[PATCH v2 1/3] libstdc++: Make atomicity helpers use unsigned arithmetic [PR121148]
Jonathan Wakely
jwakely@redhat.com
Fri Sep 12 07:02:25 GMT 2025
On Fri, 12 Sept 2025 at 06:44, Dimitar Dimitrov <dimitar@dinux.eu> wrote:
>
> On Mon, Sep 08, 2025 at 01:35:59PM +0100, Jonathan Wakely wrote:
> > The standard requires that std::atomic<integral-type>::fetch_add does
> > not have undefined behaviour for signed overflow, instead it wraps like
> > unsigned integers. The compiler ensures this is true for the atomic
> > built-ins that std::atomic uses, but it's not currently true for the
> > __gnu_cxx::__exchange_and_add and __gnu_cxx::__atomic_add functions
> > defined in libstdc++, which operate on type _Atomic_word.
> >
> > For the inline __exchange_and_add_single function (used when there's
> > only one thread in the process), we can copy the value to an unsigned
> > long and do the addition on that, then assign it back to the
> > _Atomic_word variable.
> >
> > The __exchange_and_add in config/cpu/generic/atomicity_mutex/atomicity.h
> > locks a mutex and then performs exactly the same steps as
> > __exchange_and_add_single. Calling __exchange_and_add_single instead of
> > duplicating the code benefits from the fix just made to
> > __exchange_and_add_single.
> >
> > For the remaining config/cpu/$arch/atomicity.h implementations, they
> > either use inline assembly which uses wrapping instructions (so no
> > changes needed), or we can fix them by compiling with -fwrapv.
> >
> > After ths change, UBsan no longer gives an error for:
> >
> > _Atomic_word i = INT_MAX;
> > __gnu_cxx::__exchange_and_add_dispatch(&i, 1);
> >
> > /usr/include/c++/14/ext/atomicity.h:85:12: runtime error: signed integer overflow: 2147483647 + 1 cannot be represented in type 'int'
> >
> > libstdc++-v3/ChangeLog:
> >
> > PR libstdc++/121148
> > * config/cpu/generic/atomicity_mutex/atomicity.h
> > (__exchange_and_add): Call __exchange_and_add_single.
> > * include/ext/atomicity.h (__exchange_and_add_single): Use an
> > unsigned type for the addition.
> > * libsupc++/Makefile.am (atomicity.o): Compile with -fwrapv.
> > * libsupc++/Makefile.in: Regenerate.
> > ---
> >
> > v2: no changes to this first patch in the series.
> >
> > Tested powerpc64le-linux.
> >
> > .../cpu/generic/atomicity_mutex/atomicity.h | 5 +--
> > libstdc++-v3/include/ext/atomicity.h | 35 +++++++++++++++++--
> > libstdc++-v3/libsupc++/Makefile.am | 5 +++
> > libstdc++-v3/libsupc++/Makefile.in | 5 +++
> > 4 files changed, 44 insertions(+), 6 deletions(-)
> >
> > diff --git a/libstdc++-v3/config/cpu/generic/atomicity_mutex/atomicity.h b/libstdc++-v3/config/cpu/generic/atomicity_mutex/atomicity.h
> > index b1328c025bcc..7d5772d54840 100644
> > --- a/libstdc++-v3/config/cpu/generic/atomicity_mutex/atomicity.h
> > +++ b/libstdc++-v3/config/cpu/generic/atomicity_mutex/atomicity.h
> > @@ -44,10 +44,7 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
> > __exchange_and_add(volatile _Atomic_word* __mem, int __val) throw ()
> > {
> > __gnu_cxx::__scoped_lock sentry(get_atomic_mutex());
> > - _Atomic_word __result;
> > - __result = *__mem;
> > - *__mem += __val;
> > - return __result;
> > + return __gnu_cxx::__exchange_and_add_single(__mem, __val);
> > }
> >
>
> Hi Jonathan,
>
> This broke the build for pru-unknown-elf target. I suspect it's either
> because it uses newlib instead of glibc. Or because this target does not
> define its own atomic ops.
>
> The error is:
> atomicity.cc: In function ‘_Atomic_word __gnu_cxx::__exchange_and_add(volatile _Atomic_word*, int)’:
> atomicity.cc:47:49: error: invalid conversion from ‘volatile _Atomic_word*’ {aka ‘volatile int*’} to ‘_Atomic_word*’ {aka ‘int*’} [-fpermissive]
> 47 | return __gnu_cxx::__exchange_and_add_single(__mem, __val);
> | ^~~~~
> | |
> | volatile _Atomic_word* {aka volatile int*}
> In file included from atomicity.cc:25:
> /home/dinux/projects/pru/testbot-workspace/pru-gcc-build/pru/libstdc++-v3/include/ext/atomicity.h:95:43: note: initializing argument 1 of ‘_Atomic_word __gnu_cxx::__exchange_and_add_single(_Atomic_word*, int)’
> 95 | __exchange_and_add_single(_Atomic_word* __mem, int __val)
> | ~~~~~~~~~~~~~~^~~~~
Ugh, there's really no reason for these functions to take a volatile pointer.
I'll fix it today...
More information about the Libstdc++
mailing list