hno changed the topic of #linux-sunxi to: /Allwinner/sunxi development discussion - Don't ask to ask. Just ask and wait! - See | | Logs at
naobsd has quit [Quit: Page closed]
mcbrick has quit [Quit: Leaving]
Gerwin_J has quit [Quit: Gerwin_J]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
ykchavan has joined #linux-sunxi
hramrach has joined #linux-sunxi
deasy has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
<steev> the only box i know of that has decent ram is the openblocks axX
<steev> and that's just because it has a sodimm slot
drachensun has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
hramrach has quit [Remote host closed the connection]
hramrach has joined #linux-sunxi
hansg has joined #linux-sunxi
soul has joined #linux-sunxi
soul is now known as Guest4430
Soru___ has quit [Ping timeout: 245 seconds]
atiti has joined #linux-sunxi
akaizen_ has joined #linux-sunxi
akaizen has quit [Ping timeout: 264 seconds]
Soru has joined #linux-sunxi
Guest4430 has quit [Ping timeout: 245 seconds]
rellla has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
<oliv3r> interesting factoid; it appears that A20 dropped the PATA controller. Instead, it put EMAC pinmux on PH mux2. e.g. 100mbit + gbit is possible in a dual setup
<mnemoc> :o
<mnemoc> but but but... do they use different registers too?
n01 has joined #linux-sunxi
huosen has joined #linux-sunxi
ynezz_ has joined #linux-sunxi
Soru has quit [Remote host closed the connection]
jlj_ has quit [*.net *.split]
WarheadsSE has quit [*.net *.split]
Amitabha has quit [*.net *.split]
ynezz has quit [*.net *.split]
<oliv3r> Turl: have you seen PI12 and PI13 on A20? checkout mux 5
<oliv3r> PI14 and PI15 might be interesting, not sure what those mean
WarheadsSE has joined #linux-sunxi
massi_ has joined #linux-sunxi
<oliv3r> mripard: the cubietech,cubietech thing was stupid :S
FR^2 has joined #linux-sunxi
rellla2 has joined #linux-sunxi
rellla2 has quit [Quit: Nettalk6 -]
panda84kde has joined #linux-sunxi
Soru has joined #linux-sunxi
jlj_ has joined #linux-sunxi
<oliv3r> has anybody ever built stage-3.4 with ahci enabled?
* mnemoc only builds defconfigs :(
<oliv3r> yeah ahci isn't in the defconfig
<oliv3r> and it fails
<mnemoc> fix it :)
<oliv3r> i will ;)
<mnemoc> \o/
popolon has joined #linux-sunxi
<oliv3r> and commit the patch immediatly too
<mnemoc> good
notmart has joined #linux-sunxi
notmart has quit [Changing host]
notmart has joined #linux-sunxi
<oliv3r> so buzz is making us a 3D case if I send him my CT?
<mnemoc> oliv3r: I'll push the sanitized source-drop/linux-3.4/a20-2.0 branch in a few
<mnemoc> E: modules/nand/libnand: application/x-archive; charset=binary
<mnemoc> obviusly :(
<oliv3r> gah
<oliv3r> well we have sources, but if taht's a different verseion from what we had, lkcl may have to get the source for it, again.
<oliv3r> btw, copared to the Cubieboard, the cubietruck is a beautiful board imo
<oliv3r> connector wise that is
<mnemoc> +1
<oliv3r> the cb was a little chaotic with ports on every side
<oliv3r> the CT could go into a set-top-box right away as is
<oliv3r> we just need to design a VGA -> Composite; VGA -> SVHS and VGA -> YUV extension set of cables
<mnemoc> that was the initial intention afaik
<oliv3r> since the RGB wires of the VGA port, can be pinmuxed as either/or
<mnemoc> including a VESA compatible case....
<mnemoc> but then radxa happened...
<oliv3r> what about radxa?
<mnemoc> tom moved to shenzhen
<oliv3r> ohh yeah
<oliv3r> Benn will have to relearn how to communicate with the community etc
hansg has quit [Quit: Leaving]
<mnemoc> yup
vicenteH has quit [Ping timeout: 240 seconds]
<oliv3r> what kinda bothers me is that it's labeled 'open mini-pc'
<oliv3r> how is it open atm really?
<mnemoc> and things that were agreed with tom got forgotten
<oliv3r> besides that it doesn't have a case
<mnemoc> oliv3r: software wise
<mnemoc> but well....
<oliv3r> but why print that on the PCB
<oliv3r> i can install wince on it, and how is it then an 'open mini-pc'
<mnemoc> marketing?
<oliv3r> lies :p
<oliv3r> well lets ask benn for the schematics next week
<mnemoc> oliv3r: why does it have an ISBN in the box bothers me even more
<oliv3r> LOL i noticed
<oliv3r> i was 'is this a book'?
<mnemoc> i just laughed
DigiDaz_ has quit [Remote host closed the connection]
<mnemoc> but it's nice packaging, nice board, etc...
jinzo has joined #linux-sunxi
<mnemoc> but yes, please bother benn about schematics etc
<oliv3r> yeah overal it's pretty good
<mnemoc> initial source-drop/linux-3.4/a20-2.0 up. will try to find the right android-3.4 base
<oliv3r> oh nice one
<oliv3r> you doing that from your phone?
<mnemoc> tethering :p
jemk has joined #linux-sunxi
<mnemoc> but it's mostly about writting and using scripts
<oliv3r> :p
<mnemoc> find * .gitignore .mailmap -type f -print0 | xargs -r0 ../
<mnemoc> git ls -o | xargs git add -f ... etc
Black_Horseman has quit [Remote host closed the connection]
<oliv3r> did we change mach/sys_config -> plat/sys_config or was it vice-versa?
<oliv3r> i see 2 in our git tree, sun7i uses mach; sunxi uses plat
<mnemoc> plat/ is ours
<mnemoc> part of the mach unification
<oliv3r> ok
<mnemoc> plat-sunxi is intended to help making mach-sunxi
<oliv3r> so we should include mach/sys_config.h in new code
<oliv3r> erm old code*
<oliv3r> since it simply can't find plat/sys_config.h
<oliv3r> maybe i shoudl do a git pull
<mnemoc> because Kconfig is wrong
<mnemoc> sun[67]i need to enable plat-sunxi too
<rellla> mnemoc: where is this drop from?
<oliv3r> rellla: A20 sdk
<oliv3r> rellla: it's old-ish
<mnemoc> rellla: olimex's a20 2.0 sdk
<mnemoc> from the last days of august
<mnemoc> oliv3r: just see how the other mach-sunNi do it
<mnemoc> but enable plat and minimize mach-sunNi
<rellla> ok. a few new things in there like i saw in
<oliv3r> well this is a defconfig, so shouldn't it include most stuff?
<mnemoc> sun7i defconfig is crap
<rellla> e.g. drivers/char/sunxi_mem
<oliv3r> mnemoc: so /silo/build/sunxi-bsp/linux-sunxi/drivers/input/input.c:106:7: warning: unused variable ‘not_gsensor’ is also from sun7i being crap?
<oliv3r> erm
<oliv3r> wrong paste
<oliv3r> linux-sunxi/drivers/hwmon/mma7660.c:39:25: fatal error: mach/system.h:
<oliv3r> that one
<oliv3r> so it's really that crap?
<oliv3r> ouch
<mnemoc> nah, it's aw vs sunxi code
<mnemoc> sunxi-3.4 uses plat, aw doesn't like to unify
<mnemoc> they prefer to copy&paste
<mnemoc> just change those mach/ to plat/
<oliv3r> yeah that i knew
<oliv3r> i think it IS play allready in the sw driver
<mnemoc> and remove/unify mach/ headers
<oliv3r> but i'll go over the config and see if there's useless crap enabled
<mnemoc> Kconfig should enable PLAT_, not defconfig
<oliv3r> yeah i added it there
<oliv3r> sunxi-pwm also gives a fatal error
<oliv3r> we deffo need a better defconfig :)
<oliv3r> even if it's just a 'quick-n-dirty' one
<mnemoc> rellla: we should import those into our 3.4
<mnemoc> iirc they use it in the new cedarx driver
<oliv3r> shouldn't the depend read (ARCH_SUN4I || 5i || 7i ) && ARCH_SUNXI?
<mnemoc> if the archs selects the sunxi, we only need to depend in the sunxi
<rellla> mnemoc: drivers/char/sunxi_mem is used by the new memory allocator, yes:
<mnemoc> oliv3r: making it closer to mach-sunxi
<oliv3r> gah, this is to messy to work with :p
<mnemoc> :(
<oliv3r> we need some devs who want to port drivers to 3.10;
<oliv3r> anyhow, i need a kernel to see if ahci works, and compare registers with
<rellla> i was building a kernel yesterday with sun7i-cedar and sunxi_mem - imported to stage/sunxi-3.4 from 3.3-cb2
<rellla> i'll check if it's the same in the drop ... and hopefully can test it...
<mnemoc> rellla: olimex's drop include 3.3 and 3.4
<oliv3r> gah i hate all these 'person X's kernel'
<mnemoc> please feel free to push to stage commits you find worthy
<mnemoc> that's why it's a community repo and not a person X repo
<oliv3r> yeah
<oliv3r> but there's how many kernels that we know off? patwood kernel, rm kernel; any others?
<mnemoc> focus in the ML
atiti has quit [Remote host closed the connection]
<mnemoc> patwood has sent patches... rm rarely does
<oliv3r> we should get daily kernel builds out so we can offer a stable/stage kernel for download so that it atleast wil be know as the 'sunxi-kernel'
<oliv3r> kwown*
<oliv3r> known*
<rm> mnemoc, oliv3r, that's because there is no "community kernel" that people can download (binary) and just start using
<rm> and expecting everyone to build their own is unrealistic
<mnemoc> so let's solve that
<oliv3r> there's a kernel right there ;)
<oliv3r> but it's not updated since apr
<mnemoc> that's a bug it needs to be fixed
<oliv3r> we'll probably want nightly-stage and nightly-non-stage
<mnemoc> and 3.0, 3.4 and 3.10....
<oliv3r> hehe aye
<mnemoc> and for each defconfig
<oliv3r> aye
<mnemoc> but we first need good usable defconfigs!
<oliv3r> right, back to hacking ;)
<mnemoc> :p
<rellla> also i successfully built xbmc for a20 with the new sources and blob - but no test at all. first build a kernel with new cedar_dev. atm there is compilation running for xbmc for a10... let's see how well the new blobs work :p
<oliv3r> pray to the gods of cedarx
<rellla> lol
<mnemoc> I'll get the automated builds of each defconfig of each relevant branch tonight
<mnemoc> that might increase the changes of people helping to improve the defconfigs :p
<oliv3r> i'm sending a sun7i defconfig to ML in 10-60 minutes
atiti has joined #linux-sunxi
<Montjoie> Does the mali gpu have some public datasheet available ?
<popolon> on my latest test, in the sun7i_deconfig, CONFIG_SUNXI_EMAC=m could it a good idea to change it to y instead of m, isn't it ?
<rellla> ok. sun7i_cedar and sunxi_mem are (almost) complete identical in 3.3-cb2 and 3.4
<popolon> in emac_readme.txt : 2) using cpu to read/write fifo, not dma
<mnemoc> popolon: in those case the driver should have a depends on MYARCH; default m and y in the defconfig
<popolon> mnemoc, what is MYARCH ? Can't found it
vicenteH has joined #linux-sunxi
<popolon> do you mean depend on personal arch ?
<oliv3r> yep
<oliv3r> ARCH_SUN47I in this case
<popolon> ok, thanks
<popolon> perhaps lot of sun7i doesn't need it by default after all
<popolon> I didn't added this detail in the linux building tuto
<popolon> need to add it in the case of device with ethernet port
<mnemoc> or PLAT_SUNXI :)
<mnemoc> by default we should build (at least as m) everything sunxi specific.... what to make built-in instead of module is a more complex topic
eebrah|noPC has joined #linux-sunxi
<mnemoc> but sunxi-3.{0,4,10} aren't mainline... so we can have useful/ready-to-use defconfigs instead of bare minimums
<mnemoc> and generate good nightly builds
eebrah|noPC has quit [Read error: Connection reset by peer]
eebrah|noPC has joined #linux-sunxi
<oliv3r> oh crapster; i was using my sunxi_defconfig :)
<aep> at which point does the fb driver switch to the configured screen resolution?
<aep> i tried setting a bootlogo the size of the screen, but linux claims its too big
<mnemoc> monster penguin!
<oliv3r> bah, i probably have to redo the defconfig now
<oliv3r> a nother day :)
<oliv3r> i need 2 or 3 BSP dirs ;)
eebrah|noPC has quit [Ping timeout: 248 seconds]
ykchavan has quit [Ping timeout: 264 seconds]
woprr_ has quit [Quit: Ex-Chat]
woprr has joined #linux-sunxi
jemk has quit [Ping timeout: 264 seconds]
eebrah|noPC has joined #linux-sunxi
eagles0513875_ has quit [Ping timeout: 240 seconds]
<hramrach> plat is used for common drivers
<hramrach> heh, scrollback
<focus> popolon: let me know when the EN translation is done - i put a link for you :)
\\Mr_C\\ has joined #linux-sunxi
<oliv3r> hramrach: your patch foobared
eebrah|noPC has quit [Ping timeout: 268 seconds]
Gerwin_J has joined #linux-sunxi
Gerwin_J_ has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 264 seconds]
Gerwin_J_ is now known as Gerwin_J
eebrah|noPC has joined #linux-sunxi
<popolon> focus, I just done it
<popolon> and added the MyPaint video on youtube :
<Montjoie> helllo hwclock fail to run on cubieboard2 (A20) does I need to use any other rtp driver sent on the mailing list for that ?
eagles0513875 has joined #linux-sunxi
eebrah|noPC has quit [Read error: Connection reset by peer]
eebrah|noPC has joined #linux-sunxi
eagles0513875 has quit [Changing host]
eagles0513875 has joined #linux-sunxi
eagles0513875 has joined #linux-sunxi
<oliv3r> Montjoie: rtc* :)
<oliv3r> Montjoie: and i think it sould work just fine
<focus> thank you very much
<Montjoie> oups yes rtc oliv3r
naobsd has joined #linux-sunxi
<n01> Montjoie: try to uso the driver I sent on ml. It is not mainlined but it works
<popolon> <=another one for the GLmark2 vs Glmark2-es2 test
<Montjoie> I dont understand, yesterday hwclock fail forever, and now for getting the error output and show it, it works...
<n01> Montjoie: which kernel are you using?
<Montjoie> latest sunxi 3.4
* n01 wonder why people keep using the sunxi kernel
<Montjoie> because it is the one I found in lots of doc
<Montjoie> which source do you think its better to use ?
<n01> nvm
<mnemoc> .oO
<oliv3r> n01: what other kernel would you recommend?/
<n01> mainline -.- I know, it doesn't boot, not everything is supported, bla bla
<oliv3r> haha
<oliv3r> ok, i think you pretty much answered your own question, why people aren't using mainline ;)
<oliv3r> or rather, why they use 3.4
<Montjoie> I know that mainlining is the goal, but for the moment I have too few time
Soru has quit [Ping timeout: 240 seconds]
<bfree> am I right that sun4i and sun7i need seperate kernels on 3.4?
<n01> cb is not a board for people with few time IMHO
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier.]
<mnemoc> bfree: still yes
<bfree> mnemoc: thanks ... demotivates me a little to update my 3.4 packages, but I'll get there ... now I'm back home to my cubies and sun7i support is in sunxi-3.4 ;)
<mnemoc> bfree: 3.4 already has soc-detection code, so we CAN improve it to the point of supporting multi-sunxi sun[4567]i in a single bin
<mnemoc> but not there yet
<mnemoc> such refactoring/cleanup also helps 3.10 and then mainline
<bfree> yep, I thought I knew the answer but that it was also a bit of a WIP, hence optimistically asking before I did anything ;)
<popolon> n01 : it is from reaching the goal for people with only fewtime
<popolon> +near
ganbold_ has joined #linux-sunxi
<popolon> only need to make public my µSD image ^^
<popolon> or to download focus one
<popolon> it works out of the box now
<Montjoie> n01, what driver need to be mainlined in priority ?
<Turl> oliv3r: what's on those pins?
vicenteH has quit [Ping timeout: 240 seconds]
<Turl> oliv3r: EINT mux? what's so interesting? :P
<n01> Montjoie: ask oliv3r, he knows better
<hramrach> probably mmc and/or nand (in progress) see
<Montjoie> good link to know
<hramrach> sata is almost complete so you can like flash a kernel into nand and have it boot from a SATA disk provided you can make the experimental SATA support work
<oliv3r> Turl: no, CLOCK_IN_A etc
<hramrach> what do I need to use the VGA output?
<oliv3r> probably a fex file
<oliv3r> doesn't the mele a1000 use the VGA by default?
<hramrach> dunno
<hramrach> I cannot boot with the CT fex file
<hramrach> which seems to support VGA
<Turl> oliv3r: that's mux 4 I think :p
<hramrach> but screen0_output_type = 3 (HDMI) screen1_output_type = 1 (LCD??)
<Turl> oliv3r: see p92/93
<hramrach> I get these sdc_d0 = port:PF01<2><1><2><default> sdc_det = port:PH01<0><1><default><default> lcd_bl_en = port:PH07<1><0><default><1> lcd_power = port:PH08<1><0><default><1> lcd_pwm = port:PB02<2><0><default><default> lcdd0 = port:PD00<2><0><default><default>
<hramrach> and somehow it conflicts
<hramrach> perhaps it's the PH part?
wingrime has joined #linux-sunxi
<hramrach> hello
<Montjoie> it is unfortunate that doesnt have links toward current patch status
<hramrach> sdc_det = port:PH01<0><1><default><default> lcd_bl_en = port:PH07<1><0><default><1>
<hramrach> is that different mux settings? <0><1> vs <1><0>
<Turl> oliv3r: tclkin are for timers, see eg p121
<Turl> hramrach: I think first one is mux and second is pullup
<Turl> Pxx<mode><pull><drive><data>
<Turl> yep
<hramrach> and you set all the pins in the same group (eg PH) to same function?
<Turl> hramrach: you can set functions to individual pins
<Turl> the mode 0 is input and 1 is output (gpio)
<hramrach> so that should, theoretically, not cause any problems to enable the lcd. Except the board does not boot
<Turl> time to enable earlyprintk? :)
<hramrach> I do have earlyprintk
<hramrach> the problem whan I copy the display settings from cubietruck fex file is I never get mmc0 so no boot
<Turl> "I never get mmc0"?
<hramrach> yes, the device does not come online and the kernel forever waits for it
<Turl> hramrach: can you post your full fex?
<hramrach> the cubietruck fex in sunxi-boards
<hramrach> can't boot with that one
akaizen has joined #linux-sunxi
akaizen_ has quit [Ping timeout: 245 seconds]
<Turl> hramrach: you've got a conflict with one of the lcd thingies
<Turl> +lcdd1 = port:PH01<2><0><default><default>
<Turl> it's marked as lcd_used=0, but I don't trust disp much :)
<hramrach> I have that section as well
<hramrach> but maybe enabling any lcd stomps on those pins, hmm
<hramrach> any other sane value to set those pins to
<hramrach> or just leave them out?
<Turl> we'd need the schematics
<hramrach> presumably any unused pingroup is fine
<Turl> for testing I suppose
<Turl> or just remove the full lcd thing
<hramrach> I guess that's not going to work
eebrah|noPC has quit [Ping timeout: 264 seconds]
<hramrach> because the second output is set to lcd
<Turl> disable it? :)
<hramrach> but maybe removing just lcd1 is ok
<hramrach> no vga?
<Turl> lcd != vga?
<hramrach> because I tried just about anything and the vga aoutput is not working
<hramrach> and the cubietruck nand firmware uses lcd for the second output and vga does work on it
<Turl> that fex has hdmi + lcd configured
<hramrach> yes, that's what ships in hte nand, basically
<hramrach> and VGA works with that
<hramrach> but mmc does not
* Turl scratches head
<Turl> so when configured for hdmi+lcd, vga works? :p
<Turl> try +screen1_output_type = 1 -> +screen1_output_type = 4
wingrime has quit [Ping timeout: 264 seconds]
eebrah|noPC has joined #linux-sunxi
<hramrach> that's what I did. no vga
eebrah|noPC has quit [Client Quit]
<Turl> what about mmc0?
eagles0513875 has quit [Ping timeout: 248 seconds]
<hramrach> it works with lcds disabled
<hramrach> did not try with ldc1 removed and lcd0 enabled yet
Soru has joined #linux-sunxi
eagles0513875 has joined #linux-sunxi
vicenteH has joined #linux-sunxi
<oliv3r> Turl: yeah but that's new on A20, is that easily useable with the clk framework?
<Turl> oliv3r: which one? clkin?
<oliv3r> clk_out_A
<Turl> oliv3r: that one smelled a lot like a mod0 clock
<Turl> need to check the offsets
<Turl> oliv3r: do you need an external clock for sth? max that one was able to provide was 24MHz
<oliv3r> Turl: actually I do for something else ;)
<oliv3r> but it caught my interest as a new 'feature' of the A20 when i updated the PIO section for A20
<oliv3r> btw, did you know that EMAC is also muxed to the old ATA pins? so we can have dual ethernet ;)
<Turl> oliv3r: :O really
<Turl> poke tom :p maybe he can make it happen for his board
<Turl> 100+1000 would make a nice server board
<Turl> firewall/router kind
<oliv3r> Turl: Tsvetan tested gbit on his A20-som, he said 330 MBIT was pretty much the max
eebrah|noPC has joined #linux-sunxi
<oliv3r> but yeah dual ports isn't a horrible idea
<oliv3r> Turl: but poke tom? tom is working on RK now; Benn would do that ;)
<Turl> oliv3r: maybe with OC'ed mbus it goes higher :p
<oliv3r> :p
<oliv3r> and proper driveres
<Turl> oliv3r: I thought tom was going to make another AW board after his rockchip one
<oliv3r> ah well if his next board is a80 based; it'll be sad :(
<Turl> mnemoc: ^ confirm/deny? :p
<oliv3r> i doubt they'll make a nother A20 one; 3 A20 boards?
