ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
torqu3e has quit [Ping timeout: 245 seconds]
torqu3e has joined #linux-sunxi
torqu3e has quit [Read error: Connection reset by peer]
torqu3e has joined #linux-sunxi
<drachensun> turl: For allwinner specifically for just in general? I used it for a previous embedded project
<Turl> in general I guess
<Turl> I built it for cortex-A8 but my init was dying with SIGILL
<Turl> with linaro toolchain
<Turl> I then did a generic_arm build with codesorcery toolchain and it worked
<Turl> I can live with generic_arm for my debug env so I guess I'll stick to it for the time being
<Turl> good night :)
tinti has joined #linux-sunxi
<drachensun> yeah when I was using it for an atmel chip we had a lot of random build failures that could be fixed by weird stuff like switching compilers then just never running clean again...
ZaEarl has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
tinti has quit [Quit: Leaving]
theborger is now known as y0gurt
hipboi has joined #linux-sunxi
titankiller has quit [Ping timeout: 256 seconds]
jochensp has quit [Ping timeout: 260 seconds]
jochensp has joined #linux-sunxi
hipboi has left #linux-sunxi ["Leaving"]
<mripard> Turl: as a matter of fact, I'm also a buildroot developper (probably not the most active one though)
rellla has joined #linux-sunxi
ZaEarl has quit [Ping timeout: 264 seconds]
rellla2 has joined #linux-sunxi
rellla3 has joined #linux-sunxi
rellla has quit [Ping timeout: 256 seconds]
rellla3 is now known as rellla
rellla2 has quit [Remote host closed the connection]
n01 has joined #linux-sunxi
hurtigbuffer is now known as jelly-home
<oliv3r> mnemoc: I found that my script.fex is the same as 'inet97f-ii.fex'. Very similllar anyway. The commit log does say it's for the 'Bora' Tab.
<oliv3r> my file actually even has some chinese comments left behind ;)
<mnemoc> many livesuit images come with the .fex template of the SDK
<oliv3r> anyway, not sure how to proceed now. I would say, that inet97f-ii would be the 'original' as that is the name of the fex (inet97f isn't used anywhere on the wiki)
<mnemoc> but you can't trust there is any similarity with the actual .bin
<oliv3r> yes, my next point is indeed, that the script.fex is a default for 'PAD application'. It is named as A10-EVB-V1.2, e.g. A10 evaluation board
<oliv3r> brb, $work
wingrime has joined #linux-sunxi
<oliv3r> mnemoc: well since they are all 'rebranded' having a similar fex makes sense. anyway to start with, I will compare the fex file more closely (I think even ram parameters could be the same, but that I need to verify)
<mnemoc> techn__: will you send the v2 of the CEC patch including copyright header?
<oliv3r> one major difference is the effinition of the GPIO's, mine adds <default><default> at the end of each port (I saw some talk abou thtat yesterday?) should that be ommited?
<oliv3r> and I'm thinking of renaming the BORA page to inet97f-ii, since that appears to be 'the mother' :p
<mnemoc> hramrach: what's the companion of '[linux-sunxi] [PATCH 2/2] Suspend framebuffer on sustem suspend.' ? can't find the 1/2 :<
<mnemoc> oliv3r: trailing <default> can be omited, yes
<oliv3r> mnemoc: can or should
<mnemoc> can
<oliv3r> what's better and why would I? :)
<mnemoc> it looks cleaner trimmed. BUT fexc needs to be improved to cut that out
<oliv3r> so there's no added value in keeping it in, so it's better to remove iiiiit?
<mnemoc> not sure if it's a good idea to make a style change for something the tool will destroy
<oliv3r> well inet97f-ii doesn't have it, mine does have it, so i'll have to use either/or :)
<mnemoc> I think the "minimize the diff" rule wins here
<oliv3r> then i'll remove it from my file
<mnemoc> ack
<oliv3r> btw, I can't find 1/2 either
<hramrach> mnemoc: you don't need. it applies on wingrimes' suspend V2
<mnemoc> ok. I'll merge those 2 first then
<mnemoc> no one wanting to fix usb gadget so we can deprecate 3.0? :<
hansg has joined #linux-sunxi
<mnemoc> specially now with sun6i/sun7i based on 3.3
<mnemoc> wb hansg
<oliv3r> btw, what is the cardscript.fex used for? seems like for boot0, but can't see how it relates to anything
<mnemoc> hramrach: wingrime's V2 2/2 doesn't apply . am I missing some other patch?
<oliv3r> or do several stages each rely on their own fex file?
<wingrime> mnemoc: you can drop 2/2 this is cleanup
<wingrime> mnemoc: but I tested on 3.0
<mnemoc> wingrime: so 1/2 of your V2 and hramrach's 2/2
<wingrime> yep
<wingrime> mnemoc: I do cleanup more one time on fresh cut
<wingrime> mnemoc: but I have to finish zet6221 driver
<wingrime> mnemoc: I now know finger conut on touch , now I must extract coordinates and send it to userspace , also do suspend/resume
<mnemoc> nice work :)
<mnemoc> you could then write a tutorial on how to write ts drivers ;-)
<mnemoc> that's the weakest point in sunxi openness currently
<mnemoc> as every manufacturer adds their own
<wingrime> mnemoc: ts i2c simply have registers you simply must dump registers and see what happend ,
<mnemoc> still, a good thing to tutorialize ;-)
* mnemoc annoyed by the dominance american-english has on his spelling
<oliv3r> heh
<oliv3r> the British did bring us american-english :p
<mnemoc> :)
<oliv3r> I think it's fair to say, the American's brought us world-wide use of english
<oliv3r> Hate or love the yanks, it's their english that's dominating :p
<wingrime> mnemoc: USA and britush so far from there
<wingrime> mnemoc: my english based on internets
<oliv3r> imagine if we wouldn't have ha WW2, we'd all most likly be communicating in german now!
* mnemoc learned english by reading programming books :|
* oliv3r tv
<oliv3r> :p
<mnemoc> oliv3r: .de ended up conquering europe anyway
<oliv3r> I remember watching 'The Super Channel' which was I think an american-british thing?
<oliv3r> yes, but it was because our recentment, and their lack of scientific progress after, that it hasn't dominated since
<oliv3r> German was the scientific language before WW2
<oliv3r> then again, the internet is an american invention I guess
<mnemoc> also, most germans speak english just fine. while most american can hardly speak their own language
<oliv3r> lol, true
<oliv3r> is there any use of having more then 1 touch screen driver in a fex file?
<oliv3r> I guess it does make it more flexible
<oliv3r> 1 fex file for more devices, but I'd think it would make more sense to have 1 fex per deviceconfiguration?
<oliv3r> ctp_para does allow up to 4? definitions however
<mnemoc> probably there to make the life of the manufacturers easier
<wingrime> oliv3r, only when drivers patched
<wingrime> oliv3r, only one driver can have number
<wingrime> ft5x = 4
<wingrime> zte = 3
<wingrime> for example
<oliv3r> let me rephrase, how shall we deal with it :)
<oliv3r> wingrime: yeah, i see ctp_used, ctp1_used, ctp2_used, with each ctp having their own i^2 address and driver
<oliv3r> so 0 is ft5x, 1 is goodix-ts etc etc
<oliv3r> and while I can see 1 device having multiple touch screensont and rear, how hould we deal with it i sunxi-boards.git? fork a file (is that possible) per device?
<wingrime> oliv3r you can easy bring this feature addint number to CTP params in driver
<oliv3r> i know, I see this happening int hhe fexx file :p my question is, how shall we handle this in sunxi-board.git repostiory. Do we want 'general' files that are as common as possible (e.g. have 4 or more ctp parameters that define the touch screens) or have only what that board is supposed to have.
<oliv3r> to be honest though, while the latter sounds logical and 'better', with different spins (combinations) having a different version number I guess? But isn't all this i2c stuff autodetected by the linux kernel anyway? So why would it be needed? u-boot doesn't use the TS anyway?
<oliv3r> so what is the point of having ctp_para if the kernel probes the i2c bus based on the installed driver.
<hramrach> wingrime: mnemoc: no, the 2/2 is not cleanup. It is the sole patch changing pm.c with the cleanup and fixup intermixed. And yes, it does not apply for me either.
