This is the mail archive of the
mailing list for the GCC project.
Re: RFA [reload]: Fix PR other/58545
- From: Jeff Law <law at redhat dot com>
- To: Joern Rennecke <joern dot rennecke at embecosm dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 30 Oct 2013 14:05:57 -0600
- Subject: Re: RFA [reload]: Fix PR other/58545
- Authentication-results: sourceware.org; auth=none
- References: <20131003162313 dot 00ny1vjalss8sskw-nzlynne at webmail dot spamcop dot net> <52577630 dot 1090405 at redhat dot com> <CAMqJFCrOf_BJd8mP7SRuTE9PUv4By7WEb5q_Bd50qMihSKbVBQ at mail dot gmail dot com>
On 10/11/13 08:40, Joern Rennecke wrote:
You're right -- I wasn't aware of Bernd's work to reuse alignment
paddings for real objects. It was added a few months after I'd been
poking at problems in that loop of reload. Let me look at this stuff again.
On 11 October 2013 04:53, Jeff Law <firstname.lastname@example.org> wrote:
> With your change it seems to me that we do a single round of spilling &
caller-save setup, then align the stack, then restart. The net result being
we align the stack a lot more often.
Yes, but AFAICT, that should not result in more space being used,
because assign_stack_local uses ASLK_RECORD_PAD, so
assign_stack_local_1 will record any space
added for alignment purposes with add_frame_space, so it can be used
for a subsequent spill of suitable size.