linux-kernel - Re: [PATCH] HID: get rid of HID_QUIRK_NO_INIT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 6 Mar 2017 13:15:27 +0100 (CET)
From:   Jiri Kosina <jikos@...nel.org>
To:     Benjamin Tissoires <benjamin.tissoires@...hat.com>
cc:     Dmitry Torokhov <dmitry.torokhov@...il.com>,
        David Herrmann <dh.herrmann@...glemail.com>,
        Oliver Neukum <oneukum@...e.de>,
        Jason Gerecke <killertofu@...il.com>,
        linux-input@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-usb@...r.kernel.org
Subject: Re: [PATCH] HID: get rid of HID_QUIRK_NO_INIT_REPORTS

On Thu, 5 Jan 2017, Benjamin Tissoires wrote:

> For case 1, the hiddev documentation provides an ioctl to do the
> init manually. A solution could be to retrieve the requested report
> when EVIOCGUSAGE is called, in the same way hidraw does. I would be
> tempted to not change the behavior and hope that we won't break any
> userspace tool.

I'd like to be applying the HID_QUIRK_NO_INIT_REPORTS removal as soon as 
possible so that it gets exposure in linux-next over the whole development 
cycle.

I am however too conservative to ignore the potential hiddev breakage, I 
am afraid. This has a real potential of breaking systems, and 
administrators having hard time figuring out of happened; essentialy, this 
is userspace-visible behavior change (regression) for which we haven't 
done any long-term depreciation (such as printing a warning "please talk 
to your hiddev driver vendor" in case the driver seems to assume 
initialized reports) at least for a few years.

I think that either doing it at a connect time, or during first 
EVIOCGUSAGE ioctl() call is a must.

Otherwise, I'd be super-happy to finally get rid of this giant PITA.

Thanks!

-- 
Jiri Kosina
SUSE Labs