npcomp has quit [Read error: Connection reset by peer]
bsdfox has quit [Ping timeout: 246 seconds]
lucaswang has joined #linux-sunxi
npcomp has joined #linux-sunxi
cnxsoft1 has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 265 seconds]
p1u3sch1_ has joined #linux-sunxi
naobsd has quit [Ping timeout: 250 seconds]
<Turl>
cnxsoft1: it has been discussed heavily on the list, have a look
p1u3sch1 has quit [Ping timeout: 265 seconds]
naobsd has joined #linux-sunxi
<cnxsoft1>
Turl: I can see lots of discussions for the February "release", but nobody replied anything to Allwinner update on March 16. (Code submitted on March 13).
khuey is now known as khuey|away
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
vpeter has quit [Ping timeout: 265 seconds]
vpeter has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 252 seconds]
Snackman has joined #linux-sunxi
Snackman has left #linux-sunxi [#linux-sunxi]
konradoo77 has joined #linux-sunxi
Snackman88 has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 264 seconds]
kurain has quit [Read error: Connection reset by peer]
kurain has joined #linux-sunxi
kaspter has joined #linux-sunxi
reinforce has joined #linux-sunxi
kaspter has quit [Ping timeout: 252 seconds]
domidumont has joined #linux-sunxi
domidumont has quit [Remote host closed the connection]
domidumont has joined #linux-sunxi
RSpliet has quit [Read error: Connection reset by peer]
RSpliet has joined #linux-sunxi
<oliv3r>
mripard_: am I right in saying that the a80-optimus board has its dts revised to the 'new structure' and the old ones should also head that way (like the lime2 for example)?
cubeast has joined #linux-sunxi
simosx has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
<oliv3r>
cnxsoft1: the latest blobs are still binary only, but they seem to have debug symbols, though the encoder seems extremly small
sehraf has joined #linux-sunxi
<plaes>
cnxsoft1: unfortunately it doesn't seem to be as clean as Allwinner is claiming..
f15h has joined #linux-sunxi
<cnxsoft1>
Allwinner may be under the impression that what they did is OK though, since nobody replied in their thread.
iamfrankenstein has joined #linux-sunxi
<plaes>
it takes some time to analyze that mess :S
lucaswang has quit [Ping timeout: 250 seconds]
selfbg has joined #linux-sunxi
<mripard_>
oliv3r: ?
<mripard_>
oliv3r: all of them are converted.
<oliv3r>
ah then i'm seeing some old ones
<oliv3r>
i'm on 4.0-rc4 (vanilla)
Andy-D has joined #linux-sunxi
<oliv3r>
mripard_: but i see it proper in your sunxi-next repo; good, then I did it right :)
ricardocrudo has joined #linux-sunxi
FreezingCold has joined #linux-sunxi
<ssvb>
wens: 144mhz instantly deadlocks on a10-lime and 312MHz is just unreliable
<ssvb>
I'll reply to the mailing list as soon as claws-mail finishes synchronizing lakml, right now it is totally DoSed :(
<plaes>
ugh.. 5k messages so far this month
Night-Shade has quit [Read error: No route to host]
<wens>
ssvb: it's because i copied the wrong settings
<wens>
a10's don't go that low
<ssvb>
wens: the allwinner 3.4 kernel can go as low as 60mhz on a10
<wens>
maybe the instability is due to the low voltage then
<ssvb>
the 3.4 kernel does not reduce the core voltage below 1.25V though
<ssvb>
btw, I'm running tests with dts modified to have 1.4V for all operating points and it still fails
<ssvb>
and you mentioned that the voltage regulators are not supported yet anyway
<wens>
also possible is ahb/apb (derived from cpu) were too slow
<ssvb>
60mhz was where the cpu was sitting with the ondemand governor on a10, resulting in absolutely horrible system responsiveness, as observed by many users :)
ssspiff has quit [Quit: leaving]
awe00 has joined #linux-sunxi
<ssvb>
low clock frequency operating points are very much harmful
<wens>
probably not so useful
gianMOD has quit [Remote host closed the connection]
<alexvf>
even if i continue calling the ioctl, it never seems to get the vsync
<alexvf>
i just wanted to know what event triggers the vsync, i looked into linux-sunxi code and i found that it needs a change in g_fbi.wait_count[sel] variable but i can't find what function makes that change
<hno>
alexvf, probably not implemented properly.
<alexvf>
hno: but it is weird, because it works up to some point
<alexvf>
hno: do you know what piece of code should be failing to change that variable? i could look into it and fix it is nor propery implemented
<hno>
alexvf, I don't know that code too well, but I suppose it should be triggered by a TCON vertical blank or line interrupt.
<hno>
but maybe that is only for LCD/composite, and not HDMI. Not sure.
