On Fri, Nov 9, 2012 at 3:40 AM, Ansis Atteka <aatt...@nicira.com> wrote:
> On Fri, Nov 9, 2012 at 3:30 AM, Jesse Gross <je...@nicira.com> wrote:
> > On Thu, Nov 8, 2012 at 5:14 AM, Ansis Atteka <aatt...@nicira.com> wrote:
> >>
> >> diff --git a/datapath/actions.c b/datapath/actions.c
> >> index 8ec692d..317aa4a 100644
> >> --- a/datapath/actions.c
> >> +++ b/datapath/actions.c
> >> +static void set_ipv6_addr(struct sk_buff *skb, u8 l4_proto,
> >> + __be32 addr[4], const __be32 new_addr[4],
> >> + bool recalculate_csum)
> >> +{
> >> + if (recalculate_csum)
> >> + update_ipv6_checksum(skb, l4_proto, addr, new_addr);
> >> +
> >> + skb_clear_rxhash(skb);
> >> + memcpy(addr, new_addr, sizeof(__be32[4]));
> >
> >
> > The sizeof(__be32[4]) seems a little overdone here - you could just use
> > sizeof(addr).
> My understanding is that in C/C++ arrays decay into pointers, if they
> are passed as function arguments. Try:
>
> #include <stdio.h>
> void fn(int a[4]) {
> printf("%lu vs %lu vs %lu!\n", sizeof(a), sizeof(*a), sizeof(int[4]));
> }
> int main() {
> int in[4];
> fn(in);
> }
>
> Output is:
> 8 vs 4 vs 16!
>
> Because sizeof(addr) would be system pointer size, but sizeof(*addr)
> would be size of a single array element.
You're right, sorry about that.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
