[Bug bootstrap/64102] New: ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[Bug bootstrap/64102] New: ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

            Bug ID: 64102
           Summary: ARM bootstrap fails with segfault with -mapcs in
                    BOOT_CFLAGS
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ktkachov at gcc dot gnu.org

I'm seeing a bootstrap segfault in stage2 on arm-none-linux-gnueabihf when
given BOOT_CFLAGS="-O2 -mapcs":

$BUILD/./gcc/xgcc -B$BUILD/./gcc/ -B$BUILD/armv7l-unknown-linux-gnueabihf/bin/
-B$BUILD/armv7l-unknown-linux-gnueabihf/lib/ -isystem
$BUILD/armv7l-unknown-linux-gnueabihf/include -isystem
$BUILD/armv7l-unknown-linux-gnueabihf/sys-include    -g -O2 -O2  -g -O2
-DIN_GCC    -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem
./include   -fPIC -fno-inline -g -DIN_LIBGCC2 -fbuilding-libgcc
-fno-stack-protector   -fPIC -fno-inline -I. -I. -I../.././gcc -I$SRC/libgcc
-I$SRC/libgcc/. -I$SRC/libgcc/../gcc -I$SRC/libgcc/../include  -DHAVE_CC_TLS 
-o _eprintf.o -MT _eprintf.o -MD -MP -MF _eprintf.dep -DL_eprintf -c
$SRC/libgcc/libgcc2.c -fvisibility=hidden -DHIDE_EXPORTS

$SRC/libgcc/libgcc2.c: In function '__eprintf':
$SRC/libgcc/libgcc2.c:2114:3: internal compiler error: Segmentation fault
   fprintf (stderr, string, expression, line, filename);
   ^
0x69c0cf crash_signal
        $SRC/gcc/toplev.c:359
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.
(kyrtka01@pdtl-cb1)~/builds/gcc-bootstrap-apcs-clean/

This is based on r218090 with configuration:
--with-float=hard --with-cpu=cortex-a15 --with-mode=arm --with-fpu=neon-vfpv4
--enable-languages=c,c++,fortran

Bootstrap without -mapcs works fine.
I don't know if this is a regression, didn't try released compilers.

Can someone please confirm this?

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

--- Comment #1 from ktkachov at gcc dot gnu.org ---
attaching gdb for a better back trace:
Program received signal SIGSEGV, Segmentation fault.
0xbeffe4fc in ?? ()
(gdb) bt
#0  0xbeffe4fc in ?? ()
#1  0x001824b8 in build_function_call_vec(unsigned int, vec<unsigned int,
va_heap, vl_ptr>, tree_node*, vec<tree_node*, va_gc, vl_embed>*,
vec<tree_node*, va_gc, vl_embed>*) ()
    at $SRC/gcc/c/c-typeck.c:2967
#2  0x001a433c in c_parser_postfix_expression_after_primary(c_parser*, unsigned
int, c_expr) () at $SRC/gcc/c/c-parser.c:7846
#3  0x001957d4 in c_parser_postfix_expression(c_parser*) () at
$SRC/gcc/c/c-parser.c:7667
#4  0x0019f318 in c_parser_unary_expression(c_parser*) () at
$SRC/gcc/c/c-parser.c:6554
#5  0x0019ff00 in c_parser_cast_expression(c_parser*, c_expr*) () at
$SRC/gcc/c/c-parser.c:6392
#6  0x001a00bc in c_parser_binary_expression(c_parser*, c_expr*, tree_node*) ()
at $SRC/gcc/c/c-parser.c:6207
#7  0x001a0bd0 in c_parser_conditional_expression(c_parser*, c_expr*,
tree_node*) () at $SRC/gcc/c/c-parser.c:5983
#8  0x001a11c0 in c_parser_expr_no_commas(c_parser*, c_expr*, tree_node*) () at
$SRC/gcc/c/c-parser.c:5901
#9  0x001a18e8 in c_parser_expression(c_parser*) () at
$SRC/gcc/c/c-parser.c:7974
#10 0x001a23dc in c_parser_expression_conv(c_parser*) () at
$SRC/gcc/c/c-parser.c:8007
#11 0x001b3e10 in c_parser_statement_after_labels(c_parser*) () at
$SRC/gcc/c/c-parser.c:5067
#12 0x001b58c4 in c_parser_compound_statement_nostart(c_parser*) () at
$SRC/gcc/c/c-parser.c:4676
#13 0x001b6054 in c_parser_compound_statement(c_parser*) () at
$SRC/gcc/c/c-parser.c:4513
#14 0x001b1a2c in c_parser_declaration_or_fndef(c_parser*, bool, bool, bool,
bool, bool, tree_node**, vec<c_token, va_heap, vl_ptr>) () at
$SRC/gcc/c/c-parser.c:1966
#15 0x001b7678 in c_parser_external_declaration(c_parser*) () at
$SRC/gcc/c/c-parser.c:1436
#16 0x001b804c in c_parse_file() () at $SRC/gcc/c/c-parser.c:1323
#17 0x00213024 in c_common_parse_file() ()
#18 0x0069c11c in compile_file() () at $SRC/gcc/toplev.c:570
#19 0x00139e30 in toplev::main(int, char**) () at $SRC/gcc/toplev.c:2041
#20 0x0013a9a8 in main () at $SRC/gcc/main.c:38
(gdb)

