On Wed, 2024-05-29 at 11:37 +0100, Andrew Cooper wrote:
> {cmci,lmce}_support are written during S3 resume, so cannot live in
> __ro_after_init. Move them back to being __read_mostly, as they were
> originally.
>
> Link: https://gitlab.com/xen-project/xen/-/jobs/6966698361
> Fixes: 19b6e9f9149f ("x86/MCE: optional build of AMD/Intel MCE code")
> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com>
> ---
> CC: Jan Beulich <jbeul...@suse.com>
> CC: Roger Pau Monné <roger....@citrix.com>
> CC: Sergiy Kibrik <sergiy_kib...@epam.com>
> CC: Stefano Stabellini <sstabell...@kernel.org>
> CC: Oleksii Kurochko <oleksii.kuroc...@gmail.com>
>
> We're past feature freeze and this was a silent change in a patch,
> which was
> also untested. A 30s look at mcheck_init() shows clearly that it's
> not a 30s
> job to fix.
> ---
Release-Acked-by: Oleksii Kurochko <oleksii.kuroc...@gmail.com>
~ Oleksii
> xen/arch/x86/cpu/mcheck/mce.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/xen/arch/x86/cpu/mcheck/mce.c
> b/xen/arch/x86/cpu/mcheck/mce.c
> index 1664ca6412ac..32c1b2756b90 100644
> --- a/xen/arch/x86/cpu/mcheck/mce.c
> +++ b/xen/arch/x86/cpu/mcheck/mce.c
> @@ -38,10 +38,10 @@ DEFINE_PER_CPU_READ_MOSTLY(unsigned int,
> nr_mce_banks);
> unsigned int __read_mostly firstbank;
> unsigned int __read_mostly ppin_msr;
> uint8_t __read_mostly cmci_apic_vector;
> -bool __ro_after_init cmci_support;
> +bool __read_mostly cmci_support;
>
> /* If mce_force_broadcast == 1, lmce_support will be disabled
> forcibly. */
> -bool __ro_after_init lmce_support;
> +bool __read_mostly lmce_support;
>
> DEFINE_PER_CPU_READ_MOSTLY(struct mca_banks *, poll_bankmask);
> DEFINE_PER_CPU_READ_MOSTLY(struct mca_banks *, no_cmci_banks);
>
> base-commit: 0840bc5ea114f536a4bdfb2ca095b79f2069aae6
