{{ message }}
Add deterministic cuDNN max pooling#25269
Merged
tensorflow-copybara merged 1 commit intoJan 30, 2019
Merged
Conversation
…ronment variable is 1/true
tensorflow-copybara
requested review from
timshen91
and removed request for
yifeif
Collaborator
timshen91
approved these changes
Jan 29, 2019
rthadur
added
kokoro:force-run
shahzadlone
reviewed
Jan 30, 2019
Contributor
Author
There was a problem hiding this comment.
Sorry I didn't get a chance to address this before the merge. I agree that adding the const would have been preferable.
Member
There was a problem hiding this comment.
I'm fine with either way. Not having const is shorter, and happens to be consistent with other places. The variable is used only once, and it's immediately below, so const doesn't actually help much.
Contributor
There was a problem hiding this comment.
It is better code to have const. I can fix all of them if you can guide me to them.
rthadur
added
kokoro:force-run
tensorflow-copybara
pushed a commit
that referenced
this pull request
Jan 30, 2019
…_max_pooling PiperOrigin-RevId: 231661364
copybara-service Bot
pushed a commit
that referenced
this pull request
Apr 15, 2025
Imported from GitHub PR openxla/xla#25269 Reported issue: ``` exec ${PAGER:-/usr/bin/less} "$0" || exit 1 Executing tests from //xla/service:compiler_test_gpu_amd_any ----------------------------------------------------------------------------- Running test /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any.runfiles/xla/xla/service/compiler_test_gpu_amd_any --gtest_shuffle --gtest_fail_if_no_test_linked on GPU 0 ================================================================= ==168009==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x50400002c1c0 at pc 0x7f59e50b52e7 bp 0x7ffc8c2358d0 sp 0x7ffc8c2358c8 READ of size 8 at 0x50400002c1c0 thread T0 #0 0x7f59e50b52e6 in absl::lts_20230802::container_internal::CommonFields::capacity() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 #1 0x7f59e50b52e6 in absl::lts_20230802::container_internal::probe(absl::lts_20230802::container_internal::CommonFields const&, unsigned long) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1298:41 #2 0x7f59e50b52e6 in std::pair<unsigned long, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::find_or_prepare_insert<std::tuple<std::type_index, void*>>(std::tuple<std::type_index, void*> const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2645:16 #3 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable::operator()<std::tuple<std::type_index, void*>, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(std::tuple<std::type_index, void*> const&, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>&&) const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2459:20 #4 0x7f59e50af8a8 in decltype(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>()(std::declval<std::tuple<std::type_index, void*>&& const&>(), std::piecewise_construct, std::declval<std::tuple<std::tuple<std::type_index, void*>&&>>(), std::declval<std::tuple<stream_executor::MultiKernelLoaderSpec&&>>())) absl::lts_20230802::container_internal::memory_internal::DecomposePairImpl<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::tuple<std::type_index, void*>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:140:10 #5 0x7f59e50af8a8 in decltype(memory_internal::DecomposePairImpl(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), PairArgs(std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0)))) absl::lts_20230802::container_internal::DecomposePair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:207:10 #6 0x7f59e50af8a8 in decltype(absl::container_internal::DecomposePair(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(), std::declval<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>())) absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/flat_hash_map.h:591:12 #7 0x7f59e50af8a8 in decltype(absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0))) absl::lts_20230802::container_internal::hash_policy_traits<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, void>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/hash_policy_traits.h:134:12 #8 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::emplace<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, 0>(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2064:12 #9 0x7f59e50af8a8 in absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::insert(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1991:12 #10 0x7f59e50af8a8 in stream_executor::gpu::GpuKernelRegistry::RegisterKernel(std::type_info const&, void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/gpu/gpu_kernel_registry.cc:67:45 #11 0x7f59e50d1982 in absl::lts_20230802::Status