This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Add new built-in: __builtin_unreachable()


On Mon, Jun 8, 2009 at 3:34 PM, David Daney<ddaney@caviumnetworks.com> wrote:
> Richard Guenther wrote:
>>
>> On Mon, Jun 8, 2009 at 2:34 PM, David Daney<ddaney@caviumnetworks.com>
>> wrote:
>>>
>>> Ian Lance Taylor wrote:
>>>>
>>>> David Daney <ddaney@caviumnetworks.com> writes:
>>>>
>>>>> +@deftypefn {Built-in Function} void __builtin_unreachable (void)
>>>>> +This function is used to indicate to the compiler that control flow
>>>>> +will never reach the point of the @code{__builtin_unreachable}. ?If
>>>>> +control flow does reach @code{__builtin_unreachable}, program behavior
>>>>> +is undefined. ?The only valid use of this builtin is immediately
>>>>> +following an @code{asm} statement that either never exits or transfers
>>>>> +control elsewhere never returning.
>>>>
>>>> I have wanted __builtin_unreachable for a while, but never got around to
>>>> doing anything. ?But I only want it if it works in general, not just
>>>> after an asm statement. ?I think it should mean "if control flow reaches
>>>> this point, the program is undefined." ?Ideally there should be a
>>>> command line option to control it, so that it is possible to compile
>>>> __builtin_unreachable as expanding to abort, or something like that,
>>>> when debugging.
>>>>
>>> Ok, then how about this version? ?All my original commentary still
>>> applies:
>>> http://gcc.gnu.org/ml/gcc-patches/2009-06/msg00317.html
>>>
>>> I have added a new command line flag (-funreachable-traps) that causes
>>> __builtin_unreachable() to be expanded exactly as if it were a
>>> __builtin_trap().
>>
>> I wouldn't add a new flag to do this. ?In fact preprocessor wrapping
>> around the two builtins can do this. ?Or was there a particular reason?
>>
>
> My reason was that I thought iant requested it. ?Although
> '-D__builtin_unreachable=__builtin_trap' might accomplish the same thing
> without polluting the flag namespace.

Ian may want to clarify here.  I have my don't-add-new-parameters hat on,
so without a compelling reason I would not add it.

Richard.

> David Daney
>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]