https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127336
Bug ID: 127336
Summary: wrong code with __atomic_add_fetch() at -O1 and above
on x86
Product: gcc
Version: 17.0
Status: UNCONFIRMED
Keywords: wrong-code
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: zsojka at seznam dot cz
Target Milestone: ---
Host: x86_64-pc-linux-gnu
Target: x86_64-pc-linux-gnu
Created attachment 65564
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=65564&action=edit
reduced testcase
Output:
$ x86_64-pc-linux-gnu-gcc -O testcase.c
$ ./a.out
Aborted ./a.out
$ x86_64-pc-linux-gnu-gcc -O testcase.c -S -masm=intel -fverbose-asm -o-
...
# testcase.c:8: u /= __atomic_add_fetch (&s, s, 0);
mov ecx, DWORD PTR "s"[rip] # s, s
lock xadd DWORD PTR "s"[rip], ecx #,, _9
add ecx, DWORD PTR "s"[rip] # tmp109, s
# testcase.c:8: u /= __atomic_add_fetch (&s, s, 0);
mov eax, DWORD PTR "u"[rip] # u, u
mov edx, 0 # tmp112,
div ecx # tmp109
mov DWORD PTR "u"[rip], eax # u, _5
...
The instruction:
...
lock xadd DWORD PTR "s"[rip], ecx #,, _9
loads the old value of "s" to ecx, and
add ecx, DWORD PTR "s"[rip] # tmp109, s
adds the sum to the addend, so we have 3*9 instead of 2*9
$ x86_64-pc-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=/repo/gcc-trunk/binary-latest-amd64/bin/x86_64-pc-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-20260911053823-g2af7be77bc6-checking-yes-rtl-df-extra-nobootstrap-amd64/bin/../libexec/gcc/x86_64-pc-linux-gnu/17.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--disable-bootstrap --build=x86_64-pc-linux-gnu --host=x86_64-pc-linux-gnu
--target=x86_64-pc-linux-gnu --with-ld=/usr/bin/x86_64-pc-linux-gnu-ld
--with-as=/usr/bin/x86_64-pc-linux-gnu-as --enable-libsanitizer
--disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-20260911053823-g2af7be77bc6-checking-yes-rtl-df-extra-nobootstrap-amd64
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 17.0.0 20260910 (experimental) (GCC)
