Re: [PATCH RFC V7 4/5] xen, libxc: Request page fault injection via libxc

["Jan Beulich" <JBeulich@suse.com>]

>>> On 27.08.14 at 13:54, <rcojocaru@bitdefender.com> wrote:
> On 08/26/2014 05:13 PM, Jan Beulich wrote:
>>>>> On 13.08.14 at 17:28, <rcojocaru@bitdefender.com> wrote:
>>> +    case XEN_DOMCTL_set_pagefault_info:
>>> +    {
>>> +        ret = -EINVAL;
>>> +
>>> +        if ( is_hvm_domain(d) )
>>> +        {
>>> +            d->arch.hvm_domain.fault_info.address_space =
>>> +                op->u.set_pagefault_info.address_space;
>>> +            d->arch.hvm_domain.fault_info.virtual_address =
>>> +                op->u.set_pagefault_info.virtual_address;
>>> +            d->arch.hvm_domain.fault_info.errcode =
>>> +                op->u.set_pagefault_info.errcode;
>>> +            d->arch.hvm_domain.fault_info.valid = 1;
>>> +
>>> +            ret = 0;
>>> +        }
>>> +    }
>> 
>> Pointless curly braces.
> 
> You're right, of course, but I've written it like that because that
> seems to be the style (even where it is not necessary / no local
> variables are introduced) in
> do_domctl(XEN_GUEST_HANDLE_PARAM(xen_domctl_t) u_domctl).
> 
> Should I break with the coding style for this switch case?

Neither do_domctl() nor x86's arch_do_domctl() really consistently
do like you say. Hence I think rather than continuing the bad habit,
making new additions do better is the right approach. (As to why
I really think this isn't just a cosmetic thing: These braces usually
get placed at the same indentation level as the containing switch
statement's, breaking consistent indentation, potentially leading to
two immediately successive closing braces at the same indentation
level.)

Jan