Support the latest version of Sony binaries

Foreword: I am not sure if we will ever see a new version of SFOS at all, let alone improved device support. Nevertheless I am asking for an important upgrade.
Text: Newer devices, especially the Xperia 10 III, are ridden with low-level, device-specific bugs (power consumption, mobile data, etc.). There is a high chance that these lie in Sony binaries, which have been updated several times after the version used by SFOS. Although this implies a kernel and an Android base upgrade, I think that supporting an updated version of Sony binaries would be very important in order to make the device usable and avoid forcing people to flee unwillingly to usable systems.


Someone with a fresh 10 III should actually try them, or else we don’t even know if maybe some of them might work without any special efforts. Probably not, but who knows. It’s just that in order to check e.g. the Android 12 based binaries it might take having the phone upgraded to Android 12 (I mean the Sony’s stock Android), to have the correct partitions layout, etc., which may differ between Android versions. And the same for Android 13, of course.

So, has anyone actually tried it yet on the 10 III with either Android 12 or Android 13 binaries, having upgraded the phone to the corresponding Android version first? I was thinking about it many times, but I have way too many modifications (and still way too stable and good running OS) to wipe it off for such a test…

1 Like

My Xperia 10 II arrived with Android 12 installed but the GPS was not able to fix any satellite. However, after an investigation about that issue I reach the conclusion that without a correct configuration of A-GPS system there is no cache to have a quick and precise geolocalisation fixing.

The thread above is the continuation of a wider product planning to complete the hardware support for Xperia 10 II e III, in particular. Then, deliver a system configuration more suitable for advanced users that do not want spend their time in struggling in the hacking playground¹.

So, it seems reasonable for me to - in the near future - to start again from fresh but using an Android 12 as pre-installed system instead of Android 11. I have been riduculised when I supported the idea that Android system and microG was able to influence the behaviour of a much larger stack portion respect those supposed to be.

Unfortunately, there is the ideological bias about SailFish OS that it should be an indipendent system respect Android but as long as it would deliver its own hardware support with Linux kernel device drivers supported by proprietary firmware - in case of necessity - but installed and used indipendently the Android systen, there is a little chance that the jailing will work as expected without any corner case. Having a corner case, it means that the jail does not work!

However, I agreed to not propagate this PoV without reproducible proof of my claims. Therefore, consider this post a confutation trail of the SFOS architecture. In fact, in science the confutation is used to exclude false hypotehsis and to reach to the most objective factual description of the reality. It is the equivalent of reasoning for absurd in maths.

  • SailFish OS is an indipendet system


  • Android and related stuff influence SFOS

There is different experiences about using Android 12 with SFOS and there are many explanations possible that can drive torward one HP or the other.

It is an interesting and sensitive topic but a good playground to investigate the SFOS architecture and confirm or confute the theory about its indipendence.


¹ leave to the hackers those belong to the hacker, give to the users what users wish to pay for.

I tried the phone with different Android 11 and 12 versions but it made no difference.

I also tried all the SW Binaries, but the newer ones dodn’t work.

1 Like

I did. The phone stops during data encryption. Rebooting it, the encryption code is set to 00000. A simple way to avoid encryption (which is for me only a nuisance and a loss of time on boot, btw) could be useful to test if there are further issues.

1 Like

you can after flashing boot into recovery and then delete a file: rokua *without* encryption? - #33 by tombln

Too much of a hassle.

You underestimate how much work that is. it is not a little something to be done “just because”.
We know the echo cancellation is fixed in later binaries; and which does not drag everything with it.
It sounded hopeful that we could benefit eventually.

But do we actually know that any of the other issues would be fixed?

I think it needs to be quite a bit beyond “chance” territory to be done to the extent you describe. There are no silver bullets.
Some annoyances are probably similarly likely to exist in the kernel itself; and thus be subject to much less intrusive regular backporting once identified.

For completeness; i do agree that these are things where a newer base port and binaries do have an impact (we just need to categorize/attribute the issues to know where to spend effort).
But why someone would conflate these binaries with microG is beyond me.


I do not underestimate. If Sony employed men and resources to develop the updated binaries, this means that they have a purpose for the device. Whatever the purpose, it is a minimal request that improvements carried out by Sony be made available to SailfishOS users, rather than leave us with not-quite-ready software forever.


The purpose could theoretically be to simply be compatible with newer Android versions, with no actual improvements whatsoever (echo fix being a fluke). Should we then still update?

I don’t believe it to be quite that bad… but until we know; we don’t know.

You barely know what you are talking about. You find SFOS as another android based clone but you are completely wrong. It is not ideological bias but only your belief and nothing more. You were offered a good advice to learn at first but did not take a chance.


Instead of so many words, tell me where is the driver to deal with the fingerprint reader for Xperia 10 II and the software stack that manage it, so I can fix it. Thanks, R-

Here you go: GitHub - sonyxperiadev/vendor-sony-oss-fingerprint: Sony Fingerprint HAL (TZ based)

And we have come full circle, and are going for a loop.


The only sad true here is that in 3 years since Xperia 10 II entered into OpenDevice program the hardware support is still not complete. Probably because people here are more concerned to close loops than deliver a usable product. :slight_smile:

  • XA2 users still should go through a complex procedure for activating A-GPS
  • X10 III users complaining that the battery is draining out too quickly
  • X10 II - I am taking care of them because I am one of them…

Would you help me in doing something useful or do you prefer plays with loops? The choice is your.

