Re: [PATCH 4/4] arm64: traps: fix -Woverride-init warnings

[Will Deacon]

On Mon, Oct 26, 2020 at 08:28:45PM +0100, Arnd Bergmann wrote:
> On Mon, Oct 26, 2020 at 6:27 PM Will Deacon <will@xxxxxxxxxx> wrote:
> > On Mon, Oct 26, 2020 at 05:13:30PM +0000, Peter Maydell wrote:
> > > On Mon, 26 Oct 2020 at 16:23, Mark Rutland <mark.rutland@xxxxxxx> wrote:
> > > > On Mon, Oct 26, 2020 at 05:03:31PM +0100, Arnd Bergmann wrote:
> > > > > From: Arnd Bergmann <arnd@xxxxxxxx>
> > > > >
> > > > > There are many warnings in this file when we re-enable the
> > > > > Woverride-init flag:
> > > > >
> > > > > arch/arm64/kernel/traps.c:704:26: warning: initialized field overwritten [-Woverride-init]
> > > > >   704 |  [ESR_ELx_EC_UNKNOWN]  = "Unknown/Uncategorized",
> > > > >       |                          ^~~~~~~~~~~~~~~~~~~~~~~
> > > > > arch/arm64/kernel/traps.c:704:26: note: (near initialization for 'esr_class_str[0]')
> > > > > arch/arm64/kernel/traps.c:705:22: warning: initialized field overwritten [-Woverride-init]
> > > > >   705 |  [ESR_ELx_EC_WFx]  = "WFI/WFE",
> > > > >       |                      ^~~~~~~~~
> > > > >
> > > > > This is harmless since they are only informational strings,
> > > > > but it's easy to change the code to ignore missing initialization
> > > > > and instead warn about possible duplicate initializers.
> > > >
> > > > This has come up before, and IMO the warning is more hindrance than
> > > > helpful, given the prevalance of spurious warnings, and the (again IMO)
> > > > the rework needed to avoid those making the code harder to reason about
> > >
> > > FWIW in QEMU we turn the clang version of this off with
> > > -Wno-initializer-overrides because we agree that the code is
> > > fine and the compiler is being unhelpful in this case. (There's
> > > a reason gcc doesn't put it in -Wall.)
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91688 is a request
> > > for something that would catch bugs without breaking ranged-array
> > > initializer syntax usage, but the gcc devs don't seem to have
> > > responded.
> >
> > Yes, I'm inclined to agree. The code is fine, and "fixing" it just leads to
> > churn and the possible introduction of bugs.
> 
> Ok, shall we just disable it for all of arch/arm64/kernel then?
> The other parts of the kernel that follow the same line of thinking
> are drivers/gpu/drm/amd/ and drivers/ata, for which I already
> just turn them off. The rest of the kernel is mostly clean for
> the warning, or there are actual bugs that it finds.

I'm happy to take a patch disabling it, either for arch/arm64/kernel or
arch/arm64 as a whole. Thanks!

Will