[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZruGvIf5+UVEcajd@embed-PC.myguest.virtualbox.org>
Date: Tue, 13 Aug 2024 21:45:56 +0530
From: Abhishek Tamboli <abhishektamboli9@...il.com>
To: Greg KH <gregkh@...uxfoundation.org>
Cc: laurent.pinchart@...asonboard.com, dan.scally@...asonboard.com,
skhan@...uxfoundation.org, dan.carpenter@...aro.org,
rbmarliere@...il.com,
linux-kernel-mentees@...ts.linuxfoundation.org,
linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] usb: gadget: uvc: Fix ERR_PTR dereference in uvc_v4l2.c
Hi Greg,
Thank you for the feedback.
On Tue, Aug 13, 2024 at 10:11:48AM +0200, Greg KH wrote:
> On Fri, Aug 02, 2024 at 11:32:47PM +0530, Abhishek Tamboli wrote:
> > Fix potential dereferencing of ERR_PTR() in find_format_by_pix()
> > and uvc_v4l2_enum_format().
> >
> > Fix the following smatch errors:
> >
> > drivers/usb/gadget/function/uvc_v4l2.c:124 find_format_by_pix()
> > error: 'fmtdesc' dereferencing possible ERR_PTR()
> > drivers/usb/gadget/function/uvc_v4l2.c:392 uvc_v4l2_enum_format()
> > error: 'fmtdesc' dereferencing possible ERR_PTR()
> >
> > Signed-off-by: Abhishek Tamboli <abhishektamboli9@...il.com>
> > ---
> > drivers/usb/gadget/function/uvc_v4l2.c | 6 ++++++
> > 1 file changed, 6 insertions(+)
> >
> > diff --git a/drivers/usb/gadget/function/uvc_v4l2.c b/drivers/usb/gadget/function/uvc_v4l2.c
> > index a024aecb76dc..9dd602a742c4 100644
> > --- a/drivers/usb/gadget/function/uvc_v4l2.c
> > +++ b/drivers/usb/gadget/function/uvc_v4l2.c
> > @@ -121,6 +121,9 @@ static struct uvcg_format *find_format_by_pix(struct uvc_device *uvc,
> > list_for_each_entry(format, &uvc->header->formats, entry) {
> > const struct uvc_format_desc *fmtdesc = to_uvc_format(format->fmt);
> >
> > + if (IS_ERR(fmtdesc))
> > + continue;
> > +
> > if (fmtdesc->fcc == pixelformat) {
> > uformat = format->fmt;
> > break;
> > @@ -389,6 +392,9 @@ uvc_v4l2_enum_format(struct file *file, void *fh, struct v4l2_fmtdesc *f)
> > return -EINVAL;
> >
> > fmtdesc = to_uvc_format(uformat);
> > + if (IS_ERR(fmtdesc))
> > + return -EINVAL;
> > +
> > f->pixelformat = fmtdesc->fcc;
>
>You are now only checking 2 of the responses here, not all of them,
I addressed only the errors reported by Smatch.
> which feels odd.
>
> Either fix all calls to this function, or none of them :)
As you suggested, I'll review the remaining calls to this function
and submit an updated patch that covers all cases.
Regards,
Abhishek
Powered by blists - more mailing lists