linux-kernel - Re: [PATCH 0/3] usb: gadget: Add a prefix to log messages

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]

Message-ID: <2026012744-disfigure-plural-5699@gregkh>
Date: Tue, 27 Jan 2026 15:22:50 +0100
From: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To: Alan Stern <stern@...land.harvard.edu>
Cc: Jisheng Zhang <jszhang@...nel.org>, linux-usb@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/3] usb: gadget: Add a prefix to log messages

On Fri, Jan 23, 2026 at 11:18:29AM -0500, Alan Stern wrote:
> On Fri, Jan 23, 2026 at 03:52:11PM +0100, Greg Kroah-Hartman wrote:
> > On Fri, Jan 23, 2026 at 10:22:18PM +0800, Jisheng Zhang wrote:
> > > I met some log as below:
> > > [  581.262476] read descriptors
> > > [  581.265558] bcdVersion must be 0x0100, stored in Little Endian order...
> > > 
> > > To be honest, if I'm not familiar with the usb gadget framework, I dunno
> > > which component is complaining. Add a prefix to log messages, so the
> > > above messages will be look as below:
> > > 
> > > [  581.262476] usb_f_fs: read descriptors
> > > [  581.265558] usb_f_fs: bcdVersion must be 0x0100, stored in Little Endian order...
> > > 
> > > Then solve similar issues for f_mass_storage and f_tcm.
> > 
> > These should all be using the dev_info() and like calls, not pr_*(), and
> > if that happens, then you will get the correct prefix.  Can you make
> > that change instead?
> 
> Unfortunately, the composite gadget driver doesn't create sub-devices of 
> the gadget for the various function drivers to bind to.  The only device 
> available is the gadget itself, and its driver's name is just 
> "composite".  Therefore dev_info() and others will not show the 
> particular function driver, which is what Jisheng wants to see.
> 
> Also, some of the calls affected by these changes occur at module 
> initialization time, before anything has been bound.  They don't yet 
> have any device to refer to.

For things that happen before a device, sure, pr_() is ok (but really
should be quiet unless it's for debugging stuff.)  The other ones should
have or get a device pointer eventually.  Using dev_info() AND the
function driver name is the best way forward here, that way you know
exactly what device it is referring to.

thanks,

greg k-h