<leming>
and traeak, no arch idiot blasted the package manager, though an arch user apparently canceled the full system upgrade to update pacman first and subsequently broke their pacman
<traeak>
leming: any sin to hit return ?
<traeak>
leming: with no warning?
<traeak>
leming: is it a problem the pacman update didn't pull in the proper dependencies?
<leming>
it specifically asks if you want to cancel the current operation (presumably a -Syu), in order to update pacman first
<traeak>
pretty much yeah
<leming>
pacman doesn't specifically depend on glibc because that's a base group package that is assumed to be there
<traeak>
so there's that little problem that the glibc install fundamentally changed
<leming>
glibc 2.16 introduced a new linker path for hard-float ARM, which you can more or less thank debian/ubuntu/linaro for obfuscating
<traeak>
which pacman depends on
<A2Sheds>
naeg, NTU and some others
<traeak>
the problem is the "workaround" resulted in pacman flat core dumping, that's it
<leming>
A2Sheds, you mean Arch Mobile then, that was a project that died, not us
<traeak>
the secondary workaroudn was to back out pacman first then do a system install then reinstall pacman
<leming>
you should be able to extract the new glibc package into /
<traeak>
there's a fundamental problem there to differing levels
<leming>
that will put things into place
<leming>
then -Syuf to fix the hanging files and clean things up
<traeak>
the little problem that the package manager busted and didn't get teh opportunity to download the required glibc
<traeak>
i'm not sure how that should be solved. in some instances pacman may have to be updated first to reflect changes in the package manager architecture
<traeak>
in other instances the package manager has to be installed last in order to not hose itself
<leming>
there is no pretty way to solve it
<leming>
but being a dick about the issue isn't one way
<Turl>
make the package manager static?
<traeak>
Turl: hehe that would be one yes
<leming>
they actually removed statically linked pacman a couple years ago
<leming>
caused more problems than it solved, iirc
<traeak>
if default actions hose the system i'd say that's a pretty massive bust
newbie12 is now known as hp_
<traeak>
i have the ability to figure out how to recover that stuff, but i'm not sure a typical user would
<traeak>
and people shouldn't take things so personally
<leming>
i believe we make clear on the front page, echoing upstream Arch's sentiments, that the distribution is designed for accommodating competent linux users
<leming>
it's really not my problem if people get in over their head and wreck their setup
<leming>
but if you have a problem, you should come to us instead of just calling us idiots somewhere else
<leming>
as you see, there is a completely valid explanation and nothing was actually hosed.. sequence and details are everything
<DonkeyHotei>
ZaEarl: i don't think immediate porting of aosp to devices is something to jump on. it should be a CM branch first
<rmull>
ManoftheSea: ping
<steev>
DonkeyHotei: that's crazy talk, aosp first, then cm
nibb has quit [Read error: Connection reset by peer]
nibb has joined #arm-netbook
<DonkeyHotei>
device-specific? seriously?
<ManoftheSea>
pong
<steev>
DonkeyHotei: device-specific?
<ManoftheSea>
er... rmull: pong
<rmull>
So regarding your email about the EOMA-68 passthrough board - I just want to make sure my head's screwed on straight
<rmull>
My understanding was that this is essentially a "breakout board" that takes the PCMCIA EOMA interface and makes the signals available on the usual connectors so that they're easier to interface with
okapi has quit [Ping timeout: 246 seconds]
<ManoftheSea>
I believe it's the opposite.
<rmull>
Oh geez.
<ManoftheSea>
A passthrough should provide signal IN to the motherboard, right?
<rmull>
well, these connectors do I and O alike...
<ManoftheSea>
indeed.
<rmull>
So are we saying the same thing?
<ManoftheSea>
I will make an example
<rmull>
But, for example, the HDMI output from the EOMA module
<ManoftheSea>
The Mobo side of SATA would have a disk.
<ManoftheSea>
The passthrough would make an eSATA available
<ManoftheSea>
for connecting to a host.
<ManoftheSea>
The Mobo side of the RGB would connect to a display (laptop, Smart-TV)
<ManoftheSea>
So the EOMA side would provide a VGA, HDMI, etc. in.
<ManoftheSea>
like a monitor.
<ManoftheSea>
Is this your understanding too?
<rmull>
I think I missed the mark entirely
<rmull>
I was thinking there was an ARM computing module inside the EOMA card, and it had a PCMCIA-style pinout
<ManoftheSea>
rmull, yes.
<rmull>
And that peripherals might be connected to this compute module through the EOMA interface
<ManoftheSea>
but a passthrough card would lack the CPU
<rmull>
So where would the CPU be for a passthrough?
<ManoftheSea>
Outside the EOMA-Motherboard assembly entirely.