stream_executor::gpu::GpuKernelRegistry::RegisterKernel<stream_executor::gpu::MakeBatchPointersKernel>(void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/gpu/gpu_kernel_registry.h:86:12 #12 0x7f59e50d1982 in RegisterKernelMakeBatchPointersKernelRocmImpl() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #13 0x7f59e50d1982 in 'lambda'()::operator()() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #14 0x7f59e50d1982 in 'lambda'()::__invoke() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #15 0x7f59e50d1982 in stream_executor::port::Initializer::Initializer(void (*)()) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/platform/default/initialize.h:26:42 #16 0x7f59e50d1982 in __cxx_global_var_init.1 /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #17 0x7f59e50d1982 in _GLOBAL__sub_I_make_batch_pointers_kernel_rocm.cu.cc /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc #18 0x7f5a5b27a47d in call_init elf/dl-init.c:70:3 #19 0x7f5a5b27a567 in call_init elf/dl-init.c:33:6 #20 0x7f5a5b27a567 in _dl_init elf/dl-init.c:117:5 #21 0x7f5a5b2942c9 (/lib64/ld-linux-x86-64.so.2+0x202c9) (BuildId: e4de036b19e4768e7591b596c4be9f9015f2d28a) 0x50400002c1c0 is located 8 bytes after 40-byte region [0x50400002c190,0x50400002c1b8) allocated by thread T0 here: #0 0x557d0f77fcdf in malloc (/root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any+0x1e8cdf) (BuildId: e96972f8c7f880083ff6ad5985d3c06d) #1 0x7f59d733098b in operator new(unsigned long) (/lib/x86_64-linux-gnu/libstdc++.so.6+0xae98b) (BuildId: e37fe1a879783838de78cbc8c80621fa685d58a2) SUMMARY: AddressSanitizer: heap-buffer-overflow /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 in absl::lts_20230802::container_internal::CommonFields::capacity() const Shadow bytes around the buggy address: 0x50400002bf00: fa fa fd fd fd fd fd fd fa fa fd fd fd fd fd fa 0x50400002bf80: fa fa fd fd fd fd fd fa fa fa fd fd fd fd fd fa 0x50400002c000: fa fa fd fd fd fd fd fa fa fa 00 00 00 00 00 fa 0x50400002c080: fa fa fd fd fd fd fd fd fa fa 00 00 00 00 00 00 0x50400002c100: fa fa 00 00 00 00 00 00 fa fa 00 00 00 00 00 00 =>0x50400002c180: fa fa 00 00 00 00 00 fa[fa]fa fa fa fa fa fa fa 0x50400002c200: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c280: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c300: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c380: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c400: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==168009==ABORTING ``` Why this fixes the issue: * Consider compiling this class into a different .so files where this function will get inlined and we will get different instances while we still want to have singleton. * In rocm compiler wrapper script we do not yet support sanitizer flags so our cu.cc files are not getting instrumented while our normal cc files do! This might cause a memory disalignment while running with asan (theory). Copybara import of the project: -- ffcd58918137191cdba6db571e0e5af0e57de2e1 by alekstheod <atheodor@amd.com>: Fix asan issue do to a singleton in header file Merging this change closes #25269 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#25269 from ROCm:ci_fix_singleton_in_header_file ffcd58918137191cdba6db571e0e5af0e57de2e1 PiperOrigin-RevId: 747815699
copybara-service Bot
pushed a commit
that referenced
this pull request
Apr 15, 2025
Imported from GitHub PR openxla/xla#25269 Reported issue: ``` exec ${PAGER:-/usr/bin/less} "$0" || exit 1 Executing tests from //xla/service:compiler_test_gpu_amd_any ----------------------------------------------------------------------------- Running test /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any.runfiles/xla/xla/service/compiler_test_gpu_amd_any --gtest_shuffle --gtest_fail_if_no_test_linked on GPU 0 ================================================================= ==168009==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x50400002c1c0 at pc 0x7f59e50b52e7 bp 0x7ffc8c2358d0 sp 0x7ffc8c2358c8 READ of size 8 at 0x50400002c1c0 thread T0 #0 0x7f59e50b52e6 in absl::lts_20230802::container_internal::CommonFields::capacity() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 #1 0x7f59e50b52e6 in absl::lts_20230802::container_internal::probe(absl::lts_20230802::container_internal::CommonFields const&, unsigned long) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1298:41 #2 0x7f59e50b52e6 in std::pair<unsigned long, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::find_or_prepare_insert<std::tuple<std::type_index, void*>>(std::tuple<std::type_index, void*> const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2645:16 #3 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable::operator()<std::tuple<std::type_index, void*>, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(std::tuple<std::type_index, void*> const&, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>&&) const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2459:20 #4 0x7f59e50af8a8 in decltype(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>()(std::declval<std::tuple<std::type_index, void*>&& const&>(), std::piecewise_construct, std::declval<std::tuple<std::tuple<std::type_index, void*>&&>>(), std::declval<std::tuple<stream_executor::MultiKernelLoaderSpec&&>>())) absl::lts_20230802::container_internal::memory_internal::DecomposePairImpl<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::tuple<std::type_index, void*>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:140:10 #5 0x7f59e50af8a8 in decltype(memory_internal::DecomposePairImpl(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), PairArgs(std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0)))) absl::lts_20230802::container_internal::DecomposePair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:207:10 #6 0x7f59e50af8a8 in decltype(absl::container_internal::DecomposePair(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(), std::declval<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>())) absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/flat_hash_map.h:591:12 #7 0x7f59e50af8a8 in decltype(absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0))) absl::lts_20230802::container_internal::hash_policy_traits<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, void>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/hash_policy_traits.h:134:12 #8 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::emplace<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, 0>(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2064:12 #9 0x7f59e50af8a8 in absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::insert(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1991:12 #10 0x7f59e50af8a8 in stream_executor::gpu::GpuKernelRegistry::RegisterKernel(std::type_info const&, void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/gpu/gpu_kernel_registry.cc:67:45 #11 0x7f59e50d1982 in absl::lts_20230802::Status stream_executor::gpu::GpuKernelRegistry::RegisterKernel<stream_executor::gpu::MakeBatchPointersKernel>(void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/gpu/gpu_kernel_registry.h:86:12 #12 0x7f59e50d1982 in RegisterKernelMakeBatchPointersKernelRocmImpl() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #13 0x7f59e50d1982 in 'lambda'()::operator()() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #14 0x7f59e50d1982 in 'lambda'()::__invoke() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #15 0x7f59e50d1982 in stream_executor::port::Initializer::Initializer(void (*)()) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/platform/default/initialize.h:26:42 #16 0x7f59e50d1982 in __cxx_global_var_init.1 /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #17 0x7f59e50d1982 in _GLOBAL__sub_I_make_batch_pointers_kernel_rocm.cu.cc /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc #18 0x7f5a5b27a47d in call_init elf/dl-init.c:70:3 #19 0x7f5a5b27a567 in call_init elf/dl-init.c:33:6 #20 0x7f5a5b27a567 in _dl_init elf/dl-init.c:117:5 #21 0x7f5a5b2942c9 (/lib64/ld-linux-x86-64.so.2+0x202c9) (BuildId: e4de036b19e4768e7591b596c4be9f9015f2d28a) 0x50400002c1c0 is located 8 bytes after 40-byte region [0x50400002c190,0x50400002c1b8) allocated by thread T0 here: #0 0x557d0f77fcdf in malloc (/root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any+0x1e8cdf) (BuildId: e96972f8c7f880083ff6ad5985d3c06d) #1 0x7f59d733098b in operator new(unsigned long) (/lib/x86_64-linux-gnu/libstdc++.so.6+0xae98b) (BuildId: e37fe1a879783838de78cbc8c80621fa685d58a2) SUMMARY: AddressSanitizer: heap-buffer-overflow /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 in absl::lts_20230802::container_internal::CommonFields::capacity() const Shadow bytes around the buggy address: 0x50400002bf00: fa fa fd fd fd fd fd fd fa fa fd fd fd fd fd fa 0x50400002bf80: fa fa fd fd fd fd fd fa fa fa fd fd fd fd fd fa 0x50400002c000: fa fa fd fd fd fd fd fa fa fa 00 00 00 00 00 fa 0x50400002c080: fa fa fd fd fd fd fd fd fa fa 00 00 00 00 00 00 0x50400002c100: fa fa 00 00 00 00 00 00 fa fa 00 00 00 00 00 00 =>0x50400002c180: fa fa 00 00 00 00 00 fa[fa]fa fa fa fa fa fa fa 0x50400002c200: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c280: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c300: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c380: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c400: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==168009==ABORTING ``` Why this fixes the issue: * Consider compiling this class into a different .so files where this function will get inlined and we will get different instances while we still want to have singleton. * In rocm compiler wrapper script we do not yet support sanitizer flags so our cu.cc files are not getting instrumented while our normal cc files do! This might cause a memory disalignment while running with asan (theory). Copybara import of the project: -- ffcd58918137191cdba6db571e0e5af0e57de2e1 by alekstheod <atheodor@amd.com>: Fix asan issue do to a singleton in header file Merging this change closes #25269 PiperOrigin-RevId: 747823659