<ssvb>
alexvf: can you elaborate on "it works up to some point"?
<alexvf>
ssvb: yes, i'm playing a video without problems and, at some point i start getting errors in the FBIO_WAITFORVSYNC ioctl
<alexvf>
i have a small player implemented for a20
<alexvf>
i call that ioctl after rendering each frame
<alexvf>
once i start getting errors, the video plays sloppy and with a lot of noise
<ssvb>
I just suggest to try the latest kernel from the stage/sunxi-3.4 branch
<alexvf>
ssvb: the kernel i have is from that branhc
<alexvf>
but i think it is not the latest
<ssvb>
if it still does not work properly, then it would make sense to debug the problem
<alexvf>
i should confirm what exactly version ...
<ssvb>
if you want a help from the others with debugging, then explaining the exact steps to reproduce it would definitely help
<alexvf>
ssvb: i know, but it is hard because i'm not debugging the kernel itself
<alexvf>
just want to know how it works to fix my code
<alexvf>
but i will try to write simple program that fails
Renard has joined #linux-sunxi
<alexvf>
ssvb: maybe i can try the vsync demo on xf86-video-fbturbo and try to modify it to get the error
hipboi_ has joined #linux-sunxi
<alexvf>
hno, ssvb: thank you for the help
hipboi has quit [Ping timeout: 256 seconds]
Seppoz has quit [Remote host closed the connection]
Seppoz has joined #linux-sunxi
kaspter has joined #linux-sunxi
Seppoz has quit [Remote host closed the connection]
Seppoz has joined #linux-sunxi
T0mW has quit [Ping timeout: 255 seconds]
T0mW has joined #linux-sunxi
speakman has joined #linux-sunxi
<speakman>
Hi folks. I've attached an pcf8563 RTC clock through /sys/class/i2c-adapter/i2c-2/new_device. Do you have any idea how to use the kernel CONFIG_RTC_HCTOSYS feature to make it set the time from the RTC at boot?
<speakman>
I'm on A20 which has RTC built-in, but Olimex didn't provide battery backup for it so I have to use an external.
<wens>
setup hwclock to use rtc1 instead of rtc0
<wens>
on debian, edit /etc/default/hwclock
<speakman>
Using the built-in one, the /dev/rtc0 is ready at boot. Using the generic i2c + pcf8563 driver one has to, AFAIK, write "pcf8563 0x51" to the 'new_device' sys file for the /dev/rtcX device to come up.
<speakman>
wens: This is about making use of kernel HCTOSYS feature
<speakman>
Doing it in a startup script (I'm not on Debian) is not a problem, but it would be more convenient letting the kernel do it.
<wens>
set RTC_HCTOSYS_DEVICE in Kconfig
<speakman>
Sure. To what? It's not available during boot. Did you even read my posts?
<atsampson>
speakman: could you add your new i2c device to the device tree instead? (and you could always disable the existing one that way too..)
libv has joined #linux-sunxi
<speakman>
atsampson: oh, I'm on old 3.4 without Device Tree support :(
<libv>
oliv3r: that email is a new low for allwinner indeed
<libv>
how anyone can today still try to defend them is beyond me.
<plaes>
?
<speakman>
I see there are Device Tree sources for Olimex A20 Micro. Which kernel is both stable and does support device trees for A20?
<libv>
plaes: it turns out that allwinner added LGPLed code in their youngest cedarx binaries
<libv>
a large part of their h264 decoder is based on libavcodec code, with the names trivially changed
<libv>
hard proof is on the ml
<oliv3r>
yeah that namechange is almost too obvious, and strongly suggest intent
<oliv3r>
obfuscation by intent i mean
<plaes>
ok.. hadn't checked mail in 2 hours.. :)
<libv>
they have zero control over the content of those binaries
<oliv3r>
so they _know_ they are doing shit wrong
<libv>
and nobody can trust them
<oliv3r>
libv: as per mail; i'm pretty sure this is just 1 engineer being an jackass
popolon has joined #linux-sunxi
<libv>
sfc was in bcc on that email, as was phoronix.
<oliv3r>
benn told me long ago, that cedarX is allwinner developed VPU code (VHDL)
<libv>
yeah, i heard that too
<oliv3r>
done by 1 engineer, so i understand why he is being 'protective of his IP'
<oliv3r>
from what I see, what they did is, take the reference implementations
<libv>
ok
<libv>
still.
<oliv3r>
like libavcodec, put half in silicon, leave the rest in cedarX
<libv>
now nothing of those binaries can be trusted, and this should definitely result in legal action
<oliv3r>
you can always respond with 'proof' on the github issue, if you could be so kind
<oliv3r>
and then get some of the ffmpeg people to sue!
