Turl changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi | #blameOliver
Gerwin_J_ has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 264 seconds]
Gerwin_J_ is now known as Gerwin_J
arokux2 has quit [Remote host closed the connection]
<Skaag> oliv3r, still here?
<Skaag> am I supposed to have /sys/devices/system/cpu/cpu0/cpufreq ?
<Skaag> because I don't...
<Skaag> trying to follow this guide: http://linux-sunxi.org/Cpufreq#Tweaks
<Skaag> I'm on A13 Olinuxino
<Skaag> ok when I edit the [usbc0] section in my .fex file and modify usb_port_type from 2 to 0 the load on my A13 suddenly drops from 1.0 to normal loads.
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
geecko has quit [Quit: Quitte]
<wens> arokux: nice!
Nyuutwo has joined #linux-sunxi
pseudomind has quit [Read error: Connection reset by peer]
guizamboni has quit [Quit: Leaving]
drachensun has quit [Remote host closed the connection]
egbert has quit [Disconnected by services]
egbert_ has joined #linux-sunxi
pfdm has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
eebrah_ has quit [Ping timeout: 246 seconds]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
deasy has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
<deasy> arokux, ! are you here ?
<deasy> i would try to improve dolphin on kde and i need your help for do it
Sonicadvance1 has quit [Ping timeout: 248 seconds]
Sonicadvance1 has joined #linux-sunxi
Soru_ has joined #linux-sunxi
Soru__ has quit [Ping timeout: 252 seconds]
Soru_ has quit [Read error: Connection reset by peer]
Soru_ has joined #linux-sunxi
Soru has joined #linux-sunxi
pfdm has quit [Ping timeout: 250 seconds]
Soru_ has quit [Ping timeout: 260 seconds]
tzafrir has quit [Ping timeout: 268 seconds]
Soru___ has joined #linux-sunxi
Soru has quit [Ping timeout: 252 seconds]
Soru has joined #linux-sunxi
Soru___ has quit [Ping timeout: 252 seconds]
Soru has quit [Ping timeout: 252 seconds]
Soru_ has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
Soru___ has joined #linux-sunxi
Soru_ has quit [Read error: Connection reset by peer]
[7] has quit [Disconnected by services]
Soru has joined #linux-sunxi
TheSeven has joined #linux-sunxi
Soru___ has quit [Ping timeout: 272 seconds]
<WarheadsSE> I think he's asleep at this time
<WarheadsSE> deasy: ^ -- but I'm headed that way too.
bsdfox\ has joined #linux-sunxi
bsdfox\ has quit [Changing host]
eebrah_ has joined #linux-sunxi
<deasy> WarheadsSE, yum what means i'm headed ?
<apo> hm
<apo> can the uboot on nand storage be configured to use a kernel on a SATA disk?
<apo> (cubietruck)
eebrah_ has quit [Ping timeout: 246 seconds]
eebrah_ has joined #linux-sunxi
n01 has joined #linux-sunxi
jeremb_ has quit [Quit: Connection closed for inactivity]
soul has joined #linux-sunxi
soul is now known as Guest66040
Soru has quit [Ping timeout: 260 seconds]
n01 has quit [Ping timeout: 240 seconds]
Soru_ has joined #linux-sunxi
Guest66040 has quit [Ping timeout: 260 seconds]
Soru_ has quit [Ping timeout: 246 seconds]
wolfy has joined #linux-sunxi
<libv> does anyone remember where there is a warning about including information on our wiki which is not appropriate, or owned by the person including that information?
<libv> ah, at the bottom of any edit page
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
HeHoPMaJIeH has quit [Changing host]
iamfrankenstein has joined #linux-sunxi
n01 has joined #linux-sunxi
apo has quit [Read error: Connection reset by peer]
apo has joined #linux-sunxi
BluesBoy has joined #linux-sunxi
Quarx has joined #linux-sunxi
apo has quit [Read error: Connection reset by peer]
apo has joined #linux-sunxi
<oliv3r> apo: nope
<apo> oliv3r: aw.
<hramrach> arokux: iirc the status of the 2GB patch for u-boot is .. waiting for upstream to figure out how they are going to do it
rellla has joined #linux-sunxi
tomboy65 has joined #linux-sunxi
vicenteH has quit [Ping timeout: 244 seconds]
eebrah_ has quit [Ping timeout: 246 seconds]
eebrah_ has joined #linux-sunxi
maz_ has joined #linux-sunxi
fluo75|2 has quit [Ping timeout: 265 seconds]
popolon has joined #linux-sunxi
<libv> oh man, which sane human being can work his way through firststeps today.
<libv> that
<libv> 's just sickening
apo_ has joined #linux-sunxi
eebrah_ has quit [Ping timeout: 244 seconds]
apo has quit [Remote host closed the connection]
<plaes> the BSP section in Device Page should mention what it is
<libv> plaes: do we even have a useful bsp howto?
<plaes> I only think that it means "Board Specific Packages" ??
<libv> the state of FirstSteps, knowing where it started out, just makes me despair.
<libv> what is the fucking point of this wiki.
<plaes> nah.. it's almost fine.. I got things running quite fast
<plaes> now need to figure out what's wrong with wifi
<libv> really?
<libv> i cannot make heads nor tails from the firststeps howto anymore
<libv> sure, the naming today is all wrong, but it was perfectly apt when i wrote it more than a year ago, as it was the only concise bit of info out there
_massi_ has joined #linux-sunxi
<libv> now instead of someone going in and properly fixing it, it's just gotten horribly maimed
<plaes> how does sunxi-boards handle identical .fex files?
<plaes> or should I just send gemei_g9.fex and add note somewhere?
<libv> plaes: it doesn
<libv> 't atm
<libv> (heh, i finally have gotten round to getting a proper us keyboard, and i am still struggling, since i haven't used one for about 2 years)
<libv> plaes: is there a zatab .fex?
<plaes> yes
<libv> plaes: or was it the uboot support for it that was missing?
<plaes> u-boot support was missing
<libv> plaes: ok
Soru has joined #linux-sunxi
<libv> afaik, it is best to just send in patches for sunxi-boards
<plaes> hrm.. it should be s/board/tablet
eebrah_ has joined #linux-sunxi
<libv> also, could you spend a bit of time extending steps 5,6,7 of new board howto?
<libv> since you just worked through that
<libv> sunxi-tablets?
<plaes> nope, the commit message :)
<libv> imho it should be sunxi-devices, but it's close enough that one shouldn't bother
<libv> plaes: close enough for a commit message :)
notmart has joined #linux-sunxi
<libv> plaes: also, but i am not sure, git emails to the ml might be preferred
<plaes> ok
<libv> i really do not know though, mnemoc should have the overview there.
boycottg00gle has joined #linux-sunxi
<oliv3r> libv: working on a .de keyboard? i make many typing mistakes on those too
<libv> i was unable to get a .us keyboard for my hp probook, at least not at a reasonable price
<oliv3r> ebay?
<libv> now they are cheaply available, either as chinese rip-offs, or as second hand ones
arokux2 has joined #linux-sunxi
<libv> but that was not the case 2 years ago
<oliv3r> i imagine
<arokux2> hramrach: ping
<libv> the laptop i had before was with a us keyboard, so i went through some pain there as well
* torbenh3 wonders whats the point with an US keyboard.
<libv> torbenh3: write some C code with a german, french or belgian keyboard and you'll know
<oliv3r> torbenh3: hat keyboard do you use
<torbenh3> i mean, when i switch to german layout, for some reason, i would be pretty lost on that.
<n01> switch to dvorak :)
<n01> no more pain in the ass
<torbenh3> oliv3r: using a german keyboard with us layout.
<oliv3r> i actually have a spare keyboard with dvorak layout (keys switched) but have to start using it
<oliv3r> torbenh3: pff, so you are using US layout too then
<oliv3r> it's nice to have the keycaps match the actual keys
<libv> that's one thing, but a .us variation is wholly incompatible with .de
<torbenh3> oliv3r: of course. i just dont understand why the keycaps should match :)
<oliv3r> sometimes it makes it easier :p
<oliv3r> beside,s my OCD would kick in ;)
maz_ has quit [Ping timeout: 268 seconds]
<arokux2> libv: "Please use the wrongly named Submitting_Boards howto " <--- rename it
<libv> the plan with the replacement keyboard for the probook is that you keep on using the bit of inter-key plastic, and just transplant that to the new keyboard
<libv> arokux2: i was going to, just pulled u-boot to cull some stupidity from one of the referencers
<libv> arokux2: this wiki is in such a shitty state that i cannot try to fix one thing without having to fix pretty much everything
<arokux2> libv: as I started I've fixed also lots of things :)
<arokux2> libv: now I just "live with it"
<libv> on the probook keyboard: i now am typing on a pretty bare .us keyboard, as the interkey plastic is wholly incompatible with the new layout
<oliv3r> i haven't touch the wiki mostly for about 6-12 month :(
<oliv3r> libv: what's an 'interrkey'?
<arokux2> oliv3r: please review GMAC
<libv> arokux2: and in the mean time, every idiot who cannot think structuredly has had a go at adding his own little not-as-useful bit of info
<mnemoc> arokux2: pong
<arokux2> mnemoc: please review GMAC
<oliv3r> arokux2: your cherry pick from benn? it needs better commit message :p
<libv> oliv3r: these new crappy apple styled keyboards have keys which only stick out out 2 mm from the surface of the keyboard
<oliv3r> arokux2: i'm not sure what there is to review, it's a simple cherry-pick; instead of a seperated patch, a hash sent to mnemoc with an ack that it works would have been fine
<mnemoc> arokux2: do you have that on a branch? for lazy pulling :p
<arokux2> oliv3r: please post all your comments as reply on the ML
<libv> they do however have 4-5mm of travel, and hp designed this thing so that you have a plastic raster around the keys
<oliv3r> libv: ugh; hate those flat keyboards
<oliv3r> arokux2: that's it, no comment, other then 'i think you did it worng' :)
<libv> oliv3r: even lenovo switched to it today
<oliv3r> libv: i don't see what's the point
<oliv3r> mnemoc: its extracted from benn's lichee-3.4 tree afaik; so should be cherry pickable
<arokux2> oliv3r: I think this should be splitted and maybe changed
<libv> oliv3r: we do not see the point, but we are apparently not relevant
<wens> arokux2: GMAC working for you?
<libv> even though we will soon end up being just about the only folk who will buy laptops
<oliv3r> arokux2: nah, first you cherry pick it from the drop; then change to make it work if needed, little changes is good, since it makes diffing easier later
<oliv3r> libv: i like keyboards with a heavier stroke; makes the muscles work out ;)
<torbenh3> pfff... be gentle to your computer :P
<oliv3r> libv: athome, I actually prefer my big box; i use my laptop only in bed/sofa
<mnemoc> <3 das keyboard
<arokux2> oliv3r: have you read the cover letter?
<oliv3r> then again, my laptop is a 2004 T42 thinkpad ;)
<oliv3r> arokux2: there was none
<arokux2> oliv3r: sure?
<oliv3r> arokux2: well not on my inbox, it just said 'from benn'
<torbenh3> wens: did you get somewhere with the stmmac thingy ?
<libv> oliv3r: i prefer to not have to switch environments at all, which is especially helpful when you are travelling to customers and such
<wens> torbenh3: well, i think the issue is to enable GMAC clock in the CCU. playing with it now.
<oliv3r> arokux2: anyway, if it is the code from lichee-3.4, it should be copied 'as is' initially for traceability/diffability, and only then fixes ontop of that
<oliv3r> libv: i'm not in that luxury position ;)
<oliv3r> libv: if I where though, I'd probably have a dockingstation at home and have the laptop closed
<torbenh3> wens: mmm... iirc you said, it doesnt enumerate the phy.
<n01> mnemoc: I bought a wasd v2 (cherry mx clear) ... honestly? better than das
<arokux2> oliv3r: please read the cover letter.
<torbenh3> wens: maybe there is some phy clock. normally the gmac clock would yield unaccessible registers of the gmac.
<mnemoc> n01: didn't know it... looks similar, how's the diff?
<mnemoc> arokux2: thanks!
<torbenh3> wens: at least on omaps its always like that.
<n01> mnemoc: I love the detachable cable and the dip switch
<oliv3r> arokux2: i don't have a coverletter in my mailbox
<n01> and I'm in love with mx clears
<arokux2> oliv3r: i'm not responsible for your mail box, please take a look at gmane or linux-sunxi google groups.
<oliv3r> just 2 mails (patches) with a commit message 'from benn'
<oliv3r> and those commit messages are inadequate as it is
<oliv3r> ok i seet he cover letter; nack
<arokux2> oliv3r: so now you can read it and clearly see what was meant by this submission.
<torbenh3> arokux2: maybe you should also mention, that its pretty similar to stmmac. this might help other folks mainlining it.
<wens> torbenh3: the setup code in gmac touches a few registers in the CCU, otherwise everything is the same as stmmac
<arokux2> guys, please all comments on the mailing list. I will take care of each and submit v2.
