infobot has quit [Read error: Connection reset by peer]
infobot has joined #neo900
lobito has quit [Quit: Leaving.]
norly has quit [Quit: Leaving.]
Humpelstilzchen has quit [Ping timeout: 276 seconds]
Humpelstilzchen has joined #neo900
modem has quit [Ping timeout: 265 seconds]
<JamesJRH>
R0b0t1: Haskell ftw.!! :-)
<JamesJRH>
And pure-functional programming in general.
rjeffries has quit [Ping timeout: 276 seconds]
fling has joined #neo900
rjeffries has joined #neo900
vakkov has joined #neo900
Pali has joined #neo900
Pali has quit [Remote host closed the connection]
arossdotme has quit [Quit: Ex-Chat]
rjeffries has quit [Ping timeout: 256 seconds]
sparetire_ has quit [Quit: sparetire_]
rjeffries has joined #neo900
rjeffries has quit [Ping timeout: 244 seconds]
lexik has quit [Ping timeout: 272 seconds]
illwieckz has quit [Ping timeout: 276 seconds]
dal has joined #neo900
dal is now known as dal2
dal2 has quit [Client Quit]
norly has joined #neo900
dal has joined #neo900
jonsger has joined #neo900
louisdk has joined #neo900
Pali has joined #neo900
louisdk has quit [Ping timeout: 265 seconds]
dal2 has quit [Ping timeout: 244 seconds]
dal2 has joined #neo900
rjeffries has joined #neo900
jonsger has quit [Ping timeout: 252 seconds]
Guest36804 is now known as FIQ
FIQ has quit [Changing host]
FIQ has joined #neo900
<R0b0t1>
JamesJRH: indeed
<JamesJRH>
R0b0t1: I want to write smartphone applications in Haskell. Do you have any idea how to do this? If I had an example application that demonstrated a few key APIs then I'd be away.
<JamesJRH>
R0b0t1: I'd like to target Dalvik (Android, Replicant, CM, etc.) and a more GNU+Linux-style mobile OS such as Maemo/Hildon.
<R0b0t1>
I don't know how well GHC supports cross compiling, but it does have an LLVM backend and should support and architecture LLVM does.
<R0b0t1>
As to whether it can link and stuff... that's another matter entirely. I've been meaning to try it, but some reading and QA I did with some gentoo people suggests GHC is a pain to set up or impossible to set up in that fashion.
<R0b0t1>
So, there's compiling natively, which takes forever.
<R0b0t1>
You can LLVM -> bytecode through a few methods, so you could target dalvik.
<Wizzup>
Hardly forever
<R0b0t1>
Ok half of forever
<Wizzup>
and android can run native applications
<Wizzup>
no need to target dalvik
<R0b0t1>
"can"
<Wizzup>
you just need some glue
<R0b0t1>
"run"
<R0b0t1>
"native applications"
<Wizzup>
Why are you putting what I am writing in quotes?
<R0b0t1>
To suggest that's a bit of a simplification. Android is a piece of shit.
<R0b0t1>
And sure, compiling natively is easier for a lot of people, but it's still the less convenient method in a lot of cases.
<R0b0t1>
Maybe not with GHC but that's GHC's fault.
vizzy has joined #neo900
<vizzy>
ha!
<JamesJRH>
15:34:49 < R0b0t1> Ok half of forever ← Lol!
<JamesJRH>
R0b0t1: I'm happy, in theory, with the NDK route if that works, though I've never even touched the NDK.
<DocScrutinizer05>
vizzy: ha too!
<DocScrutinizer05>
:-)
<JamesJRH>
15:35:51 < R0b0t1> To suggest that's a bit of a simplification. Android is a piece of shit. ← You can say that again.
<DocScrutinizer05>
android are two pieces of shit. You can't cram that much shit in one piece
<vizzy>
DocScrutinizer05, will the cpu be upgradeadable in some way? for me as a noob, i just look at the ghz, is 1ghz enough for the next years? overclocking?
<Wizzup>
1gz is likely enough, depending on what you want to do.
