Merge from upstream - #171
Merged
Merged
Conversation
Summary: Pull Request resolved: pytorch#11187 Differential Revision: D9628349 Pulled By: SsnL fbshipit-source-id: 0ff94666542ca049a6d82091bd9fc79ec1699ac6
Summary: In the state dict loading code, it would print the error message referring to the shape of the loaded parameters and the parameters in the initialised model with the formatting in the wrong order. Swapped them round to fix. Pull Request resolved: pytorch#11200 Differential Revision: D9631160 Pulled By: SsnL fbshipit-source-id: 03d9446303bd417fef67027b10d7a27de06486be
Summary: We don't generate a corresponding Type implementations for them, so this doesn't do anything at the moment. We don't plan on supporting complex32 in the near future, but it is added to reserve the name and number in case we do at some point in the future. Pull Request resolved: pytorch#11173 Reviewed By: SsnL Differential Revision: D9627477 Pulled By: ezyang fbshipit-source-id: f49a44ab1c92d8a33130c249ac7b234f210a65e6
Summary: Pull Request resolved: pytorch#11208 Differential Revision: D9632216 Pulled By: SsnL fbshipit-source-id: b181f3ce114474e171146cd2ac5de150b0e23f75
Summary: Example: ```sh python run_test.py -i sparse -- TestSparse.test_factory_size_check -f ``` With this, the `--verbose` option is redundant (one can call `python run_test.py -- -v` instead of `python run_test.py -v`. But since this is (probably) a frequently used flag, I didn't remove the existing easier-to-use option. cc ezyang Pull Request resolved: pytorch#11209 Differential Revision: D9632215 Pulled By: SsnL fbshipit-source-id: ff522802da11ef0a0714578be46e4a44f6343d44
…h#11189) Summary: Pull Request resolved: pytorch#11189 Replaces it with an operator TensorOptions() method on Type, reestablishing the implicit conversion. I originally wanted to get rid of the implicit conversion entirely, but there were a *lot* of use-sites, so I added it back to avoid a huge codemod. In this patch, I only had to fix sites that used the optional device_index API. Reviewed By: cpuhrsch Differential Revision: D9628281 fbshipit-source-id: 5fe2a68eefb77a3c9bb446f03a94ad723ef90210
lcskrishna
pushed a commit
to lcskrishna/pytorch
that referenced
this pull request
May 15, 2023
When tensor is resized, reference array to it's sizes may become invalid. Make a copy in advance.
<details>
<summary>ASAN report</summary>
```
=================================================================
==1115867==ERROR: AddressSanitizer: heap-use-after-free on address 0x61000013d790 at pc 0x03ff8e7da360 bp 0x03fff53c83a0 sp 0x03fff53c8390
READ of size 8 at 0x61000013d790 thread T0
#0 0x3ff8e7da35f in c10::SymInt::is_heap_allocated() const /home/user/pytorch/c10/core/SymInt.h:154
ROCm#1 0x3ff8e7da35f in c10::SymInt::maybe_as_int() const /home/user/pytorch/c10/core/SymInt.h:215
ROCm#2 0x3ff8e7d0a6d in c10::SymInt::sym_eq(c10::SymInt const&) const /home/user/pytorch/c10/core/SymInt.cpp:69
ROCm#3 0x3ff7a9ab0bd in c10::SymInt::operator==(c10::SymInt const&) const /home/user/pytorch/c10/core/SymInt.h:177
ROCm#4 0x3ff7a9aaedd in bool std::__equal<false>::equal<c10::SymInt const*, c10::SymInt const*>(c10::SymInt const*, c10::SymInt const*, c10::SymInt const*) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-
v11/bits/stl_algobase.h:1162
ROCm#5 0x3ff7a9aae4b in bool std::__equal_aux1<c10::SymInt const*, c10::SymInt const*>(c10::SymInt const*, c10::SymInt const*, c10::SymInt const*) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/
stl_algobase.h:1211
ROCm#6 0x3ff7a9aae05 in bool std::__equal_aux<c10::SymInt const*, c10::SymInt const*>(c10::SymInt const*, c10::SymInt const*, c10::SymInt const*) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/s
tl_algobase.h:1219
ROCm#7 0x3ff7a9aad97 in bool std::equal<c10::SymInt const*, c10::SymInt const*>(c10::SymInt const*, c10::SymInt const*, c10::SymInt const*) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_alg
obase.h:1556
ROCm#8 0x3ff4b23c771 in c10::ArrayRef<c10::SymInt>::equals(c10::ArrayRef<c10::SymInt>) const /home/user/pytorch/c10/util/ArrayRef.h:188
ROCm#9 0x3ff4cb91bc1 in bool c10::operator!=<c10::SymInt>(c10::ArrayRef<c10::SymInt>, c10::ArrayRef<c10::SymInt>) /home/user/pytorch/c10/util/ArrayRef.h:341
ROCm#10 0x3ff6d1b57ff in torch::ADInplaceOrView::resize_(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pytorch/torch/csrc/autograd/Variab
leTypeManual.cpp:408
ROCm#11 0x3ff6d1e59c7 in c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c1
0::MemoryFormat>), &torch::ADInplaceOrView::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>
> >::operator()(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pytorch/aten/src/ATen/core/boxing/impl/WrapFunctionIntoFunctor.h:13
ROCm#12 0x3ff6d1e59c7 in c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10:
:ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>), &torch::ADInplaceOrView::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::Sy
mInt>, c10::optional<c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::call(c10::OperatorKernel*, c10::Disp
atchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pytorch/aten/src/ATen/core/boxing/impl/make_boxed_from_unboxed_functor.h:480
ROCm#13 0x3ff51ca5129 in at::Tensor const& c10::callUnboxedKernelFunction<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(void*, c10::OperatorKernel*,
c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>&&, c10::optional<c10::MemoryFormat>&&) /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:50
ROCm#14 0x3ff51ca6e8f in at::Tensor const& c10::KernelFunction::call<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::OperatorHandle const&, c10::D
ispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:90
ROCm#15 0x3ff51ca6e8f in at::Tensor const& c10::Dispatcher::redispatch<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::TypedOperatorHandle<at::Ten
sor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)> const&, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)
const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:656
ROCm#16 0x3ff5182006b in c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::redispatch(c10::DispatchKeySet, at::Tensor const&, c
