Success!!! Got it working! I was wasting time in wrong places! Since the problem is strictly and exclusively with AAS’ access to SIM / radio, obviously the file to modify is /opt/appsupport/vendor/etc/vintf/manifest.xml and not the LOS one in /vendor/! That’s the file that ofono-alien-binder-plugin uses to query for the number of slots/modems, and then it compares it with /etc/ofono/binder.conf for a matching number of slots. If they don’t match, the result is broken SIM card access in AAS as ofono-alien-binder-plugin probably goes bezerk.
The /opt/appsupport/vendor/etc/vintf/manifest.xmloriginally configures TWO slots (IRadio/slot1 and IRadio/slot2 as well as IOemHook/slot1 and IOemHook/slot2). Removing both slot2 entries, so that the number of slots matches /etc/ofono/binder.conf (i.e. only slot1 is configured on devices with only one physical SIM) made the SIM card instantly visible and mobile data fully working with AAS. Of course, in the future when the SFOS side gets eSIM support working, it’ll need to be modified on eSIM models for getting eSIM to work, but for now I guess it’s more than OK.
Now I’m not even sure if adding those kernel options or patching the vendor.img manifest.xml file were neccessary. So, @bundyo, just edit that single file to have a fully working mobile data in AAS
EDIT: The above is obviously only needed on SINGLE PHYSICAL SIM models, as only in that case the number of slots in that manifest file isn’t correct. On dual physical SIM models two slots are correct.
Thank you The 1 & 5 IV definitely provide the best SFOS experience ever, as it’s what seems to be an almost perfect combination of very powerful and beautiful (and now also affordable financially) devices with an already now extremely mature, stable and fully functional OS port. That’s why it is so motivating to tinker with it - it’s actually not impatience but very strong motivation because it is simply worth every effort. So thank you so much @rinigus and @vlagged for all your splendid work allowing us all to be able to enjoy such a fantastic port!
Yep, I currently have a glimpse of what it might be like with X5 II.
Unfortunately Thaodan is out of steam for the port, and since everything up to gen V is a real light-saber e-waste risk, so I doubt I will be SFOS-ing it before J2 arrives.
But LOS 23 on X5 II is super smooth, these are indeed great devices.
I keep having problems with charging limit being ignored (wired charging). It worked fine for quite a long time, and problems started at the beginning of May when I updated from the adaptation-community repo those new gst-droid packages related to mediabuffers (some other hal and config packages were also updated at that time). It was that exact day after updating those packages that for the first time the charging limit wasn’t respected and since then it’s been failing more often than not. Actdead charging seems to be OK but normally charging a booted up OS usually goes up to 100% and ignores the thresholds. Actually, only ones or twice since early May it stopped at 80%.
In the meantime, while playing with the AAS mobile data issue, I reinstalled SFOS, and on pure .71 OS charging limit worked OK. Then, after upgrading to .76 (which also installed those gst-droid packages from May 8th), it broke again.
Seeing your progress and persistence in the last weeks I’ve been anxiously waiting for this message. This is awesome!
I’ve not done anything to my phone except what we discussed in the past (I’ve been too busy lately and I’m also not knowledgable enough) and I just commented out the two slot2 entries that you pointed out and after restarting AAS mobile data worked.
So thanks again, I’m going to ditch my backup Android phone and try daily driving this phone!
Just commented those two settings out and it started to work, thanks a lot! I’ve rummaged quite often in this file due to the camera problem, but didn’t occur to me to check the sim slots.
It might be that the original install image was created with a custom mce, such as this from this PR ? I don’t remember exactly, maybe @rinigus remembers.
mce itself seems to be the same, its modules too. I’ll play with it some more as for now I can’t see any pattern for when it respects the limit and when it doesn’t. Last night it correctly stopped charging at 80%, only to ignore it merely a few hours later. I will see what might help: restarting mce prior to charging, maybe just switching the options back and forth on the Settings screen, or whatever else comes to my mind.
@vlagged Just a short note that your new build of droidmedia from devel (i.e. https://repo.sailfishos.org/obs/nemo:/devel:/hw:/sony:/nagara/sailfish_latest_aarch64/aarch64/droidmedia-0.20260517.0-1.aarch64.rpm ) seems to work perfectly fine! Smooth playback, no green lines or any other artefacts! I’ve tested multiple videos which previously had problems and all of them now play just fine! That of course with re-enabled avc decoder in gstdroid.conf and removed the no-media-buffers quirk. I will test it with some more videos, but already now it looks that it really solves the problem entirely. Congratulations and thank you so much for fixing it!
Thanks for testing - this is timely, as mal has merged and I have added 0.20260522.0 build to devel and testing just the same. So it will look like an update, but for you it shouldn’t change anything.
I couldn’t resist and I manually downloaded from devel and installed that single package as soon as I noticed that PR. Seems to have been worth it as now I have a fully working video playback in SFOS. Literally all videos are played OK.
I’m still struggling with AAS video and audio playback, though, still no clue what breaks it. Applications using their own decoders like e.g. VLC, or audio players with their own MP3 or AAC decoders, play everything just fine, while the system itself as well as all apps relying on default Android audio and video codecs fail to play anything. Not even system sounds, ringtones, etc. No hardware codecs are listed at all, so it looks like broken codec configuration.
@wetab73 if you (or anyone with a 1 IV) updated in the last 8h, I have accidentally re-built work-in-progress configs that may lock you out. Please check if you see a 2nd sim in configs and disable it before rebooting or update again.
I manually downloaded and locally installed that single droidmedia package yesterday, and then I haven’t updated anything from the repo, so I guess I’m safe
If you don’t want to wait, you can just manually download and install that single droidmedia rpm (linked above). I’ve been using it for 5 days now and enjoying 100% correct video playback without any side issues. You’ll just need to manually re-enable avc decoder (if you disabled it) in /etc/gst-droid/gstdroidcodec.conf and then manually delete ~/.cache/gstreamer-1.0/registry.aarch64.bin to refresh the gstreamer cache. This package alone doesn’t touch any configs or anything not related to droidmedia, so it is safe to give it a go
award for persistence. Keep it up.
The 1 & 5 IV definitely provide the best SFOS experience ever, as it’s what seems to be an almost perfect combination of very powerful and beautiful (and now also affordable financially) devices with an already now extremely mature, stable and fully functional OS port. That’s why it is so motivating to tinker with it - it’s actually not impatience but very strong motivation because it is simply worth every effort. So thank you so much 
