mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
tuliom has quit [Quit: Konversation terminated!]
tuliom has joined #arm-netbook
popolon has quit [Quit: Quitte]
<drachensun> I'm trying to install these mali libs and I dont understand something. The instructions list mali_drv.so which I think is what X needs
<drachensun> but its not in the binary downloads
<drachensun> is that only a r3pX thing?
<drachensun> rm: You mean because debian doesn't have the ...eabihf cross compile package? You can get the code sourcery package, I think that will build it just as well
<drachensun> rm: seems like if you built the kernel with an hf compiler it still worked fine with a hf rootfs
<drachensun> nm find the info on mali_drv
<Turl> rm: is there such a thing as an 'armhf kernel'?
<Turl> rm: it was my belief the kernel didn't do any floating point operations - I might be plain wrong though
CaCtus491 has quit [Quit: leaving]
tuliom has quit [Quit: Konversation terminated!]
CaCtus491 has joined #arm-netbook
CaCtus491 has quit [Client Quit]
CaCtus491 has joined #arm-netbook
tekzilla has quit [Ping timeout: 260 seconds]
tekzilla has joined #arm-netbook
mysteryname_uni has joined #arm-netbook
mSquare has joined #arm-netbook
xxiao has quit [Quit: WeeChat 0.3.7]
Kraln has joined #arm-netbook
Kraln- has quit [Ping timeout: 264 seconds]
avernos has joined #arm-netbook
avernos has quit [Changing host]
avernos has joined #arm-netbook
drachensun has quit [Quit: Leaving]
Kraln- has joined #arm-netbook
Kraln has quit [Ping timeout: 260 seconds]
Brandon15811 has quit [Read error: Operation timed out]
Brandon15811_ has quit [Read error: Operation timed out]
Brandon15811_ has joined #arm-netbook
Brandon15811 has joined #arm-netbook
gimli has joined #arm-netbook
mysteryname_uni has quit [Remote host closed the connection]
rellla has joined #arm-netbook
rellla has quit [Quit: rellla]
Quarx has joined #arm-netbook
arete74 has quit [Read error: Connection reset by peer]
merbzt has joined #arm-netbook
dfletcher_ has joined #arm-netbook
acerimmer has quit [Ping timeout: 245 seconds]
<rz2k> hm
<rz2k> looks like odroid-x doesnt have hw3d on native linux
<rz2k> I thought they had mali400 working.
<rm> Turl, I thought the ABI also differs
<rm> register usage conventions and the like
<rm> or maybe that only applies to MIPS o32/n32, but not here
mysteryname has joined #arm-netbook
popolon has joined #arm-netbook
rellla has joined #arm-netbook
rellla has quit [Remote host closed the connection]
rellla has joined #arm-netbook
rellla has quit [Quit: rellla]
rellla has joined #arm-netbook
rellla has quit [Client Quit]
popolon has quit [Quit: Quitte]
mSquare has quit [Quit: Leaving.]
penguin42 has joined #arm-netbook
tuliom has joined #arm-netbook
<mnemoc> rm: the kernel is float-free
gimli has quit [Remote host closed the connection]
tuliom_ has joined #arm-netbook
tuliom has quit [Ping timeout: 245 seconds]
plan_b has joined #arm-netbook
alcides` has joined #arm-netbook
<techn_> mnemoc: but there is divisions so hf has some effect
<mnemoc> doubles :)
alcides has quit [Ping timeout: 244 seconds]
<mnemoc> uhm... actually not sure about that one. 1m
<mnemoc> no doubles either. everything is ULL
<mnemoc> but it would be nice too compare the results of using arm-linux-gnueabi- vs arm-linux-gnueabihf- ....
<mnemoc> s/too/to/
<ibot> mnemoc meant: but it would be nice to compare the results of using arm-linux-gnueabi- vs arm-linux-gnueabihf- ....
<plan_b> does the latest linux-sunxi-3.4 compile with sun4i_defconfig?
<mnemoc> sun4i_defconfig hasn't been "updated" yet
<mnemoc> none of the _defconfigs actually
<plan_b> ok ... thanks
<mnemoc> plan_b: I'm using https://filetea.me/t1s6fbba if you want to use it as start point
<plan_b> mnemoc: I will try this config ... thank you
<mnemoc> btw, tagged sunxi-v3.4.5-r2 this morning. going to test 3.4.12 now
tuliom_ is now known as tuliom
mysteryname has quit [Read error: Connection reset by peer]
popolon has joined #arm-netbook
tuliom is now known as tuliom_
ZaEarl has quit [Read error: Connection reset by peer]
tuliom_ has quit [Ping timeout: 246 seconds]
rsalveti has quit [Ping timeout: 240 seconds]
plan_b has quit [Quit: plan_b]
tuliom_ has joined #arm-netbook
<Turl> mnemoc: did my email arrive @ ML?
<mnemoc> Turl: no, didn't.
* mnemoc sending a test mail to dev@
<Turl> send it from an @linux-sunxi address
<mnemoc> ;linux-sunxi.org.INMX
<mnemoc> wth
<Turl> linux-sunxi.org.3600INMX10 mx0.pointhq.com.
<mnemoc> yes, just re-added them
<mnemoc> no clue why it got removed
<Turl> wonder if it's gonna retry sending it
<Turl> I didn't get any mailer errors, but then 'from' was on linux-sunxi too :P
<mnemoc> good. my test arrived
<mnemoc> Turl: can you resend the patch?
<Turl> mnemoc: sure
<mnemoc> the lack of a bounce confuses me
<Turl> mnemoc: there was no MX to bounce it either :P
<Turl> so bounce bounced
<mnemoc> :D
<mnemoc> good point
<Turl> mnemoc: sent
<mnemoc> arrived to moderation, but looks broken
<mnemoc> made pass
<Turl> how is it broken? (other than google saying 'unknown' o.O)
<mnemoc> in the moderation interface it was shown with double interlining
<Turl> #blamegoogle
<mnemoc> :)
<mnemoc> the delivered and the archived mails look fine
* Turl goes back to style-cleaning wemac :)
<mnemoc> Turl: did you test my 3.4 .config?
<Turl> mnemoc: heh, the bounced email arrived now :P
<mnemoc> :)
<Turl> nope, got a link handy? :)
<mnemoc> it was a temporal share :|
<mnemoc> 1m
<mnemoc> amery@cubieboard:~$ uname -a
<mnemoc> Linux cubieboard 3.4.12+ #120 PREEMPT Sat Oct 6 14:44:57 CEST 2012 armv7l armv7l armv7l GNU/Linux
<mnemoc> :)
<Turl> got it :)
<mnemoc> please try diff'ing both
<Turl> I nuked usb altogether on my config though to try 3.4 :P
<Turl> mnemoc: oldconfig'ing yours prints warning: (DRM && ION) selects DMA_SHARED_BUFFER which has unmet direct dependencies (EXPERIMENTAL)
<mnemoc> yes, should enable it
<mnemoc> stupid paste.ubuntu.com.... they want me to login to be able to see the plain text version
<mnemoc> and that obviusly doesn't work with curl
<mnemoc> Turl: http://sprunge.us/ <--- much better
<mnemoc> diff ... | sprunge -> url :)
<mnemoc> and syntax highlighting is dynamic, using ?foo
<Turl> yeah ubuntu paste sux
<Turl> but it's the default for |pastebinit on ubuntu
<mnemoc> doh, the original post arrived now
<mnemoc> #!/bin/sh
<mnemoc> exec curl -F 'sprunge=<-' http://sprunge.us
<mnemoc> make your own |pasteit
<mnemoc> simplest pastebin ever
Brandon15811 has quit [Excess Flood]
avernos has quit [Read error: No route to host]
Brandon15811 has joined #arm-netbook
<Turl> alias sprunge="curl -F 'sprunge=<-' http://sprunge.us" :)
<mnemoc> yes... that works too :p
lerc has quit [Ping timeout: 252 seconds]
lerc_ has joined #arm-netbook
Undertasker has joined #arm-netbook
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
drachensun has joined #arm-netbook
Quarx has quit []
tuliom_ is now known as tuliom
alcides` has quit [Quit: Fighter by day, lover by night, drunkard by choice! Ready to fight!]
<Turl> mnemoc: $ uname -a
<Turl> Linux mele 3.4.5+ #34 PREEMPT Sat Oct 6 17:39:21 UTC 2012 armv7l GNU/Linux :)
<Turl> mnemoc: your config doesn't have IPv6 :(
tuliom is now known as tuliom_
<rm> boo~ :)
<rm> Turl, so other than that, does everything work in 3.4?
<rm> how does next_mali relate to that, can I use 3.4 with the updated mali driver?
