Re: [GIT PULL] Second batch of KVM changes for Linux 5.6-rc4 (or rc5)

[Paolo Bonzini <pbonzini@redhat.com>]

On 01/03/20 22:33, Linus Torvalds wrote:
> On Sun, Mar 1, 2020 at 1:03 PM Paolo Bonzini <pbonzini@redhat.com> wrote:
>>
>> Paolo Bonzini (4):
>>       KVM: allow disabling -Werror
> 
> Honestly, this is just badly done.
> 
> You've basically made it enable -Werror only for very random
> configurations - and apparently the one you test.
> Doing things like COMPILE_TEST disables it, but so does not having
> EXPERT enabled.

Yes, I took this from the i915 Kconfig.  It's temporary, in 5.7 I am
planning to get it to just !KASAN, but for 5.6 I wanted to avoid more
breakage so I added the other restrictions.  The difference between
x86-64 and i386 is really just the frame size warnings, which Christoph
triggered because of a higher CONFIG_NR_CPUS.

(BTW, perhaps it makes sense for Sparse to have something like __nostack
for structs that contain potentially large arrays).

> I've merged this, but I wonder why you couldn't just do what I
> suggested originally?  Seriously, if you script your build tests,
> and don't even look at the results, then you might as well use
> 
>    make KCFLAGS=-Werror

I did that and I'm also adding W=1; and I threw in a smaller than
default frame size warning option too because I don't want cpumasks on
the stack anyway.  However, that wouldn't help contributors.  I'm okay
if I get W=1 or frame size warnings from patches from other
contributors, but I think it's a disservice to them that they have to
set KCFLAGS in order to avoid warnings.

> the "now it causes problems for
> random compiler versions" is a real issue again - but at least it
> wouldn't be a random kernel subsystem that happens to trigger it, it
> would be a _generic_ issue, and we'd have everybody involved when a
> compiler change introduces a new warning.

Yes, and GCC prereleases are tested with Linux, for example by doing
full Rawhide rebuilds.  If we started using -Werror by default
(including allyesconfig), they would probably report warnings early.
Same for clang.

I hope that Linux can have -Werror everywhere, or at least a
CONFIG_WERROR option that does it even if it defaults to n for a release
or more.  But I don't think we can get there without first seeing what
issues pop up in a few subsystems or arches---even before considering
new compilers---so I decided I would just try.

Paolo

> Adding the powerpc people, since they have more history with their
> somewhat less hacky one. Except that one automatically gets disabled
> by "make allmodconfig" and friends, which is also kind of pointless.

> Michael, what tends to be the triggers for people using
> PPC_DISABLE_WERROR? Do you have reports for it? Could we have a
> _generic_ option that just gets enabled by default, except it gets
> disabled by _known_ issues (like KASAN).
> 
> Being disabled for "make allmodconfig" is kind of against one of the
> _points_ of "the build should be warning-free".