anyone managed to get ds4 working on custom pie roms?
Trying to figure this out too. Absolutely nothing even detects the device - BT can pair of course, but zero function of the gamepad in any app. USB is useless.
EDIT: Suddenly it works (over BT)! I have no idea how though sorry lol. I do have the 1controller and Sony Framework magisk modules, though - am using TwistedKernel right now.
Sadly though, BT Autoconnect doesn't fix the lag. It does for a while, but then goes bad again. I *do* have continuous connect set to 5 seconds doesn't help.
EDIT2: Here is an interesting bit of logcat:
Code:
03-31 03:47:31.986 706 3677 I [email protected]_handler: ProcessIbsCmd: Received IBS_SLEEP_IND: 0xFE
03-31 03:47:31.986 706 3677 D [email protected]_handler: SerialClockVote: vote for UART CLK OFF
03-31 03:47:32.041 706 3677 I [email protected]_handler: ProcessIbsCmd: Received IBS_WAKE_IND: 0xFD
03-31 03:47:32.041 706 3677 D [email protected]_handler: SerialClockVote: vote for UART CLK ON
03-31 03:47:32.041 706 3677 I [email protected]_handler: ProcessIbsCmd: Writing IBS_WAKE_ACK
03-31 03:47:32.044 706 706 I [email protected]_handler: DeviceWakeUp: Writing IBS_WAKE_IND
03-31 03:47:32.045 706 3677 I [email protected]_handler: ProcessIbsCmd: Received IBS_WAKE_ACK: 0xFC
03-31 03:47:32.137 706 3677 I [email protected]_handler: ProcessIbsCmd: Received IBS_SLEEP_IND: 0xFE
03-31 03:47:33.046 706 3684 I [email protected]_handler: DeviceSleep: TX Awake, Sending SLEEP_IND
03-31 03:47:33.046 706 3684 D [email protected]_handler: SerialClockVote: vote for UART CLK OFF
03-31 03:47:33.197 706 3477 D [email protected]_lock: Release wakelock is released
The last few lines (from DeviceSleep) is exactly when the lag started again. Bluetooth Auto Connect seems to not be working with this particular driver I guess? Hmm... I have a feeling this won't be solved until dev's start releasing ROM's with custom vendors based on CAF stuff.
EDIT3: After a reboot, once again I can't get the controller working at all. Curious. I see this kind of thing looping in logcat:
Code:
03-31 04:39:44.244 1857 2533 W bt_btif : bta_hh_co_data: Error: fd = 228, ready 0, len = 10
03-31 04:39:44.288 702 2530 I [email protected]_handler: ProcessIbsCmd: Received IBS_SLEEP_IND: 0xFE
03-31 04:39:44.288 702 2530 D [email protected]_handler: SerialClockVote: vote for UART CLK OFF
03-31 04:39:44.295 1857 2533 W bt_btif : bta_hh_co_data: Error: fd = 228, ready 0, len = 10
03-31 04:39:44.313 702 2530 I [email protected]_handler: ProcessIbsCmd: Received IBS_WAKE_IND: 0xFD
03-31 04:39:44.313 702 2530 D [email protected]_handler: SerialClockVote: vote for UART CLK ON
03-31 04:39:44.313 702 2530 I [email protected]_handler: ProcessIbsCmd: Writing IBS_WAKE_ACK
03-31 04:39:44.347 1857 2533 W bt_btif : bta_hh_co_data: Error: fd = 228, ready 0, len = 10
bt_btif seems to be spitting out an error. This *is* from platform code (not a qcom blob) so maybe it can be fixed without moving to a whole CAF vendor afterall. Looking at
the source of this error, finding out why the HID device won't report ready is probably beyond my skill, let alone fixing it... *charges Google-fu*
UPDATE: Forget what I said about kernel. It would be better to use Franco's or any other kernel that upstreams from mainline Linux because kernels based on LOS/Android one use hid-sony driver from 2016; Linux kernel has had a looooot of updates to hid-sony since then so it's best to use Franco or other mainline base. Also note that both Franco kernel and TwistedKernel have the same not-ready error from above.