Thanks Ben for the explanation,
I confused myself. After running some experiment, I'm all good now.
> Do you know how i can try compile for RISC?
>
> You need a RISC platform. There are many, but none of them is really
> popular in a desktop setting.
>
> I do have an old SPARC machine (SunBlade 100), that belongs to Justin,
> in my office. (I use it to hold up my monitor, so sometimes I say
> that it is my "platform" for OVS development.) You could borrow it if
> you want. It is quite slow.
>
>
I basically want to double-check it by compile OVS in RISC platform and
make check it. Since it is your OVS development machine, could I count on
you for it? ;)
> Because we instead order the fields differently for big-endian and
> little-endian:
>
> typedef struct {
> #ifdef WORDS_BIGENDIAN
> uint16_t hi, lo;
> #else
> uint16_t lo, hi;
> #endif
> } ovs_16aligned_u32;
> > And why do we differentiate big/small endian here for
> > "put_16aligned_be32()"? Isn't be32 already big endian?
>
> Yes, be32 is already big endian, but an ovs_be32 has different values
> as seen on big-endian and little-endian machines (otherwise ntohl()
> would not be necessary).
>
Totally makes sense, I'm good with this patch.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
