Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Greg KH]

On Thu, Jul 09, 2026 at 02:23:51PM +0800, Xu Yang wrote:
> On Thu, Jul 09, 2026 at 07:33:03AM +0200, Greg KH wrote:
> > On Wed, Jul 08, 2026 at 03:54:03PM -0500, Frank Li wrote:
> > > On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > > > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > > > > From: Xu Yang <xu.yang_2@xxxxxxx>
> > > > >
> > > > > kstrndup() returns NULL on allocation failure. The condition was
> > > > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > > > would treat a successful allocation as an error and return -ENOMEM
> > > > > while leaking the string, and proceed with a NULL pointer on failure.
> > > > >
> > > > > Fix by changing the condition to 'if (!tmp)'.
> > > > >
> > > > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > > > Cc: stable@xxxxxxxxxxxxxxx
> > > > > Assisted-by: Claude:claude-sonnet-4-6
> > > > > Signed-off-by: Xu Yang <xu.yang_2@xxxxxxx>
> > > > > ---
> > > > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > >
> > > > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > index 5d201a2e30e7..e9f2632ce785 100644
> > > > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > > > >  		goto end;						\
> > > > >  									\
> > > > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > > > -	if (tmp) {							\
> > > > > +	if (!tmp) {							\
> > > > >  		ret = -ENOMEM;						\
> > > > >  		goto end;						\
> > > > >  	}								\
> > > > > --
> > > > > 2.34.1
> > > > >
> > > > >
> > > >
> > > > So this has never worked?  How has no one ever noticed that in the
> > > > decade this has been in the kernel?  Does that mean that no one uses
> > > > this file/driver and we can just remove it?
> > > 
> > > I think just no one use sys interface to change name.
> > 
> > So it can just be removed?  Please submit a patch to do that instead.
> > 
> > Ideally this whole file can be removed, does anyone still use it?
> 
> This legacy driver is still built in kernel in multi_v7_defconfig: 
> 
> ~/work/usb$ grep -nr CONFIG_USB_CONFIGFS_F_UAC1_LEGACY arch/*/configs/
> arch/arm/configs/multi_v7_defconfig:939:CONFIG_USB_CONFIGFS_F_UAC1_LEGACY=y

defconfig doesn't mean much.  Perhaps look at some of the Android
configs?

> It seems we can't simply remove it. 

Why?

> BTW, do we have other ways to figure out if it can be removed?

Remove it and see who complains?  :)

thanks,

greg k-h