<ManoftheSea>
The EOMA would breakout the ports of the EOMA header.
<rmull>
So an example from the wiki page: A tablet with a passthrough card could be used as a portable LCD monitor, including its touchscreen converted for use as a USB device.
<ManoftheSea>
I believe there is nothing "active" on the motherboard side of the EOMA-68 header.
<ManoftheSea>
The intelligence, the driving side is the CPUcard side of the EOMA68 header.
<ManoftheSea>
So, the tablet has a USB touchscreen and receives display from the RGB of the EOMA68 header.
RITRedbeard_ has joined #arm-netbook
vgrade2 has joined #arm-netbook
<ManoftheSea>
A passthrough would connect via USB and <Some display technology> to a real computer.
<ManoftheSea>
You'd plug the HDMI-out of a laptop and a USB A-A to the EOMA card front.
<ManoftheSea>
And the tablet becomes an input and display for the laptop
<ManoftheSea>
right?
<ManoftheSea>
Equally, I might want an EOMA passthrough for my (future) EOMA-complient laptop, so I can access the internal hard drive more easily.
RITRedbeard has quit [Read error: Connection reset by peer]
vgrade has quit [Read error: Connection reset by peer]
<rmull>
the motherboard, meaning the SoC and supporting circuits/
<rmull>
?
<rmull>
I really do apologize, I feel kind of stupid right now :P
<rmull>
This was my initial notion: I would eventually obtain an Allwinner A10 device in an EOMA-68 compliant module. I would have access to its IO signals, but they'd be hard to get to because they're on a PCMCIA connector. I might want to plug in peripherals, so I would connect a passthrough card which plumbs the EOMA signals to "normal connectors" like ethernet, hdmi, etc.
<rmull>
But now I'm not sure this is what was intended by the passthrough card.
<ManoftheSea>
the HDD/SSD would be connected to the EOMA68 header, the EOMA passthrough-card would put that out to an eSATA connector.
<ManoftheSea>
rmull: We're collaborating, no need to feel stupid.
<ManoftheSea>
This is my understanding.
<rmull>
You're right. I see now that an EOMA CPU module has "real" connectors on the side opposite the PCMCIA connector, so my initial notion is invalid.
<ManoftheSea>
If you agree, we're both smarter. If you disagree, make a case, maybe I learn somehting.
<ManoftheSea>
rmull, you're describing the mini-engineering board.
<rmull>
So the tablet has a bunch of peripherals (SATA disk, USB touchscreen, HDMI LCD) that are plumbed to an EOMA header so that they can be driven
<ManoftheSea>
Also something needed.
<rmull>
And ordinarily a CPU module would be needed to drive them
<ManoftheSea>
Have you seen that page?
<rmull>
But the passthrough card would allow a standard desktop to drive them by passing desktop-class connectors through to the eoma header
<ManoftheSea>
For the other direction, the mini-engineering board would take the EOMA68 header out to a "spread out bench" kind of setup.
<ManoftheSea>
You'd plug disks, displays, USB devices into the mini-engineering board, and drive them from the EOMA CPUCard.
Mazon has joined #arm-netbook
<rmull>
Okay, so now I think we're on the same page. However, is the CPU module in the mini-eng board driving two HDMI outputs, one through a standard connector and one through the EOMA header?
<ManoftheSea>
yes. The Allwinner A10 has TWO HDMI outs.
<rmull>
Okay! I think I get it now.
<ManoftheSea>
The proposed Allwinner A10 EOMA68 CPUCard would convert one of those to RGB and feed out the header.
<rmull>
For some reason I was wrongly under the impression that there were no standard connectors on a CPU module and the only way to interface with it was via EOMA
<rmull>
And that was the source of my confusion.
<ManoftheSea>
hmm.
<rmull>
But I get it now
<ManoftheSea>
Technically, the only "standard connectors" on an EOMA complient CPUCard would be the EOM68 header.
<rmull>
What about the HDMI connector on the other side?
<ManoftheSea>
However, the front panel (which may extend out) is up to the CPUCard producer.
<rmull>
Okay.
<ManoftheSea>
The front-panel HDMI is not part of the EOMA spec.
<rmull>
Great. Thanks for the consultation.
<ManoftheSea>
It's a bonus on the A10.
<ManoftheSea>
I believe the iMX53 would not have it.
<rmull>
Therefore, what you were saying in the beginning is still true - the passthrough card must take an INPUT on its EOMA header and put it out on an HDMI port.
<ManoftheSea>
no, to RGB
<ManoftheSea>
wait.
<ManoftheSea>
the passthrough card should be outputting to the EOMA header.