<oliv3r> hipboi: ^ confirm/deny :p
<Turl> oliv3r: directly from the source :P
<Turl> oliv3r: A23 maybe :pp
<oliv3r> well no ethernet/sata there
<oliv3r> so no point for that :)
<Turl> really? :(
<arete74> allwinner a10 domotic:
<oliv3r> Turl: does A13 have ethernet/sata?
<Turl> oliv3r: via USB :p
<oliv3r> Turl: er, but then how are the extra pinmuxes helping us :p
<Turl> oliv3r: I know, I know :p
<mnemoc> Turl: I have no clue what Tom is making. but he wanted to be vendor-agnostic
<Turl> that ^ :)
aalm has quit [Ping timeout: 264 seconds]
<oliv3r> a100!
* mnemoc still hopes for an A20 successor
<popolon> oliv3r, what for do you think a80 will be sad ?
<popolon> have you more information about it ?
<popolon> mali or powevr ? A15 or 12 for the big.LITTLE ?
aalm has joined #linux-sunxi
<mnemoc> popolon: next is confirmed to be 4xA15+4xA7. but silence regarding the GPU
<oliv3r> popolon: powervr
<mnemoc> but from the description it smells to an A31 successor, not A1x/A20's
<popolon> oh :(
Vespine has joined #linux-sunxi
<popolon> perhaps the rumoured A40 then ?
<popolon> ^^
<Sonicadvance1> Nothing like a PowerVR to ruin your day
<popolon> I suppose PowerVR sell license to a really low price to keep fabless chip maker continue to use them
cubear has quit [Ping timeout: 245 seconds]
<mnemoc> but android-only licenses
<mnemoc> while mali's are android+linux
<Sonicadvance1> Qualcomm is Android only as well and they are smashing the market right now :P
Gerwin_J has quit [Quit: Gerwin_J]
<Turl> that's because of the LTE radio integration
<Sonicadvance1> Yep
Gerwin_J has joined #linux-sunxi
<popolon> don't understand why Samsung does each time a qualcomm version for US and own samsung processor for the rest of the world ?
<popolon> at least the rest of the world generally has Mali :)
<oliv3r> A40 might be a nice name for a quad sun4i based design
<Turl> popolon: LTE
eebrah|noPC has quit [Ping timeout: 260 seconds]
<Turl> most carriers only have certified qualcomm LTE modems
<popolon> oh, I see
<oliv3r> ah crapster
<oliv3r> wills wangs ahci patch doesn't even work :S
<oliv3r> [ 1.338099] ata1: SATA link down (SStatus 0 SControl 300)
<oliv3r> does more then my work though
<oliv3r> double crap anyhow
<oliv3r> more digging tomorrow :S
<popolon> nextgen chromebook will have an exynos 5250
<popolon> so mali GPU :)
<mnemoc> is that the octa?
<Sonicadvance1> popolon, Already it is announced
<Sonicadvance1> and available
<popolon> the second version of octa
<popolon> oh !
<popolon> Sonicadvance1, url ?
<popolon> what is the name ?
<mnemoc> hopefully "pixel" display
<Sonicadvance1> 5250 is the same SoC that is in the current Samsung Chromebook
<mnemoc> uhm...
<popolon> sorry Exybnos 5420
<Sonicadvance1> New one is a HP chromebook that is 11" with the 5250
<popolon> 5410 is powerVR version (on galaxy S 4)
<popolon> and 5420 mali T6xx (on Note III, note 10.1 2014 edition and chromebook gen 2
* mnemoc wants an exynos octa with the display of the pixel
<popolon> HP :(
<Sonicadvance1> popolon, You were meaning Peach Pit?
<Sonicadvance1> Yea, The peach pit has the Exynos5420 in it
<Sonicadvance1> uses the same LCD as the current Samsung Chromebook
<Sonicadvance1> Still 2GB of RAM
<Sonicadvance1> 1.8Ghz CPU clock instead of 1.7Ghz
<Sonicadvance1> Mali-T628MP6 has max clock speed of 700mhz, but samsung has limited it to a little lower than 500mhz max
<popolon> oh !
<juanfont> they are fanboys
* mnemoc tired of 1366x768 displays
<Sonicadvance1> Me too
<popolon> don't care about apple :D
<popolon> and still the 8 core was not used due to software management in the S4
<popolon> who want to dev free software for so expansive /not compatible hardware, that will be out-of-date in few month as everyone :)
<Sonicadvance1> Using Big.little MP in a smartphone sounds like a terrible idea
<juanfont> yup, Sonicadvance1
<popolon> someone tiped me a good point for apple : stereo jack *in*
<popolon> most jack in are mono
<popolon> and stero jack in is really a good point to make it a portable good sound recorder
<popolon> Sonicadvance1, this is clearly not the HP one that is targeted in the git commit
<Sonicadvance1> popolon, I know, peach pit is a different device
<popolon> do you have some refs ?
<Sonicadvance1> It was all leaked information from gerrit logs and source
<popolon> ah ok
<popolon> have to wait then
discopig has quit [Read error: Connection reset by peer]
discopig has joined #linux-sunxi
<WarheadsSE> maybe i missed something.. bu isnt the XE303 also the 5200 series
piscodig has joined #linux-sunxi
piscodig has quit [Read error: Connection reset by peer]
<Sonicadvance1> WarheadsSE, Yep the XE303 is the 5250
<Sonicadvance1> So there are two chromebooks using the 5250, and one coming out,speculated this month, with the 5420
mouchon has quit [Ping timeout: 252 seconds]
<WarheadsSE> Waiting on that one..
mouchon has joined #linux-sunxi
discopig has quit [Ping timeout: 264 seconds]
wingrime has joined #linux-sunxi
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-sunxi
<popolon> there is a note 12.2 too
<popolon> note 12.2 with 5420 => add a keyboard and you have a netbook :)
<popolon> (you can get some tablette pockets including keyboard for ~10$ on chinese stores
<popolon> the note 10.1 gen 2014 include microUSB, no more proprietary connector :)
<popolon> next step => several usb connector, like on cheap chinese tablets
<atsampson> popolon: any recommendations for half-decent ones?
<popolon> no, there are so much :)
<popolon> some site are specialized about this
<popolon> i Know at least two in french, but alreasy seen in english too
<popolon> didn't bookmarked
ykchavan has joined #linux-sunxi
<popolon> (first entry on my search engine)
<popolon> french one I know have blog with deep test, videos ...
EGM has joined #linux-sunxi
<popolon> generaly camera are not good
<popolon> and don't know but battery is something to have information on for this kind of tablets
<EGM> Hi everyone. Have anyone tried using USB OTG on Cubieboard2 with sunxi kernel? I am interested in accessing data throught mass storage backend.
<popolon> too sad, on cpu : allwinner A10, A13, 31, 31s, but not A20 :(
<EGM> I am sorry, is that related to my question, or are you explaining your trouble?
<popolon> not at all
<popolon> :)
<popolon> I understand your question
<popolon> do you mean: plug the cubieboard on a computer
<popolon> and read cubieboard data from the computer ?
<EGM> Well, not exactly.
<popolon> using mass storage
<popolon> else there are 2 standard USB ports
<EGM> I need to be able to switch some kind of storage between linux system and USB OTG mode, so it can be written in linux, and then read as mass storage from Windows.
