I think it really is general though. In:
The same problem can occur if one output parameter (@var{a}) allows a
register constraint, is updating the parameter value, and references an
address while another output parameter (@var{b}) allows a memory
constraint.
Yep, you're right. I have trimmed this back to just:
The same problem can occur if one output parameter (@var{a}) allows a
register constraint
and another output parameter (@var{b}) allows a memory constraint.
Seems like that's the last of it then. I've updated the web pages
(http://www.LimeGreenSocks.com/gcc/Using-Assembly-Language-with-C.html)
if you want to look at how it turned out.
Unless you have something else, I'm going to start composing the email
for gcc-patches.
Thanks for the help.
dw