10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:492
ROCm#17 0x3ff5182006b in at::_ops::resize_::redispatch(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) aten/src/ATen/Operators_4.cpp:2144
ROCm#18 0x3ff6d1d5e07 in at::redispatch::resize__symint(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) aten/src/ATen/RedispatchFunctions.h:2847
ROCm#19 0x3ff6d1bbb67 in torch::autograd::VariableType::(anonymous namespace)::resize_(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pyto
rch/torch/csrc/autograd/VariableTypeManual.cpp:243
ROCm#20 0x3ff6d1bd197 in c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c1
0::MemoryFormat>), &torch::autograd::VariableType::(anonymous namespace)::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10
::optional<c10::MemoryFormat> > >::operator()(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pytorch/aten/src/ATen/core/boxing/impl/WrapFu
nctionIntoFunctor.h:13
ROCm#21 0x3ff6d1bd197 in c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10:
:ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>), &torch::autograd::VariableType::(anonymous namespace)::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor
const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::call(c
10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) /home/user/pytorch/aten/src/ATen/core/boxing/impl/make_boxed_from_unboxed_functor
.h:480
ROCm#22 0x3ff51ca5129 in at::Tensor const& c10::callUnboxedKernelFunction<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(void*, c10::OperatorKernel*,
c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>&&, c10::optional<c10::MemoryFormat>&&) /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:50
ROCm#23 0x3ff5181ead1 in at::Tensor const& c10::KernelFunction::call<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::OperatorHandle const&, c10::D
ispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:90
ROCm#24 0x3ff5181ead1 in at::Tensor const& c10::Dispatcher::call<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::TypedOperatorHandle<at::Tensor co
nst& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)> const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const /home/user/pytorch/at
en/src/ATen/core/dispatch/Dispatcher.h:639
ROCm#25 0x3ff5181ead1 in c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::call(at::Tensor const&, c10::ArrayRef<c10::SymInt>,
c10::optional<c10::MemoryFormat>) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:487
ROCm#26 0x3ff5181ead1 in at::_ops::resize_::call(at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) aten/src/ATen/Operators_4.cpp:2137
ROCm#27 0x3ff79b44fcf in at::Tensor::resize__symint(c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const aten/src/ATen/core/TensorBody.h:2452
ROCm#28 0x3ff79a802db in torch::autograd::THPVariable_resize_(_object*, _object*, _object*)::$_0::operator()(at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const /home/us
er/pytorch/torch/csrc/autograd/generated/python_variable_methods.cpp:13417
ROCm#29 0x3ff7999f1eb in torch::autograd::THPVariable_resize_(_object*, _object*, _object*) /home/user/pytorch/torch/csrc/autograd/generated/python_variable_methods.cpp:13419
ROCm#30 0x3ffa2c9b009 in method_vectorcall_VARARGS_KEYWORDS Objects/descrobject.c:344
ROCm#31 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#32 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#33 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#34 0x3ffa2dff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#35 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#36 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#37 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#38 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#39 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#40 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#41 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#42 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#43 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#44 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#45 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#46 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#47 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#48 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#49 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#50 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#51 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#52 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#53 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#54 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#55 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#56 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#57 0x3ffa2dff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#58 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#59 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#60 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#61 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#62 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#63 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#64 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#65 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#66 0x3ffa2dff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#67 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#68 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#69 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#70 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#71 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#72 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#73 0x3ffa2dff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#74 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#75 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#76 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#77 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#78 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#79 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#80 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#81 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#82 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#83 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#84 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#85 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#86 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#87 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#88 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#89 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#90 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#91 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#92 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#93 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#94 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#95 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#96 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#97 0x3ffa2c8ab9b in PyVectorcall_Call Objects/call.c:267
