha't a branch. want a pull request?
<hramrach__> but look at the code befor you consider merging ;-)
<mnemoc> hramrach__: not only nand. all the pending stuff
<mnemoc> i can't review :(
<mnemoc> just want to merge stuff to show it's still alive
<hramrach__> it did not break the kernel for me ;-)
<mnemoc> there is tons of hansg stuff pending
<mnemoc> is someone can make a branch to merge on one pass *safe* stuff, it would be gratly appreciated
<mnemoc> need to do tons of things before leaving .es this thrusday
<hramrach__> heh, leaving on wed so not going to do much stuff either
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
<hramrach__> maybe leave it for later. it's summer holidays so people are offline a lot
<mnemoc> great time for job hunting :-/
<Black_Horseman> great time for lives
<hramrach__> I don't think that's too much of an impediment. They have dedicated departements just for that in most companies ;-)
<mnemoc> but if the guy who signs the contract is in the caribian, your are screwed
<hramrach__> I am sure they have procedures for that. can't stop the company because somebody ins in the carribean
<mnemoc> :)
<oliv3r> bus-factor 1
hramrach__ has quit [Ping timeout: 240 seconds]
hramrach__ has joined #linux-sunxi
<oliv3r> Turl: ping
hansg has joined #linux-sunxi
jemk has joined #linux-sunxi
<Turl> oliv3r: pong
<oliv3r> hey
<oliv3r> good morning :)
<oliv3r> Turl: i would like to use your expert opinion on something
<Turl> oliv3r: :p sure
<Turl> rellla: wiki spam :( I changed the questions like 3 days ago
<Turl> mripard: do you have any news from mdp on DMA?
<oliv3r> oh dma
<oliv3r> Turl: let me prepare a paste
<mripard> Turl: nope, pinged him like last friday, he's still off the radar
<oliv3r> Turl: currently, we have a table with fixed magic values, hpcr_value[32];
<oliv3r> Turl: i found a way to give those magic values meaning
<oliv3r> the least intrusive (littlest changes) is to use those defines instead of the magic values
<oliv3r> i opted for assiging it directly as can be seen in the paste
<oliv3r> any other possibility way of doing this, keeping the descriptivness?
<oliv3r> i guess the only other way, is to set those registers one at a time
<oliv3r> yeah, but then i loose the meaining of the index i'm affraid
<oliv3r> i was thinking of doing that too
<Turl> oliv3r: it goes in the comment
<oliv3r> but it's 'fixed order' then
<oliv3r> sun5i and sun7i have different array entries
<oliv3r> indexes*
<oliv3r> and the array becomes sensitive to ordering
<oliv3r> which is why i kinda try to stay away from that method
<Turl> then I guess your solution is the way to go
<oliv3r> :S
<oliv3r> i don't like mine though :)
<Turl> yeah it's a bit verbose :)
<oliv3r> i know
<oliv3r> but i want to be able to quickly group identical settings for the various socs
<oliv3r> ok then i'll keep it that way
<Turl> maybe you can make a helper macro
<Turl> all of them have XS_EN for example
<oliv3r> like?
<oliv3r> except the ones that are '0'
<oliv3r> :)
<oliv3r> but yeah, xs_en is a little reduntant
<oliv3r> IF you set anything, you need to enable it;
<Turl> #define HPCR_DEFINE(name, val) hpcr_value[(name)] = ....XS_EN | (value)
<oliv3r> it strips the XS_EN
<oliv3r> it's a few characters shorter
<oliv3r> #define XS_EN(name, val); :p
<Turl> #define HPCR_DEFINE(name, val) hpcr_value[(DRAM_HPCR_ ## name)] = ....XS_EN | (value)
<Turl> (untested, but might work :p
<oliv3r> :p
<oliv3r> lets try that
<oliv3r> if it saves 5 characters, it'll be good
<oliv3r> i'm acking of putting that define above the declration
<oliv3r> aye read that today too
<oliv3r> and will be lowered more again jul 1st
<hno> hi mnemoc!
<oliv3r> Nelly kroes, our dutch hero-es, said 'there are no telecom borders in europe'
<oliv3r> hi hno :)
<Turl> oliv3r: today is jul 1st :p
<oliv3r> Turl: every year, every jul 1st the rates will be dropped
<Turl> oh cool :)
<oliv3r> and in a few years, all roaming costs should be gone in the EU
<oliv3r> Turl: ok it shortedned by over half the line, very good
<hno> hi oliv3r
<oliv3r> enjoying your vacation?
<Turl> oliv3r: does it build? :P
<oliv3r> erm
<oliv3r> dunno yet :)
<oliv3r> lets find out
<oliv3r> there might be a space too much in your define
<wingrime> Turl: may be disable wiki registation for while
<oliv3r> not an option really
<oliv3r> manually accepting new registrants maybe
<wingrime> ssvb: why aw set a20 clocks as 920 Mhz?