<Turl> rm: well I'm running a minimalistic server-ish mele, can't comment on anything other than it boots and wemac+usb host works :P
<rm> well that's already a good start! :D
<Turl> rm: mnemoc mentioned the android usb gadget wasn't ported/working/something though
<Turl> supposing the rest of the stuff is there it should work
<Turl> and you should be able to move the next_mali patches to 3.4 without much hassle either
<mnemoc> Turl: can you comment on the ticket?
<Turl> mnemoc: which one?
<Turl> [ 387.470000] >>> no handle, treat it handle over
<Turl> yay, arch broke
<Turl> WarheadsSE: ^
rsalveti has joined #arm-netbook
plan_b has joined #arm-netbook
<mnemoc> nah, that's sort of known
<Turl> mnemoc: it's spamming my serial terminal with that line over and over and over
<Turl> mnemoc: also, no idea if you noticed
<Turl> mnemoc: but if I'm booting from sdcard
<Turl> stuff fails
<Turl> I unplug mele's AC but leave serial connected
<Turl> it boots from nand and not SD
<Turl> after a reboot arch is back to working
<mnemoc> there are a couple of gremlins in that area
<zub> w.r.t. kernel - allwinner-v3.0-android-v2 is the recommended branch?
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
<zub> should CONFIG_SPARSE_IRQ be enabled for sun4i? defconfig has it on
<Turl> zub: if you plan on using it for something, I'd recommend it, yes
<zub> hm, I don't really understand what it does. I turned it off, and my latest build dies after several hours, so I'm wondering if it could be related
<zub> anyway I leave it on
<Turl> I run with it off I think
<zub> what exactly does SUNXI_IGNORE_ATAG_MEM do?
<Turl> it ignores the ATAG that indicates memory size
<zub> hm, and what's "ATAG"?
dfletcher__ has joined #arm-netbook
<Turl> zub: it's a bunch of configuration stuff the bootloader puts on RAM for linux
<zub> so if the ATAG is wrong on my board, it's caused by my wrong setup? does uboot put it there? is it this "other" file uboot loads to RAM?
dfletcher_ has quit [Ping timeout: 246 seconds]
<Turl> the proprietary uboot that ships on the stock firmware to boot android from NAND doesn't offer a proper mem ATAG
<Turl> and allwinner's kernel guesses the RAM instead of relying on the ATAG info
<rm> zub, did you try my kernels or configs?
<zub> rm: nope, but I might be interested in your configs
<rm> http://media.romanrm.ru/a10/server/3.0.39-rm2/ works well for me on mk802
<zub> where can I get them?
<rm> no graphics
<zub> ok
<zub> yeah, I'm running headless, at least for now. but I'd like something stable
<zub> I've been having periodical WiFi disconnects, and occasional system freezes :(
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
acfrazier has quit [Quit: ZNC - http://znc.sourceforge.net]
sspiff has joined #arm-netbook
acfrazier has joined #arm-netbook
acfrazier has quit [Excess Flood]
acfrazier has joined #arm-netbook
revident has joined #arm-netbook
<plan_b> rm: I'm not at home at the moment so I can not test your kernel config ... do you have the full 1024 MB ramd with this config?
<rm> no
<rm> I only have Mem: 443464
<plan_b> ah .. you have a 512 MB device
<rm> actually I have two MK802s which are supposed to be 1GB
<rm> one only sees 512MB even in Android
<plan_b> strange
<rm> and this one also sees only 443MB with my kernel, but I did not check if it can access the full GB in android
<rm> toggling SUNXI_IGNORE_ATAG_MEM does not help
<rm> it was $33 for me on a sale and I don't need a lot of ram; so not overly bothered by this
<plan_b> I saw a config (don't know where) with the full 1024 MB on a mk802+
<plan_b> I have removed all the X11 stuff, framebuffer and much more ... but still not 1024 MB on my mk802
<Turl> zub: the realtek chips these things run with aren't too good
<Turl> (wifi)
<rm> plan_b, there was a patch recently on the mailing list, "reserve framebuffer only if mali is enabled"
<rm> could be what you need
<plan_b> Turl: is the chip not good or is the antenna of the mk802 bad?
<Turl> plan_b: no idea, I just know the device usually drops the connection or is apparently connected but traffic doesn't flow
<Turl> mnemoc, hno does uboot pass memory info on an atag?