<EGM> I know that Cubieboard2 has OTG port (miniUSB), but I have no idea if there is support for that in sunxi kernel.
<popolon> any external USB HDD is seen as mass storage
<EGM> I am sorry, but I believe no do not understand my question.
<popolon> what should be mass storage ?
<popolon> the HDD/usb key/...
<popolon> or the cubieboard ?
<EGM> The Cubieboard should act as mass storage.
<popolon> aah ok
<popolon> and have a HDD pluged in
<popolon> (or something like that
<EGM> No, no HDD. Just create filesystem in file on NAND or SD card, and mount it in linux, so I can save some files there.
<popolon> ok
<EGM> After that, unmount it from linux and connect it as mass storage using OTG port to computer.
<popolon> so you should not umount it
<popolon> but instead, share it as mass storage ?
<EGM> Well, I believe that there is kernel module that takes file as image and exports it as USB mass storage via OTG port.
<popolon> Or could be possible to share it mounted/ unmounted depending on software
<ykchavan> EGM, I do this all the time on default installed android.
<ykchavan> I do not think there will be any problem doing with sunxi kernel.
rz2k has joined #linux-sunxi
<EGM> Yeah, I know that it works on android, so I suppose that it's hardware support needed. But I would like to do it in linux.
<EGM> Unfortunately, I have not found any good sources explaining how to do that...
<ykchavan> I see. Let me try on my cubieboard.
<EGM> IIRC, when I tried to load module called file_storage (or mass_storage, I believe that both should do basically the same), I got some error which did not help me at all.
<Turl> EGM: you just need to load the gadget module
<Turl> then echo the file path of your block device to the lun file on /sys
<EGM> Which gadget module are you talking about?
<Turl> I suppose the android gadget module works, or the pure mass storage one too
<Turl> look around the kernel menuconfig in drivers -> usb support -> usb gadget support
<EGM> Ok, there are g_file_storage and g_mass_storage modules. Since g_file_storage is deprecated, I should use the g_mass_storage, right?
<EGM> What did you mean by "android gadget module"?
<Turl> there is a gadget module named "android"
<Turl> that can handle mtp, ptp, mass storage, adb, serial and a bunch of other stuff
jinzo has quit [Quit: Leaving]
<popolon> ./kernel/drivers/usb/gadget/g_mass_storage.ko
<popolon> oh, that's compiled by default with sun7i_defconf
leowt has joined #linux-sunxi
<leowt> is it possible to boot from a root on usb?
<popolon> g_acm_ms.ko g_audio.ko g_dbgp.ko g_ffs.ko g_mass_storage.ko g_multi.ko g_printer.ko g_webcam.ko
<popolon> gadgetfs.ko g_cdc.ko g_ether.ko g_hid.ko g_midi.ko g_ncm.ko g_serial.ko
<popolon> even webcam ?
<popolon> :
<popolon> )
<EGM> Turl: I can not find the android module, but I will try the g_mass_storage
<EGM> But I believe I tried to modprobe it, but it failed with some strange error.
<EGM> Do I need to enable USB Dual Role Controller to get OTG working?
<popolon> EGM, modprobe gadget_fs
<Turl> leowt: yes
<popolon> EGM, modprobe gadgetfs (sorry)
<Turl> leowt: just pass root=/dev/whatever on kernel cmdline
<leowt> Turl, =)
<Turl> leowt: and maybe rootwait too if your device is slow
<EGM> popolon: Why not the g_mass_storage one?
<popolon> seems to be a submodule ?
<popolon> perhaps need to be inserted with insmod ....g_mass_storage.ko (with some args ?)
<EGM> popolon: Have you tried it, or are you just googling around? No offense, but I have done that already.
<popolon> I'm actually trying it
<EGM> Ok, thanks for that.
<popolon> there are some informations in linux-sunxi/Documentation/usb/gadget_multi.txt, but that's really not clear for me
HeHoPMaJIeH has quit [Remote host closed the connection]
<popolon> musb = (Multipoint) Highspeed Dual-Role Controller
<popolon> should be needed to use OTG
rellla has quit [Quit: Nettalk6 -]
<EGM> popolon: I will compile it in and try it
<popolon> there is an musb_gadget too (in sources)
<ykchavan> I am adding this to FAQ
<popolon> I suppose in menuconfig : *** OTG and related infrastructure ***
<popolon> dollowing this line
<popolon> should have the stuff
<popolon> [ ] SUNXI USB2.0 Dual Role Controller support <= not selected on sun7i
<EGM> Should be?
<EGM> The fact that is not selected in default config means nothing...
n01 has quit [Ping timeout: 245 seconds]
<popolon> could mean it doesn't work at all, don't know, I didn't hacket the kernel
<popolon> hacked
<popolon> but other could perhaps have the answer here ?
<popolon> ykchavan, do you have a FAQ ?
<EGM> I successfully inserted gadgetfs module, but I can not insert the g_mass_storage one
<EGM> ERROR: could not insert 'g_mass_storage': No such device
<ykchavan> I was referring cubieboard FAQ.
<popolon> oh :)
<ykchavan> Just realized that this will not cover all sunxi chips.
<ykchavan> Is there any FAQ covering all?
<popolon> is more general sunxi related
<popolon> but no faq
<popolon> I just suggested yesterday it could be fine to have one, everyone agree
<popolon> but didn't take time to start a FAQ page :)
<ykchavan> There exists for cubieboard
<popolon> oh !
<EGM> Turl: BTW, where are you from? I know someone using your nick
<popolon> a general FAQ with a link to this FAQ could help
<Turl> EGM: .ar
<ykchavan> popolon, yes
FR^2 has quit [Quit: Connection reset by peer]
<EGM> Turl: Ok, just same nick then.
<ykchavan> popolon, let's start building now. I can fill it as many as I can think of.
<EGM> Turl: Do you know if modprobing g_mass_storage should work right away? I got "No such device" error.
<EGM> Dunno why.
<EGM> I tried passing path to image file in "file" param
<Turl> EGM: check dmesg, it may have a hint
<popolon> nice :)
<EGM> Nothing in dmesg, nothing at all.
<Turl> EGM: does your kernel have OTG enabled?
<EGM> Last message is about gadgetfs, which I inserted before.
<Turl> and is it configured on fex as such?
<EGM> zcat /proc/config.gz | grep OTG # OTG and related infrastructure CONFIG_USB_OTG_UTILS=y # CONFIG_USB_OTG_WAKELOCK is not set
<EGM> Dunno if there should be anything else.
<EGM> Oh, it has to be allowed somewhere in FEX?
<EGM> usbc0 has usb_port_type = 2, is that it?
<EGM> Looks like that should be correct, considering the fex guide.
<Turl> EGM: your kernel probably needs USB_SW_SUNXI_USB0_OTG or sth like it
<Turl> check those options on menuconfig
<EGM> Oh, looks like enabling Inventra Highspeed Dual Role Controller (as popolon suggested) hides SUNXI dual role controller
<popolon> # CONFIG_SUNXI_NAND is not set (in default sun7i_defconf
<popolon> egm: really sorry :)
<popolon> at least we know that dual role is related to OTG
<EGM> np, I will try it without that...
<popolon> so that why I didn't see the nand at all :)
<ykchavan> Cubian on cubieboard1, $ zcat /proc/config.gz | grep OTG # OTG and related infrastructure # CONFIG_USB_OTG_WAKELOCK is not set
<EGM> And does OTG work?
<ykchavan> no..
