Re: [RFC PATCH 13/13] cpus: Constify @cpu in SysemuCPUOps::has_work() handler

On 20/8/26 12:47, Philippe Mathieu-Daudé wrote:
> All functions called by target has_work() handler take a const @cpu
> argument. We can now fulfill the comment added in commit c2804566f6
> ("target/arm: do not clear halting reason in has_work helper"),
> qualifying the handler const to denote its idempotency.
> 
> Signed-off-by: Philippe Mathieu-Daudé <[email protected]>
> ---
>   include/hw/core/sysemu-cpu-ops.h | 4 ++--
>   target/loongarch/internals.h     | 2 +-
>   target/riscv/internals.h         | 2 +-
>   target/s390x/s390x-internal.h    | 2 +-
>   target/alpha/cpu.c               | 2 +-
>   target/arm/cpu.c                 | 4 ++--
>   target/avr/cpu.c                 | 5 +++--
>   target/hexagon/cpu.c             | 4 ++--
>   target/hppa/cpu.c                | 2 +-
>   target/i386/cpu.c                | 2 +-
>   target/loongarch/cpu.c           | 4 ++--
>   target/m68k/cpu.c                | 2 +-
>   target/microblaze/cpu.c          | 2 +-
>   target/mips/cpu.c                | 4 ++--
>   target/or1k/cpu.c                | 2 +-
>   target/ppc/cpu_init.c            | 2 +-
>   target/riscv/cpu.c               | 2 +-
>   target/rx/cpu.c                  | 2 +-
>   target/s390x/cpu-system.c        | 2 +-
>   target/sh4/cpu.c                 | 2 +-
>   target/sparc/cpu.c               | 4 ++--
>   target/tricore/cpu.c             | 4 ++--
>   target/xtensa/cpu.c              | 4 ++--
>   23 files changed, 33 insertions(+), 32 deletions(-)


> diff --git a/target/arm/cpu.c b/target/arm/cpu.c
> index 9761bb173c1..ba66359d410 100644
> --- a/target/arm/cpu.c
> +++ b/target/arm/cpu.c
> @@ -141,9 +141,9 @@ int arm_cpu_mmu_index(CPUState *cs, bool ifetch)
>    * CPU_INTERRUPT_*NMI anyway. So we might as well accept NMI here
>    * unconditionally.
>    */
> -static bool arm_cpu_has_work(CPUState *cs)
> +static bool arm_cpu_has_work(const CPUState *cs)
>   {
> -    ARMCPU *cpu = ARM_CPU(cs);
> +    const ARMCPU *cpu = ARM_CPU(cs);
BTW this is why I said "Not sure if this is a good idea so marked as
RFC" in the cover letter. While the left assignment is correct, the
right QOM-generated macros happily cast without checking const-ness.
We can see that as a pre-existing risk.

Still overall enforcing const-ness on has_work() is a long-term win.
Maybe a simple comment here is sufficient to warn.