<oliv3r> Turl: yes, yes it does :)
<oliv3r> wingrime: 912 I think is the 'stock' a20 clock
<oliv3r> wingrime: did you get your c2 yet?
<wingrime> oliv3r: nope
<oliv3r> :((
<Turl> oliv3r: heh great :) I'm surprised I didn't mess up the ##concat
<Turl> oliv3r: neither have I :P
<oliv3r> i thought it would have been DRAM_HPCR_##name :)
<oliv3r> gawd i hate the 80 lines limit
<oliv3r> linus is right, checkpatch should be far more flexible there
<ssvb> wingrime: nobody knows for sure, but I guess that initially they want to play safe and avoid any potential problems
<Turl> I think this new spammers are humans, and it's hard to separate humans from humans with a captcha :/
<ssvb> wingrime: or maybe they have problems with their manufacturing process and the chips just can't work at high clock frequencies
<ssvb> wingrime: the CPU in my CubieBoard2 can be overclocked to 1.15GHz, but is really unstable at 1.2GHz even with higher voltages
soul has joined #linux-sunxi
<wingrime> ssvb: I saw somewhere that current tech process ar 40nm
<wingrime> for a20
<wingrime> but RockChip's IC now use 28nm
<ssvb> wingrime: btw, A10 in my Mele A2000 also can't be clocked higher than 1.15GHz, so A10 and A20 seem to be pretty similar with the regards to the CPU clock frequency limit
naobsd has joined #linux-sunxi
<Turl> mnemoc: what do you think? should we disable registrations?
<wingrime> ssvb: it can be related curent RTL (litography standard block library)
<rellla> Turl: so you can only separate them with manual account confirmation. but that maybe keeps someones from contributing...
<wingrime> ssvb: AW may be can't use 28nm with current price policy
<oliv3r> wingrime: i still think a20 is a 'cheap stop gap' while we wait for a40 :)
<Turl> rellla: yeah :( sucks that spammers exist
<ssvb> wingrime: it looks like Allwinner is losing this round to the competition, the others have Cortex-A9 SoCs with more cores and higher clock frequency
<Turl> rellla: elinux has registrations disabled too
<oliv3r> it's better to make us work harder in removing spam accounts, then it is to not let people contribute
<oliv3r> contribution is top prio
<oliv3r> ssvb: i kinda agree
<oliv3r> ssvb: but even with a worse soc, they could have still made a killer, if they had their oss work in order (e.g. have support mainlined before silicon leaves the factory)
<Turl> oliv3r: yeah but they sell premade device designs pretty much
<Turl> it's not a need they have
<oliv3r> true, they do buy a lot of IP
<wingrime> ssvb: but AW have good market place with 'Cheap' SOC , I think that AW can't make better SoC generaly related TSMC prices for better tech-process
<Turl> and judging by the code quality, looks like they're fans of 'it works? SHIPIT!'
<wingrime> ssvb: SoC simply not fit cheap price policy with better techprocess
<wingrime> Turl: china logic
<Turl> wingrime: s/china/most companies'/
<wingrime> Turl: it possible make manual conformation ?
<wingrime> Turl: you simply can drop 'Crap' usernames
<rellla> Turl, mnemoc, auto delete of new accounts with no edit within eg. a week would be a first step. and i would delete all accounts without contribution till now. i searched for that but i haven't found out to do that without direct manipulation of the db
<Turl> mostly prop software companies
<Turl> rellla: yeah but there's that :P how do you do it?
<Turl> rellla: there's no concept of deleting accounts afaik
<Turl> just blocking
<rellla> db tables are very connected afaik :(
<Turl> wingrime: they already have it enabled I think, you get a mail and need to click a link or sth
<wingrime> Turl: I talking about That after email activaion, some one from linux-sunxi must aprove it
<wingrime> like invite
Black_Horseman has quit [Quit: Zwi se logou mou!!!]
vicenteH has quit [Ping timeout: 252 seconds]
<Turl> mnemoc: ^
<rellla> Turl: looks good. so try it and delete us all ;)
<Turl> yeah I'm not gonna run it as is :P need to take a db backup first if anything
<Turl> but I've got class in like 15m, can't do it now :P
<oliv3r> Turl: go to your class then :)
<oliv3r> this is how it'll look like
<oliv3r> i like it
n01 has joined #linux-sunxi
<rellla> need a mediafile tester with either vlc or xbmc. any voluntaries?
n01 has quit [Client Quit]
n01 has joined #linux-sunxi
<oliv3r> i have no sd card with xbmc on it
<oliv3r> still waiting for you to create a universal image like hasng's :)
zrafa has quit [Ping timeout: 240 seconds]
<rellla> ah. nice to know ;)
<rellla> i want to know, how a 1080i file plays on a10
<rellla> hramrach__ ^
wingrime has quit [Ping timeout: 264 seconds]
eebrah is now known as eebrah|away
wingrime has joined #linux-sunxi
zrafa has joined #linux-sunxi
<naobsd> hmm, there are some forks on github, how can I find better one...