<rmull>
Argh
<rmull>
You're right
<ManoftheSea>
(which means accessing a disk, or driving a USB device)
<ManoftheSea>
Hmm... is it easier to say the Master side is on the EOMA Card, and the Slave on the motherboard?
<ManoftheSea>
in the case of passthrough, the Master is Outside of the assembly?
<ManoftheSea>
I believe there is a reference in the wiki that confuses the direction of signalling. I can't find it now, though.
<ManoftheSea>
rmull, if you're capable of it, the project needs an engineering board.
<ManoftheSea>
Even seeing that produced (sold as kit, etc) would be a morale boost for us all.
<rmull>
So I work for an embedded design consultant who would be happy to do this for money, though I dabble on the side for fun
<rmull>
This is not something that's backed by money, right?
<ManoftheSea>
You saw Gordon and LKCL's back and forth, right?
<rmull>
Haha, yeah :)
<rmull>
Point taken
<ManoftheSea>
As I understand it, lkcl and all have put in time, but not money.
<ManoftheSea>
Tom Cubie works for a chinese factory, where we got the sources for the A10.
<ManoftheSea>
I believe they produce the A10?
<rmull>
So my experience is generally with microcontroller-level stuff, not microprocessors needing high-speed routing
<ManoftheSea>
rmull, that's some of the beauty of the connectors.
<rmull>
Eh?
<ManoftheSea>
As lkcl would say, USB, SATA, 10/100 ETH are all a decade old.
<ManoftheSea>
They aren't what one might call "high speed"
<ManoftheSea>
And often, are differential signalling.
<ManoftheSea>
The SDRAM routing on the A10 CPUCard is hard. Luke reference it around "engineering team 3"
<ManoftheSea>
but the EOMA-complient motherboard is expected to be easy.
<rmull>
I could probably handle 10/100 and USB2
<rmull>
SDRAM, probably not...
<ManoftheSea>
well, if you do a passthrough or mini-engineering board, you don't have to.
<rmull>
Yep
<rmull>
I think I'd rather start with the passthrough board as a warm up, if you don't mind. I've only done one other design with kicad so far, and it was not complex.
hipboi has joined #arm-netbook
<ManoftheSea>
I wouldn't presume to TELL you what to do.
<rmull>
Haha.
<rmull>
Also, FWIW, unless it's BGA stuff, I can build short runs of boards if needed
<lundman>
pull request done, bored again
<ManoftheSea>
rmull: oh cool.
<ManoftheSea>
how many layers?
<rmull>
I would assemble the boards, not fab the PCB.
<ManoftheSea>
I think luke's hinted that the A10 board is like... 8-10 layers.
<ManoftheSea>
ah.
<rmull>
We have some preferred vendors for PCB fab but they wouldn't be the cheapest, since they're USA-based
<rmull>
And I assume cost is a priority
<ManoftheSea>
I'd kinda prefer it as an unassembled kit. But certainly others want it soldered on.
<ManoftheSea>
I believe you'd find takers.
<rmull>
Who gets to manufacture/build/sell these things if I put together a design?
<ManoftheSea>
(my opinion) You do. But I ask that you make your design open.
<rmull>
okay
<ManoftheSea>
There's a patent on the EOM68 header.
<rmull>
oh... owned by whom?
marcan has quit [Ping timeout: 272 seconds]
<ManoftheSea>
Luke has stated that if you are FSF endorseable, you get licensed for free.
<ManoftheSea>
lkcl and associates.
piezo has quit [Ping timeout: 272 seconds]
<ManoftheSea>
Thus, if you have no software in your board, I think it's easy to get endorsed.
<ManoftheSea>
As... all the source is available.
<ManoftheSea>
haha
<ManoftheSea>
please confirm with luke and your own lawyer, if it's important to you.
<rmull>
Hmm.
<hipboi>
anyone is building android jb?
<rmull>
Well, it's not a big deal. I don't mind making an open design. It'd be nice if we could manufacture it and sell it on digikey or something, since we have infrastructure for it, but shit, I don't care.
<ManoftheSea>
rmull: I can't think of a reason you couldn't.
<rmull>
No big deal.
<rmull>
Anyway - gotta tend to other stuff, but really, thanks very much, you've been very helpful.
<ManoftheSea>
Thank you, too, rmull.
<ManoftheSea>
bye
piezo has joined #arm-netbook
<ManoftheSea>
hipboi: not me.
<lundman>
not I
<hipboi>
well, i am syncing
<lundman>
hipboi: so, 10bit video isn't handled by A10, have you released enough that I can add that, or is that deep in the binary blobs?
<hipboi>
what do you mean by 10bit video
<hipboi>
lundman: do you mean the lcd driver?
