https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89436
Bug ID: 89436
Summary: [8/9 Regression] ICE in
glibc/sysdeps/ieee754/dbl-64/e_pow.c (internal
compiler error: in decompose, at rtl.h:2266)
Product: gcc
Version: 9.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: przemyslaw.wirkus at arm dot com
Target Milestone: ---
# GCC Version
revision r268990
# Host system type
Linux machine 4.15.0-43-generic #46-Ubuntu SMP Thu Dec 6 14:45:28 UTC 2018
x86_64 x86_64 x86_64 GNU/Linux
Ubuntu 18.04.2 LTS
# Options given when GCC was configured/built (gcc2 stage)
--target=arm-none-linux-gnueabi \
--enable-shared \
--disable-libatomic \
--without-cloog \
--without-isl \
--disable-libssp \
--disable-libgomp \
--disable-libmudflap \
--disable-libquadmath \
--enable-checking=yes \
--enable-languages=c \
--with-float=softfp \
--with-fpu=crypto-neon-fp-armv8 \
--with-mode=thumb \
--with-arch=armv8-a \
--with-arch=armv7-a
# Reduced testcase and command line that triggers the bug
$cat e_pow_cpp.c
typedef double a;
b;
c() {
a d, e;
b = d + d * e;
}
$ ./arm-none-linux-gnueabi-gcc -O2 -g e_pow_cpp.c -S -w -march=armv8-a
-mfloat-abi=softfp -mfpu=crypto-neon-fp-armv8
e_pow_cpp.c:2:1: warning: data definition has no type or storage class
2 | b;
| ^
e_pow_cpp.c:2:1: warning: type defaults to ‘int’ in declaration of ‘b’
[-Wimplicit-int]
e_pow_cpp.c:3:1: warning: return type defaults to ‘int’ [-Wimplicit-int]
3 | c() {
| ^
e_pow_cpp.c: In function ‘c’:
e_pow_cpp.c:6:1: error: qsort comparator non-negative on sorted output: 0
6 | }
| ^
during RTL pass: cprop_hardreg
e_pow_cpp.c:6:1: internal compiler error: qsort checking failed
0x5a3764 qsort_chk_error
$src/gcc/gcc/vec.c:214
0x14f0b09 qsort_chk(void*, unsigned long, unsigned long, int (*)(void const*,
void const*))
$src/gcc/gcc/vec.c:257
0x151d1b1 gcc_qsort(void*, unsigned long, unsigned long, int (*)(void const*,
void const*))
$src/gcc/gcc/sort.cc:238
0x7a6fa8 vec<df_mw_hardreg*, va_heap, vl_embed>::qsort(int (*)(void const*,
void const*))
$src/gcc/gcc/vec.h:1114
0x7a6fa8 vec<df_mw_hardreg*, va_heap, vl_ptr>::qsort(int (*)(void const*, void
const*))
$src/gcc/gcc/vec.h:1898
0x7a6fa8 df_sort_and_compress_mws
$src/gcc/gcc/df-scan.c:2252
0x7a6fa8 df_canonize_collection_rec
$src/gcc/gcc/df-scan.c:2281
0x7a932f df_insn_refs_collect
$src/gcc/gcc/df-scan.c:3229
0x7a973a df_insn_refs_verify
$src/gcc/gcc/df-scan.c:4088
0x7ab403 df_insn_rescan(rtx_insn*)
$src/gcc/gcc/df-scan.c:1065
0xb44897 confirm_change_group()
$src/gcc/gcc/recog.c:503
0xb47c1c apply_change_group()
$src/gcc/gcc/recog.c:516
0xb4bbc8 copyprop_hardreg_forward_1
$src/gcc/gcc/regcprop.c:1001
0xb4bfb9 execute
$src/gcc/gcc/regcprop.c:1307
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