<oliv3r> notmart: what are you after
vicenteH has joined #linux-sunxi
<oliv3r> hno: i think there's some flaws in the hpcr tables w ehave
stekern has quit [Ping timeout: 246 seconds]
<oliv3r> ah, yeah either mine or hno's u-boot
<oliv3r> both work if you have the right branch
<oliv3r> mine is very volatile though
<oliv3r> as i randomly push --force stuff :)
<oliv3r> but hno hasn't gotten all my patches worked in yet
<oliv3r> there's a small performance bug in hno's branch, it's nothing major :)
stekern has joined #linux-sunxi
<naobsd> dram init code is tooo cryptic for me ;)
<oliv3r> naobsd: yeah it's horrible
<oliv3r> but if you look at my last patches, i'm working on cleaning it up
<oliv3r> or atleast give meaning to things, cleanup follows
<naobsd> mbus clock is not fixed yet?
<naobsd> sorry
<naobsd> there is #ifdef CONFIG_SUN7I
<oliv3r> naobsd: a20 yeah :)
<oliv3r> a20 has mbus of pll6; a13 has pll5 (pll6 not available/working) and a10 has unchanchable mbus :)
<oliv3r> e.g. all settings are ignored.
<naobsd> mmm
<oliv3r> so you can't tune mbus for a10, a13 a little (but not really) and a20 (default) is pll6 which gives you about 20% performance increase
<oliv3r> boot0 has this by default obviously
Superpelican has quit [Read error: Operation timed out]
* oliv3r starts to yell *arrrrggh*U
[hawk] has quit [Remote host closed the connection]
[hawk] has joined #linux-sunxi
jemk has quit [Read error: Connection reset by peer]
<naobsd> oliv3r: thank you
[hawk] has quit [Remote host closed the connection]
[hawk] has joined #linux-sunxi
Superpelican has joined #linux-sunxi
\\Mr_C\\ has quit []
<auxym> im using a debian image from the miniand forums and its missing the kernel headers (to compile modules)
<auxym> any idea how I could get my hands on them? kernel 3.4.29+
<hramrach__> rellla: 1080 generally plays for me. I can tell you if it crashes the player but not what the output is atm ;-)
<hramrach__> also post a sample of the file otherwise nobody can test ;-)
\\Mr_C\\ has joined #linux-sunxi
<rellla> hramrach__: it's the new one (9th) sample of
<rellla> i'm interested, how the a10 goes with 1080 interlaced
<rellla> as for my pvr-streaming client i need fluid playback of 720p and 1080i dvb-s/s2 streams
<rellla> currently streaming of 1080i channels isn't that perfect
hansg has quit [Quit: Leaving]
jemk has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
rellla has quit [Quit: Nettalk6 -]
^hawk^ has joined #linux-sunxi
^hawk^ has quit [Remote host closed the connection]
[hawk] has quit [Ping timeout: 260 seconds]
^hawk^ has joined #linux-sunxi
eebrah|away is now known as eebrah
<wingrime> oliv3r: you are remeber where that tool for android
<wingrime> oliv3r: for dram
<wingrime> oliv3r: I just bye chep tablet for a10 testing
Superpelican has quit [Ping timeout: 260 seconds]
hipboi has joined #linux-sunxi
<wingrime> trul: do you remember where that tool
jemk has quit [Ping timeout: 248 seconds]
<naobsd> oliv3r: I got ramdom crash with your u-boot wip/a20
<naobsd> random
<naobsd> e.g. [ 31.127569] Unable to handle kernel paging request at virtual address aaaaaaa8
<naobsd> it seems stack is broken...
ykchavan has joined #linux-sunxi
stekern has quit [Ping timeout: 260 seconds]
<naobsd> hno's u-boot with fix for mbus_clk_cfg is working and better performance w/o fix... so something may be wrong in another place :(
stekern has joined #linux-sunxi
auxym has quit [Quit: Lost terminal]
wigyori has quit [Ping timeout: 260 seconds]
wigyori has joined #linux-sunxi
<hramrach__> hmm, rellla gone
<hramrach__> currently my disp is broken and cannot play anything with vlc
<hramrach__> my kernel was too experimental after all it seems
<hramrach__> but not sure what I did to break disp
^hawk^ has quit [Remote host closed the connection]
^hawk^ has joined #linux-sunxi
tinti has quit [Read error: Connection reset by peer]
tinti_ has joined #linux-sunxi
<hramrach__> maybe it was broken by one of the recent merges. need to investigate a bit more. I am quite sure the part with tracking disp layers stil worked since I used it to track layers
^hawk^ has quit [Remote host closed the connection]
^hawk^ has joined #linux-sunxi
leowt has joined #linux-sunxi
n01 is now known as n01_away
Superpelican has joined #linux-sunxi
hipboi has quit [Quit: Leaving]
rz2k has quit []
leowt has quit [Quit: leowt]
notmart has quit [Quit: notmart terminated!]
leowt has joined #linux-sunxi
leowt has quit [Client Quit]
auxym has joined #linux-sunxi
Superpelican has quit [Read error: Connection reset by peer]
Superpelican has joined #linux-sunxi
ibrah has joined #linux-sunxi
<auxym> i just made a bootable sd card with bsp/hwpack. does it install kernel-headers to the rootfs? if not can I get them somehow?
<auxym> i need to compile a module
ykchavan has quit [Ping timeout: 248 seconds]
^hawk^ has quit [Remote host closed the connection]
^hawk^ has joined #linux-sunxi
<mnemoc> auxym: clone the tree and build your own custom kernel
<mnemoc> those hwpacks include exclusively the defconfig
<auxym> ok... so I need to edit the kernel config so it produces headers when compiling?
<auxym> never compiled a kernel before
<Turl> auxym: do you need an out of tree module?
<mnemoc> /porc/config.gz is usually the simplest starting point
<Turl> mnemoc: don't we have a make linux-config?
<mnemoc> yes
<Turl> you can use that to enable the module if it is in-tree
<mnemoc> yes
<Turl> but if you need sth external you'll need the headers
<auxym> yeah. as we talked ysterday im going to try the latest driver from realtiek for the wifi, so I need to compile it from source i downloaded from realtek's site
<Turl> auxym: what version is it?
<auxym> 3.4.4_4749 released nov 11 2012, according to rt
<auxym> and yeah, thats the version your commit has
<auxym> too bad :( was my last hope
<Turl> yeah but it's not merged on linux-sunxi
<auxym> i see. can I somehow merge into my local copy? sorry im not terribly familiar with git
<Turl> so applying those commits to your tree and rebuilding could be a way to go for testing the new release
<Turl> auxym: yes, sec
notmart has joined #linux-sunxi
<auxym> awesome, trying it out
^hawk^ has quit [Remote host closed the connection]
^hawk^ has joined #linux-sunxi
<auxym> Cannot write to `-' (Success).
<auxym> guess thats not good?
<Turl> is that the only thing that's printing?
<auxym> bah, ill just wget them manually and cat them into git am
<auxym> git gives me this though, normal?
<auxym> previous rebase directory /home/francis/sunxi/sunxi-bsp/.git/modules/linux-sunxi/rebase-apply still exists but mbox given.
<Turl> yeah not good
<Turl> the first one doesn't apply any longer, let me refresh them
<auxym> oh :\
<Turl> meanwhile run git am --abort
<auxym> done
<oliv3r> wingrime: a10-meminfo?
<oliv3r> naobsd: that's almost impossible, since u-boot only ... boots :)
dwilkins has quit [Ping timeout: 240 seconds]
<oliv3r> naobsd: i still have to confirm, but my branch should setup the memory identically as boot0
<Turl> mnemoc: /q?
<Turl> auxym: try wget -O-|gunzip|git am
jemk has joined #linux-sunxi
<Turl> should apply the 4 patches
<auxym> looking good
rellla has joined #linux-sunxi
<auxym> so.. make clean, make the hwpack and I should have a kernel with the latest module?
<Turl> in theory, yes
Superpelican has quit [Quit: Konversation terminated!]
<auxym> great. compilin' (and crossing fingers)
<auxym> bah. huge thanks for your help Turl, but no go. still same issue with peap authentication. ill have to find a workaround i guess
<mnemoc> Turl: type freely
<mnemoc> Turl: i'm flying out of .es this thursday at 6am.... so i'm a bit short of spare time. but i can read and reply async ;-)
<Turl> auxym: hm, well, it was worth a shot in any case :)
<Turl> mnemoc: going back to .cl?
<mnemoc> Turl: i'll try .de first
<mnemoc> there is zero job for C apes in .cl
<mnemoc> VB or java only
<Turl> mnemoc: no apes programming PICs? :P
<oliv3r> samsung is looking for people in south korea
<mnemoc> Turl: chile doesn't make anything. sells cooper, apples and fish
<mnemoc> nothing made, only collected
dwilkins has joined #linux-sunxi
<mnemoc> everything else is imported
<mnemoc> not even a f* bike factory
<oliv3r> sprechen sie deutch
<auxym> would it be worth to either post the issue (with wpa_supplicant logs) somewhere, or email realtek?
<Turl> mnemoc: we are assembling our stuff at least :P
<mnemoc> oliv3r: I lived a year in berlin back in 2006..... but i wouldn't try to speak it loud :p
<rellla> *deutsch btw ;)
<mnemoc> but to be honest, i don't speak spanish either....
<mnemoc> just type :|
<oliv3r> rellla: i was born in austria, i'm half austrian :) left when i was 10 though
<rellla> your last name sounds austrian ;)
<Turl> mnemoc: how come you don't speak spanish?
<mnemoc> i've spoken more english in the last decade than spanish even when living in spanish speaking countries 33 of my 35 years
<mnemoc> Turl: social problems
<oliv3r> no hables espagnol?
<Turl> the same probably holds true for me wrt writing
<oliv3r> move to s. korea, work for samsung
<mnemoc> sure i can, if i know the person in front of me and there are les than 4 people total
Dreadlish is now known as PanBozio
<Turl> oliv3r: 8/10 nice try :P
<oliv3r> 8/10?
<mnemoc> but in english it tends to be work-related, and that's simpler to deal with
<Turl> oliv3r: we grade stuff on a 1-10 scale, with 10 being excellent
<oliv3r> oh
<oliv3r> hah
PanBozio is now known as Dreadlish
<oliv3r> parles vous francais
<auxym> oui :)
<oliv3r> ich bin ein berliner!
<mnemoc> love that city
<Turl> oliv3r: I'm a es-en guy, I need to learn cn and then I'd have covered the majority of the world's population :P
BJfreeman has quit [Ping timeout: 252 seconds]
<oliv3r> LOL
<oliv3r> yeah i wanna learn cn
<auxym> sounds hard as hell though
<Turl> I don't :P I can hardly distinguish the characters
<mnemoc> it's easier to teach chinese to speak english than an argentinian to speak anything.....
<Turl> mnemoc: lol
<mnemoc> like with spaniards
<mnemoc> they just can't
<mnemoc> .es people can't even pronnounce two consonants together in spanish words
<mnemoc> and god save us from spaniards trying to talk in english
<Turl> mnemoc: no creo :P
<mnemoc> septiembre -> se-tiembre
<Turl> mnemoc: I had a english teacher from spain once :D
<mnemoc> sexo -> seso
<mnemoc> victoria (my daughter) -> vi-toria
<mnemoc> madrid -> madri'
<mnemoc> they can't even speak their own language
<Turl> I hate 'setiembre', but RAE says it's ok
<mnemoc> sure it's "ok"
<mnemoc> they can't bad spain's spanish
<mnemoc> ban
<Turl> on the other hand, all of those are pretty normalized
<mnemoc> even with football teams -> atletico -> a-letico
<mnemoc> seso? aletico? that's not valid anywhere
<mnemoc> and not even jouranlists can pronounce them
<Turl> mnemoc: well, that must be Only in Spain (c)
<Turl> seso over here is what you have inside your head :p
<mnemoc> yes, .es problem. the rest of the spanish speaking counties only have troubles with the s/z and b/v
<Turl> that's not a problem, that's a solution ;)
<Turl> y/ll
<mnemoc> :p
<mnemoc> Turl: y/ll is written. s/z and b/v sound differently
<Turl> on some places of spain (most iirc) b/v is written too
<Turl> I'll give you s/z :P
<mnemoc> the rule is c sounds like k or z.
notmart_ has joined #linux-sunxi
<mnemoc> in the south of spain they make s sound like z
<mnemoc> in latin maerican and canaria, z sounds like s
<mnemoc> try to survive 10 minutes talking with someone from cadiz and you'll want to suicide
<Turl> haha
notmart has quit [Ping timeout: 248 seconds]
<mnemoc> the "german" sch sound is in half of the words
auxym has quit [Ping timeout: 246 seconds]
<mnemoc> and all 's' sounds like an strongly marked z
<Turl> sch sounds like y?
<mnemoc> think in chile pronounces as ssshile
<oliv3r> schoen stoppen
<mnemoc> pronounced*
eebrah is now known as eebrah|away
<Turl> like the y sound in "lluvia" then
<mnemoc> Turl: yes, like the argentinian lluvia
<mnemoc> while everyone else lluvia starts like the english "you"
<Turl> well, that sound is every ll/y on the words
<Turl> we should just write them with y, yuvia, yamada, yoyo, yo, yavero :P
<mnemoc> or teach argentians to pronounce :p
<Turl> that'd create a tear in space-time that would eat us all
<mnemoc> :)
<Turl> we could teach .es people "simplified spanish" too :P
<mnemoc> each province has an strongly differenciated accent
<mnemoc> and in my region in particular, the "local" language was standarized very short ago
<mnemoc> so galician speaking children don't understand theyr grandfathers or the other way around
notmart_ has quit [Quit: notmart terminated!]
<mnemoc> kind of shameful
auxym has joined #linux-sunxi
<Turl> chilean teens need some spanish classes too, you can hardly understand what they write when there's a "wn", "wea" and similar words mixed with contractions, flogger speech and such :p
<mnemoc> .cl has a VERY poor vocabulary. like 500 words total
<Turl> mnemoc: standarized as making its teaching compulsory?
