On Wed, 13 Dec 2017 20:44:45 +0000 Colin Ian King <colin.k...@canonical.com> wrote:
> On 13/12/17 20:38, Boris Brezillon wrote: > > On Wed, 13 Dec 2017 20:30:04 +0000 > > Colin Ian King <colin.k...@canonical.com> wrote: > > > >> On 13/12/17 20:24, Boris Brezillon wrote: > >>> On Wed, 13 Dec 2017 20:17:43 +0000 > >>> Colin King <colin.k...@canonical.com> wrote: > >>> > >>>> From: Colin Ian King <colin.k...@canonical.com> > >>>> > >>>> The check of len being zero is redundant as it has already been > >>>> sanity checked for this value at the start of the function. Hence > >>>> it is impossible for this test to be true and so the redundant > >>>> code can be removed. > >>> > >>> Nope, it's not the same test, the initial test is > >>> > >>> if (len && !buf) > >> > >> Ah, the current tip from linux-next has: > >> > >> 1912 if (!len || !buf) > >> 1913 return -EINVAL; > >> > >> ..so I guess that's why it got picked up by static analysis. > > > > Hm, that's weird, that's not what I see [1] in linux-next. > > I see my mistake, I fixed the *wrong* function, I'll send a v2. Doh. Yep, just noticed that too. No need to send a patch though.