<popolon> haaa
<popolon> ubuntu broken dependencies for gcc-4.8 arm :(
<popolon> I've no mre gcc 40.8 compiler
<popolon> 4.8
<popolon> (for arm)
<popolon> depend on dnlcpp (looks like a typo ???)
<popolon> oh, this is a proposed version that is buggy :/
mcbrick has joined #linux-sunxi
<EGM> But nothing changed
<EGM> ERROR: could not insert 'g_mass_storage': No such device
<popolon> but with insmod
<popolon> you should precise the full path and name
<popolon> of the .ko
<EGM> modprobe g_mass_storage file=/root/testfile
<EGM> I did that this way, should work too.
<EGM> Will try with insmod, gimme sec
Soru has quit [Quit: No Ping reply in 180 seconds.]
soul has joined #linux-sunxi
<EGM> Error: could not insert module g_zero.ko: No such device
soul is now known as Guest94458
<EGM> I believe that zero gadget module should work everytime, since it is example how to implement USB gadgets.
<EGM> :(
<EGM> Same result when using insmod
<EGM> insmod g_mass_storage.ko file=/root/testfile
<EGM> Do I need to connect cable before that or something like that?
<EGM> The USB port is currenctly unoccupied, dunno if it matters.
<Turl> EGM: try plugging it to your pc first
<EGM> With any cable?
<EGM> Well, I plugged it in. Host system (ubuntu) does not see it (no record in lsusb). And nothing changed.
<EGM> Is there any image in which OTG works as I need? I am willing to download it and see how it is configured for yourself.
<Turl> nothing on the device's dmesg?
<EGM> No, nothing.
<Turl> EGM: an android one maybe?
<EGM> Is there /proc/config.gz available?
<Turl> EGM: what device do you have?
<EGM> Cubieboard2
partoa has joined #linux-sunxi
<Turl> EGM: it has android preinstalled on nand doesn't it?
<partoa> Hello
<EGM> Yeah, it is.
<EGM> But I never used that, since I am interested in Linux.
<partoa> We have created a custom PCB but we have since established that our NAND connection used an reference that's in accurate
<partoa> reference design that is.
<partoa> It boots from SD but fail
<partoa> Right after loading the PMU drivers.
<EGM> Turl: I believe that I need monitor and mouse/keyboard for using that android image, right? Seems like there is no shell on serial port...
<Turl> you can use ADB shell via the usb otg port :)
<partoa> Can anyone help with that? We suspect that it may have something to do with the PMU and clock...given that we are presently not able to read the Boot1 file header and therefore can't get a custom uBoot created for our board.
_BJFreeman has joined #linux-sunxi
<EGM> Turl: Well, I dont have android sdk installed. I will look into that tomorrow probably.
<EGM> In fact, only thing I need to know today is if it is possible to use USB mass storage gadget via OTG port.
<Turl> partoa: panic=rw? did you actually write that?
_BJFreeman is now known as BJfreeman
<Turl> partoa: otherwise your ram is corrupting random bits
<partoa> Nope....PengPod image...
<partoa> Those were the only images that we found that were able to boot.
<Turl> partoa: you have 1G of ram on the device right?
<partoa> Yes, 1G
<Turl> as 2x512?
<partoa> No, 4X256
<Turl> hmm
<Turl> hno: ^ is that supported? any board using that?
<Turl> partoa: I'm not too familiar with the memory controller, but hno or oliv3r might be able to help
<partoa> Turl: Thanks...lets hope hno: or oliv3r: pick up on this :-)
<Turl> partoa: you can try replacing uboot with the mele a1000 one, it should be kind of the lowest common denominator
<partoa> Ahh...thanks
<partoa> Let me try that.
popolon has quit [Quit: Quitte]
<EGM> popolon, Turl: Guys, thank you for the help today, I appreciate it. Since it's getting late, I will leave now (so I can leave work and get home), but I will try to get it working tomorrow and after that. I believe that we will meet here again.
<Turl> :)
<Turl> you're welcome EGM
<EGM> Bye :)
EGM has quit []
massi_ has quit [Quit: Sto andando via]
wingrime has quit [Ping timeout: 240 seconds]
<partoa> Turl: seems like I will have to continue with the results tomorrow. My Internet connection is not too fast, being in Kenya and all. I'll leave the image downloading overnight. Thanks again.
partoa has quit [Quit: Page closed]
wingrime has joined #linux-sunxi
panda84kde has quit [Ping timeout: 245 seconds]
atiti has quit [Read error: Operation timed out]
wingrime has quit [Ping timeout: 240 seconds]
BJfreeman has quit [Read error: Connection reset by peer]
partoa has joined #linux-sunxi
wingrime has joined #linux-sunxi
rz2k has quit []
n01 has joined #linux-sunxi
n01 has quit [Client Quit]
n01 has joined #linux-sunxi
leowt has quit [Quit: leowt]
<hramrach> Turl: so the script.bin without the lcd1 section does not break mmc0
<hramrach> but vga still does not work
<hramrach> so there must be something different in the kernle on nand or android sets up something which Linux does not
ykchavan has quit [Ping timeout: 248 seconds]
<hramrach> there are 8 fb devices :s
<hramrach> and hte kernel actually has 2G ram unlike the sunxi kernel
ibu_ has joined #linux-sunxi
ibu_ has left #linux-sunxi [#linux-sunxi]
cubear has joined #linux-sunxi
<cubear> hey, can I get some help with setting up nand-boot on CB1?
<cubear> seems like the bootloader is rather disobedient
<oliv3r> Turl i think there's even 4x 128M; i think my tablet does that
<cubear> I've used nand-part to setup 2 partitions - nanda and nandb. nandb is the rootfs
<cubear> I've been told to put the uImage, script.bin and uEnv.txt to the nanda partition root folder
<cubear> but the bootloader fails to load them
<cubear> I can load them manually and it boots fine
<cubear> the bootloader tries to load linux/uImage and linux/env.txt files
<cubear> I can't modify the bootloader commands because it doesn't support the "saveenv" command
<cubear> naither I can put uImage and other files into the linux folder because they... magically disappear :-/
<cubear> any ideas?
<partoa> oliv3r: any idea what is happening with out boot process?
<partoa> It always stops at that exact same position, no matter what uImage we use
<oliv3r> also cubietruck ha 4 mem chips
<oliv3r> 2gigs
<hramrach> oliv3r: sent the patch on github
<oliv3r> hramrach: thanks i'll apply it
<hramrach> thanks
<hramrach> cubear: you should update the bootloader too. but saving diles to linux folder should work too. it's plain FAT partition
<cubear> hramrach, I used this one:
<cubear> I copied that to the linux folder
<cubear> but if I copy any other file like uImage it disappears
<oliv3r> cubear: what board do you have?
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<cubear> cubieboard A10
<oliv3r> sec
<cubear> I'm sure I copied uImage into that folder too
<partoa> Has anyone had any problem reading the boot1 file header getting "libusb usb_bulk_read error -7" as an error message after running "./fel read 0x42400000 0x82d0 boot1.header "
<partoa> A grep on both fel and libusb-1.0 for "usb_bulk_read" does not turn anything up...
<cubear> oliv3r, if it matters, here's a forum post:,860.msg5097.html
<oliv3r> not really :)
<cubear> I'm just puzzled by this... why can't I do saveenv???
<cubear> says command not supported
<oliv3r> get that
<oliv3r> then do dd if=cubie1.bin of=/dev/sdcard bs=8 seek=1024
<cubear> oliv3r, I'm doing a nand install
<cubear> sdcard boots fine
<oliv3r> but we don't have u-boot support for nand
<oliv3r> ohh, your doing something unsupported
<oliv3r> who was asking for 4 memory chip enabeld u-boot?
<oliv3r> hmm, your fatls look odd
<cubear> oliv3r, patwood said "The second thing you'll need is a u-boot that can load the uImage from nand. You can find one here: Put it in the /linux directory on the vfat partition."
<oliv3r> oh thats the directory named 'linux'
<cubear> so that's what I used
<oliv3r> ah ok
<cubear> I manually loaded the uImage
<oliv3r> well normally, whatever is preinstalled in nanda is what you want
<cubear> sun4i#setenv boot_normal 'fatload nand 0 42000000 uImage; bootm 42000000'
<cubear> that works fine and I can boot the linux like that
<oliv3r> isn't your script.bin at 43000000
<oliv3r> make sureyour not overwriting your script.bin in memory
<oliv3r> load your kernel to 48000000
focus_it has joined #linux-sunxi
<cubear> that works fine and I can boot normally
<cubear> problem is I can't do "saveenv"
<cubear> command not supported
<cubear> so I need to write that in serial each time I boot
<cubear> here's my printenv:
<steev> u-boot probably has ENV_IS_NOWHERE defined somewhere
<cubear> it looks for an uImage inside the linux folder
<oliv3r> nand enabled u-boot is VERY limited
<oliv3r> it's an extremly old hacked up u-boot
<oliv3r> but you can manually change your env
<cubear> where?
<oliv3r> extract the env partition
<oliv3r> convert it to txt, edit it, compile it to bin, dd it back
<cubear> that's an option yes
<steev> wow, that IS hacky
<oliv3r> steev: you onyl change the env once, its' allowed to be hacky: )
<cubear> where is this "env" partition? I assume it's at an address before the first nanda vfat partition
<oliv3r> nandb or nandc
<oliv3r> don't recall
<oliv3r> i think b
<cubear> I formatted the nand with nand-part, so there's only 2 partitions now - nanda and nandb. nandb is the rootfs
<cubear> anyway I think I've found the bootloader command
<cubear> seems to be hardcoded inside the u-boot.bin file I downloaded earlier
jinzo has joined #linux-sunxi
<cubear> I'll see if I can modify it with a hex editor
<oliv3r> hramrach: did you test your fex with fex2bin?
<cubear> okay, I take it if I don't modify the length of the strings it should work fine. assuming there isn't a checksum
\\Mr_C\\ has quit []
\\Mr_C\\ has joined #linux-sunxi
<oliv3r> nvm i used bin2fex
<oliv3r> cubear: you can hexedit too yeah
<cubear> stupid procedure though :(
<cubear> alright I'll give it a go. will report back whether it works. brb
<cubear> found the string at 0x0002AD68
<oliv3r> cubear: aye; and that is still only a guess as to why it shouldn't work ;)
notmart has quit [Quit: notmart terminated!]
deasy has joined #linux-sunxi
<cubear> doesn't...
ykchavan has joined #linux-sunxi
<hno> Turl, partoa 4x256 is quite common. Noting odd.
<hno> partoa, you can easily create a custom u-boot SPL. The dance about reading boot1 is only when not knowing the proper parameters like suitable DRAM clock etc and most people are not good at guessing these right..
<partoa> Ahh...
<partoa> I'd place us at among the people who are not too good...I have seen that the mele a1000 uses the same IC as us, what are the odds that their DRAM clock has the same frequency as ours?
<cubear> I think I bricked it, now even using the original u-boot.bin doesn't run
<cubear> mmc still boots though
<partoa> hno: could corruption of random bits, as suggested by Turl explain our current predicament? Also, does that clock setting affect communications with the PMU?
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
<hramrach> oliv3r: I copied the leds section I use and changed the default to match the the other sunxi-boards
<hramrach> fex2bin does not report any issue
eebrah has joined #linux-sunxi
<hno> partoa, incorrect DRAM parameters means unstable system.
<hno> And PMU, CPU clock and DRAM clock are all closely related.
<cubear> okay... can anyone explain to me why do the files I copy over to /dev/nanda disappear over reboot?
<partoa> hno; would it still be unstable if it always fails the moment the PMU "comes into play"?...I ask because we were finding it rather consistent that we even imagined that the PMU could be faulty.
<Turl> cubear: /dev/nanda is a block device
<partoa> hno; do you have a DRAM clock range that we can try play with? Say a minimum and maximum.
<cubear> Turl, could you elaborate?
<Turl> cubear: you need to mount it somewhere and then copy files
<Turl> mount -t vfat /dev/nanda /mnt, copy your stuff to /mnt; sync; unmount /mnt
<cubear> like this: #mkdir /tmp/nanda #mount /dev/nanda /tmp/nanda #cp /boot/uImage /tmp/nanda/linux
<cubear> that's what I did and it vanished
<Turl> cubear: then sync; umount /tmp/nanda
<Turl> then try mounting it again, your file should be there
<cubear> Turl, can I assume that sync gets called automatically upon unmount?
<cubear> nope, they vanished again
<cubear> alright, I mounted, copied, sync, unmounted
<cubear> then I remounted and the files are all there
<cubear> now let's reboot again (from sd-card)
<cubear> nope, they're all gone
<cubear> seems like the nanda partition reverts to its original state upon reboot
<oliv3r> hramrach: yeah but i used bin2fex :p
<oliv3r> it segfaulted
<oliv3r> Turl: i just mounted an old iPhone 3GS from the parts bin onto my CT
<oliv3r> it'll fit also on the olimexino
<oliv3r> it's charging now, was dead, i'm curious how much i can power from that :)
<oliv3r> i bet the axp won't upconvert to 5V for the ahci
Guest94458 has quit [Read error: Connection reset by peer]
<cubear> I give up :(
<hno> partoa, I don't know if that PenPod u-boot SPL configures the PMU or not.
<oliv3r> cubear: what's wrong
<oliv3r> hno: hi!
<hno> hi oliv3r
<oliv3r> hno: i've pushed cubietruck support to u-boot (limited to 1GB)
<oliv3r> and i've pushed to my personal github a 2GB enabled branch
<oliv3r> since u-boot doesn't (properly) support 2GB
<hno> What is the problem wit 2G?
<oliv3r> i think it 'works' most of the time, phys_size_t is typedefed as a unsigned long
<oliv3r> pyhs_size_t is used for ram_size that is passed onto the linux kernel
<oliv3r> afaik
<partoa> hno, we actually have established that we can't do much debugging without using building our u boot SPL. We'll just have to work out the collect clock setting.
<hno> unsigned long is fine for 2G.
<oliv3r> however, get_ram_size is retunring a long (signed)
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
<oliv3r> and we everbody does (inc sunxi) gd->ram_size = get_ram_size(base, size);
<oliv3r> parameters are (signed) long; return is (signed) long
<hno> oliv3r, ok, that would be a bit of a problem.
<partoa> Thanks a lot. hno, Turl. I will be sure to blog or contribute in some way as soon as we get a solution.
<oliv3r> so representation goes a little bad; but cramming that (negative) number into the phys_size_t kinda works
<oliv3r> hno: check out the patches on my github branch; i did bring the issue up on the u-boot ML; but it got ignored
<hno> but do get_ram_size actually test the 2GB when size < 0?
<oliv3r> so far :p
<oliv3r> i doubt it, how can you ever have a negative ram size
<oliv3r> i guess, the function was written with 32bit in mind, but completly ignoring the fact that a SoC could do 2 GiB's
<atsampson> I feel like I've got a negative RAM size some mornings before I've had my first cup of tea...
