<wpwrak>
Oksana: thanks ! you are officially amazing :)
<DocScrutinizer05>
yeah, thanks a lot Oksana!
<Oksana>
:-) I am greedy like that. I doubt that there is anybody else thinking about attaching one Neo900-gadget-mouse to another Neo900-small-computer :-) | Off-topic: And software here is thinking, and giving incorrect answer, and not explaining why...
nox- has quit [Quit: Leaving]
<DocScrutinizer05>
Oksana: at your end of the ball it must be just in time to have a coffee, or depending on your personal preferences, afternoon tea
<Oksana>
Yes, going for lunch soon.
<DocScrutinizer05>
ooh, that early still?
<Oksana>
((preferences)) Milk.
<Oksana>
((early)) More like I do not see the time.
<DocScrutinizer05>
:nod:
<Oksana>
It should have been an hour ago :-)
<DocScrutinizer05>
time for myself to try and have a nap
<Oksana>
Good night
wicket64 has quit [Quit: WeeChat 1.0.1]
paulk-collins has joined #neo900
freemangordon_ has joined #neo900
freemangordon has quit [Ping timeout: 245 seconds]
freemangordon has joined #neo900
paulk-collins has quit [Quit: Quitte]
sparetire_ has quit [Quit: sparetire_]
jonwil has quit [Remote host closed the connection]
Kabouik has quit [Ping timeout: 255 seconds]
Kabouik has joined #neo900
ecloud_wfh is now known as ecloud
Pali has joined #neo900
sixwheeledbeast has quit [Quit: Leaving.]
sixwheeledbeast1 has joined #neo900
freemangordon has quit [Ping timeout: 245 seconds]
modem has quit [Read error: Connection reset by peer]
modem has joined #neo900
<FIQ>
interesting
<FIQ>
there was some accident on TI's end before regarding powerVR documentation, but iirc that wasn't the source...
<FIQ>
(basically they accidentally made a NDA doc about it public)
sixwheeledbeast1 is now known as sixwheeledbeast
che1 has quit [Ping timeout: 272 seconds]
jonwil has joined #neo900
<Pali>
freemangordon: I CCed LKML mail days ago about problems with green camera on n900... do you remember something about that problem?
<freemangordon_>
Pali: yes, I saw the mail, I just can't organize (and find some spare time) to answer :)
che1 has joined #neo900
<freemangordon_>
Pali: will try to reply soon. though sakari should be pretty much aware of the problem (no IRQ generated iirc)
<Pali>
ok
<DocScrutinizer05>
moin folks!
<freemangordon_>
hi doc
freemangordon_ has quit [Quit: Leaving.]
* DocScrutinizer05
puts on hat with horns and the long tail, then drafts a statement like "we will probably be able to source 50 pieces 1GB RAM + 512MB NAND. We will raffle those among all donations of 15 EUR coming in until next Sunday. The winners can have their chips mounted on Neo900 instead of the 512MB RAM for no extra charge"
<DocScrutinizer05>
;-)
<DocScrutinizer05>
I got no hat with horns, so never mind the above :-D
<jonwil>
Looks like that pvr stuff is for omap4 not omap3 (what we have)
<DocScrutinizer05>
the diffs are not THAT huge
<DocScrutinizer05>
you probably can downsize easily when you actually got complete sources and even some specs of PVR3
<DocScrutinizer05>
unless of course when they implemented highly sophisticated stuff in driver for 3 which got done in hw of PVR4
<DocScrutinizer05>
if it's only for more polygons/renderers/whatever, I guess it's a matter of changing a few build constants
<jonwil>
Knowing imgtech though they will probably sue anyone doing anything remotely connected to PowerVR if they think they can either convince a judge that the people being sued used the leaked source or if they think they can convince the people being sued that its better to settle than to go to court
<DocScrutinizer05>
I wouldn't worry too much about that
<jonwil>
why do you say that?
<DocScrutinizer05>
anonymous publish of RE'ed source is always a possibility
<jonwil>
imgtech are very protective of their IP
<DocScrutinizer05>
then re-import those REed sources and imgtech can't sue you for that, aiui
<DocScrutinizer05>
but that's only my 2 uneducated Euro-cent
<DocScrutinizer05>
as long as we didn't know the sources we used are definitely rogue/tainted, we may use them to base own work on them, unless a licence they are explicitly under forbids
<bencoh>
what was that about the 50 1GB chips ?
<bencoh>
just an evil-troll-statement ?
<DocScrutinizer05>
when 1337hax0r publishes some "look what I coded this weekend, seems it might work for some OMAP3 chips gfx. Feel free to use in whatever way you like", we are not doing anything illegal when we base own work on that
<bencoh>
or semi-serious ? :)
<DocScrutinizer05>
bencoh: semi
<DocScrutinizer05>
just I wouldn't dare to do a thing like that, it produces lots of bad fellings among users
<bencoh>
yeah
<bencoh>
how much would it cost to buy 300 of those ?
<bencoh>
(I mean, now)
<DocScrutinizer05>
it's however a fact that we possibly could source 50some chips, and don't have a source for more yet
<bencoh>
oh, okay
<DocScrutinizer05>
and it's also a fact that we are very short on money
<DocScrutinizer05>
so the idea is somehwat tempting, from my POV
<DocScrutinizer05>
and doing a first come first served doesn't sound fair either
<bencoh>
yeah
<DocScrutinizer05>
funny sidenote: the 1GB are prolly cheaper than the 512MB
<bencoh>
meh
<bencoh>
and what if we find a source for a few hundreds (if that's possible) ?
<bencoh>
+could
<DocScrutinizer05>
then I would need money for a few hundred, which is definitely not in our funds kitty right now
<bencoh>
(I think it would be fine then to ask people for donations to secure it, one per donation - basically buy it)
<bencoh>
everyone would have its own, no first-come-first-served issue
<DocScrutinizer05>
yes, that's what I'm slowly pushing to start during last few days anyway
<DocScrutinizer05>
see microblog and wikiwide's last post on main tmo thread
<DocScrutinizer05>
a "purchase your RAM" campaign
<bencoh>
yeah
<DocScrutinizer05>
or even "purchase your Neo900 risk parts now"
<jonwil>
well I dont have the time OR the kernel/X/GPU/etc knowledge to do anything really with that pvr stuff
<jonwil>
besides, if I start on something new, it would mean stopping work on pulseaudio
<DocScrutinizer05>
I'm pretty sure *somebody* will, sooner or later
<jonwil>
which freemangordon wouldn't be happy :P
<bencoh>
not that sure
<DocScrutinizer05>
there's quite a number of OMAP3 PVRs out there which lack FOSS drivers
<bencoh>
yeah but I think the gnu-related project wont want to touch it
<DocScrutinizer05>
that's why I said "somebody"
<DocScrutinizer05>
once there's enough info available, preferrably even in form of sourcecode, sooner or later somebody will download that info and start to learn how to use it
modem has quit [Ping timeout: 240 seconds]
<DocScrutinizer05>
there's a freakin lot of FOSS kernel drivers for chips that have NDA-only datasheets
<DocScrutinizer05>
the PVR situation is only marginally different
<jonwil>
I am on powervr-devel mailing list for that project and people are talking about it
<jonwil>
possibly about using the code and pdfs to produce some clean-ish-room docs that could be used as a base for drivers
<DocScrutinizer05>
that's exactly the way
<DocScrutinizer05>
once the knowledge / info is out in the wild, it *will* get used
<DocScrutinizer05>
unless the topic is boring, dead, rotten and lost in oblivion
<DocScrutinizer05>
which PVR isn't
<jonwil>
that said, I bet the guys who do custom Android ROMs for whatever hardware these drivers apply to will have no qualms about using thre drivers as-is given how the Android community likes to just rip everything off and not care
<jonwil>
not that there are many Android handsets out there with omap4 chips in em
<DocScrutinizer05>
guess why ;-)
* DocScrutinizer05
points at former rants about non-public SiErr in OMAP4 IRQ/WAKE handling
<DocScrutinizer05>
it's quite funny if you build a periperal (let's say, a modem) and hotline forwards prio1 tickets from 5 customers like RIL, LTE aso with about same wording "your peripheral stops working randomly, doesn't talk to our OMAP4 anymore" and your lab can track down the error to OMAP4 losing the IRQ your peroipheral sent when there been new data to send from periperal to OMAP
<DocScrutinizer05>
I mean, your peripheral is NOT supposed to send a second IRQ
<DocScrutinizer05>
actually it hardly could, when the IRQ is level triggered
<MonkeyofDoom>
heh
<DocScrutinizer05>
and of course your peripheral's interface FIFOs and buffers etc are of finite size and eventually *will* overflow
<DocScrutinizer05>
in which case what you gonna do, since even kernel OOPS would have to get printed to $somewhere and that "somewhere again is on APE aka OMAP4
<DocScrutinizer05>
s/"/$/
<DocScrutinizer05>
one of the reasons why I strictly refused to base Neo900 on OMAP4
<DocScrutinizer05>
since unlike in above told case, Neo900 UG cannot ask modem manufacturer to add a botch hotfix to assert IRQ again every 100ms
<DocScrutinizer05>
and I don't like to have to deal with tickets like "when Neo900 Application CPU goes to zero clock while DSP renders some video, the modem stalls randomly and cannot get reanimated by any other means than power cycle (of modem)"
modem has joined #neo900
jonwil has quit [Quit: ChatZilla 0.9.91 [SeaMonkey 2.30/20141013232806]]
<kerio>
DocScrutinizer05: won't triggering IRQs drain the battery
<kerio>
?
