Ok..i think i got it.
The code snippet you have given is clearly used to initialise a doubly linked list.Now the whole thing is defined as a macro.
But as you know macros are quite tricky and nasty stuff...and that is the reason why.
CODE
#define MUL(a,B) a*b
and
#define MUL(a,B) (a)*(B)
can give quite different results,depending upon the arguments passed to it.If you have got your lesson in C,you can easily figure out how?
So sometimes it becomes quite necessary to have parenthesis around your statements.Now in our case the whole macro is a bit complex and thus just as a workaround and to make sure that statements stay together for any kind of argument,they have used do{}while().
Otherwise for normal arguments,the do{}while() is redundant.It is just like making the code bulletproof.If you remove the do{}while() then you will have two statements for a single macro and it is quite obvious that only one will be used.of course they could have done something like this..
CODE
#define INIT_LIST_HEADER(ptr) {
(ptr)->next = (ptr); (ptr)->prev=(ptr);
}
But i think having workarounds is a matter of choice ,the guy who has written that preferred the earlier method.And may be it is more bulletproof.