<hramrach> the event is assigned in aw_pm_enter which is not patched in 1/2 V2
<mnemoc> hramrach: ow
<mnemoc> hramrach: and you submit a v2 of 2/2?
<mnemoc> err.... v3
<mnemoc> s/and/can/
* mnemoc goes for another cup of coffee
<wingrime> hramrach: you not right my 1/2 actualy ADD resume event
* mnemoc scratches his head
<wingrime> hramrach my 2/2 also add printing resume reason in dmesg
<wingrime> hramrach: also you can look up your driver and move some earlysuspend to simple suspend
<wingrime> as early suspend must be disabled on debian
<wingrime> i mean fix #ifdef some where
<hramrach> wingrime: 2/2 also changes what wakeup event mask is assingned in aw_pm_enter which is probably what makes the device wake up
<wingrime> hramrach: look file carefully
<hramrach> well, that's what I see but it's really hard to navigate that bit pile of a patch
<hramrach> ok, it will probably work
<mnemoc> i'm try to finish and push the sanitized variant of lichee/a10-dev today
<mnemoc> s/i'm/i'll/
<mnemoc> it makes diffs saner to read
<oliv3r> awesome :D
<mnemoc> I'm stuck merging some awful changes in drivers/net/usb/asix.c :<
<oliv3r> anybody know what: "blue and black" pins are in the [target] section on the FEX guide wiki?
<oliv3r> mnemoc: aweful code writing style like, or aweful actual technical content.
<mnemoc> content
<mnemoc> they updated the driver with vendor code, but mainline has kept improving the old driver
<mnemoc> but without the original vendor code I can't know what was "tweaked" by allwinner and what by asix
<mnemoc> i'm already immune to their stick development and coding style
<mnemoc> s/stick/sick/
<oliv3r> lucky you :p
<oliv3r> well atleast we have more code to improve the driver with (hopefully) :D (always look at the bright side of things?)
paulk-desktop has joined #linux-sunxi
andoma_ has joined #linux-sunxi
hramrach has quit [*.net *.split]
andoma has quit [*.net *.split]
ganbold_ has joined #linux-sunxi
andoma_ is now known as andoma
<oliv3r> mnemoc: so they use an AXIS IP core for their ethernet? I thought it may be davicom, so this shows it's likly axis?
<mnemoc> oliv3r: nah, it's part of the commit that adds drivers/net/usb/qf9700.c
<mnemoc> i assume it's for a vendor
<oliv3r> ah, USB driver
<oliv3r> yeah, axis is a USB mac vendor amongst things, I have a few axis USB dongles
<oliv3r> well USB drivers probably can be ignored evenutally, as those need to be mainlined seperatly anyhow
<oliv3r> :D
<mnemoc> oliv3r: for the purpose of the import/lichee/* branches, I need to preserve it all
paulk-desktop has quit [Read error: Connection reset by peer]
paulk-desktop has joined #linux-sunxi
<oliv3r> mnemoc: i know, i was just saying, luckly, it's something in the end, needs to be less worrid about
<oliv3r> seriously though, all these chip manufactures need to watch GKH's linaro video, the first 11 minutes are REALLY good for them. they should get subbed in chinese and virally spread there :)
<mnemoc> why do you think linaro makes connect in .hk? :)
<oliv3r> right, taking a little break from edting http://linux-sunxi.org/Fex_Guide I took a broom through the first half. If anbyody could be so kind and proof read/glance over it? i stopped at ic2
<Turl> got a link oliv3r?
<oliv3r> er, didn't know they where doing it :p I do know GKH was tellin' linaro devs!
<oliv3r> Turl: its on cnx-software, but i'll link it here :)
<oliv3r> linaro connect HK ... HK ... d'oh
<oliv3r> we all know this naturally, nothing new there
<oliv3r> but it was a really good talk, he's a pretty ok speaker
ganbold__ has joined #linux-sunxi
rz2k has joined #linux-sunxi
ganbold_ has quit [Remote host closed the connection]
ZaEarl has joined #linux-sunxi
torqu3e has quit [Read error: Connection reset by peer]
torqu3e has joined #linux-sunxi
gzamboni has joined #linux-sunxi
hansg has quit [Quit: Leaving]
bfree__ is now known as bfree
ganbold__ has quit [Remote host closed the connection]
<oliv3r> right, I cleaned up the Fex_Guide up to, excluding locks_para
<oliv3r> if someone could take a look, that'd be much appreciated
ZaEarl has quit [Ping timeout: 264 seconds]
<oliv3r> (and change anything that could be better :p)
jukivil1 has joined #linux-sunxi
jukivili has quit [Ping timeout: 264 seconds]
<mnemoc> ~seen hramrach
techn_ has joined #linux-sunxi
ZaEarl has joined #linux-sunxi
techn__ has quit [Ping timeout: 264 seconds]
<rz2k> anyone played with sunxi watchdog?