<lundman>
newer mkv, like anime, uses 10bit per RGB instead of 8, for better mixing
<lundman>
very few things play.. xbmc, vlc.. but not mediaplayers like NMT/PCH/roku etc
<hipboi>
you mean to use the vpu - cedarx under linux?
<lundman>
I'm testing in android
<lundman>
TvdVideo just gives up, BSPlayer and MXPlayer will play at 5fps (software only) and the others just blackscreen
<lundman>
let me know if you want a sample file
<hipboi>
what about the 2160p player
<lundman>
I've yet to make that do anything but show me Albums
rzk has quit [Ping timeout: 245 seconds]
<lundman>
i wonder if I even have 2160p player, since the one thats on the mele just show art, and images, when I click play it gives me the usual players as option (mx, Tvd etc, no 2160p)
<hipboi>
really?
<hipboi>
wait
<hipboi>
what is Tvd
<lundman>
either way, I would assume 2160 uses the android playback, like all other players
<lundman>
TvdPlayer is what comes with mele
<hipboi>
i think it's the so called 2160p player
<lundman>
ah then yeah, doesnt play :)
<hipboi>
can you give me a sample video
<lundman>
the english in tvd is such that i think it is a chinese apps yeah
<hipboi>
l will report it
rzk has joined #arm-netbook
<lundman>
i know a few people that would buy mele if it did 10bit
<hipboi>
i takes me half an hour to download it....
<hipboi>
s/i/it
<hipboi>
s/i/it/
<ibot>
hipboi meant: s/it/it
<lundman>
guess the china-japan link is not great
ZaEarl has quit [Remote host closed the connection]
ZaEarl has joined #arm-netbook
<lundman>
brb, time for lunch
Turl has quit [Excess Flood]
Turl has joined #arm-netbook
hipboi has quit [Ping timeout: 264 seconds]
hipboi has joined #arm-netbook
lupul has quit [Ping timeout: 245 seconds]
QingPei has left #arm-netbook [#arm-netbook]
<ZaEarl>
hipboi, is allwinner going to get a jelly bean port going?
<lundman>
why not
<hipboi>
of course
<ZaEarl>
I guess I mean soon.
<lundman>
already runs 4.0.8, so the import should be quite straight forward
<hipboi>
ZaEarl: yes, soon
<lundman>
package in shenzen!
nibb has quit [Read error: Connection reset by peer]
nibb has joined #arm-netbook
<lundman>
i can practically (not at all) wave to those guys
<ZaEarl>
hipboi, thanks, that's good to hear. I'm wondering if jelly bean will require driver updates. if so, we're dependant on allwinner to provide those.
<hipboi>
ZaEarl: i think no driver updates is needed
<hipboi>
ZaEarl: anyway this will be results soon
<Turl>
hwcomposer might need changes
<Turl>
but we have source for hwcomposer
<Turl>
hipboi: any chance we can get a newer mali release binaries?
<hipboi>
yes
<hipboi>
r3p0
<hipboi>
maybe higher
<Turl>
r3p0 is the latest one on the arm site
<Turl>
r3p0-04rel0 - [08 June 2012]
<hipboi>
yes, now i have r3p0
<Turl>
if you can provide us with r3p0 binaries for android I'll update the kernel land to match :)
<lundman>
mali is for 3d only?
<Turl>
we will need also for linux, so we can keep a ingle kernel release merged
<Turl>
lundman: yes
<lundman>
ok
<Turl>
maybe 2d too, idk
<Turl>
but not video, that's for sure
<hipboi>
Turl: i can provide it later today
<hipboi>
both for linux and android
<lundman>
that is exciting
<Turl>
nice, thanks
Quarx has joined #arm-netbook
<Turl>
hipboi: later today I'm probably going to be sleeping or studying though :) you have my email however if you want to attach them there
<hipboi>
Turl: sure, no problem, i will upload to pulic space
<Turl>
that works too :)
<lundman>
yeah just make it public, we'll spread it like swine-flu
<mnemoc>
Turl: any plan of making an AOSP/JB for A10? or you'll wait for CM10?
<Turl>
mnemoc: well, I'm definitely not starting today, I have a final test tomorrow :P
<Turl>
mnemoc: but I synced the code and did a maguro build to ensure it's all right
<Turl>
mnemoc: hipboi__ promised updated mali binaries to the latest release, so I guess we're covered there as much as we can be :)
<hipboi__>
i think i will have to wait to tomorrow
<hipboi__>
i just edited the wiki, and planning upload it to the wiki
<Turl>
hipboi__: ok, no problem
<lundman>
hipboi: if you give those of us who can mirror, we can get the binary and help
<lundman>
mele releasing ICS was crazy, took 48 hours to download :)
<Shivan>
any update on Mali/X11/OpenGL ES guys?