I barely have an idea of what I am dealing with because before 26th May 2023, I have never get into a smartphone but I was just one of billion that were using a on-the-shelf technology. Despite this, I managed to summirise a post-installation procedure that shrink the worktime from a week or two in one day.

At the moment there are at least 4 changes proposal that are waiting for someone that takes care of them:

  • an easy deploy of the A-GPS config for X10 II and III (here)
  • energy saver at 50%, 100% battery threshold and histeresys at 87% is too much, 85% at least (here)
  • device lock adding 5/10/30/60 seconds, also (here)
  • fingerprint reader and bluethoot awakening or a fingerprint restart icon/link in lockscreen (here)

Possilbly also these two:

  • add an option for disabling the Android Support autostart and kept it sleeping/ready to start (here)
  • add an option in the native broswer to let it keeping alive a backgroud tab that is doing some taks like playing music (here).

The first four are simple tasks that require few hours/days of a SFOS hacker and greatly smooth the user experience of everyone intersted in adopting SFOS with X serie.

It is your choice, it always been since 2004 four years before Android debuted. :blush:

@robang74 While I think I get your intentions right, there are some things that you should be aware of:

  1. Some people in this community may have a deeper understanding of how Linux libc distributions (SFOS in particular) run on hardware delivered originally with Android (hint: by commiting hybris)

  2. You may come up with fresh ideas, but some of them may have been battered for ages. Please consider using a question instead of a “task” for them.

  3. Try to come up with solutions which aren’t sourced only from this community - since some of us are older than this forum and know every topic that was made here, it may be that you look like repeating something “everyone” knows - well, it’s a search away

Again, I hope I get your intentions right and am encouraging you to continue doing what you’re doing, BUT with a friendlier tone and by putting more work into understanding the innards (sorry if it’s a steep curve since you’ve only started 20 days ago;)


OK, thanks. So at least we know for sure.

So it is exactly the same moment as on the XA2 Ultra when flashed with Android 9 binaries… It would be interesting to see what happens if encryption is disabled, e.g. the way that @smatkovi suggested. Maybe I’ll try it with my XA2U.

I do not use file encryption.

First of all, thanks for the positive feedback.

Please note that after being a hacker myself - proficiency in other specific sectors not about smartphone OS, yet - I am also specifically educated in project, product delivery and business managements. This brings into the picture a complete different perspective.

  1. In fact, I highlight a particular phrase in your statment above. Sometimes, there is not a solution that fits 100% of requisites or said in another words there is not the perfect solution. Debating for ages about how to implement something is a waste of time, resources and in particular brings to the failure in delivering a product.

  2. On the other side the debate, positive and costructive, is also the reason because open source solutions are much better than the proprietary counterparts. The numbers of people and PoVs involved are much higher, many more contributions, etc.

Unfortunately, the statment #2 (positive contribution by many) contrasts with the statmente #1 (need for delivery a product). In the Linux kernel community this contrast have been solved letting Linux Towards performing the role of guide. This role has been granted along decades because he knows how to deliver a product.


You may argue that delivering a product does not matter for the community. I agree that many people overlook this aspect. Unfortunately, missing this point fails to encurage the adoption of the technology (SFOS) or blocks it completely. Without a relatively fast adoption curve and large adoption share of a technology (SFOS) every related business strives to survive and on the long run are doomed to fail.

Again, you might not care about business expecially if it is others people business. Unfortunately, others people business can/might be our own business as much as our goals are aligned or complementary.

All of these is about product management. Not about the leadership, not about the community, not about the business. It is about winning or losing the refering market (previosly written as technology adoption rate and share).

This also means that Russian and Chinese people have not a trustful OS to use for their everyday life unless they are hackers. The same for Sirian and Ukrainian people like everyone living in a battle field or in a nation at war.

So, you might not care about Jolla business or about product delivery but you should consider that missing to align/complement with these two “value propositions” also means leaving all those people above in the cold, unless another viable solution exists until it cope to exist (competitor product/offer).

You might hate words like “business”, “product”, “marketing”, etc. and I understand your point of view. I invite you to reconsider that this words are just tools not aims.

These tools can be used for the good or for the bad, either. Instead, ignoring these word or even worse contrast these tools brings nowhere. Nowhere is a place in which everybody feel safe because every confrontation and every effort is meaningless but just self-complacent. The same gap between having sex with others and sex with themselves.


Those - who got the real meaning of my overview (named vision in the companies), the deep implications of it and the transparency/clarity in which I expose it without trying to fool people but acknolege them - are prone to accept and complete tasks necessary to deliver a product.

That solutions might not be the best and the perfect solution may not exists either

Under these condition an “arbitrary” decision should be taken or ages passed in vain. A decision that goes in the direction of a fast rate adoption and large enough share. A decision that do not interfere with the technical implementation nor with the future improvements/developments nor with the possibility of having hackers-friendly options/customisations available within a reasonable range.

This brings us to please the 80% of our possible adopters and put the others 20% in condition of decide their own options and customisations. This is about product delivering, not about leadership, not about technology excellence, not about business nor money. It is about technology (SFOS) adoption and thus about technology (SFOS) success.

I hope to have clarifies enough my point of view. Therefore, I am attaching an updated list of tasks. Feel free to do whatever you like/want with this list. :blush:


EDIT - this list has been moved into quick start quide in its new extended version.


Could you please stop posting nearly the same things everywhere. Nearly all of this is totally off-topic…