Hi all,
On Mon, 25 Jan 2021 20:07:56 +0000,
Oliver Upton <oup...@google.com> wrote:
>
> > That means we have two options:
> > (a) define __hyp_panic_string in a different .c file in all pre-5.9
> > branches, or
> > (b) revert the backported patch.
> >
> > The patch was needed in 5.9 and should stay there. It wasn't needed in
> > earlier
> > versions because the symbol was being kept alive by another user. It did
> > "fix"
> > the inline asm semantics, but the problem was never triggered in pre-5.9.
> >
> > Sasha, with this and the GCC bug in mind, would you agree that (b) is the
> > better course of action?
>
> Sasha,
>
> Any chance we can get this patch reverted as David has suggested? It
> was backported to 5.4 LTS in commit 653ae33b030b ("KVM: arm64: Fix
> symbol dependency in __hyp_call_panic_nvhe") and is causing build
> issues with a 4.9.4 vintage of GCC.
Absolutely. This issue has been lingering for some time now, and needs
addressing.
I'm definitely in favour of reverting this patch, although there are
two possible alternatives:
- Cherry-pick 9fd339a45be5 ("arm64: Work around broken GCC 4.9
handling of "S" constraint"), which works around this particular GCC
bug
- Cherry-pick dca5244d2f5b ("compiler.h: Raise minimum version of GCC
to 5.1 for arm64"), which forbids GCC 4.9 as it has been
demonstrated to mis-compile the kernel (and this patch is targeting
stable anyway)
Thanks,
M.
--
Without deviation from the norm, progress is not possible.