<oliv3r> anyway I think i properly fixed it, and i can merge it over; once I receive your blessing ;)
<hno> There is several that do 2G today, so can't be alone in this problem.
<oliv3r> well i noticed it because the SPL reported a negative memory size, u-boot did report it properly (because it does ramsize >> 20 to print)
<oliv3r> and I think the negatie sign drops?
<hno> ARM spec even supports PAE with huge amounts of RAM even in 32-bit.
<hno> >> sign extends, so it must have been unsigned at that time.
<oliv3r> well I think i prpoperly fixed it
ynezz_ is now known as ynezz
<hno> Ok. I'll try to look into it on saturday. Now I need to collect a bit on the sleep account.
cubear has quit [Quit: Leaving]
<oliv3r> hno sleep well :)
<mnemoc> aaaaarg... 3580 unread sunxi mails again :'(
<oliv3r> mnemoc: :( you really should focus on doing 200 a day
<oliv3r> :p
<oliv3r> we only get like 10 msgs a day
<oliv3r> :p
<mnemoc> it includes sunxi related mails in linux-arm
<mnemoc> but yes
<oliv3r> make a folder; filter all lkarm to that one
<oliv3r> read that at a lower pace later
<hno> inbox only says 5980 unread. I'll catch up later...
<oliv3r> you two poor guys :(
<oliv3r> 4 unread :) but those are unread on purpouse
<oliv3r> my OCD goes wild if i have unread mails :)
<hno> haven't been reading my email at all for some weeks.
<oliv3r> i can imagine
<oliv3r> sometiems its too much
<hno> barely seen my wife and kids for the last weeks.
<oliv3r> well i lied, i have 100 unread now; but most is from u-boot ML; which i hope to unsubscribe from fast
<oliv3r> go leave the PC now and go spend time with your wife!
<hno> I know.. but been home whole day today, had to rest a bit.
<hno> and no large fires to extinguish today.
n01 has quit [Read error: Connection reset by peer]
<oliv3r> yeah; burn out immenent! warning!
<oliv3r> hno: small interesting update on a20 in case you missed it
<oliv3r> we know there's gmac mixed to the emac pins, appearantly they removed the ATA mux and put the emac there as second mux
<oliv3r> so we could, in theory, have 2 lan ports
n01 has joined #linux-sunxi
<hno> Oh? emac + gmac config?
<oliv3r> yeah
<oliv3r> technically possible i think
<oliv3r> i went through the PIO in the manual and updated the wiki akkordingly
* hno tries to remember what else is on the ATA pins..
<oliv3r> lots, 5 muxes in use
<oliv3r> ATA = PH
ykchavan has quit [Read error: Operation timed out]
<oliv3r> LCD1; KP/PS2; SMC; EINT and CSI1
<oliv3r> nothing interesting
<hno> link?
<hno> Missing a link from
<oliv3r> oh yeah
<hno> oliv3r, I assume the I2²_... in PA is actually I²S_..?
<oliv3r> bad typo :S
<oliv3r> i first had i2s; then copy/pasted the 2
<hno> So they killed PATA? No simple CF interface any more?
<oliv3r> looks like
<oliv3r> PH mux 3 was ata
<hno> But we did not have any driver for the ATA I think. So nothing lost really.
<oliv3r> haha, yeah
<oliv3r> maybe in sun3i or sunii sources
<oliv3r> melis*
<hno> quite likely in melis, but have not seen any in linux.
<oliv3r> well if there is ever any need/desire to get it, there's a potential source i suppose
<oliv3r> but no hardware, no reason i guess
<oliv3r> dual lan is nice
<oliv3r> Tsvetan2: ^ :)
<hno> we don't have melis sources.
<oliv3r> i know wingrime found 'some' sunii sources
<oliv3r> no clue what it was for
<hno> not sure we want to see it, even remotely...
<oliv3r> hehehe
<hno> Ok, interesting quality check boot loader in the A10 sources.. if(1)
<hno> else
<hno> __dinf("test ok\n");
<hno> __dinf("test fail\n");
<hno> wonder how many times they had that fail.
<oliv3r> LOL
<oliv3r> oul ploduct velly good
<oliv3r> zelo fail
<hno> err, that's the A20 version. A10 don't have a quality check boot loader at all.
<oliv3r> oh i got a battery connected to my ct now; its charging will run some tests soonish
<hno> tried connecting a 2.5" HDD to mine.. not very successful.
<oliv3r> it needs power
<hno> I know.
<oliv3r> i got a 60 gb 350mA ssd connected
<oliv3r> seems to work fine
<oliv3r> i'm powering via usb -> power
<hno> was powering via USB->DC jack as well. And I think there was sufficient 5V. But will check again.
<focus_it> hno: check amp rating on the disk - it will say 0.55A or 1A etc. beside the 5V
<hno> the drive do spin up and stays spinning, but do not want to talk to the CT.
<oliv3r> hno: what kernel?
<oliv3r> i THINK nand one should talk to it; our current 3.4 stage does
<oliv3r> willis wangs patches atm do not entirly
<hno> Whatever is in the Android on NAND. Drive says 800ma.
<oliv3r> that's pushing it
<oliv3r> usb should only deliver 500mA
<oliv3r> and then there's the power for the CPU too
<hno> USB capable of delivering far more than 500ma. No overcurrent triggering on the host.
<oliv3r> well nand did seem to work
<hno> the 500mA is the minimum for a powered USB host.
<hno> and specs also say there should be a reasonable overcurrent protection to protect from fires.
<hno> but I will try with another power source, later..
<hno> now sleep.
<oliv3r> oh it ran off battery for a sec
<oliv3r> ok nn
<oliv3r> i'm booting nand now :)
<Turl> night hno
<oliv3r> hmm not seeing it at all
<oliv3r> <5>[ 1.182512] sd 0:0:0:0: [sda] 117231408 512-byte logical blocks: (60.0 GB/55.8 GiB)
<oliv3r> yeah does work
eebrah has quit [Ping timeout: 268 seconds]
partoa has quit [Quit: Page closed]
focus_it has quit [Quit: Leaving]
hramrach has quit [Ping timeout: 240 seconds]
hramrach has joined #linux-sunxi
tinti has joined #linux-sunxi
<oliv3r> rz2k, slapin_ and other flash coders ^
<oliv3r> this is from stock nand on a fairly new CT, so flash can't be that horrible
n01 has quit [Ping timeout: 240 seconds]
<oliv3r> i think that indicates that there's the same sort of problems with the nand driver too?
popolon has joined #linux-sunxi
Vespine is now known as cubear
Tartarus_ has joined #linux-sunxi
Tartarus_ has quit [Read error: Connection reset by peer]
<wingrime> oliv3r: 64 Gb mmc works?
<popolon> SDXC ?
<wingrime> yep
ganbold_ has quit [Remote host closed the connection]
akaizen has quit [Remote host closed the connection]
Tsvetan3 has joined #linux-sunxi
Tsvetan2 has quit [Ping timeout: 240 seconds]
wingrime has quit [Ping timeout: 240 seconds]
piscodig has joined #linux-sunxi
tinti has quit [Ping timeout: 264 seconds]
piscodig is now known as discopig
akaizen has joined #linux-sunxi
tinti has joined #linux-sunxi
popolon has quit [Quit: Quitte]
CoSM0 has quit [Read error: Connection reset by peer]
akaizen has quit [Remote host closed the connection]