Re: [lwip-users] Struct packing/alignment problems

[Timmy Brolin]

Mountifield, Tony wrote:

> Hi All
>
> I don't have a strong opinion one way or the other about removing the packing 
> macros, but I wanted to reply to one point:
>
> > The macros will have to remain in the code for the Ethernet
> > ARP header, if I understand that correctly, so it really
> > isn't worth worrying about.
> >    
> This is no longer correct. Since the sipaddr and dipaddr fields were changed 
> from struct ip_addr to struct ip_addr2, they now only have 16-bit alignment 
> requirements (this is why they were changed) ...
>
> ... so long as compilers don't try to insert padding before and after struct A 
> in struct B in the following example:
>
> struct A {
>         u16_t x[2];
> };
>
> struct B {
>         u16_t p;
>         struct A q;
>         u16_t r;
> };
>
> Hopefully:
>
> offsetof(p) is 0
> offsetof(q.x[0]) is 2
> offsetof(q.x[1]) is 4
> offsetof(r) is 6
> sizeof(struct B) is 8
>
> Cheers,
> Tony
Actually, I think you are on to something here... Someone said that gcc 
for arm puts all structs on 4 byte boundarys in order to use the 
efficient load and store multiple instructions... I have not considered 
this. It would surely break the etharp struct.
Of course the problem can be fixed by simply replacing the ipaddr2 
struct with a u16_t ip[2] array, so it's not that hard to solve.

Timmy