Re: [PATCH 3.18 104/105] IB/nes: Fix a compiler warning

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Sep 25, 2018 at 04:11:38AM -0700, Joe Perches wrote:
> On Tue, 2018-09-25 at 10:55 +0200, Greg Kroah-Hartman wrote:
> > On Mon, Sep 24, 2018 at 10:39:53PM +0000, Sasha Levin wrote:
> > > On Mon, Sep 24, 2018 at 11:03:25AM -0700, Joe Perches wrote:
> > > > On Mon, 2018-09-24 at 19:59 +0200, Greg Kroah-Hartman wrote:
> > > > > On Mon, Sep 24, 2018 at 09:38:26AM -0700, Joe Perches wrote:
> > > > > > On Mon, 2018-09-24 at 13:34 +0200, Greg Kroah-Hartman wrote:
> > > > > > > 3.18-stable review patch.  If anyone has any objections, please let me know.
> > > > > > 
> > > > > > Why should this sort of change be applied to a stable release?
> > > > > 
> > > > > Originally I was just going to drop this as it's not fixing something.
> > > > > 
> > > > > But it might be, if that macro is used in a if() statement, or something
> > > > > like that, it could be doing something unintended.
> > > > 
> > > > No it couldn't.
> > > > An empty macro is equivalent to a single statement.
> > > > 
> > > > > So I don't feel like auditing all 500+ instances where this is used,
> > > > > it's easier to just accept the patch.
> > > > 
> > > > It's not a bug fix.
> > > 
> > > This question came up a few months ago. Greg suggested that we should be
> > > pulling in warning fixes to get the stable kernels warning-free similar
> > > to upstream.
> > > 
> > > The reasoning behind it was similar to the "no warnings" reasoning of
> > > upstream: there might be real issues hiding in the sea of "harmless"
> > > warnings, so we want to get rid of all of them to catch real issues.
> > 
> > No warnings is great,
> 
> I believe that is not necessarily true.

For me, it is essencial.

As proof of this, I found an actual bug in a patch that added a warning
to the build.  If my scripts hadn't shown that we had gone from 0 to 1
warnings, then I would have missed that.

So I want to keep the stable trees at 0 warnings if at all possible, for
x86-64 at the least.

> Change to a new compiler version and new warnings could be
> added somewhat arbitrarily.

That's true, and is why I am stuck at gcc7 at the moment, as gcc8 does
horrid things to older stable kernels :)

thanks,

greg k-h



[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux