[PATCH] Fix ia64 .copy_state/.body directive emission (PR target/32338, take 2)

Jim Wilson wilson@specifix.com
Wed Sep 12 21:51:00 GMT 2007


On Wed, 2007-09-12 at 12:35 -0400, Jakub Jelinek wrote:
> Ok, I have bootstrapped/regtested the following patch.

Thanks.  This is OK to install.

> The scheduler moves the sp restore because it sees no harm in doing so,

Ah, right.  I'm a little rusty here.

> doesn't do DCE on its own.  But even if the scheduler eliminated the second,
> after the move redundant restore, would unwind info be valid?

If the unwind directives are in the right place, I don't see why not.
However, it doesn't seem worthwhile to worry about this, since the only
issue here is that the scheduler was smarter than I was.

> BTW, could you please also review PR32337 fix?  I have
> bootstrapped/regtested that fix today as well.

Yes, this is next on my list.  I will get to this today.
-- 
Jim Wilson, GNU Tools Support, http://www.specifix.com




More information about the Gcc-patches mailing list