Fix some ARM builds due to HID brokenness
PATCH 1/1 Fix some ARM builds due to HID brokenness
The new location for HID is extremely annoying:
1. the help text implies that you need to enable it for any keyboard or mouse attached to the system. This is not correct.
2. it defaults to 'y'. When you have input deselected, this causes the kernel to fail to link:
drivers/built-in.o: In function `usb_hidinput_input_event':
hid-input.c:(.text+0x55054): undefined reference to input_ff_event' drivers/built-in.o: In function hidinput_hid_event':
hid-input.c:(.text+0x6446c): undefined reference to input_event' hid-input.c:(.text+0x644f8): undefined reference to input_event' hid-input.c:(.text+0x64550): undefined reference to input_event' hid-input.c:(.text+0x64590): undefined reference to input_event' hid-input.c:(.text+0x645b8): undefined reference to input_event' drivers/built-in.o: In function hidinput_disconnect':
hid-input.c:(.text+0x64624): undefined reference to input_unregister_device' drivers/built-in.o: In function hidinput_report_event':
hid-input.c:(.text+0x64670): undefined reference to input_event' drivers/built-in.o: In function hidinput_connect':
hid-input.c:(.text+0x64824): undefined reference to input_allocate_device' hid-input.c:(.text+0x675e0): undefined reference to input_register_device' hid-input.c:(.text+0x67698): undefined reference to input_free_device' hid-input.c:(.text+0x676b8): undefined reference to input_register_device'
make: *** [.tmp_vmlinux1] Error 1
Fix the second problem by making it depend on INPUT. The first problem is left as an exercise for the HID maintainers to solve.
drivers/hid/Kconfig +1 -0
@@ -6,6 +6,7 @@ menu "HID Devices" config HID tristate "Generic HID support" + depends on INPUT default y ---help--- Say Y here if you want generic HID support to connect keyboards,
Jiri Kosina
AppliedQuoted 1 line · Russell King
The new location for HID is extremely annoying:
Hi Russell,
well, the location itself was a subject to various discussions. At the end, drivers/hid won against drivers/input/hid, because, as Marcel correctly pointed out, in case of embedded systems there might be a need to have an abstraction for non-input devices and then we don't want to bundle it to tightly to input. There is no such implementation yet, but moving this code around is quite painful, so we don't want to do it more than once.
Quoted 3 lines
1. the help text implies that you need to enable it for any keyboard or mouse attached to the system. This is not correct.
This help text has been there for ages, it was duplicated from previous help text from usb hid implementation, but you are right that the wording might confuse someone. Will change it eventually, thanks.
Quoted 4 lines
2. it defaults to 'y'. When you have input deselected, this causes the kernel to fail to link:
Fix the second problem by making it depend on INPUT.Signed-off-by Russell King <[email protected]>
Applied, thanks.
Signature · 2 lines
Jiri Kosina SUSE Labs