[ovs-dev] [PATCH 10/11] tunnel: Geneve TLV handling support for OpenFlow.

[Jesse Gross]

On Wed, Jun 24, 2015 at 7:09 PM, Ben Pfaff <blp at nicira.com> wrote:
> On Wed, Jun 24, 2015 at 06:34:05PM -0700, Jesse Gross wrote:
>> On Wed, Jun 24, 2015 at 3:57 PM, Ben Pfaff <blp at nicira.com> wrote:
>> > On Fri, Jun 19, 2015 at 04:13:24PM -0700, Jesse Gross wrote:
>> > I like the implementation approach used in ULONG_FOR_EACH_1 better than
>> > the one in PRESENT_OPT_FOR_EACH, since the user doesn't have to provide
>> > a scratch variable.  Also it protects the macro arguments better.
>> > Actually ULONG_FOR_EACH_1 could be made to work with 32- or 64-bit maps
>> > just by changing "unsigned long" to "uint64_t" in its definition; maybe
>> > we should.
>>
>> I looked at converting ULONG_FOR_EACH_1 but decided against it since
>> it would introduce an additional branch (to decide whether to use
>> __builtin_ctz or __builtin_ctz_ll)
>
> Really?  It shouldn't, "__builtin_constant_p(n <= UINT32_MAX) && n <=
> UINT32_MAX" should be a compile-time constant.\

You're right, it doesn't. I guess I didn't have enough faith that the
compiler would be able to carry over the max value across an
assignment to a larger variable but I checked and it was fine. I sent
a patch to do it this way instead.