<DocScrutinizer05>
well, >>On Copyleft Hardware<< sounds a tad confused. Like the author fell into the same trap of mixing sw concepts with the hw world, despite they claim they don't (>>I get it. Hardware is not free or libre<<)
qwazix has joined #neo900
qwazix has joined #neo900
qwazix has quit [Changing host]
<DocScrutinizer05>
what got the statement >>We should continue the efforts of reverse engineering and of pushing hardware vendors to not only ship free software for their hardware, but also release their design under free licenses<< to do with copyleft hw ?
<jonwil>
I got a response from the librem5 people. They confirmed that librem5 will have no blobs for the cellular radio on the main CPU at all (they couldn't tell me how well documented that code would be though as that's not something they know at this point)
<DocScrutinizer05>
sure, that's about as good as 30% of arbitrary Android phones then
<DocScrutinizer05>
and sub-par to maemo back when the ISI specs were "public"
<jonwil>
They said they aim to to have a completly libre OS with NO blobs at all on the main CPU for anything.
<jonwil>
which is better than any Android phone to date.
<Wizzup>
Not that they will manage of course.
<Wizzup>
they seem to be as clueless as they were about the IME/AMT
<Joerg-Neo900>
yeah, sounds like fuzzy buzzword marketing blabla, not backed by much competence. As I already stated a few days ago
<knttl>
what do they say about the memory isolation between baseband and main cpu?
<Joerg-Neo900>
there's a nice anecdote from OpenMoko, where I been Electronics Engineer: one day management came to R&D and told us "Device needs OLEDS!" - we had nfc where and for what, particularly since the display been chosen and fixed to be LCD since long
<jonwil>
The GPU in the SoC they are using is a Vivante which has FOSS drivers (unlike PowerVR or others)
<jonwil>
So I think getting full GPU support without any blobs is definitely possible (unlike Neo900)
<Wizzup>
not that you really need the gpu :)
<Joerg-Neo900>
tbh I don't care too much about evaluating the Lebrem5 in detail here, particularly given the fact they obviously do _NOT_ follow an open hw and open r&d approach
<Wizzup>
maybe you can use it to plug neo900 a bit more somehow
<jonwil>
Anyhow, I will stop talking about librem5 and go do some N900 work (more specifically I am going to update the root CA certificates in maemo-security-certman to the latest ones from Mozilla :)
<Wizzup>
sweet, then it will not say the cert is not trusted, but instead say there's no common cipher :P
<jonwil>
The root CA certificates in Git are already up-to-date with the Mozilla CA certs as of the 4th of February and I have been running that set of root certificates on my own N900 since then without any issues.
<jonwil>
So unless something major has changed in the Mozilla tree since then I see no reason why my update should cause any issues.
<Wizzup>
did you connect to wikipedia?
<jonwil>
Yes I connected to all sorts of https sites
<jonwil>
including wikipedia
<jonwil>
wikipedia worked just now when I tested it
* Wizzup
tests
<jonwil>
what version of maemosec-certman-common-ca are you running?
<jonwil>
That will affect the outcome of any test you run
<Wizzup>
seems to work now. weird.
<Wizzup>
I'm pretty sure that last week I got 'no common ciphers'
<Wizzup>
which is just an indication that our tls libs are outdated
<Joerg-Neo900>
I seem urged to once more explain that a 100% "open libre" OS and userland still doesn't mean you own your device when you got no decent comprehensive docs about how the hw works. Some manufacturer throwing a "libre" >>write 4711(long) to 0xBADF00D; then keep GPIO42 low for 3.1415 seconds ...<< over the wall doesn't mean you control/own the device
<Joerg-Neo900>
the fallacy of RYF
<jonwil>
AMDs GPU division understands what that means. They publish nice docs, answer technical questions (and explain things that aren't explained well enough) and contribute lots of FOSS code to the FOSS drivers.
<jonwil>
Intel seems to do well with their GPUs as well on that score in terms of how well they are documented.