Brandon15811__ has joined #arm-netbook
<rm> Turl, plan_b - afaik there are some fixes to the newer Realtek chipsets in the driver
<rm> in the newer kernel versions
<rm> so a thing to try when facing wifi issues would be that 3.4 kernel
<plan_b> rm: I have tested linux-sunxi-3.4 today ... I wasn't able to build the Realtek module ... a known problem
<Turl> plan_b: build the other driver, the mainline one
<plan_b> Turl: I will try when I'm back (tomorrow)
<plan_b> thanks for the hint
Brandon15811__ has quit [Excess Flood]
Brandon15811__ has joined #arm-netbook
<plan_b> oh .. nice .... sunxi: configs: add cubieboard_defconfig ... https://github.com/amery/linux-allwinner/commit/cbf16cd2e2ac9285aefbde95a358612596a1b195
<plan_b> I wish I could order one
<plan_b> bloody chinese holidays
IEFtm has joined #arm-netbook
<rm> looks like hipboi now takes pre-orders
<rm> > Ships out within 60 days
IEF has quit [*.net *.split]
NAiL has quit [*.net *.split]
projectgus has quit [*.net *.split]
libv has quit [*.net *.split]
Brandon15811_ has quit [*.net *.split]
captainigloo has quit [*.net *.split]
cat1 has quit [*.net *.split]
L84Supper has quit [*.net *.split]
traeak has quit [*.net *.split]
IEFtm is now known as IEF
libv has joined #arm-netbook
techn_ has quit [Read error: Connection reset by peer]
NAiL__ has joined #arm-netbook
Brandon15811_ has joined #arm-netbook
captainigloo has joined #arm-netbook
cat1 has joined #arm-netbook
L84Supper has joined #arm-netbook
traeak has joined #arm-netbook
techn has joined #arm-netbook
<plan_b> rm: within 60 days ... that's the problem ... you don't know when you will get it
<hno> Turl, yes.
<mnemoc> Turl: u-boot passes the info. but it might need special configuration to find your memory correctly
<Turl> the issue I'm seeing is that, on kexec, it ignores atags completely
<Turl> on a normal boot they're read correctly and dumping to /proc/atags works
<hno> Turl, on kexec you need to set up the atag.
libv has quit [*.net *.split]
libv has joined #arm-netbook
<rm> plan_b, but I don't think this is because of the holidays
<Turl> hno: doesn't the kexec tool copy them for me?
<hno> No idea.
<rm> he could be just ordering them from the factory according to the demand
<RaYmAn> it should - are you using regular kexex or kexec-hardboot?
<Turl> RaYmAn: regular one
<Turl> and /proc/atags has the atags
<RaYmAn> then yes, it should do it for you
<RaYmAn> however, allwinner is weird, so it might not place them the right palce?
<RaYmAn> place*
<hno> Allwinner likes hardcoding stuff.
<hno> such as the address of the atag.
RITRedbeard has quit [Read error: Connection reset by peer]
RITRedbeard has joined #arm-netbook
RITRedbeard_ has joined #arm-netbook
RITRedbeard has quit [Ping timeout: 246 seconds]
<hno> Turl, kind of.. except that allwinner hardcodes r1 & r2 in arch/arm/kernel/head.S
<hno> same effect.
<hno> mnemoc, can we please kill that shit?
revident has quit [Quit: Combustible lemons? Bah, I bring you weaponized asparagus!]
plan_b has quit [Quit: plan_b]
<mnemoc> hno: do we need the asm reserves in the same file (io and script.bin) even when we have proper iomaps?
gimli has joined #arm-netbook
<mnemoc> i don't understand how those work at all :|
<hno> mnemoc, the reserve in there is for early access. The only early i/o access is the dramc memory probe.
<hno> don't remember what accessed script.bin early.
* hno is happy tonight. PMU code seems to work.
<mnemoc> \o/
<hno> now only need to beef up the CPU clock again.
<mnemoc> :)
<hno> and verify sun4i support. havent tried that in a long time.
<mnemoc> the version i run dies on reboot... but it's almost a week old
<hno> do you know which git version?
<mnemoc> 99cf18c86319cf66e088ffda30d2d27c565cab27
<mnemoc> Mon Oct 1 03:24:59 2012 +0200
<hno> mnemoc, good. There is no dram or clock changes after that. Only I2C and PMU.
<mnemoc> there is a simple captcha.. and even when bots register, very few write something...
<mnemoc> hno: I have it modified to use cubieboards' dram_para
<mnemoc> not sure if that matter
<mnemoc> Turl: do we really want to add the complexity (for humans) of a harder captcha?
<hno> mnemoc, did anything work before you changed dram parameters?