<oliv3r>
(i'm not an ffmpeg copyright holder)
<libv>
there was solid proof attached for one of the h264 symbols
<libv>
with a link to the original code
<plaes>
damn idiots
<oliv3r>
libv: i'll copy paste it into the github issue tracker :)
<libv>
what then?
<oliv3r>
your comment about it being plainly copy/pasted
<libv>
ah, i just meant: what good will that do
<libv>
that's just one allwinner employee afaict
<libv>
this will hit phoronix now
<libv>
Many factions in allwinner were made well aware of the LGPL issues in cedarX in august.
<libv>
Quink stated that cedarx 2.0 was done 3 months ago
<libv>
that's 4 months where lgpl issues were known, but in which time more lgpled code was added
<plaes>
cnxsoft1: ^^
<plaes>
answers to your question earlier
<libv>
i should actually make 2 emails of the exchanges i had in august/september/october after i discovered libvp62 symbols
<libv>
... public
Black_Horseman has quit [Remote host closed the connection]
<plaes>
(y)
<libv>
the information given to allwinner then, and the replies i got then, compared to the actions taken: namely a kernel with all binaries and code removed, and a continuation of the cedarx binary mess; that very clearly justifies my actions today and my strong stance.
naobsd has joined #linux-sunxi
bfree has quit [Quit: leaving]
lucaswang has joined #linux-sunxi
bfree has joined #linux-sunxi
ssvb has quit [Ping timeout: 255 seconds]
naobsd has quit [Quit: naobsd]
Andy-D has joined #linux-sunxi
leviathancn has joined #linux-sunxi
leviathancn has quit [Ping timeout: 256 seconds]
leviathancn has joined #linux-sunxi
<cnxsoft1>
plaes: thanks! So they just renamed the functions to make it appear like it was their code...
sillysnowflake has joined #linux-sunxi
<libv>
cnxsoft1: latest emails on our ml have all the details :)
lucaswang has quit [Quit: lucaswang]
Andy-D has quit [Ping timeout: 245 seconds]
f15h has quit [Remote host closed the connection]
awe00 has quit [Ping timeout: 244 seconds]
cubeast has quit [Read error: Connection timed out]
cubeast has joined #linux-sunxi
ricardocrudo has quit [Ping timeout: 265 seconds]
mozzwald has joined #linux-sunxi
mozzwald has quit [Client Quit]
awe00 has joined #linux-sunxi
mozzwald has joined #linux-sunxi
mozzwald has left #linux-sunxi [#linux-sunxi]
cosm has joined #linux-sunxi
nove has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
leviathancn has quit [Ping timeout: 265 seconds]
Skylark has joined #linux-sunxi
HeHoPMaJIeH has quit [Remote host closed the connection]
<ssvb>
libv: "the top of the list of companies hated by the open-source community", as if we did not have enough drama already...
* ssvb
expects even more wannabe freedom fighters in the mailing list
kurain has quit [Ping timeout: 250 seconds]
<hno>
I would not place them at the top of the list..
<plaes>
mhm
<simosx>
echo-chamber stuff
reinforce has quit [Quit: Leaving.]
<ssvb>
libv: and technically, the Allwinner's statement that "there is none GPL issues in the media-codec-lib" might be in fact true :-)
afaerber has joined #linux-sunxi
<ssvb>
assuming that they did not "borrow" some GPL code too, in addition to LGPL and BSD stuff
<simosx>
phoronix will publish anything with little checking. Here they talk about "UbuTab" (http://www.phoronix.com/scan.php?page=news_item&px=MTg1Mjc), a campaign for an Ubuntu tablet on Indiegogo. It was a single-person project from someone not related to Canonical.
leviathancn has joined #linux-sunxi
<hno>
LGPL issues is by extension GPL issues. Same rules apply wrt providing access to code. Only slightly different level of "infection".
<simosx>
Those that gave money to the campaign reading the phoronix.com article, will be waiting for long...
<hno>
what campaign?
<simosx>
hno, was refering to something else (re: phoronix)
simosx has quit [Quit: Leaving]
<ssvb>
anyone else getting very annoying "windows server" (!) popup advertisements at the phoronix website?
cnxsoft1 has quit [Ping timeout: 265 seconds]
<ssvb>
I would guess, it's a nice migration path for the "haters" :)
<hno>
ssvb, ads is from doubleclick. You have obviously been interested in Windows Server and that's why you get them there. I got an ad about cleaning paper, guess they know the number of kids over here..
<ssvb>
hno: they are not supposed to know any information about me other than the ip address, and since I have no interest in the Windows Server, I feel a bit relieved that they seem to have no clue :)
<hno>
libv, have it been ruled out that both CedarX and the too similar code in ffmpeg is sharing a common ancestor? And this whole thing should really have gone by the lawyers rather than public mailinglist.
