From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4475 invoked by alias); 29 May 2014 18:06:13 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 4451 invoked by uid 48); 29 May 2014 18:06:10 -0000 From: "sje at gcc dot gnu.org" To: gcc-bugs@gcc.gnu.org Subject: [Bug rtl-optimization/61357] New: Patch for 60969 causes MIPS regressions in register allocation Date: Thu, 29 May 2014 18:06:00 -0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: new X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: rtl-optimization X-Bugzilla-Version: 4.9.0 X-Bugzilla-Keywords: X-Bugzilla-Severity: normal X-Bugzilla-Who: sje 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-Flags: X-Bugzilla-Changed-Fields: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter Message-ID: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/ Auto-Submitted: auto-generated MIME-Version: 1.0 X-SW-Source: 2014-05/txt/msg02501.txt.bz2 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61357 Bug ID: 61357 Summary: Patch for 60969 causes MIPS regressions in register allocation Product: gcc Version: 4.9.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: rtl-optimization Assignee: unassigned at gcc dot gnu.org Reporter: sje at gcc dot gnu.org This patch (GIT hash c5d7f2f65740ad3659ec083daa6af1df3a6473a7) 2014-05-22 Vladimir Makarov PR rtl-optimization/60969 * ira-costs.c (record_reg_classes): Process NO_REGS for matching constraints. Set up mem cost for NO_REGS case. Causes some test suite failures for MIPS including gcc.target/mips/umips-lwp-3.c Here is a modified version of umips-lwp-3.c: void foo (int *r4) { int r5 = r4[511]; int r6 = r4[512]; r4[2] = r5 * r5; { register int r5asm asm ("$5") = r5; register int r6asm asm ("$6") = r6; asm ("#foo" : "=m" (r4[3]) : "d" (r5asm), "d" (r6asm)); } } When compiled with: -S -mgp32 -fpeephole2 -mtune=m14k -mmicromips -O2 it used to generate a lwp (load word paired) instruction. But after this check in it generates two lw instructions instead.