<mnemoc> n01: i'll keep it in mind if I need to buy another keyboard :)
<oliv3r> torbenh3: you suggesting they use the same IP; that sounds pretty awesome
<arokux2> wens torbenh3 mnemoc oliv3r: guys, please all comments on the mailing list. I will take care of each and submit v2.
<torbenh3> arokux2: i am not on the ML yet :S
<wens> torbenh3: so far without it, i can't detect any phys
<torbenh3> oliv3r: register layout looks the same. stmmac has a few registers not mentioned in the gmac driver.
<arokux2> oliv3r: I do not need your NACK, I have given the NACK to it myself, I need your comments.
<torbenh3> oliv3r: lets hope they work. we get VLAN and PTP timestamping then :D
<n01> mnemoc: I was thinking to make an ergodox
vicenteH has joined #linux-sunxi
<mnemoc> n01: that's too fancy for me
<mnemoc> n01: I need a keyboard looking and feeling like a keyboard
<oliv3r> mnemoc: DAS keyboard!
<mnemoc> oliv3r: that I have, but n01 pointed me to http://www.wasdkeyboards.com/index.php/wasd-v2-87-key-custom-mechanical-keyboard.html and he says it feels better
<n01> I prefer cherry mx clear to blue
<n01> (I have both)
<mnemoc> n01: what's the dip switch for?
<n01> mnemoc: I use it to remap caps lock to ctrl and to have dvorak layout in hw
<mnemoc> :o
<oliv3r> i should really get a new 'awesome' keyboard
<n01> oliv3r: one day I'll take a hhkb
<oliv3r> hhkb?
<arokux2> mnemoc: plz give GMAC some priority, as network is something essential, if it is absent users will rush to cubietech kernels :(
<n01> happy hacking keyboard
<arokux2> mnemoc: and then to us to support them
<oliv3r> i hacked it though
<oliv3r> added EL glow wire behind the keys (that i never use) and have red numlock etc leds :)
<mnemoc> arokux2: I am using my CT now, so it surely has prio :p
<n01> cherry blue?
<oliv3r> n01: nah, cherry rubber keys, non mechanical ' cheap' keyboard, but with heavy stroke
<n01> yeah, I have a sort of fetish for keyboards :)
mkutsevol has left #linux-sunxi [#linux-sunxi]
<mnemoc> n01: how much was the shipping of the wasd to .it?
<n01> _a lot_
<mnemoc> :<
<mnemoc> can't find an EU distributor
<n01> a think I paid 80e in shipping and customs
<mnemoc> :/
<mnemoc> so 200E keyboard
<n01> yes
<oliv3r> arokux2: i know why i'm getting only tyhe patches, and not the cover letter; somewhere your mailer is setup wrong; X-Amavis-Alert: BAD HEADER SECTIon, Duplicate header field: "Referneces"
<oliv3r> the patches have bad headers, the cover-letter is fine
<rellla> arokux2: but gmac won't work until u-boot driver is also present, right?
<arokux2> oliv3r: i'm using git send-email...
<arokux2> rellla: no, gmac works as is.
<arokux2> rellla: read my cover letter:p
<oliv3r> it's somehow messing up your references, ironically i don't see all patches on linux-sunxi as gmail probably delays/ignores them aswell due to the bad-headers
<arokux2> oliv3r: I see everything in the web interface of my gmail account.
<oliv3r> n01: omg wasd v2 doesn't have numpad!
<oliv3r> arokux2: they are your mails! :p
<n01> oliv3r: you can choose the 105 version
<oliv3r> arokux2: i didn't get the sun7i_defconfig patch on the ML for example
<arokux2> oliv3r: I seen them in the inbox, since I've sent them to myself
<oliv3r> custom keyboard designer :)
<oliv3r> i pieced it togeter now :)
<arokux2> oliv3r: (palm) you've missed another e-mail
Soru has quit [Write error: Connection reset by peer]
<oliv3r> arokux2: yeah, i really have to finish setting up postfix and mailman, google is just horrible
<oliv3r> arokux2: i see the 'ignore this mail' now :)
<wens> still no phy :(
Soru has joined #linux-sunxi
<oliv3r> n01: oh backlit keys :D
<oliv3r> oh wasd keyboards are PS/2!
<oliv3r> good
<oliv3r> i want
<arokux2> wens: where?
<oliv3r> and i can swtich keycaps to dvorak (this sometimes wont' work due to keycaps being slightly different, very annoying)
<boycottg00gle> will gmac work on olinuxino a20 micro?
<n01> oliv3r: no, the wasd code has backlight
<oliv3r> wtf, colemak keyboard layout
<oliv3r> n01: the 'code' keyboard i ment has backlit keys
<oliv3r> i like
<wens> arokux2: with stmmac
<wens> arokux2: phy probing returns all zeros (and stmmac_mdio doesn't think there's a problem?)
<arokux2> wens: 21:45 <arokux2> sun7i# mii info 1
<arokux2> 21:45 <arokux2> PHY 0x01: OUI = 0x0732, Model = 0x11, Rev = 0x05, 100baseT, FDX
<arokux2> this is my result in U-Boot
<wens> arokux2: do you have a u-boot tree i can use?
<arokux2> wens: I can push that awful code........
<wens> arokux2: well, working code is good enough for me
notmart_ has joined #linux-sunxi
<wens> i can't even get passed initialization
FunkyPenguin has quit [Read error: Connection reset by peer]
<arokux2> wens: sun7i# mii info 1 <--- worked.
<mnemoc> arokux2: want a nightly of that?
<arokux2> wens: but soft reset won't work. -- it will loop forever
<arokux2> mnemoc: not yet.
<arokux2> mnemoc: it does not work yet.. only phy info can be dumped.
notmart has quit [Remote host closed the connection]
<mnemoc> arokux2: i see
FunkyPenguin has joined #linux-sunxi
<oliv3r> here i go thinking that dvorak is the all and awesome new layout, but here comes colemak, even more awesome!
<arokux2> keyboards are flawed essentially.
<n01> oliv3r: colemak is usefull if you want to stick to qwerty somehow
<oliv3r> appearantly there's even less movement then dvorak
<oliv3r> and being compatible with qwerty is understandable
dapsaille has joined #linux-sunxi
<n01> colemak is born to make the switch from qwerty less traumatic
<n01> because swithing to dvorak _is_ traumatic
<oliv3r> mx clear, those new switches?
<n01> I'm using clear right now
<n01> I love the stiffness of the keys
<oliv3r> on codekeyboards.com it says 'ultra-rare cherry mx clear mechanical keyswtiches'
<oliv3r> wanted to see if that's just marketing at work
<n01> yes, clear are rare
<oliv3r> i don't mind the traumatic swtich from dvorak; but if colemak is slightly more efficient to boot; why not
<n01> your call :)
forcev has joined #linux-sunxi
<oliv3r> n01: have you tried both?
FunkyPenguin has quit [Quit: No Ping reply in 180 seconds.]
<oliv3r> i just want the 'most efficient layout' whether that's dvorak or colemak doens't matter to me, but right now, statstics at carplax (sp) say colemak is slightly better
<oliv3r> ok ONE little thing they could have done better on a 180 USD keyboard http://www.codekeyboards.com/ here is
<oliv3r> RGB leds over white :)
<wens> arokux2: sun7i# mii info
<wens> PHY 0x01: OUI = 0x0020, Model = 0x20, Rev = 0x01, 100baseT, FDX
<n01> oliv3r: you mean dvorak and colemak? nop, just dvorak
<oliv3r> n01: bummer, would have been an interesting question; so now i'm debating coleemak or dvorak
<n01> :D
<arokux2> wens: yep. that is what I've seen too.
<wens> arokux2: I can get the phy, connections is properly negotiated, but dhcp fails
<wens> arokux2: it doesn't seem to loop though?
<arokux2> wens: exactly, it hangs, right?
<wens> arokux2: no, it just retrys a couple times, and falls back to nand
<arokux2> wens: hm.. was different at my place, will take a look at it again tonight, thanks for the input!
Black_Horseman has quit [Remote host closed the connection]
<arokux2> wens: now you can see what is missing in the kernel so that phy is visible.
<arokux2> wens: wait, have you seen something like "eth0" in the u-boot command prompt?
<wens> arokux2: eth0 Speed: 100Mbps, Mode: Full duplex, Loopback: NO
<wens> arokux2: I'm running on a cubieboard2, not a cubietruck.
<arokux2> wens: ah!!
<arokux2> wens: me on cubietruck.
<wens> arokux2: I did add some pinmuxs, the ones not used with RGMII
<arokux2> wens: to my u-boot branch?
<oliv3r> not sure if i like the optimus too much
apo_ has quit [Read error: Connection reset by peer]
apo has joined #linux-sunxi
<arokux2> oliv3r: you'd buy it on russian web page, it is half the price there :)
<oliv3r> wens: the gmac driver should work just fine on the cubieboard with the 100Mbit PHY
<oliv3r> arokux2: i'm not sure if i would ike it to type on
<arokux2> oliv3r: scrooll down there and switch between: Arabian Photoshop Russian color English-Russian Russian English Hebrew
<wens> arokux2: the GNULL pins are still needed when running MII
<wens> oliv3r: I suppose it should
<arokux2> wens: can you publish your patches somewhere, so I can test if they have something to do with CT?
<arokux2> mnemoc: I do not understand your comment :(
<mnemoc> arokux2: default m in Kconfig
<mnemoc> with proper depends on
<arokux2> mnemoc: ah, ok.
<arokux2> mnemoc: will take a look
apo has quit [Read error: Connection reset by peer]
apo has joined #linux-sunxi
<oliv3r> wens: from what I saw the pins are fully compatible, depending on which mode they are in
paulk-collins has joined #linux-sunxi
<oliv3r> wens: the different here is if the IP is in MII or RMII mode
<oliv3r> arokux2: nice, people are allready sending in patches
<wens> arokux2: https://github.com/wens/u-boot-sunxi # sunxi-gmac-cb2
<wens> arokux2: just 1 line different
<arokux2> thanks wens
arokux2 has quit [Remote host closed the connection]
<wens> oliv3r: maybe i'll try gmac on mainline first, hopefully wont be too much work
enrico_ has joined #linux-sunxi
<oliv3r> i'm curious abou tthis stmmac thing :)
<oliv3r> oh it's deleted in later kernels?
<mnemoc> sunxi_gmac, sunxi_emac
<wens> oliv3r: renamed to net/ethernet/stmicro/stmmac/stmmac_*
<oliv3r> wens: thanks
<mnemoc> is it similar? registers wise
<torbenh3> oliv3r: thats the relevant header file with the registers.
<torbenh3> mnemoc: looks identical from the registers.
<mnemoc> :o
<oliv3r> if that's the case, that should be the primary angle of attack :)
<oliv3r> do we have stmicro in uboot?
<torbenh3> i didnt see it.
<wens> oliv3r: there is designware.[hc]
<oliv3r> but is designware in the stmicro?
<oliv3r> time to investiage; and i realyl don't hav etime :D
<wens> oliv3r: it's the same thing
<oliv3r> so stmicro uses a designware gmac
<oliv3r> sounds very plausible
<oliv3r> since AW uses more designware IP's
<wens> oliv3r: correct. stmmac actually has 2 supported IPs
<oliv3r> i saw dwmac1000.h and swmac100.h
<oliv3r> d*
<mnemoc> emac/gmac ? :p
<oliv3r> note the name 'dw mac', designware mac? :)
<wens> we can move the pinmux and clock setting stuff into the board init part, and test designware?
<oliv3r> http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/net/designware.h;h=e80002a0e4409fad8f46c7beef79f12143918a17;hb=HEAD#l3
<oliv3r> also a clear giveaway
<wens> mnemoc: didn't compare emac
<oliv3r> what's even more interesting, is if our current 100mbit EMAC, is also designware
<torbenh3> indeed. disignware.h looks identical to stmmac
<oliv3r> torbenh3: it does so indeed
<wens> emac registers aren't the same
<oliv3r> wens: maybe shifted a litte around
<oliv3r> well
<oliv3r> designware.h from u-boot only seems to do the dwmac1000.h
<oliv3r> so i don't think we have an easy emac replacement thing, but the DW1000 looks very promessing
<wens> emac control flags don't even look close
<oliv3r> dwmac100.h vs sunxi.h?
<wens> oliv3r: yeah
<oliv3r> diff IP then
<arokux> wens: it can mean lots of work, what is your problem now?
<oliv3r> though I think in the past it was remminiscent that it's very likly to the davicom IP
<torbenh3> http://git.denx.de/?p=u-boot.git;a=blob;f=drivers/net/sunxi_wemac.c;h=699a3815824956ba761a7301d2f838d183de240f;hb=HEAD but the emac is this....
<oliv3r> that's really old, since it was renamed for a while to sunxi_emac.c
<oliv3r> didn't know we actually had stuff upstream, in any case, it is linked somewhat to davicom IP iirc
<mnemoc> allwinner -> davicom -> dw ?
