[Bug libquadmath/55198] New: [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[Bug libquadmath/55198] New: [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

             Bug #: 55198
           Summary: [4.8 Regression] libquadmath/math/fmaq.c:233:7:
                    internal compiler error
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libquadmath
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: danglin@gcc.gnu.org
              Host: hppa2.0w-hp-hpux11.11
            Target: hppa2.0w-hp-hpux11.11
             Build: hppa2.0w-hp-hpux11.11


libtool: compile:  /test/gnu/gcc/objdir/./gcc/xgcc
-B/test/gnu/gcc/objdir/./gcc/
 -B/opt/gnu/gcc/gcc-4.8/hppa2.0w-hp-hpux11.11/bin/
-B/opt/gnu/gcc/gcc-4.8/hppa2.0w-hp-hpux11.11/lib/ -isystem
/opt/gnu/gcc/gcc-4.8/hppa2.0w-hp-hpux11.11/include
 -isystem /opt/gnu/gcc/gcc-4.8/hppa2.0w-hp-hpux11.11/sys-include
-DHAVE_CONFIG_H
 -I. -I../../../gcc/libquadmath -g -O2 -MT math/fmaq.lo -MD -MP -MF
math/.deps/f
maq.Tpo -c ../../../gcc/libquadmath/math/fmaq.c  -fPIC -DPIC -o
math/.libs/fmaq.o../../../gcc/libquadmath/math/fmaq.c: In function
'fmaq':../../../gcc/libquadmath/math/fmaq.c:233:7: internal compiler error: in
expand_asm_operands, at stmt.c:912
       asm volatile ("" : : "m" (v.value));      
^../../../gcc/libquadmath/math/fmaq.c:233:7: internal compiler error: Aborted

Probably introduced by:

2012-10-31  Tobias Burnus  <burnus@net-b.de>
            Joseph Myers <joseph@codesourcery.com>
            David S. Miller <davem@davemloft.net>
            Ulrich Drepper <drepper@redhat.com>
            Marek Polacek <polacek@redhat.com>:
            Petr Baudis <pasky@suse.cz>

        * math/complex.c (csqrtq): NaN and INF fixes. 
        * math/sqrtq.c (sqrt): NaN, INF and < 0 fixes.
        * math/expm1q.c (expm1q): Changes from GLIBC. Use expq for
        large parameters. Fix errno for boundary conditions.
        * math/finiteq.c (finiteq): Add comment.
        * math/fmaq.c (fmaq): Changes from GLIBC. Fix missing underflows
        and bad results for some subnormal results. Fix sign of inexact
        zero return. Fix sign of exact zero return. 
        Ensure additions are not scheduled after fetestexcept.

-bash-3.2$ ./xgcc -B./ -v
Reading specs from ./specs
COLLECT_GCC=./xgcc
COLLECT_LTO_WRAPPER=./lto-wrapper
Target: hppa2.0w-hp-hpux11.11
Configured with: ../gcc/configure --with-gnu-as --with-as=/opt/gnu/bin/as
--enable-shared --with-local-prefix=/opt/gnu --prefix=/opt/gnu/gcc/gcc-4.8
--with-gmp=/opt/gnu/gcc/gmp --enable-threads=posix --enable-debug=no
--disable-nls --without-cloog --without-ppl
--enable-languages=c,c++,objc,java,ada,obj-c++
Thread model: posix
gcc version 4.8.0 20121103 (experimental) [trunk revision 193124] (GCC)

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|libquadmath                 |middle-end
   Target Milestone|---                         |4.8.0

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-11-04 02:38:37 UTC ---
>Probably introduced by:

Exposed as this is a change in the library and the compiler is crashing with a
valid input that the change introduced.

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[dave.anglin at bell dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #2 from dave.anglin at bell dot net 2012-11-05 00:20:16 UTC ---
On 3-Nov-12, at 10:38 PM, pinskia at gcc dot gnu.org wrote:

> Exposed as this is a change in the library and the compiler is  
> crashing with a
> valid input that the change introduced.


The ICE occurs because we get a TF mode REG instead of a MEM.

--
John David Anglin    dave.anglin@bell.net

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #3 from John David Anglin <danglin at gcc dot gnu.org> 2012-11-22 03:13:32 UTC ---
In expand_expr_real:

(gdb) p debug_rtx (orig_op0)
(mem/c:BLK (plus:SI (reg/f:SI 91 virtual-stack-vars)
        (const_int 48 [0x30])) [6 v+0 S16 A64])
$25 = 10
(gdb) p debug_rtx (op0)     
(mem/c:BLK (plus:SI (reg/f:SI 91 virtual-stack-vars)
        (const_int 48 [0x30])) [6 v+0 S16 A64])
$26 = 10
(gdb) p mode1
$27 = TFmode
(gdb) p must_force_mem
$28 = 1
(gdb) p offset
$29 = (tree) 0x0
(gdb) p mode2
$30 = BLKmode
(gdb) p bitpos
$31 = 0
(gdb) p bitsize
$32 = 128
(gdb) p modifier
$41 = EXPAND_MEMORY
(gdb) p target
$42 = (rtx) 0x0
(gdb) p ext_mode
$43 = TFmode