<mnemoc> the rest, wildcards
<mnemoc> Turl: inventing dictionaries and grammar standarizing a language that used to defer from town to town
<Turl> that's a major fail on their part
mturquette has joined #linux-sunxi
<mnemoc> Turl: in .cl over half of the 18yo children end public education without been capable to read
<mnemoc> and that is very clear when they try to express themselves spoken
<Turl> I know they were pushing for asturiano too (on asturias obviously)
<Turl> mnemoc: ouch
auxym has quit [Client Quit]
<mnemoc> asturia's language doesn't have legal priviledges. only basque, galicia and catala/valencian
<mnemoc> valencia people keeps fighting to be a real language and not a dialect, but they are mostly ignored :p
<Turl> iirc all the software used in public administration must be translated to asturianu, and stuff like that
<Turl> their OS translating community is pretty active
<mnemoc> Turl: nationalism, but no legal support
<mnemoc> bable (asturian) is not protected/supported
<Turl> bable?
<mnemoc> that's how they call their language
<mnemoc> no clue why
<Turl> I thought they called it asturianu with u
<mnemoc> they were not considered to have enough "own history" to deserve cultural protection :p
ibrah has quit [Ping timeout: 256 seconds]
<oliv3r> mnemoc: so what is your next move thursday, randomly go to germany and apply for jobs?
wingrime has quit [Ping timeout: 246 seconds]
<mnemoc> oliv3r: i'll leach some days in the house of a friend and try to apply as C ape to his company (in postdam or luneburg/hamburg) and if fails, keep searching in the berlin area
<mnemoc> and finish a webshop project in the meantime
<mnemoc> but aiming at working as c/linux ape with corworker with which to be able to talk while taking a coffee
<mnemoc> 7y fully remote is more than enough to make anyone nuts
<mnemoc> if berlin doesn't work.... probably hannover next
<mnemoc> at least that's the current plan
<hramrach__> you speak german?
<mnemoc> i understand and sort of read it
<mnemoc> can't talk
<mnemoc> so i;m limited to english speaking envs
<hramrach__> if you stay there you will learn fast :)
<mnemoc> i survived a year in 2006... hope i'll recover at least a very basic level soon enough
<Turl> do you need special visas or stuff to work on another eu country?
<oliv3r> germany isn't very english friendly
<mnemoc> Turl: I'm .it
<oliv3r> though I don't know where linux C apes can work in NL
<oliv3r> red hat :)
<mnemoc> oliv3r: it's fdifferent to apply to work on a supermarket than to apply to code network apps in c
<mnemoc> i'll obviusly won';t fake a CV in germna. it will be in english (as always)
<mnemoc> if they accept a CV and presentation letter in english, they are willing to accept english speaking engs
<mnemoc> if not, they won';t reply
<hramrach__> hehe, german have their own computer parts
<hramrach__> Plattenspeicher
<oliv3r> lol
<oliv3r> what i ment is, that in germany, they prefer german
<oliv3r> just like in france, they prefer french :)
<oliv3r> exceptions apply of course
<mnemoc> sure. and need to learn german if i plan to work there
<oliv3r> in NL, if your an engineer, english isn't weird
<mnemoc> isn;t dutch like tisk? that not even yourself understand eachouther?
<mnemoc> err
<mnemoc> dansk
<oliv3r> lol
<oliv3r> yes we understand eachother fine
<mnemoc> :)
<Turl> mnemoc: you got a referal from the inside? :)
* rellla feels linux-sunxi transforms to "find the best language" channel ;)
<hramrach__> chinese :)
<mnemoc> for my initial try (, yes
<mnemoc> chinese doesn't exist :p
<hramrach__> it does to a point. you just need to specify which chinese you mean ;-)
<oliv3r> most importantly
<oliv3r> how much time and WHEN can you work on sunxi agian :)
<Turl> mandarin is #1 :P
<hramrach__> ttly
<Turl> but only useful in .cn :P
<hramrach__> not that I know any word of any chinese :p
<mnemoc> oliv3r: i have a 3M 40h/w web project to do while searching for a job. the rest is for interviews, and sunxi
<mnemoc> once getting a job. sunxi will take all not-work time.... as i'm moving alone with zero interest in starting anything
<mnemoc> just worki with humans, eat and send money back to spain
<mnemoc> and got my a20 and a10s olinuxinos today :)
tinti_ has quit [Remote host closed the connection]
<mnemoc> unfortunatelly lost the cb2..... shipped to my former employer
<mnemoc> but the cubietruck looks nicer ;-)
<oliv3r> mnemoc: c2?
<oliv3r> ah
<oliv3r> and former employer won't get it to you?
<mnemoc> cubieboard+a20
<mnemoc> no. he hates me
<oliv3r> yeah i ment 'what about cb2
<oliv3r> why, really why
<oliv3r> well a20 olinuxino is better :p
<hramrach__> some people are just like that :S
<oliv3r> more pins
<oliv3r> 'you get sick and cost me lots of money'
<mnemoc> because i was on the hospital nad he lost money due to his beloved bus factor = 1 rule
<oliv3r> fu
<oliv3r> if it makes you feel happy
<oliv3r> it'll cost him 10x that to get his next bf=1
<mnemoc> i was sort of happy to send him the ssh keys for him to make his own backups and learn to build his products instead of "training" him
<mnemoc> after years begging to increase the bus fastor
<hramrach__> bus factor?
