Can you please confirm you have created the /config/usb_gadget/g1 directory? The error CONFIGFS not detected is resolved by that directory being present on the filesystem in my testing.
I executed the commands, connected the USB cable, the phone beeped two or three times, the connection method menu appeared, I selected MTP, the menu was repeated, I selected MTP again, but the connection did not occur. 
This is not that important, you can put it off, I can live with it. 
[root@Xperia10IV defaultuser]# systemctl stop usb-moded
[root@Xperia10IV defaultuser]# mkdir -p /config/usb_gadget/g1
[root@Xperia10IV defaultuser]# mkdir -p /config/usb_gadget/g1/configs/c.1
[root@Xperia10IV defaultuser]# mkdir -p /config/usb_gadget/g1/strings/0x409
[root@Xperia10IV defaultuser]# systemctl start usb-moded
[root@Xperia10IV defaultuser]# journalctl -u usb-moded -b
-- Logs begin at Thu 2026-08-06 08:22:57 CEST, end at Thu 2026-08-06 08:24:34 C>
aug 06 08:23:39 Xperia10IV systemd[1]: Starting usb-moded USB gadget controller>
aug 06 08:23:39 Xperia10IV usb_moded[15391]: usb_moded 0.86.0+mer69 starting
aug 06 08:23:39 Xperia10IV usb_moded[15391]: init_done -> reached
aug 06 08:23:39 Xperia10IV usb_moded[15391]: CONFIGFS detected
aug 06 08:23:39 Xperia10IV usb_moded[15391]: /sys/kernel/config/usb_gadget/g1/f>
aug 06 08:23:39 Xperia10IV usb_moded[15391]: /sys/kernel/config/usb_gadget/g1/f>
aug 06 08:23:39 Xperia10IV usb_moded[15391]: usb-moded started after init-done;>
aug 06 08:23:39 Xperia10IV systemd[1]: Started usb-moded USB gadget controller.
do you think you will take this further into the 10 series?
i.e. the 10vi/vii for example
Hey! You have CONFIGFS detected now, and no more all bets are off! My research shows it should be working! @juz has also reported it appears to work on his 10V the same way.
As a last-ditch attempt, try zypper ref && zypper dup and reboot and try again when that is done. Maybe my system is more up to date than yours, but I donât think anything should have changed.
2 Likes
Hmm, I donât own any other Xperia besides the 10IV. The only reason I was also able to do the 10V was because the phones are so similar and use the same SoC and kernel.
Perhaps one day if I ever purchase a 10Viii or something to replace my 10IV, but I do not have the desire to purchase a new phone when there is nothing wrong with this one. That said, porting these devices was quite fun! So whatever phone I get next I will certainly try to port as well.
4 Likes
I canât get out of the lock screen after this. It looks like the touch screen stopped working, it doesnât respond to touches.
EDIT: the sounds also stopped working
To make it more precise, I used the commands pkcon refresh and pkcon update
I specifically said zypper ref && zypper dup for a reason. You have updated your modules directory now without reflashing the updated kernel.
Read up at sailfish-on-murray/TROUBLESHOOTING.md at main ¡ sharks-dev/sailfish-on-murray ¡ GitHub to find out in more detail what you have done and how to fix it.
2 Likes
I think such things are normally done via systemd / udev rules and it looks like it applies here:
1 Like
Thanks, that is certainly a helpful document, and it mirrors the path I have started going down tonight. I plan on making a systemd service that will run a script before usb-moded.wants or something like that. My earlier thinking was that this wouldnât be necessary and it was something wrong in my usb-moded configs, but I am less convinced of that now.
I am still trying to actually work out what I need to do in order to get developer mode to work, but once I have a script that works every reboot to fix developer mode and MTP reliably, Iâll make the service, add it to droid-config-pdx225/sparse at master ¡ sharks-dev/droid-config-pdx225 ¡ GitHub, and push out the update to users to install with zypper ref && zypper dup.
2 Likes
The mobile responds to touches again, my faul (âA word to the wise is enoughâ), but the MPT still doesnât work (I ran the zypper commands and rebooted).
EDIT: Can I delete the previous (old) directory from /lib/modules/?
1 Like
Glad you managed to fix it.
As discussed, I am working on an easier solution so perhaps wait a moment for now. I have just published a commit that reliably solves the MTP problem using a systemd service as discussed above. In a few days I will push this out to all users, but hopefully I can also push out a fix for developer mode at the same time, weâll see.
Edit: provided you have got the new kernel version correctly reporting in uname -r, yes you can delete the old modules directory.
Okay, of course Iâll wait, I donât mind, I have 10IV for laboratory work.