linux-kernel - Re: [PATCH] misc: powermate: update the reference count of the usb interface structure

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] [day] [month] [year] [list]

Date:   Sat, 24 Jul 2021 21:35:19 -0700
From:   Dmitry Torokhov <dmitry.torokhov@...il.com>
To:     Salah Triki <salah.triki@...il.com>
Cc:     gregkh@...uxfoundation.org, linux-input@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] misc: powermate: update the reference count of the usb
 interface structure

On Sun, Jul 25, 2021 at 12:56:37AM +0100, Salah Triki wrote:
> Hi,
> 
> On Sat, Jul 24, 2021 at 04:08:02PM -0700, Dmitry Torokhov wrote:
> > 
> > On Sat, Jul 24, 2021 at 10:20:16PM +0100, Salah Triki wrote:
> > > Use usb_get_intf() and usb_put_intf() in order to update the reference
> > > count of the usb interface structure.
> > 
> > This is quite pointless as the driver will be unbound from the interface
> > before interface is deleted.
> 
> From the documentation of usb_get_intf():
> 
> [quote]
> 
> Each live reference to a interface must be refcounted.
> 
> Drivers for USB interfaces should normally record such references in their 
> probe() methods, when they bind to an interface, and release them by calling
> usb_put_intf(), in their disconnect() methods.
> 
> [/quote]

I see. In this case I recommend you submit a patch to the documentation
clarifying this, as taking a reference when binding to an interface
makes sense when dealing with a secondary interface, but not primary
one. See for example drivers/usb/class/cdc-acm.c where it binds to the
data interface, but also locates control interface and bumps up
reference to that.

As you may notice there is only a handful of drivers that use this API,
and as I am looking at them most do not actually need to use it.

Thanks.

-- 
Dmitry