@realrichardsharpe,
The Snapdragon 888 in the Xiaomi Mi 11 5G includes the WiFi, Bluetooth, cellular modem and GNSS inside the chip, so it would be impossible to have hardware kill switches that cut the electrical current to those components. In contrast, the L5 uses six separate chips (i.MX 8M Quad, RS9116, BM818/PLS8, Teseo-LIV3F, WM8962 and bq25895) in place of the standard integrated mobile SoC like the Snapdragon, and the PinePhone uses 4 chips.
There may be pins on the Snapdragon 888 to turn off those individual components or maybe it can only be done via software switches. It is hard to know, because Qualcomm doesn’t publicly release any hardware documentation for the Snapdragon 888 or its development tools. To get that info, you need to be an electronics company that signs an NDA.
Qualcomm doesn’t support the use of the Linux desktop for the Snapdragon 888, since it only focuses on Android (and Android’s modified Linux kernel), so the only practical way to make a Linux phone based on a Snapdragon would be to use libhybris which allows the standard Linux stack to run on top of an Android kernel with Android drivers. The other option is to wait 2-3 years for the mainline Linux kernel to eventually support the Snapdragon 888, but by that time, it is likely that Qualcomm will no longer be producing the chip, although Qualcomm does chose a few of its Snapdragon series to get long-term production and support for its industrial customers.
Qualcomm is better than the other manufacturers of mobile integrated SoC’s, since it releases the source code for its Snapdragon Android Linux kernels and drivers (as required by the kernel’s GPL 2.0 license) on CodeAurora with the commit history, rather than just doing tarball dumps (like Samsung does) or only releasing the code to OEMs who then violate the GPL by not releasing the code to end users (like MediaTek does).
What this means is that eventually it may be possible to run a Snapdragon using a mainline Linux kernel, but it is hard for any company to plan their products around it. Look at what happened to the F(x)tec/XDA’s Pro1 X phone. They originally planned to use the Snapdragon 835, but then Qualcomm stopped manufacturing it, so they had to switch to the Snapdragon 662, which is one of the reasons that the phone has been delayed. I assume that the Ubuntu Touch version of the phone will be using libhybris with Android drivers. The Snapdragon 835 was announced in Nov. 2016 and started shipping in Q1 2017, but it didn’t start getting mainline support until Linux 4.20 in Dec. 2018 and it wasn’t until Linux 5.4 in Nov. 2019 that most Snapdragon 835 laptops were able to run standard Linux.
It is far better for Purism to chose an SoC that will be manufactured till Jan. 2033, whose manufacturer makes commits directly to the mainline Linux kernel, has a public forum to ask questions, and releases all 6000 pages of its documentation and the source code (to anyone who has registered an email address that isn’t Google, Yahoo, etc.). Yes, the performance of the i.MX 8M Quad doesn’t compare to the Snapdragon 888 and I do think that will effect people who want convergence devices, but you have to ask what is important to you.
Do you care about transparency? I suspect that Qualcomm, MediaTek, Samsung and UNISOC won’t let any company publicly release the schematics for their phone if they use their SoC’s, and they probably wouldn’t allow them to release x-rays of their phones either, since they are based on their copyrighted reference designs. Do you care about lifetime software updates and getting firmware updates till 2033? Do you care about being able to run the latest Linux kernel on your phone, or do you want a phone like the Fairphone 2 that is still running a kernel from 2012 because Qualcomm refused to release updated drivers for the Snapdragon 801 that would run with a newer kernel.
At the end of the day, you do have ask whether you want to invest in developing mobile Linux, because frankly that is the only way that we are going to combat the kind of surveillance capitalism, planned obsolescence and tech lock-in that currently plagues the mobile industry. I have been installing LineageOS and using F-Droid for the last 6 years on my own phones, but I see far better prospects for mobile Linux to become a viable market niche than the AOSP derivatives, and more importantly, I see AOSP derivatives as only a partial solution that is still ultimately dependent upon the mercies of Google.
Let me give you an example of why you won’t be able to buy a phone with an AOSP-derivative preinstalled in the future, whereas the big phone makers may be willing to take a chance with mobile Linux. Let’s say that Sony decides to make one experimental AOSP-derivative model to test whether it is a viable market. When Google gets wind of it, Google can call up Sony and say:
Are you sure that you aren’t creating a fork of Android, because that is a violation of our Open Handset Alliance rules. We can cut off your access to Google Mobile Services for all your phone models and we can stop giving you early access to the AOSP/Android code in development, so your phones will get the next version of our OS later than all your competitors. Are you sure that you want to become the next Huawei?
Google can’t make the same kinds of threats to phone makers who decide to release a few Linux phones (without violating anti-trust laws), but it will require a lot more dev work to make mobile Linux a viable alternative to AOSP. There are really only two paths available to us. One path is to pay the high prices for the Librem 5, so that Purism has the money to develop Phosh, or the other path is to participate in the community development of Plasma Mobile, which is going to be a much slower process and is much less focused on normal (non-technical) users. Take your pick, but don’t pretend that you don’t have to make some kind of investment either in the form of money or your time in order for mobile Linux to get to a better future.
PS: Before anyone points out that you can buy the Xperia X/XA2/10/10 II with Sailfish OS from jolla-devices.com, then consider the fact that there is no phone maker currently selling new phones with Sailfish OS preinstalled and every company that has tried gave up after 1 or 2 phone models. Sailfish OS will never get any community support like Phosh and Plasma Mobile are currently getting, so I can’t see it ever attracting much developer mind share.
Yes, you can now buy the Pro1 X or Volla Phone with Ubuntu Touch preinstalled, but then you are going to be dependent on Android device drivers that won’t get updated and the work of the volunteers at UBports who are maintaining a mountain of siloed code that they didn’t write. I wish them luck, but I have looked at the number of UBports commits and I don’t think they are ever going to recruit enough volunteers to maintain all that code (but I would be happy to be proven wrong).