Re: [PATCH] PCI: fix ptr_ret.cocci warnings

From: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
To: Thierry Reding <thierry.reding@gmail.com>
Cc: kbuild test robot <fengguang.wu@intel.com>,
	Shawn Lin <shawn.lin@rock-chips.com>,
	kbuild-all@01.org, linux-kernel@vger.kernel.org,
	Bjorn Helgaas <helgaas@kernel.org>,
	Jonathan Hunter <jonathanh@nvidia.com>,
	linux-tegra@vger.kernel.org, linux-pci@vger.kernel.org
Subject: Re: [PATCH] PCI: fix ptr_ret.cocci warnings
Date: Tue, 18 Sep 2018 17:30:50 +0100	[thread overview]
Message-ID: <20180918163050.GC31440@e107981-ln.cambridge.arm.com> (raw)
In-Reply-To: <20180918144737.GA30386@ulmo>

On Tue, Sep 18, 2018 at 04:47:37PM +0200, Thierry Reding wrote:
> On Tue, Sep 18, 2018 at 03:18:38PM +0100, Lorenzo Pieralisi wrote:
> > On Thu, Jul 12, 2018 at 07:30:41AM +0800, kbuild test robot wrote:
> > > From: kbuild test robot <fengguang.wu@intel.com>
> > > 
> > > drivers/pci/controller/pci-tegra.c:1132:1-3: WARNING: PTR_ERR_OR_ZERO can be used
> > > 
> > > 
> > >  Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR
> > > 
> > > Generated by: scripts/coccinelle/api/ptr_ret.cocci
> > > 
> > > Fixes: 6e0832fa432e ("PCI: Collect all native drivers under drivers/pci/controller/")
> > > CC: Shawn Lin <shawn.lin@rock-chips.com>
> > > Signed-off-by: kbuild test robot <fengguang.wu@intel.com>
> > > ---
> > > 
> > >  pci-tegra.c |    5 +----
> > >  1 file changed, 1 insertion(+), 4 deletions(-)
> > 
> > Thierry,
> > 
> > are you OK with this change ? I will remove the Fixes: tag since
> > this does not fix anything AFAICS.
> 
> This has been proposed several times in the past and each time Bjorn and
> I have agreed that we'd rather not merge that change. I think the
> original is clearer and allows the code to be more easily extended.
> 
> I don't know if there's a way to "whitelist" certain drivers or sub-
> systems where these kinds of changes are known not to be desired. Or
> perhaps there's some way to trick coccinelle into not recognizing this
> particular instance.
> 
> Then again, it seems wrong to have to work around over-ambitious
> semantic patches...

I agree with you and Bjorn, I will drop the patches but you have a
point. Certainly these patches aren't fixing anything unless I am
missing something obvious.

Lorenzo