(gdb) bt
#0 
_Z18expand_expr_real_1P9tree_nodeP7rtx_def12machine_mode15expand_modifierPS2_
(exp=0x7af2d2a0, target=0x0, tmode=VOIDmode, modifier=EXPAND_MEMORY, 
    alt_rtl=0x0) at ../../gcc/gcc/expr.c:10006
#1  0x0045c3b8 in
_Z16expand_expr_realP9tree_nodeP7rtx_def12machine_mode15expand_modifierPS2_
(exp=0x7af2d2a0, target=0x0, tmode=VOIDmode, 
    modifier=EXPAND_MEMORY, alt_rtl=0x0) at ../../gcc/gcc/expr.c:7816
#2  0x0092f538 in
_ZL11expand_exprP9tree_nodeP7rtx_def12machine_mode15expand_modifier
(exp=0x7af2d2a0, target=0x0, mode=VOIDmode, modifier=EXPAND_MEMORY)
    at ../../gcc/gcc/expr.h:444
#3  0x00933588 in _ZL19expand_asm_operandsP9tree_nodeS0_S0_S0_S0_ij (
    string=0x7af6cf78, outputs=0x0, inputs=0x7af2e288, clobbers=0x0, 
    labels=0x0, vol=1, locus=2147483912) at ../../gcc/gcc/stmt.c:890
#4  0x00934698 in _Z15expand_asm_stmtP18gimple_statement_d (stmt=0x7af3c7e0)
    at ../../gcc/gcc/stmt.c:1153
#5  0x0028ad94 in _ZL20expand_gimple_stmt_1P18gimple_statement_d (
    stmt=0x7af3c7e0) at ../../gcc/gcc/cfgexpand.c:2150
#6  0x0028b554 in _ZL18expand_gimple_stmtP18gimple_statement_d (
    stmt=0x7af3c7e0) at ../../gcc/gcc/cfgexpand.c:2305
#7  0x00292b8c in _ZL25expand_gimple_basic_blockP15basic_block_defb (
    bb=0x7af42b00, disable_tail_calls=false) at ../../gcc/gcc/cfgexpand.c:4084
#8  0x00294ecc in gimple_expand_cfg() () at ../../gcc/gcc/cfgexpand.c:4603
#9  0x007ddd44 in _Z16execute_one_passP8opt_pass (
    pass=0x40033b20 <pass_expand>) at ../../gcc/gcc/passes.c:2327
---Type <return> to continue, or q <return> to quit---
#10 0x007de108 in _Z17execute_pass_listP8opt_pass (
    pass=0x40033b20 <pass_expand>) at ../../gcc/gcc/passes.c:2387
#11 0x002d6130 in _ZL15expand_functionP11cgraph_node (node=0x7aed22a0)
    at ../../gcc/gcc/cgraphunit.c:1641
#12 0x002d6718 in _ZL20expand_all_functionsv ()
    at ../../gcc/gcc/cgraphunit.c:1745
#13 0x002d7734 in _Z7compilev () at ../../gcc/gcc/cgraphunit.c:2043
#14 0x002d79ac in finalize_compilation_unit() ()
    at ../../gcc/gcc/cgraphunit.c:2120
#15 0x00079300 in _Z27c_write_global_declarationsv ()
    at ../../gcc/gcc/c/c-decl.c:10120
#16 0x00957404 in _ZL12compile_filev () at ../../gcc/gcc/toplev.c:559
#17 0x0095abbc in _ZL10do_compilev () at ../../gcc/gcc/toplev.c:1881
#18 0x0095aed0 in _Z11toplev_mainiPPc (argc=14, argv=0x7eff052c)
    at ../../gcc/gcc/toplev.c:1957
#19 0x00dc62c4 in main (argc=14, argv=0x7eff052c) at ../../gcc/gcc/main.c:36
(gdb) c
Continuing.

Breakpoint 9,
_Z18expand_expr_real_1P9tree_nodeP7rtx_def12machine_mode15expand_modifierPS2_
(exp=0x7af2d2a0, target=0x0, tmode=VOIDmode, 
    modifier=EXPAND_MEMORY, alt_rtl=0x0) at ../../gcc/gcc/expr.c:10011
10011            if (TREE_CODE (type) == RECORD_TYPE && BYTES_BIG_ENDIAN
(gdb) p debug_rtx (op0)
(reg:TF 767)
$44 = 10

Think the problem is this asm operand is being treated as bitfield.

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #4 from John David Anglin <danglin at gcc dot gnu.org> 2012-11-23 00:39:12 UTC ---
Created attachment 28763
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28763
Testcase

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #5 from John David Anglin <danglin at gcc dot gnu.org> 2012-11-23 00:40:06 UTC ---
Created attachment 28764
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28764
Patch

Testing.

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #6 from John David Anglin <danglin at gcc dot gnu.org> 2012-12-07 14:47:55 UTC ---
Patch is here:
http://gcc.gnu.org/ml/gcc-patches/2012-12/msg00042.html

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |build, ice-on-valid-code,
                   |                            |patch
             Status|UNCONFIRMED                 |NEW
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2012-12/msg00042.htm
                   |                            |l
   Last reconfirmed|2012-11-20 00:00:00         |2012-12-31
     Ever Confirmed|0                           |1

--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-12-31 09:43:03 UTC ---
Confirmed.

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

--- Comment #8 from John David Anglin <danglin at gcc dot gnu.org> 2013-01-02 21:14:22 UTC ---
Author: danglin
Date: Wed Jan  2 21:14:13 2013
New Revision: 194822

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=194822
Log:
    PR middle-end/55198
    * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
    BLKmode objects when EXPAND_MEMORY is specified.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/expr.c

[Bug middle-end/55198] [4.8 Regression] libquadmath/math/fmaq.c:233:7: internal compiler error

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55198

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |jakub at gcc dot gnu.org
         Resolution|                            |FIXED

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-01-02 21:17:01 UTC ---
Assuming fixed.