> > I agree too. Not to use packing would make the code bigger and all > > compilers have packing, it just a problem of syntax. > > In contrast, for most processors, the code gets bigger with structure > packing! This is because the compiler doesn't know where e.g. (for a > 32-bit processor) an U32 will be so it has to load every byte of it as > a single load and combine the 4 bytes in one register. Compared to one > single load instruction for a correctly aligned U32, this is much > bigger!
Not all 32-bit processors are code-size affected by non-aligned reads and writes. PowerPC's and Intel processors for example use the same instructions for aligned and non-aligned accesses. The accesses may be slower (quite possibly with caching it's negligible) but the program size doesn't change. Bill _______________________________________________ lwip-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/lwip-users
