Hello, in Buildroot we've seen BlueZ 5.79 builds fail if configured with --disable-hid and --enable-hog (see [1] for logs, [2] for discussion): /home/autobuild/autobuild/instance-7/output-1/host/lib/gcc/mips-buildroot-linux-musl/13.3.0/../../../../mips-buildroot-linux-musl/bin/ld: profiles/input/bluetoothd-hog.o: in function `hog_disconnect': /home/autobuild/autobuild/instance-7/output-1/build/bluez5_utils-5.79/profiles/input/hog.c:211:(.text.hog_disconnect+0x3c): undefined reference to `input_get_userspace_hid' /home/autobuild/autobuild/instance-7/output-1/host/lib/gcc/mips-buildroot-linux-musl/13.3.0/../../../../mips-buildroot-linux-musl/bin/ld: /home/autobuild/autobuild/instance-7/output-1/build/bluez5_utils-5.79/profiles/input/hog.c:211:(.text.hog_disconnect+0x44): undefined reference to `input_get_userspace_hid' collect2: error: ld returned 1 exit status Looks like profiles/input/device.c isn't compiled with --disable-hid, which is why the function is missing during linking. It appears to me that either the HOG conditional should be fixed so all required files are compiled if it's enabled, or HOG should require HID. I'm not that familiar with BT input devices, so I'm not sure which is the correct approach. Buildroot has been forbidding building HID without HOG due to a similar bug for a while [3] due to a similar bug, which makes me wonder what's the recommended solution. For Buildroot granular configuration is desirable so embedded systems with limited storage can include only the required parts. What are your thoughts on this? I'm happy to look at patching once the correct approach is clear. Best regards, Fiona [1] http://autobuild.buildroot.net/?reason=bluez5_utils-5.79 [2] https://lore.kernel.org/buildroot/4abf920a-3da8-457b-a287-cb57fc7fd20d@xxxxxx/ [3] https://gitlab.com/buildroot.org/buildroot/-/commit/9850f262fd6546d15ae27f84a0e36411abc2b7b5
