Re: [PATCH v3 20/27] KVM: s390: arm64: Introduce host definitions

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Macro `__ctxt_sys_reg` returns NULL, breaking KVM_GET_ONE_REG/KVM_SET_ONE_REG for ELR_EL1 and SPSR_EL1
--

commit e22b4e249b651c1cc6aa77134aa91b1e8f9d4063
Author: Steffen Eiden <seiden@linux.ibm.com>
KVM: s390: arm64: Introduce host definitions

This commit adds basic definitions the arm on s390 KVM host requires,
including struct kvm*arch definitions, function prototypes, and various
macros required to run arm64 guests.

> diff --git a/arch/s390/include/asm/kvm_host_arm64.h b/arch/s390/include/asm/kvm_host_arm64.h
> new file mode 100644
> --- /dev/null
> +++ b/arch/s390/include/asm/kvm_host_arm64.h
[ ... ]
> +/* Minimal (=no) vgic definitions */
> +#define KVM_IRQCHIP_NUM_PINS 1
> +#define irqchip_in_kernel(_v) false;
> +
> +#define __ctxt_sys_reg(ctx, reg) NULL

[Severity: High]
Does defining this macro to return NULL break KVM_GET_ONE_REG and
KVM_SET_ONE_REG for ELR_EL1 and SPSR_EL1?

Shared code in virt/kvm/arm64/guest.c uses this macro to resolve the storage
address of core registers ELR_EL1 and SPSR_EL1. Returning NULL causes
core_reg_addr() to return NULL, resulting in the ioctls failing with -EINVAL.

However, copy_core_reg_indices() includes these registers in the list returned
by KVM_GET_REG_LIST because core_reg_size_from_offset() returns a valid size
for them.

Advertising support for registers in KVM_GET_REG_LIST but failing to get or
set them is an API contract violation that can cause userspace VMMs like
QEMU to abort during vCPU initialization.

> +struct kvm_cpu_context {
> +	/*
> +	 * These are just for 32 bit, which we don't have, making them RES0.
> +	 * They are exposed to user space.
> +	 */
> +	u64	spsr_abt;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260529155050.2902245-1-seiden@linux.ibm.com?part=20