This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [dataflow]: PATCH COMMITTED
- From: Andrew Pinski <pinskia at gmail dot com>
- To: Kenneth Zadeck <zadeck at naturalbridge dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, "Berlin, Daniel" <dberlin at dberlin dot org>, Seongbae Park <seongbae dot park at gmail dot com>, richard at codesourcery dot com, Ian Lance Taylor <iant at google dot com>
- Date: Wed, 15 Nov 2006 08:29:16 -0800
- Subject: Re: [dataflow]: PATCH COMMITTED
- References: <455922F5.1080907@naturalbridge.com> <87fyckyf2h.fsf@talisman.home> <455B382B.4000007@naturalbridge.com>
On Wed, 2006-11-15 at 10:54 -0500, Kenneth Zadeck wrote:
>
> The conversation was made on or right before sept 21. Maybe danny can
> find the irc log.
I can remember part of the conversation. The issue is that a plain
unspec without a set is always un-deletable instruction.
So:
(insn (unspec (xxxx) )
Should not be deleted even if it does not set anything.
In the ia64, it is used like use but use are useless after reload or at
least that is what the comment says:
;; As USE insns aren't meaningful after reload, this is used instead
;; to prevent deleting instructions setting registers for EH handling
(define_insn "prologue_use"
[(unspec:DI [(match_operand:DI 0 "register_operand" "")]
UNSPEC_PROLOGUE_USE)]
Thanks,
Andrew Pinski