Effect of Structure Packing on Alignment

As long as the alignment rules discussed in the previous section, Rules for Alignment of Complex Types, are observed, no alignment errors occur. Structure packing, however, interacts with compiler alignment behavior as follows.

For each member of a structure:

• If the packsize is set equal to or greater than the default alignment, the packsize is ignored.
  
  
• If the packsize is set smaller than the default alignment, the compiler aligns according to the packsize value.

Thus, if the packsize is set to 4, data types having a size of 4, 8, or 16 bytes would be aligned on addresses that are multiples of four. However, there is no guarantee that data types 8 bytes in size (64 bits) would be aligned on addresses that are a multiple of eight. The packsize has no effect on data types outside of a structure.

The following techniques set a packsize, in bytes:

• The command-line option /Zpn sets the packsize to n, in which n can be 1, 2, 4, 8, or 16, and in which 8 is the default.
  
  
• The compiler directive #pragma pack([n]) sets the packsize to n, in which n can be 1, 2, 4, 8, or 16. If n is not specified, #pragma pack resets the packsize to its value at the beginning of compilation: either the value specified by /Zp, or the default, which uses no packsize. The pragma applies only from the point at which it occurs in the source.

  The syntax of the #pragma pack directive supports other features for optionally saving and restoring the current packing state. For more information, see “Pack” in the Preprocessor Reference.

For example, the /Zp1 option sets the packsize to 1, which causes the compiler to use no padding within structures. Consequently, any data type that has an alignment factor larger than 1 may cause an alignment fixup to occur when accessed. To avoid this problem, either turn off structure packing or use the _ _unaligned keyword as described in Alignment Issues on Alpha.