<rm> too bad Reummilla never went anywhere
<rm> or what was the name :p
<rm> was such a promising h/w company
<arokux> designware.h looks indeed similar to gmac
<oliv3r> i'd say almost identical
<arokux> funny, AW could have taken and obfuscated it, but they invented their own wheel.
<oliv3r> as always
<torbenh3> arokux: nah. the driver structure looks pretty similar do stmmac.
<oliv3r> this way it allows them to claim copywrite
<arokux> torbenh3: I'm talking about u-boot
<torbenh3> ah.
soul has joined #linux-sunxi
<arokux> oliv3r: they have it GPLed so...
<torbenh3> well... the gmac driver in the cubie kernel is pretty similat.
<oliv3r> http://git.rhombus-tech.net/?p=u-boot.git;a=blob;f=drivers/net/sun6i_gmac.h;h=d9280df31c24d9adf2e4a2e0e008e8ef6ca88be3;hb=refs/heads/allwinner-sunxi-a31#l32 a nother peak into internal names :)
<oliv3r> arokux: that may not have been their intention
soul is now known as Guest70477
Soru has quit [Ping timeout: 244 seconds]
<oliv3r> #define GMAC_RGMII_STATUS (0xD8) /* S/R-GMII status */ is missing from dmac
boycottg00gle has quit [Remote host closed the connection]
<arokux> by STMicroelectronics
<arokux> in those boards is that dw
egbert_ is now known as egbert
paulk-collins has quit [Quit: Ex-Chat]
<mnemoc> tom actually confirmed sunxi-emac is davicom based long ago
<torbenh3> wens: with which kernel are you messing around ?
eebrah_ has quit [Ping timeout: 240 seconds]
<ganbold_> mnemoc: I think they used dm9000 as a template
<ganbold_> mnemoc: it is different
<torbenh3> dm9000 is very different.
<torbenh3> at least how you access registers, and pull out the data.
<torbenh3> maybe behind that stuff its similar.
<oliv3r> sun6i header seems more complete, more bits defined
<torbenh3> although its been over a year, since i had a dm9000, and i dont remember the details :)
Guest70477 has quit [Ping timeout: 246 seconds]
<oliv3r> sun6i.h is far superiour in details
<arokux> oliv3r: where do we get the datasheet for designware GMAC?
<arokux> oliv3r: ?
panda84kde has joined #linux-sunxi
<arokux> will design ware give us the datasheet if we ask them?
<oliv3r> try :D
<oliv3r> but not really needed since we have what we need
<oliv3r> and i strongly doubt they would
jinzo has joined #linux-sunxi
<arokux> ok, I want to write it from sunxi.org, so I'd need to consult all you
<n01> burocracy
eebrah_ has joined #linux-sunxi
<arokux> datasheet is always nice it helps understand hardware and resolve errors if any.
<torbenh3> arokux: thats the spear trm. contains gmac stuff.
<arokux> JohnDoe_71Rus: git grep rt2800usb.o
<n01> JohnDoe_71Rus: look in the makefile
<wens> arokux: stmmac is not seeing any phys
<oliv3r> JohnDoe_71Rus: RAlink 28x0 iirc
FR^2 has joined #linux-sunxi
<arokux> wens: mainline?
<wens> torbenh3: sunxi-next
<arokux> torbenh3: 50 pages of GMAC, nice
<wens> arokux: i remember seeing one on scribd
<arokux> wens: pass clk_disable_unused or so
<arokux> wens: just to make sure no unused clocks are turned off
<arokux> wens: not much should be set up, till you can see phy, right? in u-boot only pinmux and GMAC gatable clock.
<arokux> anybody with account on scribd?
<JohnDoe_71Rus> oliv3r: find CONFIG_RT2x00. just add CONFIG_RT2x00=m to build module?
<wens> arokux: correct
<oliv3r> a bit more needed john
<arokux> JohnDoe_71Rus: this could work, but this is not recommended
<arokux> JohnDoe_71Rus: better find it in menuconfig
pacopad has joined #linux-sunxi
<wens> arokux: if the clock isn't enabled, mmios to gmac shouldn't respond, right?
<oliv3r> arokux: it won't work that way ;)
<arokux> oliv3r: why?
<oliv3r> arokux: missing dependancie
<JohnDoe_71Rus> arokux: menuconfig or make menuconfig?
<arokux> wens: yes, it was the cas in u-boot I think
<arokux> JohnDoe_71Rus: second obviously
<hramrach> pong
<arokux> oliv3r: he needs to figure deps too :p
<oliv3r> make menucofngi does that for you, andt he module needs several options set
<arokux> hramrach: why have you deleted bcm status from Cubietruck pages?
<hramrach> arokux: because it's supposed to work. iirc it worked with the shipped android image
Soru has joined #linux-sunxi
<arokux> oliv3r: what is you recommended way to do it? I always press "/" in menuconfig and take a look at deps and enable them.
<oliv3r> if we really have DW Gmac; then wow, 16k Jumbo frames etc
<arokux> hramrach: hm... it is status of the sunxi-3.4, not preloaded android or any other kernel.
<hramrach> arokux: then nothing works
<oliv3r> arokux: dependancies should be enabled when you select your driver in the menu
<hramrach> but we have the cubie repo
<arokux> hramrach: so please revert your change :)
<JohnDoe_71Rus> arokux: some not work http://pastebin.com/t1rxGn3D
<hramrach> which has different and supposedly working driver
<arokux> JohnDoe_71Rus: we here support only sunxi-3.4, sorry
derethor has joined #linux-sunxi
<arokux> hramrach: alright, add this as a note then.
derethor_ has joined #linux-sunxi
derethor has quit [Read error: Connection reset by peer]
<arokux> oliv3r: if deps are unmet stuff will not appear in menuconfig (maybe not always?)
<oliv3r> JohnDoe_71Rus: for help with the lichee-3.3 kernel you have to bark up to allwinner ;)
derethor_ has joined #linux-sunxi
<oliv3r> arokux: on badly written KConfig's yeah, well depends
<oliv3r> arokux: if you disable PCI, you won't see any devices that connect to the PCI bus
<arokux> who has an account to download this? ----> http://www.scribd.com/doc/80253178/Ethernet-Eval-Databook
<wens> arokux: for some hard dependencies, like MII, you might want to use "select"?
<arokux> wens: ok (we were talking about configs in general)
<JohnDoe_71Rus> oliv3r: arokux: :( i just whant build module to wifi
<wens> arokux: i see
<arokux> JohnDoe_71Rus: I see, but I cannot help you, since I only work with sunxi-3.4 and there everything just works.
<arokux> JohnDoe_71Rus: so you might want considering switching to it. in case of problems we will happily assist you.
<arokux> The software and documentation are furnished under a license agreement and may be used orcopied only in accordance with the terms of the license agreement. No part of the software and documentation may be reproduced,transmitted, or translated, in any form or by any means, electronic, mechanical, manual, optical, or otherwise, without prior writtenpermission of Synopsys, Inc., or as expressly provided by the license agreement.
<oliv3r> JohnDoe_71Rus: make menucofnig, net -> wireless -> ralink -> rt28x0 -> USB
<JohnDoe_71Rus> arokux: but 3.4 does not support cb2 android yet
<oliv3r> JohnDoe_71Rus: why not?
<hramrach> arokux: that's why we have kernel links there. because the sunxi kernel does not work
<JohnDoe_71Rus> oliv3r: see pastebin
<arokux> hramrach: but the title of the section reads: "Status of the community kernel (sunxi-3.4) / U-Boot"
<arokux> hramrach: I have update it a bit: http://linux-sunxi.org/A20-Cubietruck#Status_of_the_community_kernel_.28sunxi-3.4.29_.2F_U-Boot
<oliv3r> arokux: have you tried bugmenot.com?
<oliv3r> JohnDoe_71Rus: how is that a 3.4 + android issue?
<arokux> JohnDoe_71Rus: in your pastebin you show us lichee..
<arokux> oliv3r: I've registered, but download costs $
<arokux> oliv3r: $9 to be exact
<JohnDoe_71Rus> What's the difference 3.3 or 3.4? configuration and build should work the same way.
<arokux> JohnDoe_71Rus: please, we do not support ANY lichee kernel.
<JohnDoe_71Rus> ok
<oliv3r> that spear 6000 doc should get linked on the wiki; it has potentially valueable info in it
<arokux> oliv3r: yes
<oliv3r> JohnDoe_71Rus: 3.3 is hacked together by allwinner, based on 3.0 with their crap over it
<oliv3r> JohnDoe_71Rus: 3.4 is our stuff, where we patched and fixed tons and tons of things
<oliv3r> JohnDoe_71Rus: it's not as much as a version difference, but more an alwinner (lichee) vs sunxi kernel, one we work on , the other is an abomination
alabaster has quit [Ping timeout: 245 seconds]
<hramrach> arokux: made the section a bit more verbose so it should be clear enough now
<arokux> hramrach: oh, that looks much better
diego has joined #linux-sunxi
diego is now known as Guest35677
<torbenh3> oliv3r: i am still wondering why you are sticking to 3.4 ... to much work with DT ?
panda84kde has quit [Disconnected by services]
Guest35677 is now known as panda84kde
<hramrach> too much stuff not working in mainline
<hramrach> you can boot mainline and hopefully soon you can use USB stick or SATA s storage - but neither is bootable
<n01> too much effort on sunxi kernel IMHO
<oliv3r> torbenh3: huh? what do you mean?
<hramrach> it's a working kernel. that's something you want on your device when you are actually using it for something :p
<n01> yep but _IMO_ making dirty patches for sunxi kernel will double the work for mainlining
minos__ has joined #linux-sunxi
<oliv3r> n01: maybe, but 3.4 works today, having small fixes to keep it working is little work; getting everything mainlined, will take months/years
<n01> oliv3r: I know, keeping a dirty kernel is easy task, nevertheless the effort could have been redirected to mainline. my 2 cents
pitillo has quit [Changing host]
pitillo has joined #linux-sunxi
<oliv3r> n01: yes, but people wirting the dirty fixes sometimes aren't compentent or botehred to go the mainline route
<n01> oliv3r: apart from mripard and few others nobody here is a kernel hacker
<n01> but, again I really appreciate your work on sunxi-kernel, but I'd like to see more support on mainline. hramrach is right, mainline is far from mature
<kz1> I have a lichee kernel that is 3.4 . Does that mean they are now using sunxi-linux or are they just updating their lichee versions on a semi regular basis?
tzafrir has joined #linux-sunxi
<libv> n01: start coding.
<libv> or stfu
<n01> libv: wdt, rtc and lracd are being mainlined. So I code for mainline in my spare time
<n01> not a great contribution but I do what I can
<arokux> minos__: thanks.
minos__ has quit [Ping timeout: 250 seconds]
<kz1> is the sunxi-linux kernel more power efficient than the lichee kernel?
notmart_ has quit [Quit: notmart terminated!]
<arokux> torbenh3: we would love to use mainline and recommend it to the users, but the point is the mainline cannot have overnight all the features that sunxi-3.4 has. so sunxi.org is working on 1) sunxi-3.4, legacy, bug fixing only and unification 2) mainlinING
<dapsaille> hi, little question, when i specify lcd parameters in fex file, u-boot use them directly and pass them to kernel driver or pass them to kernel directly ?
<arokux> n01: you cannot start mainlining without having the documentation. there is no or very little documentation. so you need to gather the bits from the AW code drops. but then you realize they have different kernel even if driver has slightly changed. so to have the complete picture one should unify. it means sun4i, sun7i etc all get fused to sunxi. only THEN you have a clear picture of the hardware and CAN start mainlining.
<dapsaille> i'm trying to understand the mechanisms :)
<arokux> torbenh3: --------------------------------------------------------------------^
fredy has quit [Excess Flood]
<arokux> dapsaille: u-boot won't take a look into script.bin, just pass that blob to kernel.
<dapsaille> ty
<dapsaille> sorry .. thank you ..
fredy has joined #linux-sunxi
<dapsaille> so, if kernel can handle pixel clock with floating i need to tweak fex source code to pass floating to kernel right ?
<libv> arokux: why does this have to be explained every few minutes?
<dapsaille> because for now it's only integer
<libv> why can't people just start doing useful work?
<arokux> libv: because ppl ask :)
naobsd has joined #linux-sunxi
<torbenh3> arokux: hmm... make some sense ;)
<arokux> libv: imho, this is important to understand and should be even put on the wiki.
tzafrir has quit [Ping timeout: 240 seconds]
<libv> arokux: none of the whiny gits can be bothered with reading such things\
<n01> libv: don't be rude. I try to help as I can, just I'd like to see a working mainline kernel asap since personally I'm not interested in sunxi-kernel
<libv> n01: again, don't waste your time whining
<arokux> n01: it is the same in my case, 100% the same.
