research about free software wifi card options on the pinephone
In the posts are some pieces of information about hardware
approaches. I am not skilled such that I can gage them.
If the pinephone is a great phone and we have to settle with the
fact, that free software modems are not a thing then apart from
the wifi card requiring non free software, I consider the pinephone
a major step forward.
Therefore I have asked around on what can be done on the wifi card
matter?
The pinephone does not support pcie. According to pinephone's
irc, the ar9271 does not support sdio.
I asked both thinkpenguin and bunnie about the ar9271.
Thinkpenguin gets ar9271 cards manufactured. I asked him if
he could ask the manufacturer if ar9271 sdio wifi cards are an
option? He dismissed answering. I also contemplated reverse
engineering a wifi card. Thinkpenguin said, reverse engineering
a wifi 5 card could cost about 500000usd. Not more than 50000
people paying 10usd at a crowd funding.
Bunnie answered, he does not know about a sdio free software supported
wifi card. What he does is adding a controller between a non
free software wifi card and the mainboard's cpu. The controller chip
acts as a firewall. Providing the same security as when you
connect a device to a router running non free software.
Another option is to add a nano usb hub to the pinephone and
connect an usb ar9271 wifi card. If it will provide a well working
wifi card solution to the pinephone, I cannot tell. On replicant
phones it does not. Furthermore I have been told, an usb wifi card
on a phone consumes to much power.
Because pinephone reacts on user demands if enough users wants
a feature, then both the firewall controller chip and the
usb nano hub ar9271 wifi card are viable options. But people
have to ask for them.
I incited thinkpenguin to write
https://www.redpinesignals.com/
which he dismissed.
"we have to settle with the fact, that free software modems are not a thing then apart from the wifi card requiring non free software, I consider the pinephone a major step forward."
The problematic areas in phones have always been the hardware things like the GSM radio and WiFi. If we ignore the problematic pieces of hardware then it's just as libre as those that came before it.
It's a very hard problem to solve, which is why it hasn't been solved before. At the same time, let's not delude ourselves into thinking that the Pinephone somehow improves in this area or makes this difficult problem any easier to solve than it was before. It doesn't.
> ignore the problematic pieces of hardware
Getting familiar with the difficult matters on a field and
then decide for a best option approach is not ignorance.
> Pinephone somehow improves in this area or makes this difficult problem any easier to solve than it was before
We do not know if the steps taken improves the security and
privacy of the pinephone because we do not know if the non free software
on the pinephone is able to compromise the pinephone as such. We know that
free software can get audited and if required get modified. If
more software on a phone is free software then more software can get
audited. More free software on a device is an improvement. I
support your all or nothing view about free software. But I do not
think if we cannot do all we should do nothing. The pinephone's
cpu can work entirely on free software. Measures to rein in the
modem has been made. About the wifi card you can make use of
the hardware switch. And if better options get
available then buyers can demand they get
implemented.
"But I do not think if we cannot do all we should do nothing."
I think there may be a misunderstanding. I never said we should do nothing. But what people are working on continues the status quo; it does not advance. For an example:
"The pinephone's cpu can work entirely on free software."
Sure. My point was only: So can other phones. So the Pinephone's no different there. Continuing the status quo.
"And if better options get available then buyers can demand they get implemented."
That only works if people are working on those better options. As far as I know, they're not. Cell phones seem to have reached a "plateau" if you will where people work on (re-)inventing the easy areas and not progressing on the hard areas. The Pinephone is an example of that but it's by no means the only one. It just has the distinction of being the latest in this trend of continuing the status quo of this "plateau."
Your original post brought up good issues. No one's working on them but they need to be addressed in order to get off this plateau. They're very hard to address which is probably the reason no one's working on them and instead continuing the status quo.
> it does not advance
The pinephone utilizes hardware which purposefully is
hostile towards free software. But it is one of the
better options. And it is an improvement compared to the current
replicant compatible phones. Because its bootloader is
free software and it has hardware switches. Do you know other
phones having these features, apart from the purism phone?
If the non free wifi card on the pinephone finds a solution
then I regard the phone as a significant improvement in terms
of free software.
