linux-kernel - Re: [PATCH] usb: typec: fix potential NULL dereference

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: <ZED6KBl6HNT8D0ae@kuha.fi.intel.com>
Date:   Thu, 20 Apr 2023 11:39:04 +0300
From:   Heikki Krogerus <heikki.krogerus@...ux.intel.com>
To:     Maxim Korotkov <korotkov.maxim.s@...il.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
        lvc-project@...uxtesting.org
Subject: Re: [PATCH] usb: typec: fix potential NULL dereference

On Tue, Apr 18, 2023 at 09:56:51AM +0300, Maxim Korotkov wrote:
> On 18.04.2023 09:16, Greg Kroah-Hartman wrote:
> > On Mon, Apr 17, 2023 at 10:50:03PM +0300, Maxim Korotkov wrote:
> > > The pointer 'adev' was being dereferenced before being checked for NULL
> > > in the 'type_alt mode_enter()' and 'type_alt mode_exit()' functions.
> > > Although this is a hypothetical issue, it's better to move the pointer
> > > assignment after the NULL check to avoid any potential problems.
> > > 
> > > Found by Linux Verification Center with Svace static analyzer.
> > > 
> > > Fixes: 8a37d87d72f0 ("usb: typec: Bus type for alternate modes")
> > > Signed-off-by: Maxim Korotkov <korotkov.maxim.s@...il.com>
> > > ---
> > >   drivers/usb/typec/bus.c | 13 +++++++++----
> > >   1 file changed, 9 insertions(+), 4 deletions(-)
> > > 
> > > diff --git a/drivers/usb/typec/bus.c b/drivers/usb/typec/bus.c
> > > index 098f0efaa58d..ae0aca8f33db 100644
> > > --- a/drivers/usb/typec/bus.c
> > > +++ b/drivers/usb/typec/bus.c
> > > @@ -125,13 +125,16 @@ EXPORT_SYMBOL_GPL(typec_altmode_notify);
> > >    */
> > >   int typec_altmode_enter(struct typec_altmode *adev, u32 *vdo)
> > >   {
> > > -	struct altmode *partner = to_altmode(adev)->partner;
> > > -	struct typec_altmode *pdev = &partner->adev;
> > > +	struct altmode *partner;
> > > +	struct typec_altmode *pdev;
> > >   	int ret;
> > >   	if (!adev || adev->active)
> > >   		return 0;
> > > +	partner = to_altmode(adev)->partner;
> > > +	pdev = &partner->adev;
> > 
> > As you point out, the original code is still fine here, we check before
> > we actually use these values.
> > 
> > Also, can adev every actually be NULL?  In looking at the code paths, I
> > can't see how that could happen.
> > 
> > thanks,
> > 
> > greg k-h
> 
> I agree that the adev will most likely never be NULL, but usually this
> pointer is checked before usage (for example in typec_altmode_notify() or
> typec_altmode_vdm()). It is a little odd that in these functions it utilized
> before check. Is it just extra check that can be removed?

Please go ahead and remove it.

thanks,

-- 
heikki