<n01> libv: just don't read what I write
<n01> but you are right, I have work to do
<arokux> n01: but today I've submitted patches against sunxi-3.4. you can ask why? answer: to know the code works and start mainlining it.
pitillo has quit [Ping timeout: 260 seconds]
<n01> arokux: good, nothing against it... just my 2 cents
pitillo has joined #linux-sunxi
<arokux> n01: so now if something isn't working in mainline I go to sunxi-3.4 and compare... I do now know what is your workflow.
mouchon has quit [Ping timeout: 248 seconds]
mouchon has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<arokux> hm.. that SPEAr600 had Dual ARM926EJ-S core up to 333 MHz, was it enough to fuel GiB ethernet, or it is not relevant?
<oliv3r> arokux: not relevant; it probaby doesn't
<oliv3r> that doesn't mean the DesignWare can't handle it properly :)
<arokux> oliv3r: "it probaby doesn't"?
<arokux> oliv3r: sorry, I didn't get. my question was: were the other components of SPEAr600 powerful enough to fuel GiB Ethernet.
<arokux> STMicroelectronics seems to be very OSS friendly, we were lucky!
<arokux> only look at it! http://git.stlinux.com/
Black_Horseman has joined #linux-sunxi
<hramrach> oliv3r: maybe send the u-boot 2G patch upstream?
<mnemoc> arokux: cherry-picking directly from cubieboard's repo worked fine here. but will wait for agreement before pushing it to stage
<hramrach> I don't see the patch there
<arokux> oliv3r: does it follow that kernel will fail to see 2GiB too?
<arokux> hramrach: me too, but i thought u were missing that thread
<arokux> mnemoc: +1
<hramrach> it will see 2G if you patch u-boot to aloow for passing that much ram
<oliv3r> hramrach: i have sent it upstream :)
<arokux> hramrach: oh. then we need u-boot fixed!
<arokux> oliv3r: where is that sent patch?
<oliv3r> arokux: it probably isn't power full enough, but that doesn't matter really
<mnemoc> arokux: i hope the script.bin to platfrom_device thing I'm writting will allow us to use non-sunxi drivers in 3.4 when possible as well
<oliv3r> arokux: oh wow, st git!
<arokux> oliv3r: yeah!
<oliv3r> arokux: nah, kernel should work fine if your lucky
<oliv3r> arokux: it's a bug in u-boot so shouldn't affect the kernel mostly; IF you have a properly defined board.c
<arokux> oliv3r: I fail to find the patch on that lengthy thread at u-boot ML
<oliv3r> all we need really, is for hno to give the patch his blessing, and we can merge it into sunxi-u-boot.
<arokux> oliv3r: don't you want to push it directly upstream?
soul has joined #linux-sunxi
<oliv3r> https://github.com/oliv3r/u-boot-sunxi/commits/wip/2GiB last 5 patches from that tree
<oliv3r> arokux: i can't push upstream :)
<oliv3r> don't have commit access
soul is now known as Guest491
<arokux> oliv3r: that is clear, what about posting patches to their ML directly?
Soru has quit [Ping timeout: 248 seconds]
<oliv3r> arokux: very fine point, my patches aren't on the mail-archive link
Black_Horseman has quit [Remote host closed the connection]
<oliv3r> arokux: patchwork knows all: http://patchwork.ozlabs.org/patch/281094/
paulk-collins has joined #linux-sunxi
<arokux> oliv3r: I've updated this with 2GiB issue, please keep it up-to-date: http://linux-sunxi.org/A20-Cubietruck#Status_of_the_community_kernel_.28sunxi-3.4.29_.2F_U-Boot
<arokux> oliv3r: I'm not sure the u-boot devs use it.. just drop the patches on their ML, ping, poke etc them, so they accept :)
<oliv3r> arokux: that is picked up FROM their own mailing list :)
<oliv3r> patchwork scans the ML and adds the patch to itself
<oliv3r> and i asked like 4 times if there's anything needed to be changed etc
<oliv3r> so now i just wait for hno to review it and say 'guys, merge it'
<oliv3r> mnemoc: i would expect gmac for 3.4 to be cherrypicked from cubieboard github; with patches added ontop to 'fix' things
<arokux> oliv3r: I cannot find patches here: http://www.mail-archive.com/u-boot@lists.denx.de/msg123014.html
<arokux> oliv3r: where are they in that thread?
<mnemoc> oliv3r: so you suggest to just push the cherry-picked commit now?
<arokux> oliv3r: and what is the justification for commit noise? (comes into way if bisecting)
<mnemoc> Select the DMA TX/RX descriptor operating modes
<mnemoc> > 1. Enable Descriptor Ring Mode (GMAC_RING) (NEW)
<mnemoc> 2. Enable Descriptor Chained Mode (GMAC_CHAINED) (NEW)
<mnemoc> uh
<mnemoc> .oO(wtf is that?)o
<arokux> mnemoc: this is what should be understood before merging I think.
<arokux> mnemoc: there is another strange option *SYSCONFIG* or so
<mnemoc> sysconfig is the official name of script.bin
<oliv3r> mnemoc: yes
<arokux> mnemoc: but what is the option for...
<oliv3r> mnemoc: and fixes ontop
<oliv3r> mnemoc: for mainline, I suggest to use the stmmmac driver anyway
<mnemoc> arokux: ignore lies on script.bin and go for hardcoded values?
<oliv3r> we shouldn't be picking and poking the standard allwinner stuff, we haven't for all the other drivers either
<arokux> oliv3r: hm.. it makes sense. the commit message still needs to be fixed.
<mnemoc> oliv3r: I'll writting the generic script.bin to platform_device thing to let us use things like stmmac in 3.4 too, checked _used, doing the pinmuxing and creating the devices in the core
<oliv3r> we merge them 'as is' as it's a (partial) code drop, then mish/mash it to get it working (if needed) and fixes ontop
<oliv3r> mnemoc: yeah i'm allready excited
<mnemoc> oliv3r: hope it works...
<arokux> mnemoc: can one cherry pick and change commit message on the fly?
<mnemoc> arokux: it's called amend :)
<mnemoc> git commit --amend
<mnemoc> will let you modify the previous commit, like for adding a decent description
<mnemoc> or subtile fixes
<arokux> mnemoc: this is the amending of something which was commited
<mnemoc> commit --amend applies over the HEAD commit
<oliv3r> arokux: don't invest TOO much time in getting sun6i working on u-boot, our itme is probably better spent at getting sun7i hardware working with the designware driver allready in u-boot
<arokux> mnemoc: right.
<mnemoc> but you "can't" do that to something already pushed
<mnemoc> or people will hate you
<mnemoc> because git push -f will be required
<oliv3r> heh, that's where git pull --force is for! :p
<mnemoc> but depends on the nature of the branch if you "can" do that or not
<arokux> oliv3r: ... I only started, and were you so nice you could find it before that, but you cannot even remember where sun7i gmac is, which you claim you have seen once. so right, I'll see how I get designware.c/h working with CT, obviously.
<oliv3r> arokux: VGA - probably wired differently on this board - the original fex selects LCD and has VGA working with the kernel shipped in NAND what do you mean here?
<oliv3r> arokux: haha, it was sun7i gmac
<oliv3r> if only i could remember which repo it was in
<oliv3r> memory might even be so hazy, that it may not even be u-boot at all, and only regular linux!
<arokux> oliv3r: that was hramrach who added comment on VGA
<oliv3r> i think that's only valid for sun5i olimex micro boards
<oliv3r> sun4i and sun7i have native vga out via the 'tv encoder'
<oliv3r> arokux: i don't think all messages are archived by mail-archive btw
<oliv3r> the patches are attached there
<arokux> oliv3r: since when attaching patches is the way to go?
<arokux> oliv3r: and top posting too :p
<mnemoc> http://sprunge.us/cNgP <--- stupid mali rebuilds itself AFTER the old modules were already installed!
<hramrach> mali rebuilds itself every time
<mnemoc> i knew that... but doing it so AFTER modules_install is even more stupid
<hramrach> oliv3r: that's what I see shipped on the nand. Of course, the fex might be bogus but setting it for VGA does not give working VGA so ..
<oliv3r> arokux: git send-email did that :(
<arokux> oliv3r: agree on "mail-archive is just fucko", I've update that wiki page with ref to gmane.
<arokux> oliv3r: by having attachment you prevent a random person from commenting, but you know it yourself.
<oliv3r> hramrach: that's idiotically strange :) i know olimex configures it as such, but if oyu look at the a20-micro schematic, you see that the vga pins are coming from the tv-encoder pins, as that's an analog output, makes sense
<arokux> oliv3r: maybe you incorporate comments (if any) and start a new thread (this time with inline patches)
<oliv3r> arokux: it's how git send-email works though
Guest491 has quit [Ping timeout: 248 seconds]
<arokux> oliv3r: if the patches are too big?
<oliv3r> actually, you are making me doubt myself
<oliv3r> maybe I did attach the files seperatly, I don't remember to be honest
<arokux> ....
<oliv3r> but yeah if hno doesn't have time to look at it, maybe a resend is needed
<arokux> git send-mail never did any attachment business for me
<oliv3r> you migh tbe right
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
notmart has joined #linux-sunxi
notmart has joined #linux-sunxi
notmart has quit [Changing host]
<oliv3r> it may have been becaues the patches are more then 100k; i don't recall to be honest :)
<oliv3r> but i will do a resend soon
<oliv3r> to be fair, most mail clients properly handle x-patch mime times fine and are 'inlined' by the client
AreaScout has joined #linux-sunxi
<hramrach> I wish that was true
<mnemoc> thunderbird does
eebrah_ has quit [Read error: Connection reset by peer]
<mnemoc> when the attachment has a decent mime type
<hramrach> it cannot do something
<hramrach> save sent email in a specific folder
* mnemoc bcc himself instead of using "sent mail" folder
<hramrach> I don't have folder 'Sent' on imap server so it yells every time I send an email
<mnemoc> in the copies & folders menu of the account you can choose what to do
<mnemoc> place a copy here, or cc or bcc
<hramrach> but here is only "Sent"
<mnemoc> the place here usually has 2 options, use Sent folder or use custom folder
<mnemoc> but I prefer bcc, so I can classify my outgoing mails too
eebrah has joined #linux-sunxi
eebrah is now known as Guest26794
Guest26794 is now known as eebrah_
<hramrach> ok, I guess using bcc works
<hramrach> but it can't delete mail because my trash is named "deleted items" :s
<mnemoc> you can also configure how to delete with delete
<mnemoc> s/delete/deal/
<mnemoc> but that is the server settings part
<mnemoc> "when I delete a message:" move it to here, or mark it as deleted, or remove it immediatelly
<hramrach> yes, that one works at least, thanks
atiti has joined #linux-sunxi
<hramrach> and it can even search, that's sooo cool
<hramrach> something Google cannot do
<mnemoc> search works fine for me too :p
<arokux> n01: I've just realized yet another point. having feature full sunxi-3.4 actually attracts users, which test stuff on real hardware and report bugs to us. users would never test mainline, because it can do nothing.
<arokux> torbenh3: ^
* hramrach tests mainline
<hramrach> but I don't use the board for anything else tbh
<arokux> hramrach: you are hacker, not user :)
<arokux> or very advanced user :)
<mnemoc> arokux: that's why we were trying to create a middle ground sunxi-3.10
<mnemoc> arokux: feature full BUT dts based
<mnemoc> arokux: and with easy translation between .fex to .dts
<mnemoc> also 3.10 is LTS and androidized
<arokux> mnemoc: have you booted sunxi-3.10 on ct?
<hramrach> 3.10 is not that useful. You need mainline so you don't get ancient upstream bugs
<mnemoc> haven't tried yet
<mnemoc> hramrach: 3.10 is LTS, it will get ancient upstream bugs fixed
<hramrach> maybe not so bad idea then
<mnemoc> it's also the chosen version of LTSI
<n01> arokux: IMHO a bit too many steps toward mainline sunxi->3.4->3.10->mainline ... but i have nothing more to say on this argument
eebrah_ has quit [Ping timeout: 245 seconds]
<hramrach> yes, you would have to forward.backward port patches
<mnemoc> n01: sure it's another step, but an step that means normal people will be able to use it, test it, and report issues
<arokux> n01: I'm not arguing for 3.10, indeed I was against it, forgive me mnemoc.
<mnemoc> fine
<mnemoc> arokux: but there are several drivers which will take ages to be mainlinable
<mnemoc> and that will keep real users away of mainline
<arokux> n01: for me it is (allwinner code drops) -> sunxi-3.4 -> mainline, and IF there is a working playground - 3.10 which boots, I'll quickly backport to it.
<hramrach> as it is mainline is nearly usable as router
<arokux> hramrach: why nearly? :)
<hramrach> need some storage. netbooted router is not the way to go
<mnemoc> oliv3r: does 3.10 boot? :p
<hramrach> and need the cpufreq division by 0 resolved
<oliv3r> mnemoc: should work 'fine' for hwat it can do
rz2k has joined #linux-sunxi
boycottg00gle has joined #linux-sunxi
<hramrach> but with ehci and/or sata polished to workable state you have storage
<oliv3r> n01: 3.10 will replace 3.4 asap
<arokux> hramrach: USB EHCI works now (it is not yet added to mainline though)
<hramrach> yes, I get ports reported all right but did not try connecting anything
cajg has quit [Ping timeout: 264 seconds]
<arokux> hramrach: everything works.
