Re: [PATCH v2 1/2] cmdline_parse: Convert no- prefix into =no for OPT_CUSTOM

[Don Slutz <dslutz@verizon.com>]

On 08/01/14 02:57, Jan Beulich wrote:
>>>> On 31.07.14 at 02:59, <dslutz@verizon.com> wrote:
>> On 07/30/14 03:58, Jan Beulich wrote:
>>> --- a/xen/common/kernel.c
>>> +++ b/xen/common/kernel.c
>>> @@ -135,6 +135,13 @@ void __init cmdline_parse(const char *cm
>>>                        parse_size_and_unit(optval, NULL));
>>>                    break;
>>>                case OPT_CUSTOM:
>>> +                if ( !bool_assert )
>>> +                {
>>> +                    if ( *optval )
>>> +                        break;
>>> +                    safe_strcpy(opt, "no");
>>> +                    optval = opt;
>>> +                }
>>>                    ((void (*)(const char *))param->var)(optval);
>>>                    break;
>>>                default:
>>>
>>>
>> Without the log message, I think it might help to include the part about
>> ignoring bogus "no-<name>=<value>"... aka stacked inversions in the
>> commit message.  This is because boolean options do still support stacked
>> inversions, which is valid under the statement "Explicitly specifying any
>> value other than those listed above is undefined, as is stacking a `no-`
>> prefix with an explicit value."
> I added:
>
> Obviously that'll work only when no
> other argument was specified for the option. Command line settings of
> the form "no-<name>=<value>" will now be ignored as ambiguous (rather
> than being interpreted as "<name>=<value>", i.e. ignoring the "no-"
> prefix).
>
> Jan
>

Looks good to me.
     -Don Slutz