ROCm#98 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#99 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#100 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#101 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#102 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#103 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#104 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#105 0x3ffa2c8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#106 0x3ffa2c8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#107 0x3ffa2d3f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#108 0x3ffa2c8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#109 0x3ffa2df0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#110 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#111 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#112 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#113 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#114 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#115 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#116 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#117 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#118 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#119 0x3ffa2dff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#120 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#121 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#122 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#123 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#124 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#125 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#126 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#127 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#128 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#129 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#130 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#131 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#132 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#133 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#134 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#135 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#136 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#137 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#138 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#139 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#140 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#141 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#142 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#143 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#144 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#145 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#146 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#147 0x3ffa2c8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#148 0x3ffa2c8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#149 0x3ffa2d3f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#150 0x3ffa2c8ad17 in _PyObject_Call Objects/call.c:305
ROCm#151 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#152 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#153 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#154 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#155 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#156 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#157 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#158 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#159 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#160 0x3ffa2dff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#161 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#162 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#163 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#164 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#165 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#166 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#167 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#168 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#169 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#170 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#171 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#172 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#173 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#174 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#175 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#176 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#177 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#178 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#179 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#180 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#181 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#182 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#183 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#184 0x3ffa2dff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#185 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#186 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#187 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#188 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#189 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#190 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#191 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#192 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#193 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#194 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#195 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#196 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#197 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#198 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#199 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#200 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#201 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#202 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#203 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#204 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#205 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#206 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#207 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#208 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#209 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#210 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#211 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#212 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#213 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#214 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#215 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#216 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#217 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#218 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#219 0x3ffa2c8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#220 0x3ffa2c8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#221 0x3ffa2d3f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#222 0x3ffa2c8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#223 0x3ffa2df0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#224 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#225 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#226 0x3ffa2dffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#227 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#228 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#229 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#230 0x3ffa2c8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#231 0x3ffa2c8ac65 in _PyObject_Call Objects/call.c:290
ROCm#232 0x3ffa2c8ada9 in PyObject_Call Objects/call.c:317
ROCm#233 0x3ffa2e059c7 in do_call_core Python/ceval.c:5943
ROCm#234 0x3ffa2dffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#235 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#236 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#237 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#238 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#239 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#240 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#241 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#242 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#243 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#244 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#245 0x3ffa2c8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#246 0x3ffa2c8eddd in method_vectorcall Objects/classobject.c:53
ROCm#247 0x3ffa2df00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#248 0x3ffa2df013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#249 0x3ffa2e05447 in call_function Python/ceval.c:5891
ROCm#250 0x3ffa2dff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#251 0x3ffa2df052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#252 0x3ffa2e02b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#253 0x3ffa2c8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#254 0x3ffa2c8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#255 0x3ffa2c8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#256 0x3ffa2d3f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#257 0x3ffa2c8a933 in _PyObject_MakeTpCall Objects/call.c:215
0x61000013d790 is located 80 bytes inside of 192-byte region [0x61000013d740,0x61000013d800)
freed by thread T0 here:
#0 0x3ffa3237de5 in operator delete(void*) /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:160
ROCm#1 0x3ff8e7e3221 in c10::TensorImpl::~TensorImpl() /home/user/pytorch/c10/core/TensorImpl.cpp:75
previously allocated by thread T0 here:
#0 0x3ffa323734f in operator new(unsigned long) /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:99
ROCm#1 0x3ff4aeeb3d1 in c10::intrusive_ptr<c10::TensorImpl, c10::detail::intrusive_target_default_null_type<c10::TensorImpl> > c10::intrusive_ptr<c10::TensorImpl, c10::detail::intrusive_target_default_nul
l_type<c10::TensorImpl> >::make<c10::intrusive_ptr<c10::StorageImpl, c10::detail::intrusive_target_default_null_type<c10::StorageImpl> >, c10::DispatchKeySet&, caffe2::TypeMeta&>(c10::intrusive_ptr<c10::S
torageImpl, c10::detail::intrusive_target_default_null_type<c10::StorageImpl> >&&, c10::DispatchKeySet&, caffe2::TypeMeta&) /home/user/pytorch/c10/util/intrusive_ptr.h:498
ROCm#2 0x3ff76f79e17 (/home/user/pytorch/build/lib.linux-s390x-cpython-310/torch/lib/libtorch_cpu.so+0x2fb79e17)
SUMMARY: AddressSanitizer: heap-use-after-free /home/user/pytorch/c10/core/SymInt.h:154 in c10::SymInt::is_heap_allocated() const
Shadow bytes around the buggy address:
0x100c2000027aa0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd
0x100c2000027ab0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x100c2000027ac0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd
0x100c2000027ad0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x100c2000027ae0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd
=>0x100c2000027af0: fd fd[fd]fd fd fd fd fd fd fd fd fd fd fd fd fd
0x100c2000027b00: fa fa fa fa fa fa fa fa 00 00 00 00 00 00 00 00
0x100c2000027b10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x100c2000027b20: fa fa fa fa fa fa fa fa 00 00 00 00 00 00 00 00
0x100c2000027b30: 00 00 00 00 04 fa fa fa fa fa fa fa fa fa fa fa