<hipboi__>
lundman: yes
<hipboi__>
Shivan: i have sent a mele box to a guy from arm mali team
<hipboi__>
though we have official support from arm, we paid for it
<hipboi__>
i still ask him to help us
<Shivan>
I remember last time i tryied i was getting a (EE) MALI(0): [maliModifyPixmapHeader:165] UMP failed to retrieve secure id :/
<hipboi__>
secure id problem should be solved by modifying the fb driver
<hipboi__>
and after arm released the r3p0 driver and ddk, as they said better support for X11
<hipboi__>
i have tried it, too
<hipboi__>
but failed
<hipboi__>
it's a little mess now, i think i will sort everything and write a mail to the arm mali guy
<Shivan>
and what about wip/linux-3.4-sunxi/forward branch? it say "video: sun4i-disp: fix mali/x11 integration [Tom Cubie]" i tryied with it, but rtl8192 driver was failling to compile :/
simonckenyon has joined #arm-netbook
okapi has joined #arm-netbook
<Shivan>
another thing i wanted to ask is if someone knows where to get the Cedar libs for linux
<Shivan>
i did not find any info about it
kaspter has joined #arm-netbook
<Turl>
Shivan: the allwinerized rtl8192 was broken, yes
<Turl>
Shivan: use the one on the rtlwifi directory, that's the upstream one
<lundman>
:)
<Turl>
the allwinerization is like.. 5 lines over the free download you can get on realtek's site btw
<Shivan>
i compiled it whiout the rtl driver but the kernel fail to load anyway
<Shivan>
unless my aurora is playing with me again :/
<Shivan>
nop its the kernel, it doe snot work, at least not with default setting -rtl
vgrade has joined #arm-netbook
vgrade2 has quit [Ping timeout: 260 seconds]
<Turl>
what doesn't work, wifi?
vbmithr has joined #arm-netbook
<mnemoc>
aurora needs lcd_spi, which is not merged yet
<mnemoc>
the patch brings troubles with hdmi output
rellla has quit [Quit: ChatZilla 0.9.88.2 [Firefox 13.0.1/20120614114901]]
<rmull>
ManoftheSea: I'd like to discuss the EEPROM on the EOMA when you get a chance, ping :)
kaspter has quit [Read error: No route to host]
<traeak>
i guess they did that wiring thing since the jz4770 doesn't have hdmi out
<traeak>
and there's the other problem of using a vivante gpu
<traeak>
Turl: so with 3.4 it's not necessary to use rtl8192 ?
<mnemoc>
traeak: Turl has a hack on his tree to be able to use rtlwifi's
<traeak>
ahh "hack" not correct fix
<traeak>
ugh the arch guys, they even have problems admitting that they make poor decisions
<mnemoc>
correct "fix" would be part of the usb driver's power manager
<traeak>
ahhh
<traeak>
the level of pain on arch is about what it was on gentoo, except arch doesn't have to be continually built
<mnemoc>
too much ego in both...
<traeak>
a huge rule of software development: it all sucks, some of it sucks less than others
<mnemoc>
:)
<traeak>
and compromises are the norm. if you hit the ideal somehow you discovered the proper design (which rarely happens)
<Shivan>
To me, with an aurora, using 3.4 kernel the usb mouse/keyboard does not even turn on, thats equal no try to boot without the uImage, so i guessing the kernel is not loading at all
<Shivan>
what about the JZ4770? i have a Novo 7 Basic with a broken touchscreen around here...
<traeak>
sry, discussion way above about the new 50 + 15USD HDMI stick
<traeak>
if these mips soc companies played nicely with open source they could get a huge jump on the arm guys
<Shivan>
Ok some test the JZ4770 seens faster than A10...
<Shivan>
but that vivante gpu is crap
<traeak>
closed crap
<A2Sheds>
traeak, too much IP theft to have open hardware docs, they would end up in the sights of patent trolls
<Shivan>
as well the vpu or wharever is doing the video decoder on that chip
Turl has quit [Excess Flood]
<traeak>
A2Sheds: i'm afraid you are exactly right about that. The US patent law stupidity is out of hand
<A2Sheds>
IMHO we aren't going to see open docs until the patent system gets fixed
<Shivan>
my jz4770 could no even hadle a 10mbps 1080p file... even a +/- 8mbps 720P file gives troubles
Turl has joined #arm-netbook
<traeak>
and sadly contagious to other countries
<Shivan>
and HDMI sound is bad quality too...
<traeak>
Shivan: can't say if that's a poor video dsp or
<traeak>
from what i understand the jz4770 soc itself doesn't support hdmi
<traeak>
probably just a technicality...
