On Sat, 22 Jun 2019 12:24:10 +0300
"ni nhar" <nin...@inbox.lv> wrote:

> My understanding is replicant can run on any android phone.
> Correct?
Add support for a device in Replicant requires to have the very basic
features working in Replicant without Replicant having to ship nonfree

So for many Android phones, getting to that point would require a lot
of work, whereas for some other phones that are not very different from
the ones already supported by Replicant it's probably not a huge amount
of work.

> The reason why replicant is only available for a
> few old phones is because you have a modem isolation requirement
> about replicant.
It's because the project tried to add support for phones like the
GTA04 that have free software booloaders.

This port was to use a kernel that is closely based on upstream
Linux, and at the time the Replicant port failed because it required
more time than we had at the time. Now, the amount of work to use an
upstream kernel is way lower.

After that there was also an attempt to first port the LG Optimus black
to the upstream kernel and u-boot, for then adding support for it in
Replicant but that also took quite a lot of time, so that device was
still added upstream in u-boot with some support in Linux but, it
probably still lacks an upstream display driver for instance.

So as everybody were busy working in that direction, nobody worked to
add support for more recent devices.

> Does the pinephone meet the modem
> isolation requirement about replicant?
Assuming that they will also use an EC25 modem, like they did in the
prototypes, I think it could easily meet such requirements.

On the prototypes, the modem was connected through USB.

So, good modem isolation here is probably just a matter of
configuration either within the kernel, or through something like
usbguard, if it's possible to run it on Android, to make sure that the
modem cannot become a keyboard and take control of Replicant.

> If that is the case and the pinephone becomes a functioning phone of
> quality, I see no reason why replicant should not make the pinephone
> a priority.
What I was pointing out was that we are working to port Replicant to
Android 9, with some of the phones we currently support on Replicant 6.

So if that goes well, that work would most probably be reusable
with the PinePhone with minimal effort. However since the PinePhone
modem is different, we will need to look into it to understand how much
work would be required to add support for that modem. 

Some preliminary work has already been done to support the Galaxy SIII
4G modem, which uses the QMI protocol.

I've not looked deep enough into it to know what is the best approach
in supporting that modem protocol, but if that work is advanced enough,
adding support for that protocol might also be done faster by using the
Galaxy SIII 4G to do it.

It's also very difficult to know in advance if we will be able to add
support for a given device. To do that:
- We would need to be able to predict that some developers will have
  the time to work on it.
- We would need to make sure that the hardware doesn't have issues
  preventing it from being easily supported by Replicant.

To understand better what is possible or desirable we would need to get
more information on the PinePhone, for instance:
- What chip will it use for the WiFi, Bluetooth?
- What is the modem protocol?
- Is the bootloader fully free software for the A64?
  In u-boot, board/sunxi/README.sunxi64 has more information about
  that but we also need to look into it.

And as I understand, the PinePhone is not yet released, so some things
may change between the prototypes and the final device. I think it's
important for the Replicant project to review the PinePhone once it's

The Replicant project is looking forward since quite some time to
support devices with free software bootloaders, as nonfree bootloaders
are a huge issue for users freedom.

As devices with free software bootloader often use kernels that are
closely based on upstream Linux, adding support for any device with a
kernel closely based on upstream Linux is already a very big step in
that direction, and we are working on that.


Attachment: pgpGd3l5ZllGt.pgp
Description: OpenPGP digital signature

Replicant mailing list

Reply via email to