<libv>
ssvb: at this point, allwinner has lost all credibility. a bit of extra drama will not further hurt things
<libv>
hno: this indeed should become a lawyer thing soon
<libv>
hno: i bcced sfc on that first email
<libv>
this h264 code _is_ the same code, just has the asserts compiled out
<libv>
the asserts and the unused argument
<libv>
(which was used by the assert)
<hno>
I don't doubt that it is the same code.
<libv>
i can spend ages detailing how to convert from my pre-parsed assembly to the result i posted to the ml
<hno>
Not needed.
<libv>
but that is about 4x the amount of work and then i really do need to be very certain that i can bill those hours to someone
<libv>
preferably allwinner.
<hno>
but this is not only about showing that the code is close to equal.
<libv>
hno: what is missing then?
<hno>
if you publish any more facts, please just state & document facts. Don't make any conclusions or accusations before talking to sfc.
<libv>
perhaps i should make a big wiki write-up on what we have found in the cedarx binaries so far
<libv>
and how allwinner has responded to what
<ssvb>
leviathanch: wtf were they thinking? just like kids getting caught in this silly act is not going to make things any better
<hno>
for one thing you must also prove that there is no other possible source it can be based on before making the conclusion on inheritance.
gianMOD has quit [Remote host closed the connection]
<libv>
hno: there are many different cases at different times and in different places in both cedar and libavcodec code, where clearly libavcodec code was used
<libv>
google turns up no other sources
<libv>
for these symbols
<libv>
and the 2 cases which i manually decompiled are from different parts of both cedar and libavcodec
<hno>
which includes knowing in detail the origin of the now LGPL licensed source and assured that that the authors of that code have not also licensed it under other terms.''
<libv>
ok, will email the original authors soon.
<ssvb>
leviathanch: the best thing that they could do was to take down the media-codec repository and consult the lawyers
<ssvb>
leviathanch: and depending on what the lawyers say, either abstain from ever publishing it, or clean the license problems in a proper way
naobsd has joined #linux-sunxi
<libv>
those omx mit style licenses added just now make this code copyrighted by the linux foundation?
<hno>
what?
<libv>
and then below that license statement, there is the allwinner copyright statement
<libv>
that smells really fishy
<hno>
Ah, saw the OMX license change now.
<libv>
haha
<libv>
this is an mit style licenses before an lgpl license
<libv>
people really seem to have absolutely no idea what they are doing
<ssvb>
libv: you may probably want to talk with leviathanch because he is the author of some of the commits there (typo fixes)
<libv>
ssvb: i think that levianthanch will respond as soon as he is awake :)
<ssvb>
libv: oh, ok :)
<libv>
i also trust that he starts translating/communicating what is flying around now
<libv>
so that perhaps allwinner might start to understand the gravity of the situation, for a change
<ssvb>
libv: regarding the MIT style license in some of the source files, it is perfectly compatible with LGPL so there should be no problem
gianMOD has joined #linux-sunxi
<libv>
sure, but it is pointless to add this after the fact, on top of lgpl
<ssvb>
libv: but it's good that they seem to have started acknowledging the origin of the code, maybe they do have in fact a lawyer consulting them now
<libv>
if they stole mit code and removed the copyright before, you would expect the mit license to sit below, and a short explanation that some code was taken from ... and was originally licensed mit
<plaes>
it was probably from the original files
<libv>
ssvb: i am not sure whether they are doing exactly that
<libv>
it can be assumed
<libv>
but at this point i would not be so certain
<libv>
there is an enormous amount of inane stupidity inside allwinner
cubeast has joined #linux-sunxi
<hno>
please stick to facts and leave shit throwing to lawyers, no matter how mad you feel about it.
<ssvb>
diego71: hmm, are you sure that you need to patch it?
<T0mW>
ssvb: yes, they left the definition out of the linux-3.19 device trees for the all-winner, only put HDMI in there.
sillysnowflake has joined #linux-sunxi
<ssvb>
diego71: maybe also try a newer kernel? 4.0-rc4
<diego71>
T0mW: it means than with this patch it work with lcd, and not with hdmi?
konradoo77 has quit [Ping timeout: 264 seconds]
<diego71>
(not a big problem atm)
<T0mW>
diego71: correct, I have an LVDS LCD that I'm using on my project with the A20
<diego71>
ok
<diego71>
T0mW: thanks
gianMOD has quit [Remote host closed the connection]
<T0mW>
look in the various arch/arm/boot/dts/sun*.dtsi for 'framebuffer' and all you'll find is the definition for the hdmi.