<mnemoc> hno :)
<rz2k> oh
<rz2k> I should read drivers before asking about them :/
<wingrime> mnemoc:techn: are you tryed do suspend ?
<mnemoc> wingrime: no. haven't had time to play in weeks :<
<wingrime> mnemoc: I waana do something with battarey suspem , linaro not show right procentage
<rellla> wow. hramrach needs glass imo -> http://linux-sunxi.org/Mali400
<rz2k> lol
<mnemoc> ow
<rellla> done
<libv> hah, the dual core mali thing was also by him
<libv> *sigh*
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
<Turl> rz2k: I used it too
ZaEarl has quit [Ping timeout: 264 seconds]
provel has joined #linux-sunxi
<rm> I hate doing "make install" in a system
<rm> there is an expression, "turing your system into slackware"
rellla has joined #linux-sunxi
<rm> that's what this does
<provel> Hi! Do the hdmi CEC feature still needs testing ?
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
plan_b has joined #linux-sunxi
<rm> oh, s/turing/turning/
carlo_ has joined #linux-sunxi
n01 has quit [Disconnected by services]
carlo_ is now known as n01
carlo_ has joined #linux-sunxi
<techn_> mnemoc: I'll send that copyright to separate thread.. That design needs also fixes (dependency problem)
<mnemoc> techn_: so I apply 1-13/14?
<techn_> provel: Testing welcome.. but currently it needs most of those other patches
<techn_> mnemoc: yes please.. unless there is some comments to fix :)
<techn_> provel: but if mnemoc applies those 1-13 of that serie.. 14 should apply fine :)
<mripard> Turl: regarding your question about buildroot from yesterday, Linaro toolchains are pretty well-known to be buggy
<mripard> I only had troubles with them.
<techn_> wingrime: Booting.. or should I enable some config for it?
<wingrime> techn: enable suspend options in make menuconfig exept wakeloks
<wingrime> *locks
<wingrime> then
<wingrime> echo "mem" > /sys/power/state
<wingrime> for suspend
<techn_> wingrime: it worked with default sun4i_defconfig
<techn_> wingrime: but only power button made device to resume
<wingrime> techn: what wake sources you except?
<techn_> I tried usb keyboard, serial console and power button
<techn_> only power button resumed
<wingrime> serial can
<wingrime> no_console_suspend
<techn_> ah
<wingrime> in cmdline for it
<techn_> how about usb resume?
<wingrime> it not implemented
<wingrime> only studs
<wingrime> and more one, USB powered off
<wingrime> so it best, to save power
<wingrime> also , I can add suport UART WAKE up becose uart have IRQ
<techn_> sure.. this is better than nothing :)
<techn_> but device like mini-x.. resume on power button only is bad thing :p
<techn_> mini-x doesnt have any buttons :)
<wingrime> just add no_console_suspend ))
<wingrime> and also LRADC
<wingrime> for me I can wakeup using volume keys (LRADC)
<wingrime> also, as I seen in code , posible wakeup by RTC, Some power cases (low batt??)
<mnemoc> techn_: pushed
<techn_> mnemoc: thanks
<techn_> mnemoc: I'll try implement that https://github.com/ffainelli/linux-hdmi-cec API and resend it again.. but then it propably needs some test app
<mnemoc> techn_: can't it be hooked inside some other already-existant system instead of implementing cec as itself?
<techn_> mnemoc: ?
<mnemoc> acpi?
<techn_> ah.. it is connected to suspend/resume logic in that patch
<techn_> so that patch is testable.. but it is done ugly - disp wont work without hdmi :(
<oliv3r> techn_: I only know that that is some RFC, so it can change at any point. oth it's all we have right now
<techn_> oliv3r: you mean that v4l api proposal?