<mnemoc> hno: nope, it crashed once and the freeze in the second try
<mnemoc> then*
<hno> ok. because I tried to jtag boot the SPL on my cubie and it hang very early in dram setup.
<Turl> mnemoc: I don't care about the bots, as long as they don't spam all the things :)
<mnemoc> unless i've missed something, they don't spam
<hno> but did not spend much time with it.
<hno> can you send a pull request with your dram parameter cleanup?
<hno> mnemoc, so why do they register
<mnemoc> simple captcha is a plain text question
<mnemoc> some are smart enough to answer it
<mnemoc> hno: http://sprunge.us/CiaV <--- my diff..... but I don't know how to mange the config system to let that board_foo.c thing work properly :|
RITRedbeard_ has quit [Read error: Connection reset by peer]
<mnemoc> i mean, integrated in Makefile
<mnemoc> append ?diff to the url to get syntax highlighting
<mnemoc> or you want the pull request to have a place for discussing it?
RITRedbeard has joined #arm-netbook
gimli has quit [Quit: Verlassend]
Undertasker has quit [Remote host closed the connection]
<Turl> mnemoc: can you add arbitrary plaintext questions?
<mnemoc> don't know :|
<mnemoc> i guess so
<mnemoc> but i'm kind of alergic to php so i do as little customization as possible
<Turl> we could add some tech questions so bots can't join as easy
<Turl> some trivial logic maybe, true xor true = ?
<Turl> or "Is the sunxi family a x86 processor family?"
<mnemoc> you have `sudo` powers :)
<mnemoc> cd /srv/http/linux-sunxi.org/wiki/
<mnemoc> beware the wiki and all extensions are updated by cron
<Turl> updated by cron?
<mnemoc> there is an update.sh there, which is called at *:42
<Marex> hno: hey ;-)
<Marex> hno: any news on the upstreaming ?
<Marex> hno: -next is open for a while now
<mnemoc> Turl: it also updates the sitemap
<Turl> heh 42 :)
<mnemoc> :)
<mnemoc> Marex: there is still work to do before u-boot/sunxi becomes stable
<mnemoc> and mainlineable
mysteryname has joined #arm-netbook
rz2k has quit [Ping timeout: 245 seconds]
<Marex> mnemoc: what work ?
<hno> Marex, mainly SPL related stuff.
<Marex> whats up with SPL?
<Marex> still the memory init ?
<hno> No, was missing PMU setup so CPU ran on wrong voltage and some other details.
itamarjp has joined #arm-netbook
<Marex> bah
<hno> basic PMU support landed tonight.
<Marex> hno: heh
<Marex> hno: I wish basic documentation would land ;-)
<hno> the PMU do have quite extensive documentation.
<Marex> hno: ooh, that's good
<hno> the document for the specific chip is chnieese, but another chip in the family have english docs, and they share all registers & pins.
<hno> have not figured out how they differ yet.
<Marex> hno: you really need to learn chineese :)
<hno> translation engines and friends fill that gap sufficiently well for reading.
<hno> writing & communicating is more difficult.
<hno> Marex, how long is the merge window?
<Marex> hno: I'd expect it open till november
<hno> ok. And then closed for a month or so?
<Marex> hno: next release is scheduled for january
<Marex> hno: but then, if you get basic support into current -next, you'll be able to patch it until -rc, which is around december
<hno> mnemoc, seems sun5i also have GPS. Confusing.
<Marex> hno: sun5i is A13 ?
<hno> Marex, understood.
<hno> Marex, A13 is in the sun5i family yes.
<mnemoc> hno: i also wonder if they really removed g2d
<hno> based on
<hno> ?
<mnemoc> the reserve is merely commented out
<mnemoc> also the mixer processor is the best document part in the a10 manual.... and by the registers it seems to be what we call g2d... kind of makes me think if it's something they would suddenly consider unworthy
<hno> mnemoc, g2d is erased from all sun5i headers & code. gps have slightly different & refined clock registers in sun5i.
<hno> GPS is most likely there. But have no idicatoin g2d/mixer is kept. As mentioned current Android do not make use of it at all.
<hno> remember that the g2d memory reservation was commented out in the later sun4i code drops as well.
<mnemoc> right
<hno> it's a pity as it liits the interest in investing efforts into a g2d exa driver.
<mnemoc> yes
<hno> allwinner don't care as all their customers are happy with mali.
projectgus has joined #arm-netbook
popolon has quit [Quit: Quitte]
merbzt has quit [Ping timeout: 255 seconds]