<hramrach> sata is more interesting since u-boot has support for that
<arokux> oliv3r: how such claims "3.10 will replace 3.4 asap" with your absence of hacking time?
<mnemoc> asap doesn't mean tomorrow ;-)
<n01> arokux oliv3r in that case I'll be happy to test it
<mnemoc> as soon as *possible* :)
<hramrach> so you could like put a SD card in the board and have it load kernel from SATA drive
<hramrach> imho 3.10 is not useful. something like sunxi-next that has working and testing drivers that are meant for mainline would be nice. Rebased on otp of mainline every time they change something relevant for sunxi
<mnemoc> hramrach: sunxi-next is about already accepted commits
<oliv3r> arokux: many hackers here ;) the idea is to merge ugly 3.4 drivers into 3.10
<hramrach> I had to pick the usb and sata drivers from random repos
<mnemoc> a feature full dts based kernel including the clean drivers and the crap drivers
<oliv3r> hramrach: sata driver is under review on ML; but it's being a little annoying
<mnemoc> so people can really test what's coming next AND be able to use their devices normally
<mnemoc> we have today a legacy script.bin based full featured kernel, and a barely usable but clean code -next
<hramrach> it's getting usable already so we don't need an extra stage
<oliv3r> no its not
<hramrach> btw is there any plan to support disp in u-boot?
<hramrach> GUI menu ftw
<oliv3r> wingrime tried to do some initial work on it
<hramrach> + free graphics support in Linux
<oliv3r> anyhow, mainline won't be usuably for the majority for ages, because of things like disp
<mnemoc> tiny details that whimpful desire of normal device owners of having something on the display
<hramrach> if you get it into u-boot you don't need it in Linux for it to be usable
vicenteH has quit [Ping timeout: 240 seconds]
<hramrach> *and* you quelch whimpful desires for full-featured bootloader at the same time
<oliv3r> ok disp in u-boot will take a year or more
<oliv3r> disp in kernel, may take 1 - 1 1/2 years
<oliv3r> 3.10 could happen in 3 months if a few wwould work on it a little
<mnemoc> backporting sunxi-next to 3.10, and forward port the rest of crap drivers to 3.10, to turn it feature full dts based is much easier/faster than getting a feature-full mainline
<mnemoc> and "deprecate" 3.4
<torbenh3> whats so different with 3.10 and 3.13 ?
<mnemoc> 3.10 is LTS/LTSI/android supported
<hramrach> maybe it would help if the Chinese used 3.10 then. but you can't tell what they do
<mnemoc> 3.13 is not... and they never take odd numbers for stables
<mnemoc> also 3.10 already has mainline sunxi
dack has joined #linux-sunxi
<mnemoc> officially
<hramrach> also before this LTS stuff
<mnemoc> considering the activity in android's kernel-common android 5 will most like use 3.10
<hramrach> we have an almost working MTD driver which is better for LTS than that nand junk from AW that locks up at random
<mnemoc> LTS stuff means other people will backport bugfixes for a very long time
<mnemoc> sure we want the mtd driver, even in 3.4
<hramrach> at least with MTD you you will get diagnostics and tons of other people who know MTD and can diagnose stuff for the users
<oliv3r> and if android 5 will 'default' to 3.10, chances are, a lot of oems will use 3.10; even though the kernel is somewhat android agnostic
<oliv3r> someone needs to pick up where rz2k left off though
<hramrach> anyone understands nand here?
<hramrach> I read an article about nand and I was horrified
<mnemoc> i'm really interested in knowing if it's possible to migrate the aw nand and partitions logic on top of the low level mtd driver to ease the transition
<hramrach> the partitions are on top of a block device. MTD can provide a block device.
Fusing has joined #linux-sunxi
paulk-collins has quit [Quit: Ex-Chat]
<hramrach> so afaict if you wanted you could do that just fine
<mnemoc> sounds like a soft way of getting mtd in
<mnemoc> without breaking compatiblity
<hramrach> you would need lichee u-boot with MTD support
<hramrach> it mainline u-boot built so that it can replace lychee one
<rz2k> oliv3r: I'll find time to test things if someone is willing to continue figuring out the nand controller
<mnemoc> as an initial step you can boot from mmc and read the aw-style nand after loading linux
<rz2k> and especially the ECC part
<hramrach> and some way to coexist with the room reserved for boot0/boot1 w/e
<rz2k> and interaction with oob
<rz2k> hramrach: spl and etc is already figured out by yuq in his driver
<mnemoc> hramrach: once our u-boot learns to boot from nand, we can replace boot0/boot1 with spl
<mnemoc> rz2k: would it be to hard to port the current aw partition layout and anti-wearing stuff on top of the mtd driver?
<oliv3r> rz2k: i was hoping slapin would pick it up :( but you did great in what you did so far
<mnemoc> slapin is too busy with his 90% female young students...
<rz2k> mnemoc: it will be hard
<oliv3r> rz2k: did you read what mnemoc asked above? Is it at all possible to be able to read the data on the AW nand, with the mtd driver?
<rz2k> the current driver uses block devices and fat32 layout, adequate drivers are mtd ones with mtd partitioning and nand-flavored filesystems
<hramrach> I don't think reading AW data is really doable
<oliv3r> i don't even think layout wise that would work; but i'd be happily proven wrong
<rz2k> there are no back compatability, BUT, the mtd drivers are supprted widely
<oliv3r> or it would require a huge amount of legwork; not really worth our time
<hramrach> it uses different wear levelling and block management logic, and the whole point of MTD is to replace that
<rz2k> so you can setup your android to do the images for mtd
<oliv3r> i'm not even tlaking about 'fresh' installs, because for those it won't matter, flash new image, done
<mnemoc> but iirc our mtd driver doesn't implement wear levelling
<oliv3r> mtd nver does
<oliv3r> you use a filesystem that takes care of that, like jffs2
<rz2k> the filesystem should do that
<rz2k> mtd compatible one
<rz2k> yeah, like ubifs or jffs
<rz2k> and other ones
<rz2k> i was able to run ubifs
<mnemoc> or code split out of aw's monster driver ;-)
<rz2k> on the yuq's driver with fixes
<hramrach> isn't there a MTD compatibility layer that provides plain block device on top of MTD?
<rz2k> yes, of course
<rz2k> the mtdblock
nove has joined #linux-sunxi
<rz2k> but as hno said, using mtdblock is not a solution
<rz2k> especially with block-flavored filesystems like ext4
<hramrach> if it makes transition easier for android I don't see much problem with that
<rz2k> first thing that is needed to be done in the yuq's driver is replace this https://github.com/yuq/sunxi-nfc-mtd/blob/master/nfc.c#L211 with proper waiting tasks
<rz2k> right now it actually does what it does - waits (non preemtable) for 0xfff and if it didnt write - throws an error
<rz2k> also the whole driver is not mutex'ed and I have no idea what will happen on A20 setup with two cores if kernel decides to do something fishy with preemption of this driver.
<rz2k> hramrach: android has native jffs and etc. support
<rz2k> ton of hardware is running on that, like omap systems
<mnemoc> arokux: have a text for me to add in gmac's import commit? :)
<oliv3r> using mtdblock wil kill your flash in a week
<oliv3r> using it read-only to copy data in a transition, sure, but on top of mtdblock you need aw-ext4 layer which doubtfully will work
<rz2k> that seriously depends on the flash itself :p
<oliv3r> a day? :p
tomboy65 has quit [Quit: And remember, aal is well.]
<mnemoc> or sunxi-legacy-mtdblock implementing aw's wearing and black magic as-is
hansg has joined #linux-sunxi
<oliv3r> you could do that, but that'll be a lot of effort, and for what
<oliv3r> most people will just reinstall anyway
wolfy has left #linux-sunxi ["Dragostea e oarba? Nu. E doar putin presbita: nu vede nici un defect pana nu se indeparteaza putin. (Drew Barrymore)"]
<mnemoc> oliv3r: i hoped it to be transparent...
<mnemoc> yuq's nfc-mtd + aw compatible block device + aw compatible partition table all over the same bytes as today
<rz2k> lets hope to get it detect bad blocks first
<rz2k> the right way
<rz2k> because right now it is too late to talk about any compat. layering until this thing actually works 100/100 launches.
<mnemoc> ok
hansg has quit [Remote host closed the connection]
boycottg00gle has quit [Remote host closed the connection]
<oliv3r> mnemoc: if you have about 10000 manhours, sure no problem :)
<mnemoc> oliv3r: :p
<mnemoc> oliv3r: i hoped it was just about refactoring current code to use a different backend
<mnemoc> <6>PHY: sunxi_gmac-0:00 - Link is Up<c> - 100/Full<c>
<mnemoc> <6>ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
<mnemoc> <3
<slapin> mnemoc: WOW!!!
<slapin> mnemoc: gmac works?
<slapin> rz2k: hi!
<slapin> rz2k: long time nos ee
<mnemoc> in linux, after importing a patch arokux "found" on cubieboard's tree, yes ;-)
<slapin> rz2k: can you point me on your mtd experiments?
Gerwin_J has joined #linux-sunxi
hansg has joined #linux-sunxi
<slapin> rz2k: ping
<slapin> rz2k: I just found you!
hansg has quit [Client Quit]
* slapin just went out for a drink :(
<oliv3r> slapin: back to work, chop chop
<oliv3r> mnemoc: nah, noway :(
<mnemoc> oliv3r: reality slap :|
<oliv3r> mnemoc: no worries
<oliv3r> slapin: there's some stuff on the ML, let me see if i can find a link
<mnemoc> no power button on the CT?
<oliv3r> mnemoc: sure there is
diego71 has quit [Ping timeout: 272 seconds]
<mnemoc> oliv3r: I only see one button, and assumed it's FEL
<oliv3r> mind you those are old mails; there may be newer ones
<oliv3r> mnemoc: i have like 3 or 4
<oliv3r> check the opposite side
<hramrach> mnemoc: there are like 5 buttons
* mnemoc looks for a lamp
* slapin was denied of cubietruck :(
<oliv3r> mnemoc: power button is on the 'front' side next to the IR led
* slapin feels rejection from the bottom of heart
<oliv3r> mnemoc: FEL and reset (i think) is on the right side, next to the USB port
<oliv3r> slapin: this was first batch for most active memebers; you could be part of the second batch
<oliv3r> slapin: there is still some love left
<slapin> oliv3r: I was told I already have similar hardware, so I don't need one :(
<oliv3r> slapin: well a10 and a20 are all the same in nfc thoughts, it just needs a little finishing touches
<oliv3r> slapin: i tell you what, If you make it work and fix outsanding issues/bugs, If cubietech doesn't send you one, I will send you mine!
andoma has quit [Ping timeout: 264 seconds]
<slapin> oliv3r: I know as u-boot one works on both
andoma has joined #linux-sunxi
<slapin> oliv3r: I'm going to do that anyway and have cubie2
<slapin> oliv3r: I want CT for gmac
<oliv3r> :)
<oliv3r> if the a new batch arrises and benn asks for more names, I will make sure yours in on top of the list :)
<slapin> oliv3r: I appreciate it, will wait then
diego71 has joined #linux-sunxi
<oliv3r> in the mean time, don't sit idle; we need some extra love on mtd :D
<oliv3r> and i'm almost done with my !linux-sunxi stuff!
<hramrach> is the mtd driver for a10 or a20?
<hramrach> is there difference between a10 and a20 other than they invented new partition format
<oliv3r> both
<oliv3r> not really, maybe some small bug fixes
n01 has left #linux-sunxi [#linux-sunxi]
<oliv3r> but who knows really
<hramrach> hmmpf
<hramrach> will need to figure out a setup to boot different stuff on the boards without juggling cards
<mnemoc> is it fair to request the gpios of a script.bin _used section even if the drive has not been loaded?
<mnemoc> or the driver must to do it on it's own?
<mnemoc> (and release them on unload)
<hramrach> can you even have alternate drivers for same gpio?
<mnemoc> how does that work on mainline pinmux? driver or dtb parser?
<mnemoc> hramrach: my idea is to create normal pdevs, and then whatever driver providing the right name will get modprobed...
<mnemoc> pre-requesting the GPIOs will help cleaning driver's probe code
<torbenh3> seriously ?
<mnemoc> the aw crap drivers in 3.4 invest a lot of boilerplate on that
<torbenh3> in mainline you request gpios yourself.
<torbenh3> and pinmux is requested via pinctrl.
<mnemoc> also for pinmuxes defined on the .dts ?
<mnemoc> ok. then I'll make the driver request them on load and release them on unload
Fusing has quit [Ping timeout: 246 seconds]
<mnemoc> wanted to avoid that... as the feature specific pins are very unlikely to touched during runtime by anything else
<hramrach> problem is that sometimes you have pins defined for multiple stuff and only some is used - and logic for that is in the driver
<mnemoc> fair point
<hramrach> like you can have pins for multiple wifi cards and depending on the set card type one set is used
<arokux> mnemoc: will give you text tonight
<mnemoc> arokux: thanks
<arokux> mnemoc: after you fix dynamic page list on wiki :p
<mnemoc> arokux: what is the problem?