<oliv3r> cec v4l thing yeah
<techn_> hopefully that wont go trough.. and most likely wont, it is 2,5 years old
<oliv3r> it's a good start for sure
<oliv3r> hmm, that e-mail fro the ML isn't that old i don't think
<techn_> there was other discussion.. 5 months ago.. but it dryed up
<oliv3r> get in touch with them and see what they say
<mnemoc> techn_: I don't know much about CEC or ACPI, but if we can hook CEC as ACPI userspace support would be free....
<techn_> mnemoc: for me, most improtant thing with CEC is full remote control support. Not only power control :)
<mnemoc> iirc ACPI does deal with buttons too
<mnemoc> and many hw events are implemented as buttons
<mnemoc> but it could be all part of my imagination
<techn_> mnemoc: please find reference.. I have no idea what I'm looking for :D
<techn_> Hans Verkuil propably knows latest status
<mnemoc> event -> command. closing the lid of a laptop is an acpi "button"
<techn_> mnemoc: or /dev/lirc
<mnemoc> techn_: http://lxr.free-electrons.com/source/drivers/acpi/ .... there are tons of drivers there, for very assorted things. not only power, but also detected events
<n01> hi guys, anyone working on MTD?
<mnemoc> n01: the previous team deserted the job
<n01> gh... why?
<mnemoc> lack of documentation and time
<n01> yeah, documentation sucks
<mnemoc> but they managed to read the raw nand using dma
<n01> is there a branch or something out there?
<mnemoc> hno knows better, but he is very busy with $work$
<n01> uhm, ok, tnx for the info
<oliv3r> techn_: what's the title of that page?
<wingrime> I wanna kill that chinse that packed x y in 3 bytes in such sadistic way
<techn_> oliv3r: ?
<oliv3r> the poronix page :p
<techn_> oliv3r: "Linux Support For HDMI CEC Still In Development" , October 08, 2012
plan_b has quit [Quit: plan_b]
plan_b has joined #linux-sunxi
<oliv3r> ah, that's the mean ;) i'll look at that late rthen
<oliv3r> btw, rob clarke is awesome :) with his driver being merged into mesa now
calris has joined #linux-sunxi
<calris> Anyone running Fedora 18 image on a Merle?
<calris> Wondering what kind of performance to expect - mine seems really slow
plan_b has quit [Client Quit]
plan_b has joined #linux-sunxi
<mnemoc> calris: first of all check the cpu frequence
<calris> How?
<mnemoc> /proc/cpuinfo is a good start
<calris> Yeah, found that - no freq, but 405.52 BogoMIPS
<mnemoc> that's mostly equivalent to 400MHz
<mnemoc> much better than the 60MHz most people gets by default
<mnemoc> techn_: thanks for the pointer to that thread
<calris> Shouldn't it run more like 1GHz?
<provel> techn_: thank you, I'll wait for all the 14 patches then
<techn_> mnemoc: ? :)
<mnemoc> idling -> power saving -> frequency goes down
<mnemoc> techn_: cec thing
<calris> Ah :)
<techn_> provel: mnemoc applied 1-13 of that patch series.. now you'll need to patch only that latest
<provel> techn_: If I understood, cec patch will only drive standby, still we need to adatp libcec.... Am I right?
<techn_> latest -> last :)
<calris> X seems to run like a dog (xfce) :(
<techn_> provel: CEC implementation in that patch wont give any user space API.. It just powers of display when device goes to suspend.. and powers on when device resumes
<ssvb_> calris: 1920x1080 screen resolution?
<calris> Yeah - HDMI LCD monitor
<provel> techn_ doh.... ok, but it has some hook to use with userspace API?
<provel> I can adapt libcec, I hope
<techn_> provel: I'm currently working on it.. but what I like is to get some testers feedback of that patch
<techn_> .. I'm working on that user space API
<provel> for what I've seen it's all in 3 units in libcec
<provel> oh ok..... but I don't have suspend in my box..... or I never used
<provel> how can I test ?