<DocScrutinizer05>
we think 1GHz is just fast enough for everything you reasonably want to do with this device
<DocScrutinizer05>
OC is deprecated but possibe. We don't have number nor would we publish them if we had
<vizzy>
oh, nothing special. maybe faster browsing, somewhat fater behaviour, e.g. strting up the camera or a recorder
<DocScrutinizer05>
browsing speed is limited by RAM
<Wizzup>
That is mostly IO bound
<DocScrutinizer05>
and network BW
<DocScrutinizer05>
to speed up starting up of apps you also need fast storage more than a quad core
<vizzy>
i almost do mobile irc, jabber, ... gps and wifi.
<JamesJRH>
I hate the way that Google do things so crappily, despite how much money they have. They just throw money at things. Then again, so does Microsoft. Apple seem to do things to a high quality, then hinder it beyond it being anything near good anymore, but still market it as a designer brand.
<JamesJRH>
The OS, that is; Apple's hardware is generally underspecified for the price (or overpriced for the specification).
<vizzy>
so, instead of maemo, it will run a real debian? that would be awesome. with real x erver and all that stuff i know from my laptop?
<DocScrutinizer05>
vizzy: and no, alas we can't upgrade the CPU
<vizzy>
ok, but 1ghz is still 2 times the old n900 :D
<vizzy>
and regarding the memory.....
<DocScrutinizer05>
well, we sort of could, since we have two boards (the notorious sandwich) of which one is CPU, PMU and kbd etc, and the other has modem and audio etc
<DocScrutinizer05>
memory is 4 times N900
<vizzy>
:)
<DocScrutinizer05>
and at the max which is possible
<vizzy>
also ich will das ding auf jedenfall!
<DocScrutinizer05>
and this will speed up the whole device *a lot*
<JamesJRH>
DocScrutinizer05: Heh, “because people worship what they don’t understand.”
<vizzy>
hey, im oldschoool, i know ram i more worth than cpu ;)
<DocScrutinizer05>
yep, never been more true than for this device
<DocScrutinizer05>
so we invested quite some effort to get the 1GB RAM
<vizzy>
at first i found it totally fantastic that there will be a group of people redsigning one of the coolest phones ever. i found my n900 by accident and didnt know what a wonderful device it is before it had one.
<pigeons>
i've got 2 n900's but do you think someone who can barely even do a horrible soldering job is ok buying just the board, or should i go for the whole neo900
<vizzy>
i had the n82 before, because of the xeon flash and great camera
<vizzy>
the price differs not much imho, its worth selling or giving xour n900 to neo and get a whole neo900 then ;)
<vizzy>
but having one spare n900 for playing around is geart
<vizzy>
great
<vizzy>
i need to get another one too, replace the solar powered very unstable laptop on my garden house (weatherstation, webcams etc)
<vizzy>
or maybe a neo, i need a new flightcompuer !!!
<DocScrutinizer05>
pigeons: the swap is very "simple", no soldering needed
<pigeons>
ok sounds good
<DocScrutinizer05>
the most complex part is plugging in the display B2B connector
<DocScrutinizer05>
while taking care not to tear off or break the flex cable
<DocScrutinizer05>
when you even can do any level of soldering (electronics) I bet you got all the needed skills to do the swap
<DocScrutinizer05>
it's literally just 6 screws and a plug
<DocScrutinizer05>
*maybe* a tiny bit of cutting/filing (nothing critical). In which case we deliver the needed tool (file)
<pigeons>
i'll give it a shot
<DocScrutinizer05>
when you can trim your fingernails, you can do any such filing as well
<DocScrutinizer05>
but probably it's not even needed
<DocScrutinizer05>
unless you want to use hackerbus
Kabouik_ has quit [Read error: Connection reset by peer]
<JamesJRH>
Yeah, it seems to be more common these days.
<JamesJRH>
I think it must be quite old though because it talks of Jobs like he's still alive.
