Re: [PATCH 1/1] KVM: s390: Fix memory corruption by not reinjecting CK machine checks

On Thu, Aug 06, 2026 at 02:52:41PM +0200, Christian Borntraeger wrote:
> Channel-subsystem damage machine checks are for the host channel
> subsystem. The guest channel subsystem is emulated in the userspace VMM.
> There is no point in forwarding such machine checks into the guest.
> 
> This also simplifies the machine check reinjection and avoids kfree of a
> stack variable as reported by sashiko.  There might be still machine
> checks that have the ck bit set with another bit (like instruction
> damage), mask out the CK bit in s390_backup_mcck_info(), like the CP and
> ED bits already are.
> 
> Fixes: 4d62fcc0b692 ("KVM: s390: Inject machine check into the guest")
> Cc: [email protected]
> Signed-off-by: Christian Borntraeger <[email protected]>
> ---
>  arch/s390/include/asm/nmi.h |  1 +
>  arch/s390/kernel/nmi.c      |  4 ++--
>  arch/s390/kvm/interrupt.c   | 24 ++++++++----------------
>  3 files changed, 11 insertions(+), 18 deletions(-)
> 
> diff --git a/arch/s390/include/asm/nmi.h b/arch/s390/include/asm/nmi.h
> index 6454c1531854..dd26c20bd231 100644
> --- a/arch/s390/include/asm/nmi.h
> +++ b/arch/s390/include/asm/nmi.h
> @@ -22,6 +22,7 @@
>  #define MCCK_CODE_SYSTEM_DAMAGE		BIT(63)
>  #define MCCK_CODE_EXT_DAMAGE		BIT(63 - 5)
>  #define MCCK_CODE_CP			BIT(63 - 9)
> +#define MCCK_CODE_CK			BIT(63 - 11)
>  #define MCCK_CODE_STG_ERROR		BIT(63 - 16)
>  #define MCCK_CODE_STG_KEY_ERROR		BIT(63 - 18)
>  #define MCCK_CODE_STG_DEGRAD		BIT(63 - 19)
> diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
> index e17a59d4d5a4..652b98795243 100644
> --- a/arch/s390/kernel/nmi.c
> +++ b/arch/s390/kernel/nmi.c
> @@ -345,7 +345,7 @@ static void notrace s390_backup_mcck_info(struct pt_regs *regs)
>  	sie_page = container_of(sie_block, struct sie_page, sie_block);
>  	mcck_backup = &sie_page->mcck_info;
>  	mcck_backup->mcic = get_lowcore()->mcck_interruption_code &
> -				~(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE);
> +			~(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE | MCCK_CODE_CK);
>  	mcck_backup->ext_damage_code = get_lowcore()->external_damage_code;
>  	mcck_backup->failing_storage_address = get_lowcore()->failing_storage_address;
>  }
> @@ -357,7 +357,7 @@ NOKPROBE_SYMBOL(s390_backup_mcck_info);
>  #define ED_STP_ISLAND	6	/* External damage STP island check */
>  #define ED_STP_SYNC	7	/* External damage STP sync check */
>  
> -#define MCCK_CODE_NO_GUEST	(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE)
> +#define MCCK_CODE_NO_GUEST	(MCCK_CODE_CP | MCCK_CODE_EXT_DAMAGE | MCCK_CODE_CK)

So what you proposed in the cover-letter is to change the above to:

	mcck_backup->mcic = get_lowcore()->mcck_interruption_code & ~MCCK_CODE_NO_GUEST;

Right? Yes, pleae do that! And this can also go via kvm.

Acked-by: Heiko Carstens <[email protected]>