> if people are working on those better options
I regard the riscv as such an effort. But it should not stop us from
at the same time to follow the arm track. From what you have
written I understand the matter of a free software phone modem
is particular difficult. But are you ruling out a solution could
be found if we had adequate riscv hardware?
It is aggravating that there are probably 100 thousands people who
know about the pitfalls of hardware requiring non free software
to run, but free software people are incapable of allocating the resources
and counter.
One example is a free software wifi 5 card. If 500000usd can get
the reverse engineering job done, then why can a noted entity not
be found to manage a crowd funding?
> From what you have written I understand the matter of a free software
> phone modem is particular difficult. But are you ruling out a solution
> could be found if we had adequate riscv hardware?
The problems with the modem are inherent to the cell network. They have
nothing to do with the computer architecture, so RISC-V vs ARM vs x86
does not make a difference. The cell network is designed in such a way
that modems must run non-free software, and a modem capable of running
free software would be illegal.[1] Since the act of connecting to cell
towers allows them to triangulate your location, a modem must also be a
tracker.
The PinePhone's and Librem 5's kill switches are a mitigation, and a
mitigation is better than nothing, but it is not a step forward.
[1] http://wiki.openmoko.org/wiki/Open_GSM_modem
> It is aggravating that there are probably 100 thousands people who
> know about the pitfalls of hardware requiring non free software to run
I suspect that the number is much smaller than that.
> One example is a free software wifi 5 card. If 500000usd can get the
> reverse engineering job done, then why can a noted entity not be found
> to manage a crowd funding?
If spearheaded by the FSF then it is possible that this could work, but
it would be tough to raise that amount of money. In another message you
said that if 50000 people donated then it would only be $10 per person,
but I don't think 50000 would donate. Very few people care about WiFi
firmware.
> RISC-V vs ARM vs x86
Reverse engineering for one architecture would be doing
it for all architectures?
> modem must also be a tracker
That is another matter because it is something you know. The problem
about the none free software is the impossibility to verify
what it does.
> is not a step forward
It is because by that you can use the rest of the phone if
you do not want the phone's modem turned on.
> I suspect that the number is much smaller than that
If millions of people are familiar with gnulinux, you do not think
a fraction of them know about non free software running
devices?
> by the FSF
I have two views on fsf. One is the foot-dragging, arrogant
organization which gives the impression, give us the money and
leave us alone with what we want to do. I have aired the
crowd funding idea to fsf. And fsf did not forward it. Suggesting
it to rms did not change that. One time I suggested rms, that
fsf should promote, buy and test talos computers. The answer I
got was, the talos computers are nothing special. They are
the only intended free software computers on the market. Further
more the reply I got was, if raptorcs wanted ryf certification,
they could contact fsf. rms has said, the most important task
is getting computer hardware which does not require non
free software.
The other part is fsf's firmness about software must be
free software. You can rely on fsf not compromising. Under
rms you knew, rather would he have fsf dissolve than compromise.
> this could work
No it could not. If fsf announced a crowd funding, I am sure it would
not succeed. It would take preparation and planing should a crowd
funding have just a small change of succeeding. All major gnulinux
forums and outlets should be involved. It should be a campaign such
that wherever you were searching or informing yourself about
gnulinux, you would get to watch a banner about a wifi 5 free
software card crowd funding. People and companies with money should
be persuaded to make double up contributions.
If all that would be done following a detailed plan, then a
crowd funding might have a chance. How did fsf react when approached
about this?
> Very few people care about WiFi firmware
Part of running a crowd funding is to find out. Do people
want it? If people get the option to get a free software wifi 5
card on sale spending 10usd and people do not participate,
then they can only blame themselves.
> Reverse engineering for one architecture would be doing
> it for all architectures?
No, architecture is simply irrelevant to this discussion. Reverse engineering any architecture would not bring us closer to having a privacy-respecting phone.
> > is not a step forward
> It is because by that you can use the rest of the phone if
> you do not want the phone's modem turned on.
That's a mitigation, not a step forward. While the modem is off, the device is not a phone. This does not bring us any closer to having a privacy-respecting phone.