che1 has quit [Ping timeout: 240 seconds]
louisdk has joined #neo900
Openbot has joined #neo900
Kabouik has quit [Ping timeout: 255 seconds]
Pali has quit [Ping timeout: 265 seconds]
<DocScrutinizer05>
kerio: yes, operating electronics usually drains the battery, when the electronics is mobile and battery powered
<DocScrutinizer05>
sorry, didn't want to sound sarcastic. But triggering is what IRQs are meant to do. And to transfer data you must trigger an IRQ to wake up the far end so it starts receiving
<DocScrutinizer05>
so it boils down to "yes, data transfer increases power consumption of the system"
<DocScrutinizer05>
it's like "yes, ringing the phone will cause the far end to be more busy, and probably you will also be more busy to initiate the call"
<DocScrutinizer05>
actually the phone analogy is pretty good for the SiErr thing. When the instructions for your braindead store personal is "when there's a package addressed to A, you pick up the phone receiver, press key "A" and wait-listen until far end tells you how to ship the package" and then far end A is missing the ringing phone on their end and your personal doesn't notice that the ringing stopped after 120s and the calling tone in receiver
<DocScrutinizer05>
turned into a busy tone, then your personal will be in a deadlock waiting for A to answer which will never happen since A missed the 120s of phone ringing
<Openbot>
Clear this -Neo900 = linux capable hardware and highly open and offers (a) security feature - that is modem sandboxing and monitering right ?
<DocScrutinizer05>
A is supposed to never miss a phone call. But when they do nevertheless, you _could_ instruct your store personal to stop phone call after 120s when no answer came in, wait 10 min, then redo
<DocScrutinizer05>
Openbot: I hardly can parse your question. But probably the answer is yes
<Openbot>
Docscrutinizer05 ^^ so its not a privacy phone like blaphone
<DocScrutinizer05>
Openbot: sorry, you completely lost me now
<Openbot>
People getting confused lol
<DocScrutinizer05>
blackphone is a software solution based on pretty poor and vulnerable hardware. Neo900 is a hardware which allows user to install whatever they can make fit on that platform, even blackphone suite
<Openbot>
Confused like its a secure privacy phone like blackphone na
<bencoh>
it is what you want it to be, you have full controle over the hardware, apart from an (isolated) modem
<bencoh>
the modem being nothing more than a gsm/gprs/umts/lte dongle
<bencoh>
+usb
<Openbot>
Bencoh i know
<DocScrutinizer05>
a combination of blackphone "OS" and Neo900 hardware would probably make for a really nice secure phone
<Openbot>
Just had a lil debate and found he thought it was something like blackphone
<Openbot>
Definately !
