volatile asm fix

[Richard Henderson]

On Fri, Apr 09, 1999 at 01:21:05AM -0600, Jeffrey A Law wrote:
> cse and reload_cse will incorrectly optimize across volatile asms.  tsk
> tsk.  We have to assume that any/all machine state is changed by a volatile
> asms.

We do?  All I see in the way of documentation for volatile asms is

]    You can prevent an `asm' instruction from being deleted, moved
] significantly, or combined, by writing the keyword `volatile' after the
] `asm'.

which is not nearly the same thing.  You're thinking of unspec_volatile.

What was the example that lead you to this conclusion?


r~