Re: [PATCH] misc: suppress build warning

[Andrew Morton]

On Thu, 4 Dec 2014 08:30:32 -0800 Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:

> On Thu, Dec 04, 2014 at 03:13:00PM +0000, Prabhakar Lad wrote:
> > On Thu, Dec 4, 2014 at 2:59 PM, Arnd Bergmann <arnd@xxxxxxxx> wrote:
> > > On Thursday 04 December 2014 14:38:30 Lad, Prabhakar wrote:
> > >> this patch fixes following build warning:
> > >>
> > >> drivers/misc/ioc4.c: In function ___ioc4_probe___:
> > >> drivers/misc/ioc4.c:194:16: warning: ___start___ may be used uninitialized in this function [-Wmaybe-uninitialized]
> > >>   period = (end - start) /
> > >>                 ^
> > >> drivers/misc/ioc4.c:148:11: note: ___start___ was declared here
> > >>   uint64_t start, end, period;
> > >>
> > >> Signed-off-by: Lad, Prabhakar <prabhakar.csengg@xxxxxxxxx>
> > >
> > > Please explain why the compiler thinks there is a bug, why you
> > > are sure that there isn't, and why you picked '0' as the
> > > initialization value.
> > >
> > Its a false positive, to suppress the warning '0' was picked.
> 
> Are you _sure_ it's a false positive?  That odd do/while loop looks like
> it might just not ever initialize the start variable, are you sure the
> logic there is correct?
> 

As long as IOC4_CALIBRATE_END is greater than IOC4_CALIBRATE_DISCARD (it is),
`start' is written to.

It would be nice to simplify the code, but I'm not sure how.

And I really dislike this initialize-it-to-zero-to-stop-the-warning
thing which we do all over the place.  The reader doesn't know *why*
it's initialized to zero and the initialization can conceal bugs if we
get a code path which should have written to it but forgot to.  And it
adds unneeded code to vlinux.

I much prefer unintialized_var() which fixes the documentation issue
and doesn't add code.  But Linus and Ingo had a dummy-spit over it.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/