HID: usbhid: Fix warning caused by 0-length input reports
authorAlan Stern <stern@rowland.harvard.edu>
Wed, 1 Sep 2021 16:36:00 +0000 (12:36 -0400)
committerJiri Kosina <jkosina@suse.cz>
Wed, 1 Sep 2021 18:52:31 +0000 (20:52 +0200)
commit0a824efdb724e07574bafcd2c2486b2a3de35ff6
tree1b0db6fa8d983cb67b797336139cf8125e9f10b2
parent5049307d37a760e304ad191c5dc7c6851266d2f8
HID: usbhid: Fix warning caused by 0-length input reports

Syzbot found a warning caused by hid_submit_ctrl() submitting a
control request to transfer a 0-length input report:

usb 1-1: BOGUS control dir, pipe 80000280 doesn't match bRequestType a1

(The warning message is a little difficult to understand.  It means
that the control request claims to be for an IN transfer but this
contradicts the USB spec, which requires 0-length control transfers
always to be in the OUT direction.)

Now, a zero-length report isn't good for anything and there's no
reason for a device to have one, but the fuzzer likes to pick out
these weird edge cases.  In the future, perhaps we will decide to
reject 0-length reports at probe time.  For now, the simplest approach
for avoiding these warnings is to pretend that the report actually has
length 1.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-and-tested-by: syzbot+9b57a46bf1801ce2a2ca@syzkaller.appspotmail.com
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Tested-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
drivers/hid/usbhid/hid-core.c