<oliv3r> bf
<mnemoc> hramrach__: how many people need to be hit by a truck to kill your project
<oliv3r> how many people can be hit by a bus
<oliv3r> that :p
<mnemoc> hramrach__: he sees it as company protection. devs CANT talk to each other
<mnemoc> or they'll steal his customers
<hramrach__> funny variable
<oliv3r> mnemoc: what u-boot branch do I need to get uart -> uSD output
<mnemoc> oliv3r: that's not a branch thing, it's a board choice
<mnemoc> with different -D in boards.cfg
<hramrach__> oh, well
<hramrach__> good night
<mnemoc> good night
<hramrach__> and good luck with job hunting
<mnemoc> i'll still pretty pissed off with him. probably next week will write some tutorials for my former boss
<mnemoc> but not a prio this week
<oliv3r> why?
<oliv3r> he hates you
<mnemoc> yes, but he fed me for 7y
<oliv3r> if he sends you the cubie 2.0; you write his shit
<mnemoc> true
<oliv3r> :)
<oliv3r> a pure principle thing
<oliv3r> mnemoc: ok let me rephrase, how do I get a nand-capable u-boot that prints via uSD-uart
<mnemoc> ah, nand-capable u-boot is another story
<mnemoc> don't don;t about current dev. but you used to need the lichee-dev branch and patch some #defines
<mnemoc> hramrach, anyhow... at least for me german is FAR easier than french .... mostly due to speed
<mnemoc> and intention of comunicating
<oliv3r> :)
<oliv3r> hramrach went to bed! :p
<oliv3r> do we have wiki page of nand-capable u-boot
<oliv3r> or better, precompiled bin?
<oliv3r> i want to start debugging my tablet with 3.4
<oliv3r> ssvb: ping
<ssvb> oliv3r: ?
jemk has quit [Ping timeout: 256 seconds]
<oliv3r> ssvb: i think i found a nother bit of performance drop
<oliv3r> or not, not sure yet
<oliv3r> pll5 vs pll6 was one bit
<oliv3r> factor N = 2 vs 1 is the other
<oliv3r> the thing is, mbus has always been set, even for sun4i
<oliv3r> lkcl ported it over for sun7i but overlooked the changes on mbus; no problem
<oliv3r> but i'm sure i spotted it, anyhow, with N = 2, M = 2 and PLL6 *2 we should have identical performance/setup
<oliv3r> i still don't like how we're shooting blanks in a blackbox ::)
<oliv3r> anyway, mbus clock should be setup idential now, time to look at the rest
<oliv3r> ick, pll5 N is setup wrong aswell
<ssvb> oliv3r: are you comparing results by looking at the register dumps for boot0 and u-boot?
<oliv3r> yeah
<oliv3r> and compare them against boot0 source drop
<oliv3r> and u-boot source code
<oliv3r> needle, haystack, blackbox, lots of deep shit
<oliv3r> i know
<oliv3r> i'm open for suggestions
<oliv3r> i'm a little supprised that pll6 is used intensivly, but isn't even setup properly it seems
<oliv3r> 'we just go with the defaults'
<ssvb> well, it's better than nothing
<oliv3r> unless i'm overlooking something, it's a mess, but that'll be cleaned in time
<oliv3r> i'll gladly here suggestions :)
<oliv3r> ohh pll5 freq. might be off due to different board frequency
<oliv3r> code checks out
<oliv3r> pll5 is set depending on clk-freq from dram.c
<oliv3r> ah i'm using A20-olinixino
<oliv3r> let me switch to cubieboard
<Turl> pll5 at least on A10 is the dram pll
<oliv3r> yeah it is on a20 aswell
<oliv3r> but pll5 'n' factor is determined like such:
<oliv3r> reg_val |= ((clk / 24) & 0x1f) << 8;/* n factor */
<oliv3r> and clk is dram_clk from dram_para
<oliv3r> for cubieboard2, that should be 480 MHz
<oliv3r> ok that's odd
<oliv3r> if clk is 480, and you div that with 24, that's 20
<oliv3r> buf if you & 0x1f, it'll be gone
<oliv3r> oh wait, decimal :)
<oliv3r> so that's 0x14
<oliv3r> and that's not what's in the register :S
<oliv3r> my cubie runs at 432 MHz
<oliv3r> that explains that
<oliv3r> and someone submitted a 480 MHz dram_para
<oliv3r> right, fine with me :)
* Turl wants his cb2
<Turl> I'm missing all the fun :P
<oliv3r> Turl: :(
<oliv3r> wingrime still hasn't gotten his
<oliv3r> you got a10s though
<oliv3r> oh
<oliv3r> you want some fun?
