| Message ID | 54DD19B7.6060401@redhat.com |
|---|---|
| State | New |
| Headers | show |
On 12/02/15 13:23 -0800, Richard Henderson wrote: >When we fixed PR54005, making sure that atomic_is_lock_free returns the same >value for all objects of a given type, we probably should have changed the >interface so that we would pass size and alignment rather than size and object >pointer. > >Instead, we decided that passing null for the object pointer would be >sufficient. But as this PR shows, we really do need to take alignment into >account. > >The following patch constructs a fake object pointer that is maximally >misaligned. This allows the interface to both the builtin and to libatomic to >remain unchanged. Which probably makes this back-portable to maintenance >releases as well. Am I right in thinking that another option would be to ensure that std::atomic<> objects are always suitably aligned? Would that make std::atomic<> slightly more compatible with a C11 atomic_int, where the _Atomic qualifier affects alignment? https://gcc.gnu.org/PR62259 suggests we might need to enforce alignment on std::atomic anyway, or am I barking up the wrong tree?
On 12/02/15 13:23 -0800, Richard Henderson wrote: >When we fixed PR54005, making sure that atomic_is_lock_free returns the same >value for all objects of a given type, we probably should have changed the >interface so that we would pass size and alignment rather than size and object >pointer. > >Instead, we decided that passing null for the object pointer would be >sufficient. But as this PR shows, we really do need to take alignment into >account. > >The following patch constructs a fake object pointer that is maximally >misaligned. This allows the interface to both the builtin and to libatomic to >remain unchanged. Which probably makes this back-portable to maintenance >releases as well. > >I believe that for all of our current systems, size_t == uintptr_t, so the >reinterpret_cast ought not generate warnings. > >The test case is problematic, as there's currently no good place to put it. >The libstdc++ testsuite doesn't have the libatomic library path configured, and >the libatomic testsuite doesn't have the libstdc++ include paths configured. >Yet another example where we really need an install tree for testing. Thoughts? > > >Ok? OK for trunk. > >r~ > * include/bits/atomic_base.h (__atomic_base<T>::is_lock_free): Build > a fake pointer indicating type alignment. > (__atomic_base<T *>::is_lock_free): Likewise. > * include/std/atomic (atomic<T>::is_lock_free): Likewise.
On Thu, 12 Feb 2015, Richard Henderson wrote: > When we fixed PR54005, Hm, there's confusion. When was this fixed? (Not fixed AFAICT.) Maybe you mean PR54004, but there was no note there either. Or maybe there's a typo and you meant some other PR and that PR54005 is supposedly fixed by this patch (committed as r221701) ...but it doesn't seem right: you use a specific object when deducing the alignment for the fake-pointer, so it's used anyway and is_lock_free must not be object-specific (despite its name) and only type-specific as mandated by the standard (see PR). To wit, deduce from the known-alignment of _Tp, not known-alignment of _M_i. Or is this what happens; they're the same? Why then use __alignof(_M_i) (the object-alignment) instead of _S_alignment (the deduced alas insufficiently increased type-alignment)? > making sure that atomic_is_lock_free returns the same > value for all objects of a given type, (That would work but it doesn't seem to be the case.) > we probably should have changed the > interface so that we would pass size and alignment rather than size and object > pointer. > > Instead, we decided that passing null for the object pointer would be > sufficient. But as this PR shows, we really do need to take alignment into > account. Regarding what's actually needed, alignment of an atomic type should always be *forced to be at least the natural alignment of of the object* (with non-power-of-two sized-objects rounded up) and until then atomic types won't work for targets where the non-atomic equivalents have less alignment (as straddling a page-boundary won't be lock-less-atomic anywhere where straddling a page-boundary may cause a non-atomic-access...) So, not target-specific except for targets that require even higher-than-natural alignment. > The following patch constructs a fake object pointer that is maximally > misaligned. (With regards to the known object alignment of the _M_i object.) > This allows the interface to both the builtin and to libatomic to > remain unchanged. Which probably makes this back-portable to maintenance > releases as well. > > I believe that for all of our current systems, size_t == uintptr_t, so the > reinterpret_cast ought not generate warnings. > > The test case is problematic, as there's currently no good place to put it. > The libstdc++ testsuite doesn't have the libatomic library path configured, and > the libatomic testsuite doesn't have the libstdc++ include paths configured. > Yet another example where we really need an install tree for testing. Thoughts? brgds, H-P
diff --git a/libstdc++-v3/include/bits/atomic_base.h b/libstdc++-v3/include/bits/atomic_base.h index fe6524f..8104c98 100644 --- a/libstdc++-v3/include/bits/atomic_base.h +++ b/libstdc++-v3/include/bits/atomic_base.h @@ -346,11 +346,19 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION bool is_lock_free() const noexcept - { return __atomic_is_lock_free(sizeof(_M_i), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_i)); + return __atomic_is_lock_free(sizeof(_M_i), __a); + } bool is_lock_free() const volatile noexcept - { return __atomic_is_lock_free(sizeof(_M_i), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_i)); + return __atomic_is_lock_free(sizeof(_M_i), __a); + } _GLIBCXX_ALWAYS_INLINE void store(__int_type __i, memory_order __m = memory_order_seq_cst) noexcept @@ -653,11 +661,19 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION bool is_lock_free() const noexcept - { return __atomic_is_lock_free(sizeof(__pointer_type), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_p)); + return __atomic_is_lock_free(sizeof(_M_p), __a); + } bool is_lock_free() const volatile noexcept - { return __atomic_is_lock_free(sizeof(__pointer_type), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_p)); + return __atomic_is_lock_free(sizeof(_M_p), __a); + } _GLIBCXX_ALWAYS_INLINE void store(__pointer_type __p, diff --git a/libstdc++-v3/include/std/atomic b/libstdc++-v3/include/std/atomic index 1a17427..cc4b5f1 100644 --- a/libstdc++-v3/include/std/atomic +++ b/libstdc++-v3/include/std/atomic @@ -198,11 +198,19 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION bool is_lock_free() const noexcept - { return __atomic_is_lock_free(sizeof(_M_i), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_i)); + return __atomic_is_lock_free(sizeof(_M_i), __a); + } bool is_lock_free() const volatile noexcept - { return __atomic_is_lock_free(sizeof(_M_i), nullptr); } + { + // Produce a fake, minimally aligned pointer. + void *__a = reinterpret_cast<void *>(-__alignof(_M_i)); + return __atomic_is_lock_free(sizeof(_M_i), __a); + } void store(_Tp __i, memory_order __m = memory_order_seq_cst) noexcept