<diego71>
ssvb: tomorrow i'll try the new kernel
<T0mW>
Just copy the hdmi section and rename the pipeline to 'de_be0-lcd0'
domidumont has joined #linux-sunxi
<T0mW>
oh, and make it 'framebuffer@1'
<diego71>
in the evening I need to do a demo with the olinuxino and the lcd
gianMOD has joined #linux-sunxi
Andy-D has joined #linux-sunxi
gianMOD has quit [Ping timeout: 252 seconds]
<T0mW>
You'll find that u-boot looks into the device tree file (dtb) in ./drivers/video/sunxi_display.c, it is looking for 'de_be0-lcd0' so it can create the video node to handoff to the kernel.
gianMOD has joined #linux-sunxi
gianMOD has quit [Remote host closed the connection]
<sillysnowflake>
on the tutorial page linux-sunxi.org/Bootable_SD_card#Bootloader there's two methods and not much explanation except "as an alternative, you can use the combined u-boot-sunxi-with-spl.bin:"
libv has quit [Ping timeout: 264 seconds]
libv has joined #linux-sunxi
<sillysnowflake>
so one is sunxi-spl.bin and the other is sunxi-**with**-spl.bin. what's the difference
<diego71>
T0mW it works!
<T0mW>
diego71: you're welcome! :)
<diego71>
:)
domidumont has quit [Remote host closed the connection]
domidumont has joined #linux-sunxi
diego_r has quit [Ping timeout: 252 seconds]
<libv>
ssvb: ah, ok.
<libv>
pretty shoddy work though
domidumont has quit [Client Quit]
bonbons has joined #linux-sunxi
<ssvb>
libv: yes, the original file had the "Redistributions of source code must retain the above copyright notice" requirement written in clear English
<libv>
more proof that allwinner has absolutely no control over what code is inside cedar
<libv>
they have "appropriated
<libv>
" from all over the place, indiscriminately
<ssvb>
stripping the copyright notice had no sense at all, there is just no logical explanation (and just being "evil" is not a good explanation either)
<libv>
well, their recent actions are harder and harder to explain with "stupidity"
domidumont has joined #linux-sunxi
<libv>
there is however little credibility left for the cedarx binaries, and some independent IP reviewer needs to go in and waste many hundreds of hours reviewing all of allwinners cedarx source code and revision history
<libv>
and even then, the lgpl will apply.
<libv>
it clashes with some other licenses, but the lgpl does apply
<libv>
the only way forward to allwinner is to fully support cedrus
<ssvb>
sillysnowflake: the banana pi should be supported
<sillysnowflake>
is the "normal" way building uboot myself?
<atsampson>
sillysnowflake: the SPL is the very first bit of code that runs; it then loads u-boot proper -- you can either install the SPL and u-boot separately, or use u-boot-sunxi-with-spl.bin which combines both into one file (the result is the same either way)
<ssvb>
sillysnowflake: yes, I think that everyone is currently building u-boot himself, it is not too difficult
<ssvb>
sillysnowflake: but there are also various "bananian" sd card images floating around too
<ssvb>
atsampson: would it be a good idea to purge the wiki page a bit and remove "alternatives" in order not to confuse people?
<sillysnowflake>
my goal is to have gentoo installed somewhen in the future, but i guess installing bananian first would be a good practice..
<atsampson>
ssvb: might as well -- I always use the combined file these days
<ssvb>
sillysnowflake: well, if you are familiar with gentoo, then compiling u-boot should be very easy for you
viccuad has joined #linux-sunxi
<sillysnowflake>
atsampson: also describe how to get the binaries, either compiling oneself or getting ready-made ones.. that'd be handy..
<libv>
ssvb: their previous code was based on the h264 avc reference code, between version 9.0 and 16.0
<sillysnowflake>
ssvb, i somehow can't get crossdev to make me an arm toolchain for whatever reason and i don't feel like fiddling around with that today.. xD.. :c
<libv>
ssvb: i spent ages downloading a load of versions of that and throwing them into git for comparison
<ssvb>
sillysnowflake: let me know if this does not work
<sillysnowflake>
ssvb, i will try that. first i'll just get a simple bananian installation working because otherwise my motivation will go away.. heh..
VargaD has joined #linux-sunxi
<sillysnowflake>
thanks for the help
domidumont has quit [Remote host closed the connection]
<ssvb>
sillysnowflake: better try it the other way around
<ssvb>
sillysnowflake: the wiki page contains very detailed instructions, just follow them precisely and everything should be fine
<ssvb>
sillysnowflake: if something fails, just complain here and we will patch the wiki
domidumont has joined #linux-sunxi
<ssvb>
sillysnowflake: an no, I personally don't recommend using random sd card images from the internet :-)
<ssvb>
libv: ok
<ssvb>
libv: imho, you are spending too much time on this counterproductive struggle
cosm has quit [Ping timeout: 250 seconds]
<ssvb>
libv: especially considering that the vpu reverse engineering is mostly done and we don't gain that much from the allwinner sources
<diego71>
ssvb: the funny things they have only to gain from releasing the sources
<ssvb>
diego71: yes and no
<diego71>
ssvb: why not?
<ssvb>
diego71: as we can see, they have reasons to hide certain things
<ssvb>
it's kind of like exposing their dirty laundry :-)
* nove
loses in any way
<ssvb>
nove: why?
<diego71>
something different from opensource code?
domidumont has quit [Read error: No route to host]
domidumont has joined #linux-sunxi
<ssvb>
nove: yes, the reverse engineering efforts become kind of a wasted time if the allwinner cedar code becomes open in the end
<nove>
ssvb: anything i do, is a game that i will lose in the end
<nove>
that is why this hesitation
libv has quit [Ping timeout: 240 seconds]
libv has joined #linux-sunxi
<sillysnowflake>
ssvb: the banana pi has a cortex-a7 cpu, the wiki says EXTRA_ECONF="--with-cpu=cortex-a8... should i leave it at that or change it to ..a7 ?
<ssvb>
nove: I don't know, the most important thing when working with open source and not being paid for it is to have fun
jinzo has joined #linux-sunxi
<ssvb>
nove: if you expect people to be grateful and appreciate your efforts, then it's a sure way to a major disappointment
<ssvb>
sillysnowflake: you can keep the cortex-a8 part, this way the toolchain is compatible with all allwinner devices
<sillysnowflake>
oki
domidumont has quit [Ping timeout: 246 seconds]
<ssvb>
sillysnowflake: it's a bit confusing, but cortex-a7 is newer than cortex-a8 :-)
<sillysnowflake>
wat
<nove>
ssvb: i would not expect much, except to only, don't make the task harder
<nove>
but even that is difficult to have
<ssvb>
what are you not happy about?
afaerber has quit [Quit: Verlassend]
<sillysnowflake>
ssvb: are you sure that arm-linux-gnueabihf even works? i'm getting an error which seems to be caused because the arch specified is invalid. like "tc-arch: command not found".
<ssvb>
sillysnowflake: just ignore it
<ssvb>
sillysnowflake: do you mean that the whole crossdev invocation fails? or is it just a single message shown during build?
konradoo77 has joined #linux-sunxi
<sillysnowflake>
it's just a single message during build. crossdev will probably fail very soon, since it had to download gcc 4.8.4 (using --stable instaed of all the specific versions failed)
<nove>
well you know, users with the favorite player + licenses deregrad + complains that we(i) aren't work hard enough
konradoo87 has quit [Ping timeout: 255 seconds]
ricardocrudo has quit [Ping timeout: 252 seconds]
<sillysnowflake>
ssvb: oh my god crossdev completed without any furter errors. this is the first time crossdev ran successfully with me around xD
<sillysnowflake>
what wiki page does "Debian alike arm-linux-gnueabihf variant in order to be able to use the instructions from the linux-sunxi wiki as-is" refer to?
<ssvb>
sillysnowflake: it means that you can just copy-paste the commands from http://linux-sunxi.org/Mainline_U-boot#Build and invoke them as-is, for example "make CROSS_COMPILE=arm-linux-gnueabihf-" is going to work fine
<ssvb>
sillysnowflake: instead of having to change it to "make CROSS_COMPILE=armv7a-hardfloat-linux-gnueabi-"
<sillysnowflake>
yeah i know i wasn't sure what "the wiki" was since there was no link to "the wiki"
wigyori has quit [Remote host closed the connection]
montjoie[home] has quit [Ping timeout: 264 seconds]
<ssvb>
nove: this is understandable, but these losers can't "make the task harder" :-)
<ssvb>
but when the upstream maintainer is making ridiculous demands, then this is becoming a real problem
montjoie[home] has joined #linux-sunxi
<ssvb>
nove: is the v4l2 framework already a good fit?
<nove>
ssvb: it is
<ssvb>
then it's good
<ssvb>
does it support memory buffer to memory buffer transcoding?
<nove>
but, we have to "no bitstream parsing in the kernel", likely to be big problem to go mainline
<ssvb>
hmm, then this sounds more like "not a good fit" to me
<nove>
ssvb: in the form of v4l2 mem2mem device, yes
* ssvb
just wants to understand whether the kms driver is really a blocker for the vpu mainlining
<nove>
ssvb: the bitstream parsing can be done by the means of libv4l plugin, as (rockchip+google is doing)
<nove>
ssvb: it is not
<ssvb>
nove: ok, this is great
npcomp is now known as roboghost
_massi has quit [Remote host closed the connection]
pmattern has joined #linux-sunxi
<libv>
ssvb: i do not care about the sources, i do care about allwinner stopping to distribute these binaries
<libv>
and i would prefer it if allwinner directly supported the REed driver
<libv>
if that source code does get released, very little of it can be used
<libv>
but it might be some form of useful documentation
<libv>
i prefer actual hw documentation though
<libv>
and direct support of the cedrus guys
<libv>
nobody really wants to see that crap allwinner frankenstein code
<libv>
but allwinner must be stopped from distributing it.
