From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-348345-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 17838 invoked by alias); 24 Feb 2011 21:54:07 -0000
Received: (qmail 17830 invoked by uid 22791); 24 Feb 2011 21:54:06 -0000
X-SWARE-Spam-Status: No, hits=-2.8 required=5.0	tests=ALL_TRUSTED,AWL,BAYES_00
X-Spam-Check-By: sourceware.org
Received: from localhost (HELO gcc.gnu.org) (127.0.0.1)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 24 Feb 2011 21:54:03 +0000
From: "pthaugen at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/47862] Incorrect code for spilling a vector register
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: rtl-optimization
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pthaugen at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Changed-Fields:
Message-ID: <bug-47862-4-8V2vOwCblL@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-47862-4@http.gcc.gnu.org/bugzilla/>
References: <bug-47862-4@http.gcc.gnu.org/bugzilla/>
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Date: Thu, 24 Feb 2011 22:00:00 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
X-SW-Source: 2011-02/txt/msg02811.txt.bz2

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47862

--- Comment #1 from Pat Haugen <pthaugen at gcc dot gnu.org> 2011-02-24 21:54:00 UTC ---
Looks like the bogus spill/restore insns are being inserted via caller-save.c
routines since IRA assigned volatile regs to pseudos which are live across a
call.


190.sched1 dump for the example above (pseudo 352).

(insn 180 177 182 12 (set (reg:V2DF 352 [ vect_var_.59 ])
        (mult:V2DF (reg:V2DF 343 [ vect_var_.53 ])
            (reg:V2DF 350 [ vect_cst_.60 ]))) spill.c:21 766 {*vsx_mulv2df3}
     (nil))

... call expd2()

(insn 185 184 186 12 (set (reg:V2DF 79 2)
        (reg:V2DF 352 [ vect_var_.59 ])) spill.c:21 756 {*vsx_movv2df}
     (expr_list:REG_DEAD (reg:V2DF 352 [ vect_var_.59 ])
        (nil)))


191.ira dump, along with some additional info I dumped in
caller-save:insert_one_insn() to show insn being inserted

      Popping a139(r352,l0)  -- assign reg 43

(vector save) Inserting insn ...
(set (mem/c:DI (plus:DI (reg/f:DI 113 sfp)
            (const_int 144 [0x90])) [9 S8 A64])
    (reg:DI 43 11))

(vector restore) Inserting insn ...
(set (reg:DI 43 11)
    (mem/c:DI (plus:DI (reg/f:DI 113 sfp)
            (const_int 144 [0x90])) [9 S8 A64]))
... before ...
(insn 185 184 186 12 (set (reg:V2DF 79 2)
        (reg:V2DF 43 11 [orig:352 vect_var_.59 ] [352])) spill.c:21 756
{*vsx_movv2df}
     (expr_list:REG_DEAD (reg:V2DF 43 11 [orig:352 vect_var_.59 ] [352])
        (nil)))


---final ira dump

(insn 180 177 182 12 (set (reg:V2DF 43 11 [orig:352 vect_var_.59 ] [352])
        (mult:V2DF (reg:V2DF 103 26 [orig:343 vect_var_.53 ] [343])
            (reg:V2DF 44 12 [orig:350 vect_cst_.60 ] [350]))) spill.c:21 766
{*vsx_mulv2df3}
     (nil))

(insn 323 322 324 12 (set (mem/c:DI (plus:DI (reg/f:DI 1 1)
                (const_int 144 [0x90])) [9 S8 A64])
        (reg:DI 43 11)) spill.c:21 405 {*movdi_internal64}
     (nil))

... call expd2()

(insn 326 184 185 12 (set (reg:DI 43 11)
        (mem/c:DI (plus:DI (reg/f:DI 1 1)
                (const_int 144 [0x90])) [9 S8 A64])) spill.c:21 405
{*movdi_internal64}
     (nil))

(insn 185 326 186 12 (set (reg:V2DF 79 2)
        (reg:V2DF 43 11 [orig:352 vect_var_.59 ] [352])) spill.c:21 756
{*vsx_movv2df}
     (nil))