<mnemoc> is that plugin supposed to work on psql?
<arokux> mnemoc: no idea, sorry. the problem is this: http://sprunge.us/GIDb
<mnemoc> ok
cajg has joined #linux-sunxi
<torbenh3> mnemoc: well... device core handles detting default pinctrl from DT itself.
<torbenh3> mnemoc: although, i dont know, since when.
<torbenh3> but people are doing this nowadays:
<mnemoc> so drivers are not dealing with pinmuxing
<mnemoc> now
<torbenh3> only if you have to use non-default pinmuxing.
<torbenh3> which is really rare.
* mnemoc wonders how to do that on the legacy/script.bin world
<mnemoc> I mean, keeping both options
<mnemoc> I want to bring ugly drivers closer to what 3.10 will need
<mnemoc> by moving pdev creating into a "device core"
<mnemoc> plat-sunxi/devices_script.c
<torbenh3> mnemoc: well... maybe you could just create a DT ?
<mnemoc> not in 3.4...
<mnemoc> or I can?
<mnemoc> is DT properly support on 3.4 for arm?
<mnemoc> supported*
<mnemoc> faking DT can be a nice workaround
<torbenh3> yeah its supported. although, you would maybe need to backport the DT stuffs from 3.10 / mainline.
<torbenh3> for the individual drivers.
<mnemoc> :<
<torbenh3> so... basically creating the pdevs from script.bin might make more sense.
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<torbenh3> although, you could mix that freely.
<torbenh3> create DT for uart, with DT support backported, and pdev for sata.... etc.
<torbenh3> this might actually allow porting individual drivers to DT inside of 3.4
<slapin> oliv3r: I'm never idle, I'm overloaded most of time (I/O bound).
soul has joined #linux-sunxi
soul is now known as Guest58706
pfdm has joined #linux-sunxi
apo has quit [Remote host closed the connection]
apo has joined #linux-sunxi
Guest58706 has quit [Read error: Connection reset by peer]
Fusing has joined #linux-sunxi
vicenteH has joined #linux-sunxi
Soru_ has joined #linux-sunxi
<arokux> mnemoc: any ideas about dpl on wiki?
<arokux> mnemoc: I'll post commit text anyway.. :p
<mnemoc> arokux: will look at the wiki thing after getting the script to pdev thing working
hramrach has quit [Remote host closed the connection]
<arokux> mnemoc: alright.
<mnemoc> arokux: thanks
tzafrir has joined #linux-sunxi
dack has quit [Remote host closed the connection]
BluesBoy has quit [Quit: HydraIRC -> http://www.hydrairc.com <- s0 d4Mn l33t |t'z 5c4rY!]
Soru___ has joined #linux-sunxi
<mnemoc> arokux: http://linux-sunxi.org/User:Alejandro_Mery/Sandbox ... reproduced :)
Soru_ has quit [Ping timeout: 248 seconds]
<arokux> mnemoc: yeah. that is the problem. well maybe we do not need to use that extension, I just want to nicely render the list of tutorial on the main page - in multiple columns. maybe you have some other ideas.
<mnemoc> arokux: try the categorytree for now, will try to fix this one later. but need to focus in refreshing kerneling programming atm
FR^2 has quit [Quit: Connection reset by peer]
<mnemoc> s/kerneling/kernel/
<arokux> mnemoc: I've studied categorytree thoroughly. it doesn't have column params or similar. take your time.
notmart has quit [Quit: notmart terminated!]
eebrah_ has joined #linux-sunxi
wingrime has joined #linux-sunxi
<wingrime> mnemoc: turn back CPU table to main!
<mnemoc> eh? i'm not wiki dictator. things need to be agreed by the community.
<mnemoc> wingrime: discuss your points to the other people interested in a different kind of main page for the wiki. come into an agreement. and everyone happy.
Skaag has quit [Ping timeout: 260 seconds]
mkutsevol has joined #linux-sunxi
HeHoPMaJIeH has quit [Remote host closed the connection]
Soru_ has joined #linux-sunxi
Soru___ has quit [Ping timeout: 245 seconds]
Soru___ has joined #linux-sunxi
<mnemoc> what's more common, foo-sunxi or sunxi-foo? :(
Soru_ has quit [Ping timeout: 240 seconds]
Soru___ has quit [Read error: Connection reset by peer]
Soru_ has joined #linux-sunxi
Soru_ has quit [Read error: Connection reset by peer]
Soru has joined #linux-sunxi
eebrah_ has quit [Read error: Connection reset by peer]
Gerwin_J has quit [Quit: Gerwin_J]
eebrah_ has joined #linux-sunxi
<arokux> wingrime: why?
<arokux> mnemoc: foo-sunxi
<wingrime> arokux: OR make page with CPU list
<wingrime> arokux: don't trow my work so easy
<arokux> wingrime: nobody has thrown out your work, but please *do* take a close look at what is there now, before you complain.
<wingrime> arokux: nice, but table on main was better for my poit of view)0)
<arokux> wingrime: who comes to our wiki?
<wingrime> arokux: ?
<arokux> wingrime: what people come to our wiki?
<arokux> wingrime: (link to that page is on the main page)
<wingrime> arokux: users/devs
rz2k has quit []
<arokux> wingrime: now, what % of these users/devs have the board and want to look at some tutorial and what % wants to make decision on which board to buy?
atiti has quit [Ping timeout: 272 seconds]
deasy has joined #linux-sunxi
<deasy> yup
<arokux> wingrime: I have thrown away lots of "work" by Rocker, I hope you do not mind :)
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
geecko has joined #linux-sunxi
Quarx has quit []
<arokux> deasy: you wanted to ask me smth about KDE?
<deasy> yup your help :p
<arokux> deasy: how can I help?
<arokux> deasy: I would run from KDE as fast as I can :)
<deasy> hunt what make the start of dolphin slow :)
<arokux> deasy: go for LXDE/XFCE they are never slow and very slick.
<deasy> i have find some poll who take 13ms (4-5) but it seems not it who make the start slowdown
eebrah_ has quit [Ping timeout: 264 seconds]
<deasy> arokux, nop kde is better but dolphin need a quicky start
<deasy> i talk about warm start not cold
rellla3 has joined #linux-sunxi
<arokux> deasy: ok, ask in some #kde if there is one..
<deasy> i think dolphin take around 500ms-1s for be display
<arokux> deasy: i'm not using KDE and I do not recommend if for real hackers
<deasy> hmm lot of programmer prefer the code of kde on gnome...
<arokux> deasy: I mean not only KDE, but all those big and bloated desktop environments like KDE, GNOME, Unity etc.
<deasy> they provide what other don't provide ;)
<deasy> do you have a nice dualscreen management on others DE :)
<deasy> this is just a example
<deasy> an*
<arokux> deasy: dual screen is provided by X, DE provide GUI to manage it. I can set up it with command line.
Gerwin_J has joined #linux-sunxi
<deasy> that's not comfortable and ok for make users switch to linux :p
<arokux> <arokux> deasy: i'm not using KDE and I do not recommend if for real hackers
<arokux> deasy: read it more carefully at the end :)
<Sonicadvance1> twm yay?
<deasy> ^^
<arokux> Sonicadvance1: no, just LXDE...
pacopad has quit [Quit: pacopad]
<oliv3r> sunxi-emac says otherwise! :p
<arokux> oliv3r: common....
<torbenh3> mmm.... whats the LOADADDR for mainline on CT ?
<dapsaille> hi, when i set lcdhsync = port:PD26<2><default><default><1> and lcdvsync = port:PD27<2><default><default><1> in my fex file in xorg i can see that hsync and vsync are set to negative ... did i miss something ?
jinzo has quit [Quit: Leaving]
<oliv3r> arokux: i would say it's 50:50
<arokux> oliv3r: have you counted?
<oliv3r> lol no
<mripard> torbenh3: 0x40008000
<torbenh3> mripard: thx
<arokux> oliv3r: so I would say *-sunxi is more common than sunxi-*
<arokux> find -name *-sunxi* | wc -l
<arokux> 5
<arokux> find -name -sunxi* | wc -l
<arokux> 0
<arokux> last is wrong.
<arokux> find -name *-sunxi* | wc -l
<arokux> 5
<mnemoc> doh
<arokux> find -name *sunxi-* | wc -l
<arokux> 62
<oliv3r> arokux: coun't other manufacturers/drivers aswell, sunxi is fairly small
<mnemoc> so sunxi-foo will be, and the others can use module aliases
<oliv3r> so I was right! :p
<arokux> ?
<arokux> no!
<oliv3r> sunxi-* is 62, *-sunxi is 5
<oliv3r> i was right! :p
<oliv3r> but you can't only ook at sunxi, look at others more
<arokux> not that easy.....
<arokux> ./.git/logs/refs/remotes/origin/stage/sunxi-3.4
<arokux> that got counted to......
<arokux> find -name *sunxi-* | grep -v "git/" | wc -l
<arokux> 2
<arokux> find -name *-sunxi* | grep -v "git/" | wc -l
<arokux> 5
<mnemoc> look for .ko names
popolon has quit [Quit: Quitte]
<mnemoc> some modules have several .c or .h
<arokux> mnemoc: only mainline should be looked at, not sunxi-3.4
<arokux> at least that was my thought...
<oliv3r> don't look soley at sunxi, ther's only a very limited number of sunxi drivers
<oliv3r> for mainline it's sunxi-*
<oliv3r> but it really depends on the subsystem
<mnemoc> what about omap or fsl?
<oliv3r> pwm for example prefers pwm-sunxi; wdt, uses sunxi-wdt
<mnemoc> :|
<oliv3r> sunxi-* seems the most sane :p
<arokux> oliv3r: for mainline it is *-sunxi, please check.
<arokux> and usb people use: ehci-sunxi.c
<arokux> :)
<oliv3r> it's sunxi-wdt; sunxi-rtc; sunxi-sid
<arokux> find -name *omap-* | grep -v "git/" | wc -l
<arokux> 75
<oliv3r> pwm will be pwm-sunxi because pwm framework uses that
<arokux> find -name *-omap* | grep -v "git/" | wc -l
<arokux> 51
<oliv3r> so sunxi-* it is :D
<mnemoc> foo-mach <-- framework plugin
<arokux> the question got redefined :p
<mnemoc> mach-foo <-- standalone driver ?
<oliv3r> not really
<oliv3r> pwm is framework, so is wdt
<arokux> mnemoc: I'd do it as the others do for each framework.......
<oliv3r> they don't even match :)
<oliv3r> arokux: exactly :p
<arokux> and all the time I will ask myself why is the naming not consistent.
<mnemoc> arokux: i would like to avoid making exceptions in a generic loop spawinging pdev out of script.bin
<arokux> mnemoc: I see, sorry.
<oliv3r> sunxi-*
<mnemoc> and MODULE_ALIAS can solve the rest...
<mnemoc> or not?
<oliv3r> it's not a huge issue
<oliv3r> for 3.4(10) it doesn't matter much; pick one
<oliv3r> i'd prefer sunxi-* personally
<mnemoc> me too
<mnemoc> once working, naming can be changed or exceptions supported
<mnemoc> somehow I like sunxi-foo more, but in a sunxi world the foo part is more important than the common mach part
<torbenh3> errm... doesnt mainline have sdcard ?
<mnemoc> so ir-sunxi tells you faster that it's the ir driver
<arokux> torbenh3: no :)
<torbenh3> oh
<oliv3r> *-sunxi is yoda speak
<mnemoc> true
<arokux> torbenh3: if you like you can go for USB, there is my branch for that.
<arokux> torbenh3: what's ur hardware?
<torbenh3> CT
<arokux> torbenh3: :( not network in u-boot, this is real shit - i'll continue my work on it tonight.
<arokux> torbenh3: wait! you *can* have network in u-boot!
<torbenh3> network in u-boot doesnt help me too much
<arokux> torbenh3: I've added USB support to u-boot too, so you need usb2ethernet dongle - it works just fine with u-boot
<arokux> torbenh3: why did you need sdcard anyway?
<torbenh3> dunno. want to boot some RFS ?
<arokux> torbenh3: NFS/USB HDD/Stick
<arokux> torbenh3: or built it into the kernel...
<torbenh3> arokux: gonna try to go with nfs first.
<arokux> torbenh3: NFS is the most flexible option, imho
<torbenh3> arokux: i know how that works :)
<torbenh3> arokux: however it might be a bit difficult with the gmac :)
<arokux> torbenh3: I do not know that you know
<arokux> torbenh3: http://linux-sunxi.org/Possible_setups_for_hacking_on_mainline -- also interesting, some nice hacks
<arokux> torbenh3: do you have usb2ethernet dongle?
<dapsaille> nice tips ^^
<torbenh3> env set fdt_high ffffffff ??? whats that for.
<torbenh3> it booted here without that.
<arokux> torbenh3: workaround ...
<torbenh3> for what ?