<Shivan>
im not talking bout using hdmi out, only the tablet screen, just an example, nothing changes using hdmi either
<Shivan>
video processor on JZ4770 is bad, as well it is the vivante gpu for 2d/3d power
<Shivan>
cpu power is about an A8, maybe a bit faster in some things
<traeak>
power envelope is good though ?
<Shivan>
power usage? i remember that my basic could hit 11hs video playbackwith wifi off, my aurora is lucky if it hits 6
<Shivan>
JZ4770 is still good for a portable media player using low bitrate files, very good, but for something more that may need linux... nah i dont think it even worth the trouble... maybe as server
<Shivan>
-and about the power- both the aurora and basic has the same 3700mah bat...
eebrah has quit [Ping timeout: 245 seconds]
<traeak>
what's the cpu in the aurora?
<Shivan>
A10
eebrah has joined #arm-netbook
eebrah is now known as Guest55190
<Shivan>
jz4770 as a mips processor is very power efficient, thats why i was thinking to just chroot debian and use it as server, no need to waste time figthing with JZ4770 boot or vivante...
Guest55190 is now known as ibrah
<traeak>
i guess...but then you have tons of other options...like pogoplug, etc
lupul has joined #arm-netbook
ibrah has quit [Remote host closed the connection]
<Turl>
hipboi__: btw, do you know why allwinner implements cedar as a "media player" (parallel to stagefright) on android instead of using OMX like all the other vendors?
lupul has quit [Ping timeout: 245 seconds]
<simonckenyon>
any idea why i am getting make: *** No rule to make target `out/target/product/<whatever>/obj/lib/libMali.so'?
<Turl>
simonckenyon: because you didn't extract the proprietary files?
<simonckenyon>
i did. used extract_files.sh and they are in vendor/allwinner/<whatever>/proprietary
<mnemoc>
Turl: can we get a public android_allwinner_proprietary repo?
<Turl>
mnemoc: I'll set one up when we get the new mali libs
<mnemoc>
why not before? :p
<Turl>
simonckenyon: if you look on vendor/allwinner/whatever/proprietary/*mk you should have a line copying libMali to that location
<mnemoc>
people can always `repo sync` later ;-)
<Turl>
mnemoc: keeping the proprietary repo size down :)
<Turl>
those tend to grow huge :P
<mnemoc>
:)
<mnemoc>
or.... ./download-proprietary.sh ?
<Turl>
yeah, we'll come up with something :P
rellla has joined #arm-netbook
<simonckenyon>
can't be 100% sure but i think there was a problem with a extra \ on the end of vendor-blobs.mk line which sets the variable PRODUCT_COPY_FILES. i've removed it and the build seems to be proceeding
<simonckenyon>
anyone know what dmt0567.ko is for? is the mma7660 driver a possible substitute?
<simonckenyon>
there are two instances in setup-makefiles.sh that need to be amended.
<simonckenyon>
whois simonckenyon
<simonckenyon>
well it wasn't that. rats!
<simonckenyon>
copied them in by hand.
nibb has quit [Read error: Connection reset by peer]
nibb has joined #arm-netbook
<mnemoc>
simonckenyon: commands start with an / ;-)
tallnerd1985 has joined #arm-netbook
<simonckenyon>
learnt that the hard way :-)
<mnemoc>
:)
gimli has joined #arm-netbook
rellla has quit [Quit: ChatZilla 0.9.88.2 [Firefox 13.0.1/20120614114901]]
<ManoftheSea>
rmull: for the computer on the other side. It cares about the screen and gpios
<rmull>
So we pass the I2C through, right?
<specing>
having a dedicated eeprom on a board that already has a microcontroller just raises costs
<ManoftheSea>
My concept of a passthrough card would be to have a full computer drive the EOMA, as if it were in the sloot.
arokux_h has joined #arm-netbook
<ManoftheSea>
so, that full computer would read the i2c to learn the screen resolution and how to drive the GPIOs
<rmull>
Yeah, but there was talk on the mailing list of having a small MCU on the passthrough card, I guess to read the eeprom. It didn't make sense to me.
<ManoftheSea>
it doesn't to me either.
<ManoftheSea>
I was just about to respond to luke with more questions.
<ManoftheSea>
lkcl, ping
<rmull>
Please do, I am learning rapidly
<rmull>
And AFAIK I agree with you
<specing>
Some mcus are cheaper than eeproms ;P
<ManoftheSea>
specing, does the uC respond on the EEPROM address?
<specing>
ManoftheSea: if you program it to, sure
<ManoftheSea>
if your mobo uC responds with the device tree, I see that as being in-spec
<specing>
yeap
<ManoftheSea>
but you have to provide the sourses to be FSF-complient.