<DocScrutinizer05>
it luckily is way better than blackphone android standard crap hardware#
<DocScrutinizer05>
but it comes without any OS preinstalled, so it's up to user to turn it into andoid or into replicant or into blackphone or into whatever user wants it to be
<DocScrutinizer05>
you can even turn it into a maemo phone ;-)
<Openbot>
Na you didnt get it - people are thinking like its a something else i.e getting lost on the privacy feature and forgeting what actually it is
<DocScrutinizer05>
by installing the yet to get finalized FOSS fremantle
<DocScrutinizer05>
people are missing the fact that we don't ship a OS but simply offer complete documentation so whoever wants can adapt&install whatever she likes on the hardware
<Openbot>
I met a friend he said you know theres a cool new blackphone with bla bla i was like yeew
<Openbot>
Atleast he got to know that it dosent come with a os after i tell him lol
<DocScrutinizer05>
all those "color" phones, greenphone, blackphone, pinkphone... :-S our next few models will be UV-A, UV-B, X-RAY, GAMMA ;-P
<Openbot>
Maybe neo900.org is culprit and in need of lil makeover ...dos1
<bencoh>
DocScrutinizer05: NeoN UV sounds cool :p
<DocScrutinizer05>
yes, we need a major redesign/pimp-up of our web appearance
<bencoh>
(neon spectrum has UV lines)
<Openbot>
Yes it it required more now
<Openbot>
Catchy starting line "Thetrulyopen smartphone
<Openbot>
that cares about yourprivacy." so the visitor thinks and goes to blackphone dreamland instantly lol
<Openbot>
Some strong diffrentation is required while still hilighting the major features on front page
<Openbot>
Anyway night guys
<DocScrutinizer05>
I think we got as much differentiation as possible
<Openbot>
Na something is missing i don know what though
Openbot has quit [Quit: used jmIrc]
<DocScrutinizer05>
maybe sth like "we're not selling an intangible 'product' consisting of a hw-platform, a mostly closed OS to run on that platfowm, and some appshop to monetize the whole thing by selling fart apps to users. What we sell is a hw platform with freedom of choice to install whatever OS you like" ?
<kerio>
DocScrutinizer05: and then add a "*coughcoughjollacoughcough*"
<DocScrutinizer05>
"and _only_ that hw platform. We're supporting the genuine spirit of linux Free and Open Source Software (FOSS), that brought stuff like linux servers (80% of servers worldwide use linux) and linux PCs with KDE or Gnome (or any of the other 2 dozen) desktop to the world. We hope for same thing happening to phones. You'll have the coice of replicant, FOSS fremantle (maemo), SHR (debian based), QtMoko (debian based), probably sailfish, and
<DocScrutinizer05>
you're theoretically even free to install your own licensed version of winCE"
<DocScrutinizer05>
well, I had to drill out two screws some jerk fixed with locktite, and of course this didn't pan out as flawless as I hoped, damaged N9 polycarbonate case a bit
<DocScrutinizer05>
[2014-11-19 Wed 18:53:36] <DocScrutinizer05> ok, I would almost like to say "I knew it". Trying to disassemble my N9, it turned out that the internal TX-4 screws are so massively tightened that a alost 10EUR worth TX-4 [[edited]] high quality screwdriver was only able to damage 2 of them so now they sit in there and I have to come up with some way to get them out
<DocScrutinizer05>
when only one copy been started, it been ~12MB/s
<nox->
ah
nicksydney has joined #neo900
<DocScrutinizer05>
which kinda keeps me wonder, since one is supposed to live on uSD and the other on eMMC
<DocScrutinizer05>
(yet to get confirmed though)
<kerio>
the n900 has negative ram dude
<kerio>
what do you expect
<nox->
probably limitation of the usb not the fs?
<nox->
s/fs/"disks"/
<kerio>
usb can do much more than 12MB/s
NIN101 has quit []
<nox->
on n900?
kolp has quit [Read error: Connection reset by peer]
paulk-collins has quit [Quit: Quitte]
<DocScrutinizer05>
ok, my bad, I copied 2 volumes of my uSD
<DocScrutinizer05>
err nope
<DocScrutinizer05>
but actually the two volumes both were fat32, MyDocs and first volume of uSD
<bencoh>
writing at 15mo/s here (usb)
<bencoh>
(to emmc)
<bencoh>
it's actually way faster than I thought
<DocScrutinizer05>
now that eMMC got finished it seems, I started copying vol2 of uSD in parallel to vol1 and that's an ext3 vol and stalls a lot in "copying..." requester, while the files at top of the requester still "count up" and so does the "X of Y MB copied" counter... While the vol1 copy not maxes out at ~10MB/s at times
<Oksana>
Theoretically, since Gemalto provides firmware updates for modem, it should be able to work around the problem. Either use 64-bit time_t, or something
<DocScrutinizer05>
I don't even think modem is too intersted in date at all
<DocScrutinizer05>
except for "CMOS" clock I don't know of any AT commands that have date in them
<DocScrutinizer05>
all GSM and newer RAT is probably only using relative timers, never any absolute date
<nox->
i c
<DocScrutinizer05>
compare timestamps in dmesg which often are in µs since boot of kernel
<DocScrutinizer05>
which creates another bug known as uptime-wraparound-bug
<DocScrutinizer05>
or similar naming
<nox->
reminds me of the classic win98(?) uptime bug
<nox->
(crashed after a few days bc timer wrapped)
<DocScrutinizer05>
in linux it is (or was) after some hundred days iirc