<arokux> torbenh3: it is needed if the kernel image is too fat
<arokux> torbenh3: otherwise u-boot will relocate stuff and pieces overwrite each other
<arokux> dapsaille: how about your tips concerning fex magic? :p
<torbenh3> seems like i was lucky.
<dapsaille> for now it's fake magic :)
<dapsaille> i can't get a picture and i'm going crazy ...
guizamboni has joined #linux-sunxi
<arokux> dapsaille: this is valuable know-how, so please post, once it works or you have found a workaround.
<dapsaille> i'm affraid that it is a limitation of vga hack :x
<dapsaille> with pleasure
<arokux> torbenh3: you can say so, I've spent several days.. it booted for the others but not for me.
<arokux> dapsaille: thanks :)
<torbenh3> arokux: dunno... i have put the dtb below the uImage.
<dapsaille> do you know what is LCD0_DE ? i mean .. is it usefull for vga ?
<arokux> torbenh3: ok, not sure it helps - forgotten the details.
<arokux> torbenh3: so no usb2ethernet dongle?
<arokux> dapsaille: ping ssvb, I think he is one of the very few ppl that could help
<dapsaille> it seems but .. it seems that he didn't have many informations regarding VGA :x
<dapsaille> ssvb are you here ? ^^
Soru_ has joined #linux-sunxi
<arokux> dapsaille: they you'll become our VGA expert :)
<dapsaille> damn ... not a good idea :)
<dapsaille> i know how to plug a vga cable and .. that's all (sort of)
Soru has quit [Ping timeout: 272 seconds]
pfdm has quit [Ping timeout: 250 seconds]
eebrah_ has joined #linux-sunxi
enrico_ has quit [Quit: Bye]
awatch has joined #linux-sunxi
FR^2 has joined #linux-sunxi
_massi_ has quit [Quit: Leaving]
<torbenh3> hmm... well.. at least the gmac detects link up down. but doesnt seem to send packets.
Gerwin_J has quit [Quit: Gerwin_J]
eebrah_ has quit [Read error: Connection reset by peer]
arokux2 has joined #linux-sunxi
eebrah_ has joined #linux-sunxi
Fusing has quit [Ping timeout: 244 seconds]
panda84kde has quit [Quit: Konversation terminated!]
eebrah_ has quit [Read error: Connection reset by peer]
<arokux2> there you have it: mnemoc: oliv3r
<arokux2> this is commit message for GMAC, please improve it or take it as is.
Gerwin_J has joined #linux-sunxi
<mnemoc> arokux2: no reference to the stmmac thing?
wingrime has quit [Ping timeout: 240 seconds]
<mnemoc> only dw?
<arokux2> good point sec
eebrah_ has joined #linux-sunxi
<arokux2> mnemoc: http://sprunge.us/SdJN
fredy has quit [Excess Flood]
<arokux2> torbenh3: which gmac ? are you playing with stmmac already?
<torbenh3> arokux2: no
<arokux2> torbenh3: which one then?
fredy has joined #linux-sunxi
<torbenh3> i took that gmac driver from the cubie tree.
<torbenh3> arokux2: because wens reported problems with the phy enumeration.
<torbenh3> so i thought, i make that one run, and then we can search for the diff.
<arokux2> torbenh3: I see. it works in sunxi-3.4 - you may want to compare what fails at which point.
<arokux2> torbenh3: https://github.com/arokux/linux/tree/sunxi-3.4-gmac -- sunxi-3.4 with gmac
<torbenh3> arokux2: it seems to thinking, its online. using ip=dhcp ... turned too much debuggin on now. i am getting blinded :)
<torbenh3> arokux2: at least it can detect when i unplug the network cable. so it seems to be somewhat communicating with the phy
<arokux2> torbenh3: let me clarify: you have cherry picked gmac and put it into mainline, is that right?
<torbenh3> yes.
Soru__ has joined #linux-sunxi
<arokux2> torbenh3: how about clocks? you are poking at bits on your own?
<torbenh3> arokux2: the driver can do that on its own.
<arokux2> torbenh3: nice driver. although I had a race.
<arokux2> torbenh3: clk framework will turnoff all unused clocks
<arokux2> torbenh3: so it can disable GMAC's too
<arokux2> torbenh3: I had such a problem...
eebrah_ has quit [Ping timeout: 252 seconds]
<arokux2> torbenh3: you can try to pass: clk_disable_unused
Soru_ has quit [Ping timeout: 244 seconds]
<oliv3r> wow
<torbenh3> arokux2: didnt seem to change anything... but i see that it already seems to be receiving packet.
rellla3 has quit [Read error: Connection timed out]
Soru___ has joined #linux-sunxi
rellla3 has joined #linux-sunxi
Soru__ has quit [Ping timeout: 272 seconds]
<arokux2> torbenh3: btw, wens was testing on cb2
tzafrir has quit [Ping timeout: 264 seconds]
<arokux2> hm.. gmac is very nicely written unlike what I've seen before
ccube has quit [Remote host closed the connection]
Soru___ has quit [Read error: Connection reset by peer]
soul has joined #linux-sunxi
soul is now known as Guest47816
ccube has joined #linux-sunxi
<arokux2> torbenh3: i think gmac_resources is incorrectly defined
<arokux2> oliv3r: any comment on the commit message?
<arokux2> oliv3r: http://sprunge.us/SdJN
<arokux2> oliv3r: please repastebin, if so.
<arokux2> torbenh3: priv_clk_reg = readl(priv->gmac_clk_reg + GMAC_CLK_REG);
<arokux2> hm.. this is also incorrect, since gmac_clk_reg = CCMU_BASE + GMAC_CLK_REG
jemk has joined #linux-sunxi
<arokux2> wens: ^
* mnemoc happily doing `apt-get dist-upgrade` on his CT :)
<arokux2> mnemoc: what did you reply? :)
<mnemoc> arokux2: about?
<arokux2> mnemoc: [PATCH 1/2] gmac: add gmac support
<arokux2> mnemoc: just got an empty e-mail from u
<oliv3r> arokux2: not half bad; ok for me
<mnemoc> o_O
<mnemoc> after some --whitespace=fix, adding a description from arokux, and correcting the duplicated IRQ define. applied on stage/sunxi-3.4.
<mnemoc> waiting for improvements now.
<mnemoc> cheers,
<mnemoc> Alejandro Mery
<mnemoc> why empty??
<mnemoc> err
pfdm has joined #linux-sunxi
<arokux2> mnemoc: yes.
<arokux2> mnemoc: but google groups failed
<mnemoc> :/
<arokux2> mnemoc: :)
<oliv3r> soon my friends, very soon
<oliv3r> almost done with my other thing
* arokux2 grabs stage
* arokux2 will do it later.
<arokux2> mnemoc: you want me to submit patch for sun7i_defconfig or you'll do it yourself?
<mnemoc> as commented I prefer a better Kconfig than _defconfig
<mnemoc> i would be happy if _defconfig where only the mach choice
<oliv3r> hmm
<arokux2> mnemoc: yes, should I submit the patch about it?
<arokux2> priv_clk_reg &= (~0x00000003);
<oliv3r> Kconfig would apply to all archs though no?
eebrah_ has joined #linux-sunxi
<mnemoc> depends on ?
<arokux2> this has no influence on priv_clk_reg??
<arokux2> ugh, it has..
<mnemoc> depends on PLAT_SUNXI and default m for example
<mnemoc> but sure it deserve a patch to have an open discussion
<mnemoc> deservers
<mnemoc> deserves
<mnemoc> f*
* mnemoc goes for another cup of coffee
<arokux2> isn't it plain wrong???
<arokux2> we read off priv_clk_reg, then change its second bit, then overwrite it with zeros...
<oliv3r> it's funny to see how it's a multi-file driver
<oliv3r> just as the designware one is
<arokux2> oliv3r: take a look at the code that I've referenced
<oliv3r> yeah i know, it's just so obvious
Guest47816 has quit [Ping timeout: 245 seconds]
Skaag has joined #linux-sunxi
<arokux2> oliv3r: you agree it is wrong?
<oliv3r> them making a driver and pretend it's all allwinner IP?
<oliv3r> yeah :p
<oliv3r> don't lie, just imrpove the existing driver :)
<arokux2> oliv3r: no...
<dapsaille> dumb question .. can i check conntent of a register from linux ? :x
<arokux2> dapsaille: yes
<arokux2> devmem2
<dapsaille> thanks ^^
<arokux2> oliv3r: look at my earlier messages
<arokux2> oliv3r: and no, they were not hiding anything
<oliv3r> where does AW say: "this is designware GMAC driver"
<arokux2> oliv3r: pr_info("\tDWMAC1000 regs (base addr = 0x%p)\n", ioaddr);
<oliv3r> unintended typo
<arokux2> oliv3r: btw, you know what, I think you haven't sent those u-boot patches with git send-mail, because you have replied to the message and what is even more important quoted it.
<oliv3r> i think yoru right
<dapsaille> arokux2 ... may i bother you regarding devmem2 :x ? i'm not sure of usage ...
<dapsaille> ok thanks .. i'm trying to find a bit from lcd registers but don't know how to interpret a13 datasheet ...
* dapsaille admit that he is a noob ...
<mnemoc> arokux2: your USB work deserves to be in the [[USB]] page, not inside your personal user scope...
tzafrir has joined #linux-sunxi
<arokux2> mnemoc: yes, there is already USB page, I have just given a cached one, sorry.
<arokux2> mnemoc: btw, can I have mine personal page back? :)
<arokux2> mnemoc: can you please take a look with fresh eye at the code I've said to be plain wrong?
<mnemoc> back in 15m
enrico_ has joined #linux-sunxi
<dapsaille> arokux2 => i got it thanks for help need to convert hex to binary :)
<torbenh3> arokux2: did you turn on GMAC_RING or GMAC_CHAINED when testing gmac ?
<arokux2> torbenh3: yes
<arokux2> torbenh3: have you seen my comments? the code seems to be plain wrong
<torbenh3> there was an or.
<arokux2> torbenh3: I also think that wrong addressed is passed to gdma_init
tomboy65 has joined #linux-sunxi
<arokux2> torbenh3: http://sprunge.us/Oibd?c
<arokux2> torbenh3: shit, forget it
<arokux2> there was ~
<arokux2> torbenh3: but I still think wrong address is passed to: gdma_init, will double check now
<torbenh3> not here.
<torbenh3> i can receive packets.
<arokux2> torbenh3: it also works for me, but that doesn't mean the code is correct.
<torbenh3> its the base address of the h/w
<torbenh3> how could that be wrong, and it generates irqs ?
ccube has quit [Remote host closed the connection]
dalee has joined #linux-sunxi
<arokux2> priv->ioaddr = GMAC_BASE + 0x1054
<arokux2> torbenh3: ^
<arokux2> torbenh3: I applogize!
<arokux2> I'm stupid...... not my day.
ccube has joined #linux-sunxi
<dalee> help, UVC webcam preview working but capture gets image with green random patern (http://i.imgur.com/T9l1D.jpg) anyone has same problems?
<torbenh3> where the hell did you get that from ?
tomboy64 has joined #linux-sunxi
ccube has quit [Remote host closed the connection]
<arokux2> torbenh3: gmac_resources
<arokux2> torbenh3: I got the end.. instead of start...
<arokux2> torbenh3: they should have done -1 all the time for the end, though.
ccube has joined #linux-sunxi
<arokux2> tomboy65: or used: DEFINE_RES_MEM
<arokux2> err torbenh3
<tomboy64> ^^
tomboy64 has quit [Quit: Uhh ... gotta go.]
tomboy65 has quit [Quit: KVIrc 4.3.1 Aria http://www.kvirc.net/]
<dalee> have anyone tried UVC webcams to capture? some problems with UVC webcams? anyone can help me
Fusing has joined #linux-sunxi
<mnemoc> arokux2: should I remove the redirect form your page, or you will replace it with real content?
rellla3 has quit [Ping timeout: 260 seconds]
<arokux2> mnemoc: I'd like to have my page empty and everything moved to USB
Skaag has quit [Ping timeout: 240 seconds]
<arokux2> mnemoc: thanks.
<arokux2> mnemoc: that was wrong link, here is correct: http://article.gmane.org/gmane.comp.hardware.netbook.arm.sunxi/4464
<mnemoc> I was already wondering...
<mnemoc> i seriously hate the idea of deciding the phy type at compile time...
* arokux2 wants board with 2xEthernet
<deasy> arokux, can you analyze a strace :p or help me to do it ? someone use a good analyzer of strace(parser)
<arokux2> deasy: I can imagine strace of an KDE app will be veeeeeery noisy, so sorry, I cannot read it.
maz_ has joined #linux-sunxi
<deasy> arokux2, yes its very noisy
<deasy> 3-7Mo
<oliv3r> mnemoc: just include all PHY's
<oliv3r> shouldn't be a problem
<oliv3r> except of course with 3.4
<mnemoc> in 3.4 we can add a .fex field
<mnemoc> i want a multi-sunxi single bin kernel in 3.4 too.... :(
<mnemoc> not per-board bull crap
<libv> good tidings on the fosdem schedule system...