<Joerg-Neo900>
or rather, the fallacy of 95% of sw devels as soon as they start to think about hw
<Joerg-Neo900>
they regularly have no clue about interfaces, seems that's not a subject of education in CS study
<Joerg-Neo900>
jonwil: (AMD) nice unique example
<Joerg-Neo900>
not very common though
<jonwil>
Other than the PowerVR GPU and 1 or 2 other secret bits (security and things), the technical reference documents I have seen for the OMAP3 seem fairly good quality
<Joerg-Neo900>
the only really relevant part in it is >>They publish nice docs<< ;-)
<Wizzup>
AMD makes good drivers too
<Joerg-Neo900>
a lot of other hw manufs provide awesome amounts of FOSS code that nobody understands since the *docs* it's based on are covered by NDA
<jonwil>
All my GPUs since the GeForce 4 MX I used to have were NVIDIA GPUs but I am seriously considering buying an AMD GPU as my next choice just because of their openness (even though its going in a Windows machine and the open drivers won't benefit me)
<Joerg-Neo900>
this is arguable approach, only acceptable when the quality and completeness of the manuf-provided FOSS interface code doesn't leave any aspect uncovered
<Joerg-Neo900>
"any aspect uncovered" means there's no code covering the case when I write 0x04 ... 0xFF to that register that their code uses to control the states 0x00 ... 0x03
<jonwil>
Broadcomm is a good example of the crappier approach. They open-sourced a whole bunch of GPU driver code but had no technical documentation and no details of it all (e.g. no documentation on the instruction sets used)
<jonwil>
Leaving the people working on hw with Broadcomm gear in it to try and reverse engineer things from the crappy FOSS source code
<Joerg-Neo900>
exactly my point
<jonwil>
I just wish there was ONE company in the mobile GPU space willing to be as open as AMD is in the desktop GPU space...
<Joerg-Neo900>
broadcomn would nevertheless qualify for RYF with this crap
<jonwil>
Someone who made decent SoCs of the quality of the i.MX or OMAP SoCs
<jonwil>
Actually I dont think the Broadcomm stuff qualifies as RYF on the hardware that matters for most people (Raspberry Pi) since the driver code release was for a totally different SoC and stuff that mattered for the Raspberry Pi was missing or incomplete and had to be reverse engineered.
<Joerg-Neo900>
I meant the particular "FOSS" driver and affiliated chip
<Joerg-Neo900>
FSF_RYF doesn't really understand hw, still they claim they certify hw a Respects Your Freedom, based on a sw centric approach and understanding of how computers work
<Joerg-Neo900>
RYF feels to me like the inapt effort to find best BBQ receipe for a pudding
<Joerg-Neo900>
could be (C) Burger King
<Joerg-Neo900>
;-)
<jonwil>
FYI, the ISI documentation you mentioned before wasn't very good quality or very useful nor was it complete
<Joerg-Neo900>
sure it's a fine thing that somebody cares about "everything needs to be barbecued" just I don't care if I buy a pudding
<Joerg-Neo900>
jonwil: that ISI docs were better than almost all other modem interface specs existing, for any modem
<jonwil>
True, they were still useful
bencoh has quit [Ping timeout: 248 seconds]
bencoh has joined #neo900
<Joerg-Neo900>
what been terrible (and still is) is the ISI interface itself
<jonwil>
By far the most useful thing on that score is the cellmo-icpr82-headers package which contains the headers that match the N900 modem (or at least close to the right ones) although there are no docs for those
<Joerg-Neo900>
I think 3GPP AT specs are a good base to understand what a generic modem needs
<Joerg-Neo900>
and as it happens 3GPP AT is exactly what our (Neo900) modem speaks
<Joerg-Neo900>
the pnatd blob is a pathetic implementation of a AT<->ISI transcoder for a very few of the mandatory AT commands
<Joerg-Neo900>
heck I even had to prod the Nokia modem crew to implement USSD
<Joerg-Neo900>
and USC
<Joerg-Neo900>
err SSC?
<Joerg-Neo900>
the starhash stuff
<Joerg-Neo900>
then they implemented it neutered by config setting, and finally allowed to change the config to enable it
<Joerg-Neo900>
query form <annonymous>: >>hi. when you have time, I am still waiting summary status if possible, please. think at me that I would buy few phones and what should I expect. thank you.<<
<Joerg-Neo900>
current sttaus: we are busy to do the layout for proto_v2 (http://neo900.org/stuff/papers/v2.pdf ). Once layout is accomplished, we will build a maybe half a dozen of prototypes to allow evaluation for hw bugs and also hand out dev boards to sw devels. We also will start a kickstarter of sorts by then to rach our production volume preorder count (doesn't apply to those who already preordered!). Then (in parallel) we will start proto_v3 which
<Joerg-Neo900>
will resemble the final product - unless we find further hw bugs that need further changes to the design. For all this we can't give any binding timelines since they have always proven to be incorrect due to real life being unpredictable and Murphy throwing sticks at our legs
<Joerg-Neo900>
re timeline let me refer to Nokia who never published other timelines than "when it's ready", exactly for same reason
<Joerg-Neo900>
basically all manufacturers do
<Joerg-Neo900>
I can state *minimal* amount of time needed for particular steps. How much they are subject to parallelization and to delays from a myriad of organizational processes is up to anybody's guess
<Joerg-Neo900>
e.g when I find a layouter and that layouter claims "I need 3 weeks to complete this", there's no hint on the unfortunate finding that it takes me 6 weeks to talk them through "how to use git on windows"
bencoh has quit [Changing host]
bencoh has joined #neo900
<Joerg-Neo900>
or, for a positive example, we find an awesome guy like metacollin who only 2h after first contact already compiled fped and eeshow on OS-X which wasn't even genuinely supported by our sourcecode repos
<Joerg-Neo900>
you never know
freemangordon_ has quit [Ping timeout: 248 seconds]
err0r3o3_ has joined #neo900
jkepler has quit [Ping timeout: 255 seconds]
jkepler has joined #neo900
rhn has joined #neo900
rhn has quit [Changing host]
rhn has joined #neo900
err0r3o3_ has quit [Ping timeout: 246 seconds]
jonsger has joined #neo900
aqwsedrftgzh has joined #neo900
aqwsedrftgzh has quit [Client Quit]
illwieckz_ is now known as illwieckz
Kabouik has joined #neo900
Kabouik_ has joined #neo900
Kabouik has quit [Ping timeout: 248 seconds]
err0r3o3 has joined #neo900
err0r3o3 has quit [Client Quit]
qwazix has quit [Ping timeout: 240 seconds]
err0r3o3 has joined #neo900
err0r3o3 has quit [Client Quit]
err0r3o3 has joined #neo900
qwazix has joined #neo900
err0r3o3 has quit [Quit: err0r3o3]
err0r3o3 has joined #neo900
rhn has quit [Ping timeout: 248 seconds]
jonwil has quit [Read error: Connection reset by peer]
xmn has joined #neo900
rhn has joined #neo900
rhn has joined #neo900
rhn has quit [Changing host]
janus has joined #neo900
err0r3o3 has quit [Quit: err0r3o3]
err0r3o3 has joined #neo900
<rhn>
doesn't nvidia provide hw docs for their Tegra GPUs? if I remember right, there's not one but *two* open source drivers
<Joerg-Neo900>
possible, never looked into it
rhn has quit [Ping timeout: 260 seconds]
illwieckz has quit [Ping timeout: 255 seconds]
illwieckz has joined #neo900
jonsger has quit [Remote host closed the connection]
jonsger has joined #neo900
Pali has joined #neo900
Kabouik- has joined #neo900
Kabouik_ has quit [Ping timeout: 248 seconds]
luke-jr has quit [Read error: Connection reset by peer]
luke-jr has joined #neo900
xes has quit [Read error: Connection reset by peer]
xes has joined #neo900
Kabouik_ has joined #neo900
Kabouik- has quit [Ping timeout: 240 seconds]
vlitzer has joined #neo900
xmn has quit [Remote host closed the connection]
xmn has joined #neo900
jcarpenter2 has quit [Ping timeout: 240 seconds]
jkepler1 has joined #neo900
arnaudj has joined #neo900
jkepler has quit [Ping timeout: 248 seconds]
jkepler1 is now known as jkepler
ArturShaik has quit [Ping timeout: 240 seconds]
vlitzer has quit [Quit: ok bye]
pkicad has left #neo900 [#neo900]
rhn has joined #neo900
rhn has joined #neo900
rhn has quit [Changing host]
pagurus has quit [Ping timeout: 240 seconds]
pagurus has joined #neo900
jonsger has quit [Remote host closed the connection]
jonsger has joined #neo900
jonsger has quit [Remote host closed the connection]
jonsger has joined #neo900
rhn has quit [Ping timeout: 276 seconds]
_cris has quit [Ping timeout: 255 seconds]
vlitzer has joined #neo900
pagurus has quit [Ping timeout: 240 seconds]
pagurus has joined #neo900
Kabouik has joined #neo900
Kabouik has quit [Remote host closed the connection]
Kabouik has joined #neo900
jonwil has joined #neo900
jkepler has quit [Ping timeout: 255 seconds]
Kabouik has quit [Remote host closed the connection]
illwieckz has quit [Ping timeout: 255 seconds]
chomwitt has joined #neo900
illwieckz has joined #neo900
Pali has quit [Ping timeout: 240 seconds]
illwieckz has quit [Ping timeout: 248 seconds]
arnaudj has quit [Quit: arnaudj]
illwieckz has joined #neo900
illwieckz has quit [Ping timeout: 240 seconds]
illwieckz has joined #neo900
jonsger has quit [Ping timeout: 240 seconds]
illwieckz has quit [Ping timeout: 248 seconds]
illwieckz has joined #neo900
vlitzer has quit [Remote host closed the connection]
vlitzer has joined #neo900
vlitzer has quit [Remote host closed the connection]