[Bug 221998] HID: playstation: GameSir G7 Pro 8K (054C:09CC) reports incorrect battery level (fixed at 5%)

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=221998

Krish Gulati (krishgulati7@gmail.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |krishgulati7@gmail.com

--- Comment #1 from Krish Gulati (krishgulati7@gmail.com) ---
Thanks for the detailed report. To narrow this down we need to see the raw HID
input report bytes from the controller, since the fixed capacity of 5% strongly
suggests hid-playstation is reading status[0] as 0x00 for this device; either
the report layout differs from a real DS4, or GameSir's DS4-emulation firmware
doesn't populate that field at all.

Could you provide the following?

1. First find the correct hidraw device for this controller (don't assume the
number, it can change between boots):

   grep -l 054C:09CC /sys/class/hidraw/*/device/uevent

This prints the sysfs path for the matching hidraw device(s); the last path
component (e.g. hidraw4) is the device name to use below. If the controller
creates multiple hidraw interfaces, capture from all of them.

2. The decoded HID report descriptor, ideally with an input report capture in
the same run....this tells us where the device itself claims the battery/status
fields live:

   sudo hid-recorder /dev/<hidrawX>

Let it run for a few seconds while moving a stick/pressing a button so we get
non-idle reports, then Ctrl+C to stop. This prints the decoded descriptor
first, then the captured reports.

3. If hid-recorder isn't installed, get the decoded descriptor and a raw report
dump separately:

   cat /sys/class/hidraw/<hidrawX>/device/report_descriptor | hid-decode

   sudo xxd /dev/<hidrawX> | head -50

Please capture at a few different real battery levels (e.g. near-empty, ~50%,
full), and both with and without USB charging active, since the parsing logic
branches on cable state.

This will tell us whether the battery/status byte is at a different offset than
expected on real Sony hardware, or whether the field is genuinely always zero
on this controller. Once we can see the byte layout we can put together a fix.
I don't have this hardware myself, so I'd need your help testing any patch/BPF
program against the real device.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.