<techn_> provel: suspend support came 14 patches ago ;)
<techn_> thanks wingrime and hramrack
<calris> I thought getting Mali drivers working would help, but I've been told it won't
<techn_> <wingrime> echo "mem" > /sys/power/state
<techn_> <wingrime> for suspend
<provel> techn_ : and power button to resume ?
<techn_> provel: yes
<provel> ok
<calris> Yep
<ssvb_> calris: ok, I just confirmed this problem a few hours ago :)
<ssvb_> calris: do you have 512MB or 1G RAM in Mele A2000?
<provel> techn_ : ok, downloading stage/3.4 right now
<mnemoc> provel: don't "download", clone it
<provel> errrr.... I just checkouted from plain 3.4
<mnemoc> git checkout -b stage/sunxi-3.4 origin/stage/sunxi-3.4
<provel> yep
<provel> hum, something weird, had to reset --hard
<jelly-home> provel: which brand tv or other cec hw do you have btw?
<provel> LG
<provel> LG + simplink
<wingrime> so. ts buggy , stange , stop working by unknown reason after some time (irq not recives....) , one touch, but works!
<calris> 512
<provel> I need inverse function as LG remote has same freq as mele a2000 remote
<provel> so cec would be awesome
<ssvb_> calris: perfect, that's the same hardware as mine :) I'm going to check something and send a reply to the mailing list
torqu3e has quit [Ping timeout: 245 seconds]
<provel> lots of conflict, it will take longer
torqu3e has joined #linux-sunxi
<calris> Thanks :-)
<mnemoc> provel: if you don't have local changes just git reset --hard origin/stage/sunxi-3.4
<provel> mnemoc : did that but I think I had local changes
<provel> better a clean test then....
rz2k has quit [Ping timeout: 272 seconds]
<mnemoc> the stage branches are rebased, it's very easy to get conflicts
<mnemoc> for your own work make another branch. then, when the time of updating to stage branch comes, git-format-patch, reset and git am
<calris> ssvb_: any hints?
wingrime has quit [Read error: Operation timed out]
<provel> hum, patch 14 does not apply in a clean way to stage/sunxi-3.4
<provel> even with -F25
<techn_> provel: check end of the patch.. there could be garbage from gmail ML :/
<provel> no I cleaned that.... but I get some hunk failing and some succeeding
<provel> I'll try by hand
<mnemoc> provel: applied just fine here
<mnemoc> provel: try curl http://sprunge.us/KgNW | git am
<mnemoc> without having to remove any "gmail garbage"
rellla has joined #linux-sunxi
hramrach has joined #linux-sunxi
n01 has quit [Ping timeout: 255 seconds]
<provel> mnemoc adapted by hand
<provel> I just noticed that git diff doesn't insert the new, unstaged files
traeak has joined #linux-sunxi
rellla has quit [Quit: Nettalk6 - www.ntalk.de]
plan_b has quit [Quit: plan_b]
torqu3e has quit [Read error: Connection reset by peer]
calris_ has joined #linux-sunxi
torqu3e has joined #linux-sunxi
torqu3e_ has joined #linux-sunxi
torqu3e has quit [Ping timeout: 245 seconds]
torqu3e_ is now known as torqu3e
<provel> techn_ : just tried CEC, it worked
<provel> but didn't turned off tv
<provel> if that was the purpose
<provel> do I need to check if I have any particular CONFIG option in kernel ?
<provel> however, for what I know, it's TV remote that is considered master and other devices slaves
<provel> Other issues: suspend triggers usb bug for which my external hd is now both sda1/2, sdb1/2 and sdc1/2
<provel> seems triggered by hd having 2 partitions
<provel> every time it's disconnected or suspended, dmesg advances, from sd 0:0:0:0 to sd 3:0:0:0 and beyond
provel has left #linux-sunxi [#linux-sunxi]
<theOzzieRat> mnemoc, I have resubmitted that spi patch can you let me know if I have submitted it correctly this time :)
<calris_> ssbv_: I just read your sunixfb article (http://ssvb.github.com/2013/02/01/new-xf86-video-sunxifb-ddx-driver.html)
<calris_> ssvb_: looks promising :)
paulk-desktop has quit [Quit: Ex-Chat]
torqu3e has quit [Read error: Connection reset by peer]
torqu3e has joined #linux-sunxi