Re: [XEN PATCH 0/7] xen/arm: address violations of MISRA C:2012 Rule 16.3

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 20/12/2023 11:03 am, Federico Serafini wrote:
> This patch series addresses violations of MISRA C:2012 Rule 16.3 on the Arm
> code. No fucntional changes are introduced.
>
> Federico Serafini (7):
>   xen/arm: gic-v3: address violations of MISRA C:2012 Rule 16.3
>   xen/arm: traps: address violations of MISRA C:2012 Rule 16.3
>   xen/arm: guest_walk: address violations of MISRA C:2012 Rule 16.3
>   xen/arm: mem_access: address violations of MISRA C:2012 Rule 16.3
>   xen/arm: v{cp,sys}reg: address violations of MISRA C:2012 Rule 16.3
>   xen/arm: mmu: address a violations of MISRA C:2012 Rule 16.3
>   xen/arm: smmu-v3: address violations of MISRA C:2012 Rule 16.3
>
>  xen/arch/arm/arm64/vsysreg.c          |  4 ++--
>  xen/arch/arm/gic-v3.c                 | 30 +++++++++++++++++++++++++++
>  xen/arch/arm/guest_walk.c             |  4 ++++
>  xen/arch/arm/mem_access.c             | 12 +++++------
>  xen/arch/arm/mmu/p2m.c                |  1 +
>  xen/arch/arm/traps.c                  | 18 ++++++++++++----
>  xen/arch/arm/vcpreg.c                 |  4 ++--
>  xen/drivers/passthrough/arm/smmu-v3.c |  2 ++
>  8 files changed, 61 insertions(+), 14 deletions(-)
>

Just a couple of notes on style.  This isn't a request to change
anything in this series, particularly as most is already committed, but
bear it in mind for what I expect will be similar patches in other areas.

We explicitly permit tabulation when it aids readibility, so patch 2
could have been written:

        switch ( hypercall_args[*nr] ) {
        case 5: HYPERCALL_ARG5(regs) = 0xDEADBEEFU; fallthrough;
        case 4: HYPERCALL_ARG4(regs) = 0xDEADBEEFU; fallthrough;
        case 3: HYPERCALL_ARG3(regs) = 0xDEADBEEFU; fallthrough;
        case 2: HYPERCALL_ARG2(regs) = 0xDEADBEEFU; fallthrough;
        case 1: /* Don't clobber x0/r0 -- it's the return value */
        case 0: /* -ENOSYS case */
            break;
        default: BUG();
        }

(give or take the brace placement other style issue)  We also have cases
where a break before a new case statement is preferred, i.e.:

        ...
        break;

    case ...:

This is to prevent larger switch statements from being a straight wall
of text.

If in doubt, match the style around it.  Please don't de-tabulate
examples which are already tabulated.  (i.e. don't de-tabulate the x86
versions of patch 2.)

~Andrew