<Turl> yeah I got the A10S olinuxino :)
<oliv3r> ok here's fun
<oliv3r> mind you, i gave up for now ;)
<oliv3r> (wait more links inc)
<Turl> oliv3r: you're doing all those manually? :P
<oliv3r> Turl: what?
<Turl> converting to HPCR(..) notation
<oliv3r> obviously
<oliv3r> how would you do it manually :p
<oliv3r> it doesn't exist yet :)
<Turl> why not write code to convert then? ;)
<oliv3r> lol
<oliv3r> it's a puzzle
<oliv3r> :)
<Turl> not really
<Turl> I'll write it later today :P
<oliv3r> slow down cowboy
<oliv3r> :)
<oliv3r> if it where only that little table, it would be done
<oliv3r> right, from the sun4i code drop; we got these as defines
<oliv3r> so it's to be assumed, those are the array indexes for sun4i
<oliv3r> all kinda makes sense
<oliv3r> the values make somewhat sense
<oliv3r> undefined ones are empty (except for the last for some reason)
<oliv3r> sun5i (a10s!!) changed that
<oliv3r> the names got shuffled around a bit
<oliv3r> and the top half is missing
<oliv3r> the table kinda matches that too
<oliv3r> from what I saw, content seems to match too; but haven't delved into it
<Turl> oliv3r: you linked me to the same thing twice :p
<oliv3r> crap
<oliv3r> but the interesting bit comes with a20
<oliv3r> where the table doesn't really match the code
<oliv3r> of course this is u-boot vs kernel
<oliv3r> so it could be very much that it's a copy/paste job failing imo
<oliv3r> but it's what we have
<oliv3r> so a) what does your nand-u-boot say at: md 0x01c01290 0x10
<oliv3r> md 0x01c01250 0x20
<oliv3r> above one is wrong :)
hramrach__ has quit [Remote host closed the connection]
<oliv3r> md 0x01c01250 0x20 for a10s is?
ZaEarl has joined #linux-sunxi
hramrach__ has joined #linux-sunxi
<lkcl> oliv3r: the main thing i wanted was "something working"
<oliv3r> aye
<oliv3r> it only causes some slight performance drop
<oliv3r> ssvb pointed out performance was lower
<oliv3r> so we went hunting :)
<lkcl> oliv3r: well, performance drop is better than shit-for-brains segfaults!
<oliv3r> hmm
<lkcl> ... is this kernel or u-boot, btw?
<oliv3r> u-boot
<lkcl> because there's two sets of patches i did
<lkcl> ahh u-boot.
<lkcl> i guessed, it work.
<lkcl> ed
<oliv3r> well that's what nand-u-boot shows; so that's good
<oliv3r> but it doesn't match the code drop
<lkcl> oh those: i got them from the allwinner gpl code release
<oliv3r> they don't match though
<lkcl> i have nooooo idea what they even do :)
<oliv3r> 0x1c012c0 and 0x1c012cc are different between code drop and 'boot from nand'
<lkcl> apologies: i thought you were talking earlier about the bug in the usb 3.3 a20 linux kernel driver
<oliv3r> sets up the 'host port controller'
<oliv3r> lkcl: now that you are here
<lkcl> yep... i'm clueless. copied it verbatim from a memory-dump and/or wits-tech / allwinner gpl source code
<oliv3r> lkcl: gpl drop for a20 u-boot is not complete, some files are not properly copywritten
<lkcl> i don't do "understanding" - i just do "working"
<oliv3r> memory dump, must be
<lkcl> oliv3r: awesome
rellla has quit [Remote host closed the connection]
<oliv3r> lkcl: also, it's only a20 sources, they deleted the a10, a13 branches and most importantly a31 branch
<lkcl> copywritten, you mean "not properly marked for copyright"?
<oliv3r> ironically, the leaked a31 from march, HAS proper copyright on most files
<lkcl> oliv3r: ah
<lkcl> *smack*
<oliv3r> lkcl: yeah, some files lack gpl headers
<lkcl> ok. send me a complete list.
<oliv3r> prime example
<oliv3r> i'll compile a list of a20 boot0 files that lack proper copyright
<oliv3r> the a10, a13 and a31 branches are missing entirely
<lkcl> great.
<lkcl> and the rest would be good as well.
<oliv3r> a31 we have 'leaked', a10 would be helpfull (but should be identical as u-boot-spl that we have now for dram-init)
<oliv3r> i'll sort it tomorrow
<oliv3r> so expect an e-mail tomorrow
<lkcl> ack.
<oliv3r> actually
<oliv3r> it'll be a really long list
<lkcl> ok enough for me tonight.
<oliv3r> first 6 files in the root are allready wrongly copyright
<oliv3r> ok sleep well
Guest86750 is now known as zumbi
<oliv3r> lkcl: not gonna mail you wiht a list, it's almost everything that is NOT copyrighted properly, only a very small bit that's actually right
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
paulk-desktop has quit [Quit: Ex-Chat]
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
derethor has joined #linux-sunxi
naobsd has quit [Quit: Page closed]