<libv>
and i did that h264 reference decoder stuff in august
<ssvb>
libv: well, then it is probably better not to be beating around the bush but just say it directly
<ssvb>
leviathanch: ^
<libv>
i have.
<libv>
which is why i _now_ am making such a big stink.
<ssvb>
well, you are stating the "problem", but allwinner probably wants a "solution"
<libv>
they were given a nice list of doable todos
<libv>
and they stated that they were going to do so
<libv>
and instead that kernel and this bullshit happened
<libv>
ssvb: i even held back sfc action back in september, as things seemed to evolve positively
<libv>
i ended up wasting 4-5 months.
<ssvb>
but the kernel thing is already in the past? at least with the recent libnand release
<ssvb>
and now we have the dram initialization code for u-boot
<libv>
ssvb: check the list
<libv>
allwinner added more binaries.
<ssvb>
yes, the newer SoCs are still problematic, but this is probably going to be resolved eventually
<libv>
ssvb: still.
<ssvb>
unless we piss off allwinner and they stop cooperating completely
<libv>
and again, i gave them a nice list of doable todos
khuey is now known as khuey|away
<libv>
for kernel and uboot i told them that they should just state mea cupla publically, and that they will over the course of months resolve the issues
<libv>
no public statement happened
<libv>
and that kernel with all binaries released as thrown out
<ssvb>
I believe that with the newer SoCs the highest priority for them is time to market
<libv>
s/released/removed/
<libv>
ssvb: even then
<libv>
a license is a license.
<libv>
it's black and white.
<ssvb>
so they throw out whatever works and then clean up the license issues after the fact
<libv>
again, and again, and again: they were given a nice list of todos, which were all handleable/
<libv>
ben el-baz stated that they were working down that list of 4 items
<libv>
then kevin or shuge also stated that they were workign on resolving the gpl issues with kernel and uboot, and on lgpl issues with cedarx
<libv>
this was september.
<libv>
before that ludicrous kernel was released
<libv>
ssvb: this shit is all on allwinner.
<libv>
stop trying to shut me up on this, i gave them a bit of crap publically, and in the backchannel fed them the severe issues with cedar, and gave them 4 doable points of action
<libv>
they said they would work them through
<libv>
and then this shit happened
<libv>
now they get what they played for.
dev1990 has joined #linux-sunxi
konradoo87 has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 252 seconds]
lioka has quit [Changing host]
lioka has joined #linux-sunxi
<ssvb>
atsampson: did you have no time yet for trying a few more tests with 400mhz mbus?
<atsampson>
ssvb: I've been using the boards for other stuff this week
<atsampson>
next week'll probably be better ;-)
<ssvb>
ok, thanks
<ssvb>
one or two tests could be tried as an overnight run, if the boards are not busy 24/7
<ssvb>
once the 400mhz mbus mystery is resolved, we (you) could try to submit an "overclocked" defconfig to the mainline u-boot
<ssvb>
atsampson: maybe cooper could also help and run tests on his two pcduino3 nano boards, that would make it 4 boards total
Netlynx has quit [Quit: Leaving]
nove has quit [Quit: nove]
awe00 has joined #linux-sunxi
marcin_ has quit [Ping timeout: 264 seconds]
paulk-collins has quit [Quit: Quitte]
khuey|away is now known as khuey
pmattern has quit [Ping timeout: 272 seconds]
paulk-collins has joined #linux-sunxi
libv has quit [Ping timeout: 240 seconds]
libv has joined #linux-sunxi
pmattern has joined #linux-sunxi
konradoo87 has quit [Ping timeout: 250 seconds]
premoboss has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
simosx has joined #linux-sunxi
dack has joined #linux-sunxi
dack has quit [Remote host closed the connection]
deasy_ has joined #linux-sunxi
<deasy_>
hi!
<deasy_>
there is a thermal probe in the A10 ? don't remember/know
nicksydney has joined #linux-sunxi
<premoboss>
deasy_, yes. cpuburn rely on that.
marcin_ has joined #linux-sunxi
<deasy_>
i don't know this software, i have read on wikipedia it's write for x86 in asm
marcin_ has quit [Ping timeout: 264 seconds]
domidumont has joined #linux-sunxi
<deasy_>
ho i see a cpuburn-arm on github
<simosx>
deasy_, it's strange that the A10 datasheet does not mention such a feature.