<arokux2> hm.. how do I select between different ethernets in u-boot :$
<mnemoc> uenv.txt?
<arokux2> mnemoc: from command line, one ethernet is native the other one is usb2ethernet
* arokux2 wants board with 2xEthernet
<arokux2> oh, I've said it already
<dapsaille> arokux2 => usb ? :)
<arokux2> dapsaille: yep, why are u surprised?
<dapsaille> you want board with 2xethernet, why not use usb eth ?
<arokux2> dapsaille: $
<dapsaille> if someday an arm port of pfsense come, i would need a 4 ethernet board :)
<arokux2> dapsaille: a board with 2xEth would be cheaper then 1xEth+usb2ether. the latter looks ugly also.
Fusing has quit [Quit: Nettalk6 - www.ntalk.de]
<dapsaille> yeah of course
Black_Horseman has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
<dapsaille> quadeth with enough ram for logs in ramdisk ... well .. i must stop dreaming and try to focus on my vga problem :)
<arokux2> dapsaille: and the point is A20 has EMAC and GMAC natively.
<dapsaille> .. don't know what is emac and gmac :x
jemk has quit [Remote host closed the connection]
<arokux2> dapsaille: I'm also noob, but here it goes. support for ethernet is broken down into two parts: PHY + controller
<dapsaille> ok
<dapsaille> PHY is for physical ?
<arokux2> dapsaille: controller resides in SoC; emac - 100Mbit, gmac - 1GBit
<dapsaille> ok
<dapsaille> g for gigabit
<arokux2> dapsaille: yes. and A20 has both, so two PHYs and they could be independent
<dapsaille> but .. does an arm cpu can handle gigabit transfert rate ? or is it only to use the two nics ?
<arokux2> dapsaille: I'm not sure. truck is the first board with PHY that can support GMAC
<dapsaille> ok
<dapsaille> thanks for the informations :)
<arokux2> dapsaille: you are welcome, but do not rely on them, I'm not 100% in my understanding
<dapsaille> no problem, i understand that we live in a theory and hackable world :)
<arokux2> dapsaille: what are you working on?
<arokux2> dapsaille: what is the big goal, I mean.
<dapsaille> i would like to use old arcade cab monitors with an a13 olinuxino, using emulators with perfect modelines for getting the most accurate render
<dapsaille> so .. basically, play videogames ^^
<arokux2> I see :)
<dapsaille> seriously, i need to output a 15.6khz signal from vga out :)
<torbenh3> arokux2: you dont activate the gmac in the defconfig of your patch.
<dapsaille> and if i can do it, i will be happy for some time but .. will never play at my arcade cab, like always, i preferr hack it than play with it
<torbenh3> oh
<torbenh3> arokux2: ok. anyways . patch one activated IP_PNP and nfs...
<torbenh3> that doesnt belong into a gmac patch :P
<arokux2> torbenh3: that was cherrypick it was agreed it should be taken as is and improved.
<torbenh3> oki then :)
<arokux2> torbenh3: !!
<arokux2> torbenh3: you've missed it, right?
* mnemoc prefers an improved Kconfig so the lines in defconfig aren't needed
<torbenh3> oh well... i give up for today.
<mnemoc> the "spartan" but "feature full" sunxi defconfig should only require the MACH selection... and the rest depends on and default m/y
<torbenh3> this damn thing sent out 3 packets. for some unknown reason
<arokux2> torbenh3: are you u-boot expert?
<torbenh3> and it reveives packets just fine
<torbenh3> arokux2: just a bit.
<torbenh3> arokux2: but i am sorry, i am off now.
<arokux2> torbenh3: short question if u know it is cool. so.. I have 2xethernet, can i said to u-boot which one to use?
eebrah_ has quit [Read error: Connection reset by peer]
Skaag has joined #linux-sunxi
eltom_ has joined #linux-sunxi
naobsd has quit [Quit: Page closed]
dapsaille is now known as Dapsaille_sleepm
Dapsaille_sleepm is now known as Dapsaille_Sleep
<eltom_> Hello Guys, i need some help, i have tried booting the sunxi-next Kernel but it is hanging at "Starting Kernel" i have followed all steps from http://linux-sunxi.org/Mainline_Kernel_Howto. Was anybody successful in booting the Kernel and can maybe help me?
<eltom_> This is on Cubietruck btw. ^^
eebrah_ has joined #linux-sunxi
<arokux2> eltom_: hm.. you have messed something up, as the others booted it just fine.
<oliv3r> kernel boots fine normally
<arokux2> eltom_: but they are not here atm
<oliv3r> make sure to check your powersupply
<eltom_> The Board boots the sunxi-3.4 Kernel fine, so i don't think it is the PowerSupply
<arokux2> eltom_: did you use correct dtb?
<arokux2> eltom_: also: have you ever booted sunxi-next before?
<eltom_> yes, i even tried the pre-compiled kernel's and u-boot from http://dl.linux-sunxi.org/nightly/ to make sure i didn't do something wrong during compilations
<eltom_> No i was not able to boot sunxi-next successfully so far
<oliv3r> if sunxi-3.4 boots fine, what are you hanging on? oh sunxi-next, mainline :D
<eltom_> only sunxi-3.4
<eltom_> I wanted to try xen on the Cubietruck so i need a mainline Kernel unfortunately :)
<Dapsaille_Sleep> hypervisor ?
<arokux2> eltom_: ok, so you most probably did something wrong in u-boot prompt
<eltom_> No not yet, i first wanted to be able to successfully boot the Kernel as is and then try to boot with the Hypervisor
<arokux2> eltom_: btw, there is some work done with Xen, have you seen it?
<eltom_> arokux2: i basically used the commands from http://linux-sunxi.org/Mainline_Kernel_Howto
<Dapsaille_Sleep> well .. didn't know that a Dom will work with an arm cpu :)
<arokux2> Dapsaille_Sleep: eltom_ http://linux-sunxi.org/Xen_Howto
<eltom_> That is the Output and the commands i used in the u-boot prompt:
<Dapsaille_Sleep> but .. WHY there is always an answer on wiki ? why ? does it mean that sunxi is full of competent people ? :)
<eltom_> sun7i# ext2load mmc 0:1 0x46000000 uImage
<eltom_> 1589664 bytes read in 106 ms (14.3 MiB/s)
<eltom_> sun7i# ext2load mmc 0:1 0x49000000 cubietruck.dtb
<eltom_> 8181 bytes read in 6 ms (1.3 MiB/s)
<eltom_> sun7i# setenv bootargs console=ttyS0,115200 loglevel=7 root=/dev/mmcblk0p2 rootwait panic=10 rootfstype=ext4 rootflags=discard
<eltom_> sun7i# saveenv
<eltom_> Saving Environment to MMC...
<eltom_> Writing to MMC(0)... done
<eltom_> sun7i# bootm 0x46000000 - 0x49000000
<eltom_> ## Booting kernel from Legacy Image at 46000000 ...
<eltom_> Image Name: Linux-3.12.0-rc1-87451-g2efb1f5
<arokux2> eltom_: ...
<eltom_> Created: 2013-10-16 13:13:55 UTC
<eltom_> Image Type: ARM Linux Kernel Image (uncompressed)
<eltom_> Data Size: 1589600 Bytes = 1.5 MiB
<eltom_> Load Address: 40008000
<eltom_> Entry Point: 40008000
<eltom_> Verifying Checksum ... OK
<oliv3r> oh god please use pastebin
<eltom_> ## Flattened Device Tree blob at 49000000
<mnemoc> please, use a paste service
<eltom_> Booting using the fdt blob at 0x49000000
<Dreadlish> omg
<eltom_> Loading Kernel Image ... OK
<eltom_> Using Device Tree in place at 49000000, end 49004ff4
<eltom_> Starting kernel ...
<Dreadlish> somebody kick him
<arokux2> eebrah_: wtf? use www.sprunge.us
<arokux2> eebrah_: sorry, not for u
<arokux2> eltom_: ^^
<eltom_> sorry
<arokux2> Dapsaille_Sleep: NO! there is some work. only some start...
<Dapsaille_Sleep> :)
<arokux2> Dapsaille_Sleep: and I try to push people to document things, so that I can use pointers to wiki
<oliv3r> sfirst of all, it's quite likly your kernel is compiled with debug_ll and early printk etc enablesd
<arokux2> eltom_: this is useless: root=/dev/mmcblk0p2 rootwait panic=10 rootfstype=ext4 rootflags=discard
<arokux2> eltom_: since no uSD support in mainline
popolon has joined #linux-sunxi
<arokux2> oliv3r: ... why you never read everything? he uses the kernel from nightlies
<eltom_> Ahh, then no wonder it is not booting, is there NAND or Sata Support in Mainline?
<arokux2> eltom_: no :)
<arokux2> eltom_: USB is there, but not yet pushed so you can have it from separate branch
<oliv3r> oh precompiled; i don't know if those have debug_ll and early_printk enabled
<oliv3r> so still valid answer; it's likely it's without that stuff enabled :)
<mnemoc> just sunxi_defconfig
<oliv3r> dunno if its on there, could be
<oliv3r> you need an initramfs with that at the very least :)
<arokux2> oliv3r:
<arokux2> grep DEBUG_LL arch/arm/configs/sun4i_defconfig
<arokux2> CONFIG_DEBUG_LL=y
<mnemoc> the nightlies dir includes NAME.config.gz files
<oliv3r> sunxi or sun7 :)
<mnemoc> mainline doesn't have sunNi_defconfigs....
<arokux2> oliv3r:
<mnemoc> only sunxi_
<arokux2> grep DEBUG_LL arch/arm/configs/sun*i_defconfig
<arokux2> arch/arm/configs/sun7i_defconfig:CONFIG_DEBUG_LL=y
<arokux2> arch/arm/configs/sun4i_defconfig:CONFIG_DEBUG_LL=y
<arokux2> arch/arm/configs/sun5i_defconfig:CONFIG_DEBUG_LL=y
<arokux2> oh
<arokux2> sorry, I appologize
<oliv3r> and early_printk etc?
<mnemoc> at least we should get that enabled in sunxi-devel
<oliv3r> aye
<arokux2> I was pasting stuff from sunxi-3.4
<oliv3r> and why can't I pm mnemoc!
<mnemoc> that doesn't need to be mainline-style spartan
<arokux2> mnemoc: why have you given me that green dot?
<mnemoc> me?
<arokux2> You have been opped on #linux-sunxi by mnemoc
<eltom_> arokux2: Which branch has USB-Boot pushed?
<mnemoc> damn ChanServ
<arokux2> there is a page on the wiki called USB, read it carefully.
<mnemoc> so you can kick people flooding the channel :)
FR^2 has quit [Quit: und weg...]
<eltom_> Ok, thanks a lot for the help guys, and sorry for flooding the Channel!
<mnemoc> the freaking ChanServ should not tell who gave it the order...
<arokux2> eltom_: but cubietruck.dts needs some love before you can enjoy USB, so if you want to enjoy it - tell me.
<arokux2> eltom_: have you booted the kernel?
<oliv3r> arokux2: you sound like a porn salesmen :)
<eltom_> arokux2: No not yet
<mnemoc> oliv3r! finish what you started!
AreaScout has quit []
<oliv3r> started whut
* arokux2 mnemoc will lobby 3.10 again
<mnemoc> forward to the ML?
<mnemoc> arokux2: nah
<arokux2> mnemoc: what else oliv3r started? I know at least dozens of things......
<arokux2> lost the count actually
<mnemoc> arokux2: he was teasing me with a reply he got by email
<arokux2> mnemoc: reply to what?
<mnemoc> oliv3r: forward to the ML? =) or going to keep the excitement for yourself?
<oliv3r> for now
<oliv3r> need to sort some things first
<mnemoc> :)
<arokux2> mnemoc: is it sunxi related_
<arokux2> ?
<oliv3r> very much :)
<mnemoc> *g*
<mnemoc> but maybe 3.10 related as well :p
eltom_ has quit [Quit: eltom_]
<mnemoc> considering time constrains
<oliv3r> haha yes
<arokux2> mnemoc: who will profit from it if it comes true?
<mnemoc> everyone
<arokux2> mnemoc: everyone as in sunxi.org or as on the planet? :)
<mnemoc> between both
<mnemoc> x > everyone_in_sunxi.org && x < enveryone_in_this_planet
<arokux2> oliv3r: RGMII vs GMII, CT has the latter, right?
<oliv3r> gmac has both
<oliv3r> aPHY is rgmii
maz_ has quit [Ping timeout: 272 seconds]
Gerwin_J has quit [Quit: Gerwin_J]
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<arokux2> eth0 Speed: 10Mbps, Mode: Full duplex, Loopback: NO
<arokux2> oliv3r: there? why only 10Mbps? ^
nove has quit [Quit: nove]
geecko has quit [Quit: Quitte]
Turl has quit [Quit: >.<]
pfdm has quit [Ping timeout: 250 seconds]
arokux2 has quit [Remote host closed the connection]
deasy has joined #linux-sunxi
enrico_ has quit [Quit: Bye]