<ManoftheSea>
sorry about the spelling, I'm on a 2G and am typing faster than it's echoing
<rmull>
:D
<specing>
eeprom code would be like ~50 lines of C
<specing>
assuming you do it in an interrupt
<specing>
(I program AVRs as a hobby)
<specing>
you can store the info either in EEPROM or in FLASH
<rmull>
specing: Sure, but is the MCU necessary.
<specing>
rmull: if you have the mcu on the board for other stuff then I dont see a reason to add a dedicated eeprom
<specing>
if not, a dedicated eeprom is cheaper
<rmull>
I don't think we do have one there for other stuff on the passthrough card, which I am specifically interested in
<specing>
How big is this devicetree?
<specing>
1k? 2k? 4k? bigger?
<ManoftheSea>
rmull, the concept for the Monster Engineering Board includes a uC
<rmull>
Okay, sure
<ManoftheSea>
The mini, on the other hand, doesn't
<ManoftheSea>
rmull, the passthrough card doesn't need (nor shouldn't have) an EEPROM
<ManoftheSea>
As the EEPROM is on the device side of the EOMA header, not the Master side.
<specing>
how big is the device tree info in question?
<ManoftheSea>
and the passthrough extends the Master side
<ManoftheSea>
dunno.
<ManoftheSea>
Not clear on that one.
<ManoftheSea>
It's a human readable text file, compiled down to a binary format.
<ManoftheSea>
I expect it's < 2k, but that's a wild-ass-guess
<specing>
While I dont imagine one would put an AVR onto the cards, I can say that standard AVR mega EEPROM is 1 kilobyte, while the flash space ranges from 4kilobytes to 64 kilobytes
<Shivan>
so, OpenGL ES accelertion is possible at this point?
<specing>
I would be very interesting in working on the embedded controller
<ManoftheSea>
what about the STF32?
<specing>
So if possible, tell me which ones are likely candidates so I can acquire them
<ManoftheSea>
those initials are wrong...
<rmull>
STM32F
<ManoftheSea>
that's the suggested one!
<specing>
Ok
<specing>
Part number?
<ManoftheSea>
specing, you do the mini engineering board, and Rmull will do the passthrough card.
<specing>
Heh
<ManoftheSea>
And we'll have a completely useless but entirely within spec EOMA system
<ManoftheSea>
And honestly, that would be a huge win.
<ManoftheSea>
Because we could then replace pieces modularly.
<specing>
Why useless?
<rmull>
Gotta bootstrap it somehow.
<rmull>
This method seems fine.
<ManoftheSea>
specing, the passthrough would have USB and hdmi plugs. The mini-eng would have USB and video of some kind...
<ManoftheSea>
So, you're basically making an EOMA complient long wire.
<ManoftheSea>
Don't get me wrong, I'm entirely for it.
<ManoftheSea>
oh wait. I think I see Lukes point...
<rzk>
visual quality is crap, but no shorts/breaks on copper.
<rmull>
ManoftheSea: I don't see anything in the datasheet for the TI part that permits control of resolution
<rmull>
There are a couple fairly generic mode setting pins but nothing of the sort you describe, and nothing that seems like it would need to be driven by anything with smarts
<ManoftheSea>
email sent.
<ManoftheSea>
will be back later.
<specing>
rzk: let me guess about the components: one usb port fed into one 3v3 reg fed into one 1v8 reg that are both fed to the >100MHz micro? With one reset button and one that I dont know the function of? and the usb data is fwded to the micro?
<rzk>
nope
<specing>
:(
<rzk>
thats just xc2c128\256\384 breakout board for my university project
<rzk>
with two crystals onboard.
<specing>
well it is >100Mhz ;P
simonckenyon has quit [Ping timeout: 255 seconds]
<specing>
Its pretty much obvious when you see a 1v8 line ;)
<rzk>
I can place any crystal in that package, there is a big variety of them.
lkcl has joined #arm-netbook
<rzk>
I've used 50mhz, I think I will return to pcb stuff soon, want to conquerer ddr3 and pci-e.
<rmull>
ManoftheSea: Thanks for the mail - looks consistent with what we've been discussing.
<specing>
about the HDMI issue:
<specing>
the master key has been leaked a year ago
<rmull>
Is it toxic to provide access to it, a la decss?
<specing>
?
<rmull>
It's against the law (in the USA, I think) to host the decss code to crack DVD protection
<rmull>
There was talk about possibly including one on the passthrough board
<rmull>
But that only concerns the passthrough board
<mnemoc>
RaYmAn: picture of the wiring?
