Another piece of information which may be of interest is that by running: hexdump -v -e '49/1 "%02x " "\n"' < /dev/hidraw4 I could see that the hid report contains in fact the data for the buttons which do not respond. I think there is a quirk to the official ps3 controller that isn't working with this controller, perhaps related to the fact that the ps3 controller doesn't self-initialize (you have to press the home button), and these ones don't. Let me know if you need me to run any other test. I've been compiling kernels for testing lately, so if there is anything that can be patched-in for logging, just let me know and I'll do so. On Thu, Jan 31, 2019 at 7:27 PM daniel jimenez <daniel.jimenez.gomez@xxxxxxxxx> wrote: > > I have tested the controller adapter using the evtest utility and can > confirm that no keycodes appear when pressing the unresponsive > buttons. Every other button displays keycodes in the test. > > On Thu, Jan 24, 2019 at 6:12 PM Ken Sloat > <ken.sloat@xxxxxxxxxxxxxxxxxxxxxx> wrote: > > > > > > > > To David Herrmann, and all of the linux input mailing list, > > > > > > This is to report a bug concerning a psx/ps2 to usb game controller > > > adapter that is marketed to work on the PS3. The device presents > > > itself using the manufacturer id of sony and the device id of the > > > official sixaxis/dualshock 3 controllers, and seems to self-initialize > > > upon being connected (unlike the original controller which requires > > > that one presses the menu button). > > > > > > The problem is that the 'circle', 'square' and 'cross' buttons don't > > > work while every other button does. The adapter can emulate a home > > > button press by using the 'analog' button on the controller, but this > > > also doesn't work. > > > > > > The jstest applications confirms that the driver is presenting this > > > controller as having only 13 buttons, while a real ds3 is presented as > > > having 17. The same device id and vendor id are shown in jstest, both > > > are using the hid driver, and everything in dmesg and lsusb appears > > > the same between the original controller and the adapter. I can > > > provide every output log that is required. > > > > > > After having a look at hid-sony.c, it's clear that the buttons that > > > are not working (square, circle, cross, and home) are the last four on > > > the sixaxis_keymap[] bitmap. I speculate that there is some process > > > which is used to identify the control which fails to work with the > > > current driver, so that it ends up defaulting to a different keymap > > > with fewer buttons than needed. > > > > > > This very same problem seems to have been reported years ago (circa > > > 2012 iirc) on fedora forums for some non-official ps3 controller > > > replicas that also present themselves as the ps3 controller. The fix > > > to this problem could very well improve compatibility with such clone > > > controllers. > > > > > > I have observed it works well in both the ps3 and a windows pc with > > > the non-official drivers for the dual shock controller, so I can > > > confirm the controller work and the adapter works as well. > > > > > > Be sure that I can run any test that you believe may be useful to fix > > > this issue, and can compile and test possible solutions in my system. > > > > > > Best regards, > > > > > > Daniel Jimenez > > > > > > > Hi Daniel, > > > > Have you tried running the "evtest" utility on the device? This tool > > should be able to show any events for the device (which will include > > keycodes). It would be interesting to see if any EV_KEY events occur > > when you press the non working buttons and if so what codes they > > appear as. > > > > Thanks, > > Ken Sloat