0x100c2000027b40: 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
Shadow gap: cc
==1115867==ABORTING
```
</details>
<details>
<summary>Additional backtraces (not full)</summary>
Memory deallocation:
```
#0 operator delete (ptr=0x61000013d740) at /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:160
ROCm#1 0x000003ffa77e3222 in c10::TensorImpl::~TensorImpl (this=0x61000013d740) at /home/user/pytorch/c10/core/TensorImpl.cpp:75
ROCm#2 0x000003ff63e76e8c in c10::intrusive_ptr<c10::TensorImpl, c10::UndefinedTensorImpl>::reset_ (this=0x3ffd7ec8230) at /home/user/pytorch/c10/util/intrusive_ptr.h:291
ROCm#3 0x000003ff63e76910 in c10::intrusive_ptr<c10::TensorImpl, c10::UndefinedTensorImpl>::~intrusive_ptr (this=0x3ffd7ec8230) at /home/user/pytorch/c10/util/intrusive_ptr.h:370
ROCm#4 0x000003ff63e67240 in at::TensorBase::~TensorBase (this=0x3ffd7ec8230) at /home/user/pytorch/aten/src/ATen/core/TensorBase.h:80
ROCm#5 0x000003ff63e85ee0 in at::Tensor::~Tensor (this=0x3ffd7ec8230) at aten/src/ATen/core/TensorBody.h:90
ROCm#6 0x000003ff63f67304 in resize__functionalization (dispatchKeySet=..., self=..., size=..., memory_format=...) at /home/user/pytorch/aten/src/ATen/FunctionalizeFallbackKernel.cpp:173
ROCm#7 0x000003ff63f89258 in c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>), &(resize__functionalization(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>))>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat> > >::operator()(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>) (
this=0x6030000390a0, args=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/boxing/impl/WrapFunctionIntoFunctor.h:13
ROCm#8 c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>), &(resize__functionalization(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>))>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>)>::call(c10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>) (functor=0x6030000390a0, dispatchKeySet=..., args=..., args=...,
args=...) at /home/user/pytorch/aten/src/ATen/core/boxing/impl/make_boxed_from_unboxed_functor.h:480
ROCm#9 0x000003ff6aca560a in c10::callUnboxedKernelFunction<at::Tensor const&, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat> > (
unboxed_kernel_func=0x3ff63f88a80 <c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tenso
r const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>), &(resize__functionalization(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>))>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>)>::call(c10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<long>, c10::optional<c10::MemoryFormat>)>, functor=0x6030000390a0,
dispatchKeySet=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:50
ROCm#10 0x000003ff6aca715c in c10::KernelFunction::call<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> > (this=0x6210005e1b28, opHandle=...,
dispatchKeySet=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:96
ROCm#11 c10::Dispatcher::redispatch<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)> const&, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const (
this=0x3ff919400e0 <c10::Dispatcher::realSingleton()::_singleton>, op=..., currentDispatchKeySet=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:656
ROCm#12 0x000003ff6a82006c in c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::redispatch(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const (
this=0x3ff919a07e0 <at::_ops::resize_::redispatch(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)::op>, currentDispatchKeySet=..., args=...,
args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:492
ROCm#13 at::_ops::resize_::redispatch (dispatchKeySet=..., self=..., size=..., memory_format=...) at /home/user/pytorch/build/aten/src/ATen/Operators_4.cpp:2144
ROCm#14 0x000003ff861d5e08 in at::redispatch::resize__symint (dispatchKeySet=..., self=..., size=..., memory_format=...) at aten/src/ATen/RedispatchFunctions.h:2847
ROCm#15 0x000003ff861b579e in torch::ADInplaceOrView::resize_ (ks=..., self=..., size=..., optional_memory_format=...) at /home/user/pytorch/torch/csrc/autograd/VariableTypeManual.cpp:401
```
Memory access:
```
#0 c10::SymInt::maybe_as_int (this=0x61000013d790) at /home/user/pytorch/c10/core/SymInt.h:215
ROCm#1 0x000003ff734d0a6e in c10::SymInt::sym_eq (this=0x61000013d790, sci=...) at /home/user/pytorch/c10/core/SymInt.cpp:69
ROCm#2 0x000003ff5f6ab0be in c10::SymInt::operator== (this=0x61000013d790, o=...) at /home/user/pytorch/c10/core/SymInt.h:177
ROCm#3 0x000003ff5f6aaede in std::__equal<false>::equal<c10::SymInt const*, c10::SymInt const*> (__first1=0x61000013d790, __last1=0x61000013d7a0, __first2=0x602000015c30)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_algobase.h:1162
ROCm#4 0x000003ff5f6aae4c in std::__equal_aux1<c10::SymInt const*, c10::SymInt const*> (__first1=0x61000013d790, __last1=0x61000013d7a0, __first2=0x602000015c30)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_algobase.h:1211
ROCm#5 0x000003ff5f6aae06 in std::__equal_aux<c10::SymInt const*, c10::SymInt const*> (__first1=0x61000013d790, __last1=0x61000013d7a0, __first2=0x602000015c30)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_algobase.h:1219
ROCm#6 0x000003ff5f6aad98 in std::equal<c10::SymInt const*, c10::SymInt const*> (__first1=0x61000013d790, __last1=0x61000013d7a0, __first2=0x602000015c30)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_algobase.h:1556
ROCm#7 0x000003ff2ff3c772 in c10::ArrayRef<c10::SymInt>::equals (this=0x3ffed7c9900, RHS=...) at /home/user/pytorch/c10/util/ArrayRef.h:188
ROCm#8 0x000003ff31891bc2 in c10::operator!=<c10::SymInt> (a1=..., a2=...) at /home/user/pytorch/c10/util/ArrayRef.h:341
ROCm#9 0x000003ff51eb5800 in torch::ADInplaceOrView::resize_ (ks=..., self=..., size=..., optional_memory_format=...) at /home/user/pytorch/torch/csrc/autograd/VariableTypeManual.cpp:408
ROCm#10 0x000003ff51ee59c8 in c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c
10::MemoryFormat>), &torch::ADInplaceOrView::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>
> >::operator()(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) (this=0x6030007dca40, args=..., args=..., args=..., args=...)