<deasy_>
simosx: maybe not necessary...i don't know
<deasy_>
i would read it just for know the temp and if i will need to put a cooler or not
<deasy_>
i will test it in summer with a load of several minutes
<deasy_>
if 50 or more ===> cooler
viccuad has quit [Quit: WeeChat 1.0.1]
<simosx>
I had a look at heatsinks. There are quite small (tiny) because the surface of the SoC is small as well. Some people managed to cut big desktop heatsinks to fit on the SoC area.
<deasy_>
a little piece of Al is okay
<deasy_>
simosx: what is the difference between this and the older i have see in chinese
<deasy_>
was not manual of allwinner in chinenglish ?
konradoo77 has joined #linux-sunxi
<simosx>
deasy_, that github repo has the latest versions of the documents, it's an official release by allwinner.
<deasy_>
arokux: yup
<deasy_>
is arokux always silent ?
<deasy_>
i have not look at this since a long long time so i don't even remember the date of the document in ChinEnglish
<simosx>
Google Images show the A20 in other boards not to have that cursive T. But for the BananaPi, there is indeed a T.
<simosx>
It could be a more recent batch. It would help to compare those numbers that are found below the "Dual Core" text.
domidumont has quit [Ping timeout: 246 seconds]
bonbons has quit [Quit: Leaving]
paulk-collins has quit [Quit: Quitte]
<Skylark>
thank you - let me grab the batch (i don't have pix though)
<Skylark>
the numbers under the 'DUAL CORE T' are E2126BA 62S3 or DC053BA 6CG3
<sillysnowflake>
another person with the banana pi~
<simosx>
Skylark, as far as I understand, fabless semiconductor companies get many tens of thousands of chips per order from the foundry. They may have updated the printing or chose a different foundry.
<RSpliet>
I reckon switching foundries is more invasive than it seems... probably a more minor update
<Skylark>
simosx: that sounds like a good explanation actually
<Skylark>
also the writing is below the DUAL-CORE T as opposed to the sides
<Skylark>
no known software difference as far as you guys know, right?
Lucious has joined #linux-sunxi
<Lucious>
Hi @ all
<Lucious>
is there a german group of people for the a13
<simosx>
Skylark, there is A10 and there is A10S, and they do have differences. I do not think there are differences there because that 'T' is along the marketing text.
<Lucious>
i need help with a INet86v-rev02 tablet board
<simosx>
Lucious, you are looking for german-speaking community?
<Lucious>
yes simosx
<Lucious>
i have tried to build u-boot for that and a kernel.
<simosx>
Lucious, I don't think there is a separate channel or forum for that. If you can wait here, there might be someone that can speak german.
<Lucious>
no problem.
<Lucious>
i can understand english.
<simosx>
Lucious, ok. for U-boot, there are normally some details in timings/clocks that are specific to each device. I have the impression you can get those details from the Android image (when running in Android and have root access).
<Lucious>
yes there is root acess
<simosx>
there is no automatic tool yet to grab that info, so you need to do it manually.
<Lucious>
i unpaked the original image with awimage.
<Lucious>
and have the script.bin and script1.bin
<Lucious>
then i tooked these and bin2fex these 2 files
<Lucious>
so that i have the fex. for this board.
<Lucious>
is this the right way?
<sillysnowflake>
fyi, i speak german, but i'm new to all this embedded things.
<Lucious>
should i take pictures and make a new device page on the wiki?
<sillysnowflake>
i want to have gentoo on my banana pi. i know there is a sunxi fork of the 3.4 linux kernel. do i _have_ to use that fork otherwise it isn't gonna work, or can i use the standard kernel?
<Skylark>
simosx: is there a way to boot uboot from NAND without boot0/boot1/axf
dev1990 has quit [Quit: Konversation terminated!]
konradoo77 has quit [Ping timeout: 246 seconds]
<simosx>
Skylark, why would you need that?
cubeast has quit [Quit: Leaving]
<simosx>
sillysnowflake, many thinks in the kernel are upstream. if audio/video/etc are not a requirement, you can use that kernel.
<Skylark>
simosx: oh, i have a completely new board, and I'm trying to understand with the cleanest way of booting it... the egon code is "pretty"
<Skylark>
*what is not with
jinzo has quit [Quit: Leaving]
<simosx>
Skylark, not sure about that. You may want to ask at the list (Google Group).
<sillysnowflake>
simosx, i can switch kernels later on without problem / have multiple installed at once and select them from the boot loader, right?
<Skylark>
i've been reading the Wiki, which is extremely helpful, but i'm not 100% sure how well it reflects current project status
<jero>
hmm, my A10 based system is randomly freezing, I was able to trigger a sysrq crash trace, it points at the uart code, but i'm not sure if that's only because i'm on the sysrq console.