On 23/10/2023 17:00, Julien Grall wrote:
On 23/10/2023 15:51, Nicola Vetrini wrote:
Hi,
Hi Nicola,
while taking care of some patches regarding MISRA C Rule 2.1 (code
shouldn't be unreachable), I
came across this function:
long subarch_do_domctl(struct xen_domctl *domctl, struct domain *d,
XEN_GUEST_HANDLE_PARAM(xen_domctl_t) u_domctl)
{
switch ( domctl->cmd )
{
case XEN_DOMCTL_set_address_size:
switch ( domctl->u.address_size.size )
{
case 32:
if ( !cpu_has_el1_32 )
return -EINVAL;
/* SVE is not supported for 32 bit domain */
if ( is_sve_domain(d) )
return -EINVAL;
return switch_mode(d, DOMAIN_32BIT);
case 64:
return switch_mode(d, DOMAIN_64BIT);
default:
return -EINVAL;
}
break;
default:
return -ENOSYS;
}
}
here the break after the innermost switch is clearly unreachable, but
it's also guarding a possible fallthrough.
I can see a couple of solutions to this:
- mark the part after the switch unreachable;
- introduce a variable 'long rc' to store the return value, and
consequently rework the control flow of all the switches
(e.g. rc = -EINVAL and similar);
- remove the break, but I consider this a risky move, unless -ENOSYS
would be an ok value to be returned if some case
from the switch above does not have a return statement.
- move the nested switch in a separate function, so the code in
subarch_do_domctl() can be replaced with:
return set_address_size(...);
What would be the preferred way of addressing this violation?
I would actually prefer the 4th option I suggested.
Cheers,
Would you mind sending the patch yourself?
--
Nicola Vetrini, BSc
Software Engineer, BUGSENG srl (https://bugseng.com)