> > modem must also be a tracker
> That is another matter because it is something you know. The problem
> about the none free software is the impossibility to verify
> what it does.
That's one potential problem, and goes against part of freedom 1, but it is not the only problem. The right to use, share, and modify is also important. Even if someone wrote free software to run on a modem, it would be illegal to use it with the cell network. Even if someone released modem source code for us to expect, it would be illegal to run a modified version, at least with certain kinds of modifications. And even if we could run a modified version, we probably could not prevent it from tracking us. When you connect to cell towers, your location is triangulated *by the cell towers*.
> No it could not. If fsf announced a crowd funding, I am sure it would
> not succeed. It would take preparation and planing should a crowd
> funding have just a small change of succeeding. All major gnulinux
> forums and outlets should be involved.
Most GNU/Linux forums don't care about this. The FSF cares, Debian sort-of cares, and few others seen to.
>> Very few people care about WiFi firmware
> Part of running a crowd funding is to find out. Do people
> want it?
Probably not. The vast majority of GNU/Linux users run non-free firmware. Apart from Debian and the FSF-endorsed distros, virtually all distros come with non-free firmware by default, and their users do not complain.
> If people get the option to get a free software wifi 5
> card on sale spending 10usd and people do not participate,
> then they can only blame themselves.
I think more than $10/person will be needed in order to be successful, but I would support such a campaign if a credible organization started one.
Would the PinePhone be able to function completely free as a iPod Touch of sorts without internet? No modem, wifi, or bluetooth. I would be willing to spend 150$ for a fully free offline pocket device. It could serve as a calculator, offline map, music player, video player, ebook-reader, maybe some light offline games, alarm clock. If you could tether internet through a Librebooted x200 using usb you could even use it as a offline cryptocurrency wallet. Would my desired functionality from a PinePhone be possible with 100% free software?
> Would the PinePhone be able to function completely free as a iPod
> Touch of sorts without internet? No modem, wifi, or bluetooth.
Yes. Just leave the hardware switches for the WiFi card (which is also
what provides Bluetooth) and the modem turned off all the time. The
WiFi firmware runs on the WiFi device itself, not the main computer, and
the modem is a completely separate computer running its own OS, so as
long as those are physically turned off it's a free system as far as I'm
concerned. You could completely remove the firmware if you want, so
that it can't run even if you accidentally flip the switch. It might be
a good idea to get a microUSB-to-USB adapter and USB WiFi dongle so that
you have some way of connecting to the Internet in order to install new
software.
If it is indeed full GNU/Linux that ships on the PinePhone then wouldn't I simply be able to push software/updates/media from my PC connected by usb. Would the microUSB to USB and Wifi dongle be fully freed? In any case I think while the killswitches are an excellent idea, removing the software fully and flipping the killswitch would be better.
Thanks for the helpful comments as always chaosmonk!
> If it is indeed full GNU/Linux that ships on the PinePhone then
> wouldn't I simply be able to push software/updates/media from my PC
> connected by usb.
Media, certainly. Software and updates, I'm not sure. The phone would
need packages from the repositories of whatever distro it is running.
In theory it should be possible to download those packages on your pc,
copy them over to the Phone over USB, and manually install them, but
that sounds like a pain. Maybe some sort of tethering, so that your
Pinephone can connect to the Internet via your laptop, is possible. I'm
not sure.
> Would the microUSB to USB and Wifi dongle be fully freed?
Thinkpenguin sells several USB dongles for which there is free firmware.
I have used this model[1] and found that it works okay but is not very
reliable. If you aren't going to use it often then it's probably fine.
These[2][3] are bulkier and more expensive, but perhaps work better. I
haven't tried them myself.
[1]
https://www.thinkpenguin.com/gnu-linux/penguin-wireless-n-usb-adapter-gnu-linux-tpe-n150usb
[3]
https://www.thinkpenguin.com/gnu-linux/penguin-wireless-g-usb-adapter
> completely free
According to postmarketos' irc they will make a free software
pinephone image, except for the wifi card driver.
Compiling the image you may deselect the wifi software and
then you get a free software image and every device but the
wifi will work.