linux-kernel - Re: [PATCH v2] Input: tests: add test to cover all input_grab

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:   Wed, 17 May 2023 19:24:44 +0200
From:   Javier Martinez Canillas <javierm@...hat.com>
To:     Dmitry Torokhov <dmitry.torokhov@...il.com>,
        Dana Elfassy <delfassy@...hat.com>
Cc:     eballetb@...hat.com, linux-input@...r.kernel.org,
        linux-kernel@...r.kernel.org, Dana Elfassy <dangel101@...il.com>
Subject: Re: [PATCH v2] Input: tests: add test to cover all
 input_grab_device() function

Dmitry Torokhov <dmitry.torokhov@...il.com> writes:

Hello Dmitry,

> Hi Dana,
>
> On Wed, May 17, 2023 at 06:31:45PM +0300, Dana Elfassy wrote:
>> Currently input_grab_device() isn't covered by any tests
>> Thus, adding a test to cover the cases:
>> 1. The device is grabbed successfully
>> 2. Trying to grab a device that is already grabbed by another input
>>    handle
>> 
>> Signed-off-by: Dana Elfassy <dangel101@...il.com>
>> Tested-by: Javier Martinez Canillas <javierm@...hat.com>
>> Reviewed-by: Javier Martinez Canillas <javierm@...hat.com>
>> ---
>> 
>> Changes in v2:
>> - Use input_put_device() to decrement the refcount increased by get().
>> - Remove unnecessary struct input_handle test_handle variable.
>
> So this tests something different than what patch description states.
> You are testing that there is no "recursive" grabbing happening (an API
> could be designed to allow the same handle grab device several times).
> This is a good and useful test, but you do want to also use 2nd separate
> handle to see that it gets -EBUSY as well. And ideally we should have

That was my fault since v1 had two different handles but since it wasn't
releasing it, didn't add any value really so I asked Dana to just drop it.

> another test verifying that the 2nd handle can successfully grab the
> device once the first handle releases it.
>

That's the correct approach indeed and would make the test more useful.

-- 
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat