rtx_unchanging_p vs c++ vtable fields

Richard Kenner kenner@vlsi1.ultra.nyu.edu
Sun Dec 9 03:49:00 GMT 2001


    But cause the field is _not_ unchanging.  In the case at hand, if you'd
    like to examine it, is that there are two stores to this field.  At no
    point is it safe to have more than one store to an unchanging memory.

If the field is not marked unchanging, then why is that test saying there is
such a field?

But let's get back to the first point: why does the /u there causing
problems: that should always be the safe direction?



More information about the Gcc-patches mailing list