Re: [PATCH] ubi: wl: Silence uninitialized variable warning

From: Richard Weinberger <richard@nod.at>
To: Nathan Chancellor <natechancellor@gmail.com>
Cc: Boris Brezillon <bbrezillon@kernel.org>,
	Artem Bityutskiy <dedekind1@gmail.com>,
	kernel-janitors@vger.kernel.org,
	Marek Vasut <marek.vasut@gmail.com>,
	linux-mtd@lists.infradead.org,
	Brian Norris <computersforpeace@gmail.com>,
	David Woodhouse <dwmw2@infradead.org>,
	Dan Carpenter <dan.carpenter@oracle.com>
Subject: Re: [PATCH] ubi: wl: Silence uninitialized variable warning
Date: Thu, 28 Feb 2019 10:51:12 +0100	[thread overview]
Message-ID: <4970018.iuHtFo4yVV@blindfold> (raw)
In-Reply-To: <20190228085058.GA5694@archlinux-ryzen>

Am Donnerstag, 28. Februar 2019, 09:50:58 CET schrieb Nathan Chancellor:
> On Thu, Feb 28, 2019 at 09:35:50AM +0100, Richard Weinberger wrote:
> > Am Donnerstag, 28. Februar 2019, 06:35:51 CET schrieb Dan Carpenter:
> > > This condition needs to be fipped around because "err" is uninitialized
> > > when "force" is set.  The Smatch static analysis tool complains and
> > > UBsan will also complain at runtime.
> > > 
> > > Fixes: 663586c0a892 ("ubi: Expose the bitrot interface")
> > > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> 
> Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
> Tested-by: Nathan Chancellor <natechancellor@gmail.com>

Did you really test the code or just compile it?

> This fixes a -Wsometimes-uninitialized warning from Clang:
> 
> drivers/mtd/ubi/wl.c:1514:6: warning: variable 'err' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
>         if (!force) {
>             ^~~~~~
> drivers/mtd/ubi/wl.c:1520:6: note: uninitialized use occurs here
>         if (err == UBI_IO_BITFLIPS || force) {
>             ^~~
> drivers/mtd/ubi/wl.c:1514:2: note: remove the 'if' if its condition is always true
>         if (!force) {
>         ^~~~~~~~~~~~
> drivers/mtd/ubi/wl.c:1478:9: note: initialize the variable 'err' to silence this warning
>         int err;
>                ^
>                 = 0
> 1 warning generated.

How much false positives does this trigger?
Many useful gcc warnings are disabled because they produce too much churn.

> > > ---
> > >  drivers/mtd/ubi/wl.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/mtd/ubi/wl.c b/drivers/mtd/ubi/wl.c
> > > index 40f838d54b0f..2709dc02fc24 100644
> > > --- a/drivers/mtd/ubi/wl.c
> > > +++ b/drivers/mtd/ubi/wl.c
> > > @@ -1517,7 +1517,7 @@ int ubi_bitflip_check(struct ubi_device *ubi, int pnum, int force)
> > >  		mutex_unlock(&ubi->buf_mutex);
> > >  	}
> > >  
> > > -	if (err == UBI_IO_BITFLIPS || force) {
> > > +	if (force || err == UBI_IO_BITFLIPS) {
> > >  		/*
> > >  		 * Okay, bit flip happened, let's figure out what we can do.
> > >  		 */
> > > 
> > 
> > Good catch, Dan!
> > I thought gcc is supposed to find such issues too. :-/
> 
> This isn't the first time GCC hasn't caught something...
> 
> https://lore.kernel.org/lkml/20190221222123.GC6474@magnolia/

Compilers are not perfect. :-)

Thanks,
//richard

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/