at /home/user/pytorch/aten/src/ATen/core/boxing/impl/WrapFunctionIntoFunctor.h:13
ROCm#11 c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt
>, c10::optional<c10::MemoryFormat>), &torch::ADInplaceOrView::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<
c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::call(c10::OperatorKernel*, c10::DispatchKeySet, at::Tenso
r const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) (functor=0x6030007dca40, dispatchKeySet=..., args=..., args=..., args=...)
at /home/user/pytorch/aten/src/ATen/core/boxing/impl/make_boxed_from_unboxed_functor.h:480
ROCm#12 0x000003ff369a512a in c10::callUnboxedKernelFunction<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> > (
unboxed_kernel_func=0x3ff51ee51f0 <c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor const& (c10::DispatchKeySet, at::Tenso
r const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>), &torch::ADInplaceOrView::resize_>, at::Tensor const&, c10::guts::typelist::typelist<c10::DispatchKeySet, at::Tensor const&, c10::Ar
rayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> > >, at::Tensor const& (c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::call(c10::OperatorKern
el*, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>, functor=0x6030007dca40, dispatchKeySet=..., args=..., args=..., args=...)
at /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:50
ROCm#13 0x000003ff369a6e90 in c10::KernelFunction::call<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> > (this=0x6210005e1bc8, opHandle=...,
dispatchKeySet=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/boxing/KernelFunction_impl.h:90
ROCm#14 c10::Dispatcher::redispatch<at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat> >(c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::Arr
ayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)> const&, c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const (
this=0x3ff5d6400e0 <c10::Dispatcher::realSingleton()::_singleton>, op=..., currentDispatchKeySet=..., args=..., args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:656
ROCm#15 0x000003ff3652006c in c10::TypedOperatorHandle<at::Tensor const& (at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)>::redispatch(c10::DispatchKeySet, at::Tensor const&,
c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>) const (
this=0x3ff5d6a07e0 <at::_ops::resize_::redispatch(c10::DispatchKeySet, at::Tensor const&, c10::ArrayRef<c10::SymInt>, c10::optional<c10::MemoryFormat>)::op>, currentDispatchKeySet=..., args=...,
args=..., args=...) at /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:492
ROCm#16 at::_ops::resize_::redispatch (dispatchKeySet=..., self=..., size=..., memory_format=...) at /home/user/pytorch/build/aten/src/ATen/Operators_4.cpp:2144
ROCm#17 0x000003ff51ed5e08 in at::redispatch::resize__symint (dispatchKeySet=..., self=..., size=..., memory_format=...) at aten/src/ATen/RedispatchFunctions.h:2847
ROCm#18 0x000003ff51ebbb68 in torch::autograd::VariableType::(anonymous namespace)::resize_ (ks=..., self=..., size=..., optional_memory_format=...)
at /home/user/pytorch/torch/csrc/autograd/VariableTypeManual.cpp:243
```
</details>
Pull Request resolved: pytorch#101064
Approved by: https://github.com/Skylion007, https://github.com/albanD
alugorey
pushed a commit
to alugorey/pytorch
that referenced
this pull request
May 17, 2023
arguments() returns vector member of object returned by schema() call.
When object returned by schema() call is destroyed, the vector is deallocated as well,
it's lifetime isn't extended.
This issue detected while running `pytest -v test/mobile/test_lite_script_type.py -k test_nest_typing_namedtuple_custom_classtype` with ASAN.