Kabouik has joined #neo900
<DocScrutinizer05>
>>Without question, nearly everyone believes the equation Apple = Innovation is a fundamental truth<< well, not here. We had quite a laugh at Openmoko about the 'cool new stuff' Apple came up with on their iPhone, while we did same or better since "ages". But when apple does it, it's suddenly new and hip
<JamesJRH>
Okay, well the top comments say “7 years ago”, so must be circa 2008.
<DocScrutinizer05>
:nod:
<DocScrutinizer05>
it still is true about how Apple develops stuff
<JamesJRH>
Indeed.
<R0b0t1>
DocScrutinizer05: Do you know if the processor used has accelerated crypto
<JamesJRH>
I haven't read it all. I know that Apple do things to a high quality, even if it isn't cutting-edge, but does the article say anything about how much their products are hindered? How much good functionality is locked away, but still paid for?
<norly>
DocScrutinizer05: Do you have recent numbers on how many devices have been ordered?
<JamesJRH>
Anyhow, the long and short of it is that I don't like Apple, or Microsoft, or Google. I have, however, considered Google as the lesser of the evils for some years, but now I want my next phone to be better than that.
<JamesJRH>
I'm probably going to use Maemo or similar, or at the very least, a phone preïnstalled with Replicant.
<JamesJRH>
I've had enough of mainstream phones.
<DocScrutinizer05>
R0b0t1: AES is unclear. No other accel
<DocScrutinizer05>
ask Pali or freemangordon
<Pali>
which processor?
<DocScrutinizer05>
DM3730
<DocScrutinizer05>
GP
<Pali>
no idea
<JamesJRH>
Mainstream phone designers, manufacturers, and retailers all have their own agendae which is not even close to mine.
<Pali>
we have omap3 drivers for aes, sha, md5... but they will work only if processor has needed crypto blocks
<DocScrutinizer05>
R0b0t1: please use the fine links to datasheets in our very fine Block Diagram ;-)
<pigeons>
so if mostly in usa lte usa model? anyone know about canada?
<DocScrutinizer05>
canada recently opted for EU-LTE(!) iirc
<DocScrutinizer05>
or was it Australia?
<DocScrutinizer05>
generally USA is a mess. I for one would probbaly opt for the PXS8 when I lived in USA
<R0b0t1>
Well, assuming the omap34xx documentation applies to the omap37xx parts (and that they didn't remove features...) cryptographic acceleration can be achieved via SIMD instructions or clever shaders(?). No application specific instructions, though.
<R0b0t1>
Something far more entertaining I found is that apparently ARM TrustZone is supported, meaning the device has the capability to use Xen or KVM.
<R0b0t1>
yey
<pigeons>
is the PXS8 the
<pigeons>
UMTS "worldwide" choce on the store?
<R0b0t1>
CDMA one
<R0b0t1>
don't think so
<pigeons>
ok "UMTS plus CDMA2000"
<pigeons>
guess i need to collect 'em all
<R0b0t1>
yeah. fwiw that's the one I'll probably get, but CDMA network quality varies quite heavily by region :s
<DocScrutinizer05>
R0b0t1: trustzone only in HS devices afaik
<R0b0t1>
aw
<DocScrutinizer05>
so in 3430/3630, but NOT in 3530/3730
<R0b0t1>
I suspected it was too good to be true
<DocScrutinizer05>
I don't know how much of TZ infra is actually disabled in GP devices, and how much simply not supported by the cert infra
<JamesJRH>
DocScrutinizer05: Wow, I still have a couple of screens to go. I'm finding this article to be falsely painting Apple in an very good light.
<R0b0t1>
okay weird
<R0b0t1>
DocScrutinizer05: I think it's enabled for the whole family, I can't say anything that says otherwise
<R0b0t1>
I can't find anything that even mentions the capabilities
<DocScrutinizer05>
it's quite a few years since I read it. What I recall is Apple doing several concurrent implementations to check usability etc, then even starting all new from scratch again when needed
<JamesJRH>
“They are willing to spend the money to make sure everything is perfect, because that is their mission.” ← DRM is not a member of perfect concepts; an ideal world would not contain it. An ideal device would not contain it.