copybara-service Bot
pushed a commit
that referenced
this pull request
Apr 15, 2025
Imported from GitHub PR openxla/xla#25269 Reported issue: ``` exec ${PAGER:-/usr/bin/less} "$0" || exit 1 Executing tests from //xla/service:compiler_test_gpu_amd_any ----------------------------------------------------------------------------- Running test /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any.runfiles/xla/xla/service/compiler_test_gpu_amd_any --gtest_shuffle --gtest_fail_if_no_test_linked on GPU 0 ================================================================= ==168009==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x50400002c1c0 at pc 0x7f59e50b52e7 bp 0x7ffc8c2358d0 sp 0x7ffc8c2358c8 READ of size 8 at 0x50400002c1c0 thread T0 #0 0x7f59e50b52e6 in absl::lts_20230802::container_internal::CommonFields::capacity() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 #1 0x7f59e50b52e6 in absl::lts_20230802::container_internal::probe(absl::lts_20230802::container_internal::CommonFields const&, unsigned long) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1298:41 #2 0x7f59e50b52e6 in std::pair<unsigned long, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::find_or_prepare_insert<std::tuple<std::type_index, void*>>(std::tuple<std::type_index, void*> const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2645:16 #3 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable::operator()<std::tuple<std::type_index, void*>, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(std::tuple<std::type_index, void*> const&, std::piecewise_construct_t const&, std::tuple<std::tuple<std::type_index, void*>&&>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>&&) const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2459:20 #4 0x7f59e50af8a8 in decltype(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>()(std::declval<std::tuple<std::type_index, void*>&& const&>(), std::piecewise_construct, std::declval<std::tuple<std::tuple<std::type_index, void*>&&>>(), std::declval<std::tuple<stream_executor::MultiKernelLoaderSpec&&>>())) absl::lts_20230802::container_internal::memory_internal::DecomposePairImpl<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::tuple<std::type_index, void*>&&, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::tuple<std::type_index, void*>&&>, std::tuple<stream_executor::MultiKernelLoaderSpec&&>>) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:140:10 #5 0x7f59e50af8a8 in decltype(memory_internal::DecomposePairImpl(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), PairArgs(std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0)))) absl::lts_20230802::container_internal::DecomposePair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/container_memory.h:207:10 #6 0x7f59e50af8a8 in decltype(absl::container_internal::DecomposePair(std::declval<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(), std::declval<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>())) absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/flat_hash_map.h:591:12 #7 0x7f59e50af8a8 in decltype(absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>::apply(std::forward<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable>(fp), std::forward<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(fp0))) absl::lts_20230802::container_internal::hash_policy_traits<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, void>::apply<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>>(absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::EmplaceDecomposable&&, std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/hash_policy_traits.h:134:12 #8 0x7f59e50af8a8 in std::pair<absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::iterator, bool> absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::emplace<std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, 0>(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:2064:12 #9 0x7f59e50af8a8 in absl::lts_20230802::container_internal::raw_hash_set<absl::lts_20230802::container_internal::FlatHashMapPolicy<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>, absl::lts_20230802::hash_internal::Hash<std::tuple<std::type_index, void*>>, std::equal_to<std::tuple<std::type_index, void*>>, std::allocator<std::pair<std::tuple<std::type_index, void*> const, stream_executor::MultiKernelLoaderSpec>>>::insert(std::pair<std::tuple<std::type_index, void*>, stream_executor::MultiKernelLoaderSpec>&&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:1991:12 #10 0x7f59e50af8a8 in stream_executor::gpu::GpuKernelRegistry::RegisterKernel(std::type_info const&, void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/gpu/gpu_kernel_registry.cc:67:45 #11 0x7f59e50d1982 in absl::lts_20230802::Status stream_executor::gpu::GpuKernelRegistry::RegisterKernel<stream_executor::gpu::MakeBatchPointersKernel>(void*, stream_executor::MultiKernelLoaderSpec const&) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/gpu/gpu_kernel_registry.h:86:12 #12 0x7f59e50d1982 in RegisterKernelMakeBatchPointersKernelRocmImpl() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #13 0x7f59e50d1982 in 'lambda'()::operator()() const /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #14 0x7f59e50d1982 in 'lambda'()::__invoke() /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #15 0x7f59e50d1982 in stream_executor::port::Initializer::Initializer(void (*)()) /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/./xla/stream_executor/platform/default/initialize.h:26:42 #16 0x7f59e50d1982 in __cxx_global_var_init.1 /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc:35:1 #17 0x7f59e50d1982 in _GLOBAL__sub_I_make_batch_pointers_kernel_rocm.cu.cc /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/xla/stream_executor/rocm/make_batch_pointers_kernel_rocm.cu.cc #18 0x7f5a5b27a47d in call_init elf/dl-init.c:70:3 #19 0x7f5a5b27a567 in call_init elf/dl-init.c:33:6 #20 0x7f5a5b27a567 in _dl_init elf/dl-init.c:117:5 #21 0x7f5a5b2942c9 (/lib64/ld-linux-x86-64.so.2+0x202c9) (BuildId: e4de036b19e4768e7591b596c4be9f9015f2d28a) 0x50400002c1c0 is located 8 bytes after 40-byte region [0x50400002c190,0x50400002c1b8) allocated by thread T0 here: #0 0x557d0f77fcdf in malloc (/root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/bazel-out/k8-dbg/bin/xla/service/compiler_test_gpu_amd_any+0x1e8cdf) (BuildId: e96972f8c7f880083ff6ad5985d3c06d) #1 0x7f59d733098b in operator new(unsigned long) (/lib/x86_64-linux-gnu/libstdc++.so.6+0xae98b) (BuildId: e37fe1a879783838de78cbc8c80621fa685d58a2) SUMMARY: AddressSanitizer: heap-buffer-overflow /root/.cache/bazel/_bazel_root/f367074f9120c6f1a67d35844ac058a3/execroot/xla/external/com_google_absl/absl/container/internal/raw_hash_set.h:990:36 in absl::lts_20230802::container_internal::CommonFields::capacity() const Shadow bytes around the buggy address: 0x50400002bf00: fa fa fd fd fd fd fd fd fa fa fd fd fd fd fd fa 0x50400002bf80: fa fa fd fd fd fd fd fa fa fa fd fd fd fd fd fa 0x50400002c000: fa fa fd fd fd fd fd fa fa fa 00 00 00 00 00 fa 0x50400002c080: fa fa fd fd fd fd fd fd fa fa 00 00 00 00 00 00 0x50400002c100: fa fa 00 00 00 00 00 00 fa fa 00 00 00 00 00 00 =>0x50400002c180: fa fa 00 00 00 00 00 fa[fa]fa fa fa fa fa fa fa 0x50400002c200: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c280: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c300: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c380: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa 0x50400002c400: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa Shadow byte legend (one shadow byte represents 8 application bytes): Addressable: 00 Partially addressable: 01 02 03 04 05 06 07 Heap left redzone: fa Freed heap region: fd Stack left redzone: f1 Stack mid redzone: f2 Stack right redzone: f3 Stack after return: f5 Stack use after scope: f8 Global redzone: f9 Global init order: f6 Poisoned by user: f7 Container overflow: fc Array cookie: ac Intra object redzone: bb ASan internal: fe Left alloca redzone: ca Right alloca redzone: cb ==168009==ABORTING ``` Why this fixes the issue: * Consider compiling this class into a different .so files where this function will get inlined and we will get different instances while we still want to have singleton. * In rocm compiler wrapper script we do not yet support sanitizer flags so our cu.cc files are not getting instrumented while our normal cc files do! This might cause a memory disalignment while running with asan (theory). Copybara import of the project: -- ffcd58918137191cdba6db571e0e5af0e57de2e1 by alekstheod <atheodor@amd.com>: Fix asan issue do to a singleton in header file Merging this change closes #25269 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#25269 from ROCm:ci_fix_singleton_in_header_file ffcd58918137191cdba6db571e0e5af0e57de2e1 PiperOrigin-RevId: 747833213
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
When the TF_CUDNN_DETERMINISTIC environment variable is set to 1/true, cuDNN max pooling is performed using a deterministic algorithm.
This current pull request follows-on from pull request 24747 which dealt only with the convolution algorithms.
This current pull request completes the functionality controlled by TF_CUDNN_DETERMINISTIC, so the TODO from the previous pull request is also removed.