This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH]: Fix PR tree-optimization/21407
Daniel Berlin wrote:
struct foo a;
a.a = 5;
One important thing i forgot to point out is that at the tree level, the
above func (&a) turns into
In C, this optimization actually prevents quite a lot of call clobbering
of structure fields that is unnecessary.
But, in my opinion, it is an invalid optimization -- unless you can show
that the callee isn't going to look at those fields.
I'd like to understand why you think the code you posted is invalid.