<R0b0t1>
besides device desc
<JamesJRH>
What I find strange about this article, is that many of the things describe me – I myself am striving for perfection – but I don't consider Apple's products anywhere near ideall.
<JamesJRH>
ideal*
<DocScrutinizer05>
(modem choice) don't worry, what we do right now is basically just a poll, it's not binding. We will allow opting for a different modem variant in final order
<R0b0t1>
ah good
<DocScrutinizer05>
R0b0t1: (TZ) the whole thing is not really available in public docs afaik. Only the GP devices have public docs
<DocScrutinizer05>
IOW we were not able to find a TRM for the N900 3430
<DocScrutinizer05>
neither for 3630. But then we couldn't source DM3630 either, prolly DM* doesn't even exist
<DocScrutinizer05>
HS chips and support and docs only under NDA for huge companies
<DocScrutinizer05>
anyway the funny thing I seem to remember is: AES IP was available only on GP devices, since otherwise TZ would occupy AES and not provide it to "DOMU"
modem has joined #neo900
nox- has joined #neo900
<R0b0t1>
DocScrutinizer05: Another part, Exynos5410/5420, had someone figure out the TZ functionality by reversing a signed bootloader. Then gave it back to the company to re-sign.
<R0b0t1>
It's a mode written to a register somewhere.
<R0b0t1>
As for the cert stuff, I'm not sure it's necessary to have kvm work
<DocScrutinizer05>
merlin1991: welcome! :-)
<R0b0t1>
but not entirely sure how the parts differ,so
<R0b0t1>
oh
<R0b0t1>
@ last bit
<R0b0t1>
DocScrutinizer05: What do you mean TZ would occupy AES? AES registers were TZ registers?
<merlin1991>
dang I feel like someone is looking me over my sholder placing an order xD
<DocScrutinizer05>
yes, TZ is basically just a set of registers that configure which addr range of executable code may access which registers and which RAM ranges
<DocScrutinizer05>
merlin1991: hehehe. Just found a info mail from our shop
<DocScrutinizer05>
the welcome mail is BCCed to me
<merlin1991>
but the caps in the mail is err
<merlin1991>
THIS LOOKS LIKE SPAM TO ME!!!!
<merlin1991>
:D
<DocScrutinizer05>
yeah, all the mails are terribly formatted
<DocScrutinizer05>
alas it's a major PITA to edit them
<DocScrutinizer05>
in prestashop
<DocScrutinizer05>
and actually what I wrote above is incorrect, I get a dedicated info mail: >>Your Online store Neo900 has a new activated customer, Christian Ratzenhofer.<<
<pigeons>
yeah google/gmail decided my activation link message was spam
<R0b0t1>
DocScrutinizer05: I know how it works, just confused as to how they'd gimp it in a "exchanging features" way.
<R0b0t1>
If you remember or already have any more info about TZ on that part, please send it to me. I might call TI later because it's not documented in the slightest, from what I can tell
<DocScrutinizer05>
I don't have any good info, actually zilch info. Ask freemangordon, he's basically the expert for TZ
<dal2>
Is the Neo900 as vulnerable to a gsm knockback attack as any other phone?
<R0b0t1>
if it exploits the phone via the radio, it is not vulnerable. if it exploits the protocol, it is vulnerable.
<DocScrutinizer05>
what eactly is a "gsm knockback attack"?
<R0b0t1>
^
<DocScrutinizer05>
dal2: our GSM modem has zilch control over the application CPU
<R0b0t1>
I think it's the one where you exploit a basestation and then mess with the phones connected to that basestation. If that's the case, you could lose service or the radio could be hacked. You will also be identified by the parties who control the basestation.