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

--- Comment #2 from ktkachov at gcc dot gnu.org ---
My hunch is that this is a miscompile of something earlier in the process

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

--- Comment #3 from ktkachov at gcc dot gnu.org ---
Hmm...
Trying to bootstrap 4.9 with the same configuration and -mapcs results in an
ICE in stage2 when compiling haifa-sched.c
$SRC/gcc/haifa-sched.c:6507:1: internal compiler error: in lra_create_copy, at
lra.c:1532
 }
 ^
0x6f240b lra_create_copy(int, int, int)
        $SRC/gcc/lra.c:1532
0x70f447 process_bb_lives
        $SRC/gcc/lra-lives.c:568
0x710761 lra_create_live_ranges(bool)
        $SRC/gcc/lra-lives.c:1008
0x6f3f33 lra(_IO_FILE*)
        $SRC/gcc/lra.c:2372
0x6ad3a1 do_reload
        $SRC/gcc/ira.c:5457
0x6ad717 rest_of_handle_reload
        $SRC/gcc/ira.c:5598
0x6ad75f execute
        $SRC/gcc/ira.c:5627
Please submit a full bug report,
with preprocessed source if appropriate.


I didn't try without -mapcs on 4.9 though, so need to check that first.

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

--- Comment #4 from ktkachov at gcc dot gnu.org ---
(In reply to ktkachov from comment #3)
> Hmm...
> Trying to bootstrap 4.9 with the same configuration and -mapcs results in an
> ICE in stage2 when compiling haifa-sched.c
> $SRC/gcc/haifa-sched.c:6507:1: internal compiler error: in lra_create_copy,
> at lra.c:1532
>  }
>  ^
> 0x6f240b lra_create_copy(int, int, int)
>         $SRC/gcc/lra.c:1532
> 0x70f447 process_bb_lives
>         $SRC/gcc/lra-lives.c:568
> 0x710761 lra_create_live_ranges(bool)
>         $SRC/gcc/lra-lives.c:1008
> 0x6f3f33 lra(_IO_FILE*)
>         $SRC/gcc/lra.c:2372
> 0x6ad3a1 do_reload
>         $SRC/gcc/ira.c:5457
> 0x6ad717 rest_of_handle_reload
>         $SRC/gcc/ira.c:5598
> 0x6ad75f execute
>         $SRC/gcc/ira.c:5627
> Please submit a full bug report,
> with preprocessed source if appropriate.
> 
> 
> I didn't try without -mapcs on 4.9 though, so need to check that first.

Ah, this is a known bootstrap failure: PR 63740
so can't check whether the -mapcs issue is a regression until that's resolved

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |4.8.4

--- Comment #5 from ktkachov at gcc dot gnu.org ---
4.8.4 bootstrap works with -marm -mapcs -O2 in BOOT_CFLAGS btw

[Bug bootstrap/64102] ARM bootstrap fails with segfault with -mapcs in BOOT_CFLAGS

[ramana at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64102

Ramana Radhakrishnan <ramana at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-01-15
                 CC|                            |ramana at gcc dot gnu.org
     Ever confirmed|0                           |1