<hno>
specing, there is no need for a real eeprom on the board. But the I/O board needs to act as if there is an I2C connected eeprom, but may perfectly fine emulate one using it's EC controller if it has one.
<hno>
There is NO requirement for an I/O board to have an EC. The only requirement is to provide an I2C connected eeprom interface with board information, and power for powering the CPU board. Everything else is optional.
<hno>
specing, regardign HDCP, you only need an HDCP enabled passthrough card if you want it to be able to display HDCP protected input on the connected LCD.
gimli has quit [Remote host closed the connection]
<RaYmAn>
mnemoc: sec, just need to resize the image =P
<hno>
RaYmAn, how "quite easy"? Is the plastic easily removable so wires can be added?
<RaYmAn>
hno: yes, sec, image should make it clear
<lerc>
The Ouya seems to be popping along at $100k an hour on kickstarter.
<traeak>
that's kind of scary
<RaYmAn>
hno: the top bit basically just pulls off really easily, and then the prongs on the inside are sufficiently sized that standard jumper cables actually work :)
<traeak>
too bad it's a tegra3 though
<traeak>
i guess there's not much too much that can compete at this time though
<lerc>
Is it just openness/drivers that are a problem with the tegra3? Or does it more generally suck.
<Turl>
apparently it covers all latam
<hno>
lerc, "However, this creator cannot just be a family member or friend with US requirements; they have to be directly involved in the creation of the project itself."
<mnemoc>
going to sleep
<mnemoc>
good night!
<hno>
mnemoc, good night
mikie_w has quit [Quit: Leaving]
<Turl>
night mnemoc
mike_ has joined #arm-netbook
<lerc>
hno: Bother said pooh.
mike_ is now known as Guest49864
<hno>
Ok, reading the kickstarter creators faq a little closer.. the project creator must be permanent US resident but do not actually need to live in the US and the projet do not need to be executed in the US.
<lundman>
its US only?
eebrah has joined #arm-netbook
<traeak>
Thirty percent of revenue will go to Ouya, the rest to the developer." "every Ouya box sold includes a software development kit at no extra cost."
<lundman>
morning also
<hno>
lundman, morning and goodnight.
<RaYmAn>
traeak: easy to say when it's android ;) SO they just bundle android sdk :P
<traeak>
RaYmAn: i know...the ouya doesn't seem to offer any "technical" innovation atall
<traeak>
wasnt' there some idiot out there who wanted a super cheap handheld with a "market"
<traeak>
oh gosh, already forgot about taht from a few years ago
<traeak>
anyways
Guest49864 has quit [Quit: Leaving]
ka6sox is now known as ka6sox-away
<traeak>
the ouya guys can outsource the hardware design and construction no problem, they just need to work on a stable version of android and set up their storefront
<RaYmAn>
well, given Nexus7 is tegra3, building ICS/JB for it is extremely easy
<RaYmAn>
they really just need to add any specific drivers they require (sound etc)
<traeak>
so all they need is a store infrastructure.....and some way to generate tons of hype for first party developers
<traeak>
i wonder hwy amazon hasn't been all over this yet?
<lerc>
It'll still be a small market for the ouya. If they sell everything they are offering in the kickstarter that's still < 20,000 units.
<lerc>
Unless they up their limits they'll sell out in an hour or two.
<lerc>
Selling vapour is so much easier if you're american.
<ManoftheSea>
someone asked about kickstarter? lkcl has asked, they've given him the runaround.
<ManoftheSea>
And he says their story as to why not keeps changing.
<A2Sheds>
Amazon is probably too greedy or too controlling
eebrah has quit [Remote host closed the connection]
<A2Sheds>
plus the console is a small part of it, it's all about developing the games
<traeak>
amazon already has the vehicle for distribution in place
<traeak>
and already has publisher type stuff going, regardless that's neigther here nor there
<A2Sheds>
how will they get games devs to get on board?
<traeak>
how will ouya?
<A2Sheds>
yes, ouya
<traeak>
only through connections
<lerc>
They'll get indie releases easy. Big publishers would only do it if their android version works without effort.
<ManoftheSea>
Is Henrik on here?
fjfalcon has quit [Read error: Operation timed out]
<lundman>
he went tobed
<A2Sheds>
ok, so the OUYA is a tegra 3 PC with a TV out that runs Android.... so the games may be played on any arm Android device?
fjfalcon has joined #arm-netbook
<lerc>
A2Sheds: only in theory, in practice Input device is a much bigger deal.
<A2Sheds>
or are games somehow signed to only work on the OUYA?
<ManoftheSea>
lundman: what's his nick on here?
<lundman>
hno?
<ManoftheSea>
oh really? Hmm, I didn't put that together.