<R0b0t1>
aaand that's it
<DocScrutinizer05>
well, we can't change the way the GSM OTA protocol works, since after all it stays to be a phone
<DocScrutinizer05>
but unlike other contemporary phones we can *guarantee* our device can't get hacked from GSM network (or via any other RF technology like WiFi/BT)
<DocScrutinizer05>
worst case they can mess with the modem, and the modem is considered rogue anyway, and sandboxed
<DocScrutinizer05>
so a potential attacker could gain nothing by exploiting the modem
<DocScrutinizer05>
any possible exploit to the modem is not worth it since it could get implemented easier by simply snooping or controlling the BTS<->modem data traffic
<DocScrutinizer05>
which is sth we can't stop anyway, since we have no control over the BTS
<DocScrutinizer05>
compare to PSTN: our telephone is safe from eavesdropping in device, but we can't control when somebody is wiretapping your line, or even using the surveillance functions of the telephone company's exchange infra
<DocScrutinizer05>
dal2: ^^^
<DocScrutinizer05>
btw in that regard our device is way better even than dedicated security phones like blackphone, which has nice software to do encrypted phonecalls etc (our can do as well, when you install the software you like to use for this) but can NOT guarantee that an attacker exploiting the modem OTA has no access to the system and thus would take over the complete device and thus render all the nice security sw moot
<dal2>
back, reading
<dal2>
apologies for the delay in my response
cybiko123 has joined #neo900
cybiko123 has quit [Changing host]
cybiko123 has joined #neo900
<dal2>
a "gsm knockback attack" is what happened to my phone at HOPE X
<dal2>
I wrote an article about the incident. Ripped the battery from my phone mid-attack
<dal2>
the attacker was taking screenshots of my phone
<dal2>
I got an nandroid image of the phone, sent it to a security researcher
<DocScrutinizer05>
that's *absolutely* impossible with Neo900
<dal2>
the project's been on our backburners forever, but
<dal2>
I made a video of the imaging, and he's doing multimedia on his end as well
<dal2>
if you like DocScrutinizer05, whenever the hacker and I are able to get the follow-up article done,
<dal2>
you could take a look at it and see if you want to give a quotation about how it couldn't happen on Neo900
<DocScrutinizer05>
sure thing
<dal2>
cool. What I really wish I could do with the hacked device is a gsm wifi module ram dump
<dal2>
I took the battery out mid-hack, put it back in to do the nandroid imaging, then took it back out again
<dal2>
it hasn't been in since
<dal2>
didn't do anything with it in other than the nandroid imaging
<dal2>
I don't know anyone who can do a gsm wifi module ram dump? Maybe hackerspaces in Austin I could drive to?
<dal2>
the guy who is doing the work on the nandroid image presented at Haxpo in Amsterdam
<DocScrutinizer05>
dal2: that exploit to your phone probably been enabled by security vulnerabilities in the OS (Android/cyanogen).
<dal2>
yeah, it was a rooted and out of date CyanogenMod
<DocScrutinizer05>
dal2: please note that Neo900 not even provides any default OS, the company can't engage into sw development. We provide the hw platform for community (or whoever might care) to implement a decent vuln-free OS on it. So far Maemo is considered relatively safe, and we support fremantle porting task force and the device is dedicated to be N900/maemo compatible hw
<DocScrutinizer05>
on some android phones the replicant guys (iirc) demonstrated a backdoor/vulnerability that allows rooting the device OTA, and there are more scenarios where you can't do anything against such OTA exploit when it's starting on modem firmware level and spreading from there into the OS. Not though on Neo900
<dal2>
right, I understand about no default OS
<DocScrutinizer05>
our modem is completely isolated from the OS so any exploit of modem could not reach out into the OS, no matter what
che1 has quit [Ping timeout: 250 seconds]
lexik has joined #neo900
<dal2>
:)
<dal2>
DocScrutinizer05: Do you know where I might find someone who could do a ram dump of the gsm wifi module?
<DocScrutinizer05>
sorry, no contacts for that
<dal2>
okay thanks
<DocScrutinizer05>
maybe contact the replicant guys, they should know
<DocScrutinizer05>
probably when you enter that ADP(?) mode you could read out the image from the phone
<DocScrutinizer05>
but I'm not savvy with android phones
<dal2>
yeah we got the nandroid image
<dal2>
this is taking the gsm module hardware and somehow ram-dumping it