<details>
<summary>ASAN output</summary>
```
==1134126==ERROR: AddressSanitizer: heap-use-after-free on address 0x60d0005a5790 at pc 0x03ff844488d8 bp 0x03fff584afe8 sp 0x03fff584afd8
READ of size 8 at 0x60d0005a5790 thread T0
#0 0x3ff844488d7 in __gnu_cxx::__normal_iterator<c10::Argument const*, std::vector<c10::Argument, std::allocator<c10::Argument> > >::__normal_iterator(c10::Argument const* const&) /usr/lib/gcc/s390x-i
bm-linux-gnu/11/include/g++-v11/bits/stl_iterator.h:1028
#1 0x3ff8444293f in std::vector<c10::Argument, std::allocator<c10::Argument> >::begin() const /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_vector.h:821
#2 0x3ff84d807d1 in torch::jit::toPyObject(c10::IValue) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:617
ROCm#3 0x3ff84d80305 in torch::jit::toPyObject(c10::IValue) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:604
ROCm#4 0x3ff84856871 in pybind11::detail::type_caster<c10::IValue, void>::cast(c10::IValue, pybind11::return_value_policy, pybind11::handle) /home/user/pytorch/torch/csrc/jit/python/pybind.h:138
ROCm#5 0x3ff85318191 in pybind11::cpp_function::initialize<torch::jit::initJitScriptBindings(_object*)::$_45, c10::IValue, torch::jit::mobile::Module&, pybind11::tuple const&, pybind11::name, pybind11::is
_method, pybind11::sibling, pybind11::arg>(torch::jit::initJitScriptBindings(_object*)::$_45&&, c10::IValue (*)(torch::jit::mobile::Module&, pybind11::tuple const&), pybind11::name const&, pybind11::is_me
thod const&, pybind11::sibling const&, pybind11::arg const&)::{lambda(pybind11::detail::function_call&)#1}::operator()(pybind11::detail::function_call&) const /home/user/pytorch/cmake/../third_party/pybin
d11/include/pybind11/pybind11.h:249
ROCm#6 0x3ff85317cfd in pybind11::cpp_function::initialize<torch::jit::initJitScriptBindings(_object*)::$_45, c10::IValue, torch::jit::mobile::Module&, pybind11::tuple const&, pybind11::name, pybind11::is
_method, pybind11::sibling, pybind11::arg>(torch::jit::initJitScriptBindings(_object*)::$_45&&, c10::IValue (*)(torch::jit::mobile::Module&, pybind11::tuple const&), pybind11::name const&, pybind11::is_me
thod const&, pybind11::sibling const&, pybind11::arg const&)::{lambda(pybind11::detail::function_call&)#1}::__invoke(pybind11::detail::function_call&) /home/user/pytorch/cmake/../third_party/pybind11/incl
ude/pybind11/pybind11.h:224
ROCm#7 0x3ff82ee52e9 in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) /home/user/pytorch/cmake/../third_party/pybind11/include/pybind11/pybind11.h:929
ROCm#8 0x3ffab002903 in cfunction_call Objects/methodobject.c:543
ROCm#9 0x3ffaaf8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#10 0x3ffaaf8e919 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#11 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#12 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#13 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#14 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#15 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#16 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#17 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#18 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#19 0x3ffaaf8a615 in _PyObject_FastCallDictTstate Objects/call.c:142
ROCm#20 0x3ffaaf8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#21 0x3ffab03f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#22 0x3ffaaf8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#23 0x3ffab0f0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#24 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#25 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#26 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#27 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#28 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#29 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#30 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#31 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#32 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#33 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#34 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#35 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#36 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#37 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#38 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#39 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#40 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#41 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#42 0x3ffab0ff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#43 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#44 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#45 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#46 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#47 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#48 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#49 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#50 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#51 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#52 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#53 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#54 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#55 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#56 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#57 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#58 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#59 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#60 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#61 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#62 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#63 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#64 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#65 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#66 0x3ffaaf8ab9b in PyVectorcall_Call Objects/call.c:267
ROCm#67 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
ROCm#68 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#69 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#70 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#71 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#72 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#73 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#74 0x3ffaaf8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#75 0x3ffaaf8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#76 0x3ffab03f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#77 0x3ffaaf8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#78 0x3ffab0f0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#79 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#80 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#81 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#82 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#83 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#84 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#85 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#86 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#87 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#88 0x3ffab0ff7d7 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#89 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#90 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#91 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#92 0x3ffaaf8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#93 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
ROCm#94 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#95 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#96 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#97 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#98 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#99 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#100 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#101 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#102 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#103 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#104 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#105 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#106 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#107 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#108 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#109 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#110 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#111 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#112 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#113 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#114 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#115 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#116 0x3ffaaf8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#117 0x3ffaaf8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#118 0x3ffab03f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#119 0x3ffaaf8ad17 in _PyObject_Call Objects/call.c:305
ROCm#120 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#121 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#122 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#123 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#124 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#125 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#126 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#127 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#128 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#129 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#130 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#131 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#132 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#133 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#134 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#135 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#136 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#137 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#138 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#139 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#140 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#141 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#142 0x3ffaaf8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#143 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
ROCm#144 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#145 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#146 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#147 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#148 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#149 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#150 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#151 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#152 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#153 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#154 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#155 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#156 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#157 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#158 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#159 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#160 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#161 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#162 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#163 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#164 0x3ffaaf8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#165 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
ROCm#166 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#167 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#168 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#169 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#170 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#171 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#172 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#173 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#174 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#175 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#176 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#177 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#178 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#179 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#180 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#181 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#182 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#183 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#184 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#185 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#186 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#187 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#188 0x3ffaaf8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#189 0x3ffaaf8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#190 0x3ffab03f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#191 0x3ffaaf8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#192 0x3ffab0f0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#193 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#194 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#195 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#196 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#197 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#198 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#199 0x3ffaaf8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#200 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
ROCm#201 0x3ffaaf8ada9 in PyObject_Call Objects/call.c:317
ROCm#202 0x3ffab1059c7 in do_call_core Python/ceval.c:5943
ROCm#203 0x3ffab0ffd39 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#204 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#205 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#206 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#207 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#208 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#209 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#210 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#211 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#212 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#213 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#214 0x3ffaaf8e941 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#215 0x3ffaaf8eddd in method_vectorcall Objects/classobject.c:53
ROCm#216 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#216 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#217 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#218 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#219 0x3ffab0ff779 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#220 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#221 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#222 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#223 0x3ffaaf8a695 in _PyObject_FastCallDictTstate Objects/call.c:153
ROCm#224 0x3ffaaf8b271 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#225 0x3ffab03f307 in slot_tp_call Objects/typeobject.c:7494
ROCm#226 0x3ffaaf8a933 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#227 0x3ffab0f0081 in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#228 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#229 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#230 0x3ffab0ffa57 in _PyEval_EvalFrameDefault Python/ceval.c:4231
ROCm#231 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#232 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#233 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#234 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#235 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#236 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#237 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#238 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#239 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#240 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#241 0x3ffab0f00a9 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#242 0x3ffab0f013d in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#243 0x3ffab105447 in call_function Python/ceval.c:5891
ROCm#244 0x3ffab0ff905 in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#245 0x3ffab0f052b in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#246 0x3ffab102b67 in _PyEval_Vector Python/ceval.c:5065
ROCm#247 0x3ffaaf8aec1 in _PyFunction_Vectorcall Objects/call.c:342
ROCm#248 0x3ffaaf8ab15 in PyVectorcall_Call Objects/call.c:255
ROCm#249 0x3ffaaf8ac65 in _PyObject_Call Objects/call.c:290
0x60d0005a5790 is located 80 bytes inside of 136-byte region [0x60d0005a5740,0x60d0005a57c8)
freed by thread T0 here:
#0 0x3ffab537de5 in operator delete(void*) /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:160
#1 0x3ff55984fdb in __gnu_cxx::new_allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> >::deallocate(std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2>*, unsigned long) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/ext/new_allocator.h:145
previously allocated by thread T0 here:
#0 0x3ffab53734f in operator new(unsigned long) /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:99
#1 0x3ff5598443f in __gnu_cxx::new_allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> >::allocate(unsigned long, void const*) /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/ext/new_allocator.h:127
#2 0x3fff5849ecf ([stack]+0xb2ecf)
SUMMARY: AddressSanitizer: heap-use-after-free /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/stl_iterator.h:1028 in __gnu_cxx::__normal_iterator<c10::Argument const*, std::vector<c10::Argument, std::allocator<c10::Argument> > >::__normal_iterator(c10::Argument const* const&)
Shadow bytes around the buggy address:
0x100c1a000b4aa0: fd fd fd fd fd fd fd fd fd fd fd fa fa fa fa fa
0x100c1a000b4ab0: fa fa fa fa fd fd fd fd fd fd fd fd fd fd fd fd
0x100c1a000b4ac0: fd fd fd fd fd fa fa fa fa fa fa fa fa fa fd fd
0x100c1a000b4ad0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fa
0x100c1a000b4ae0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd
=>0x100c1a000b4af0: fd fd[fd]fd fd fd fd fd fd fa fa fa fa fa fa fa
0x100c1a000b4b00: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x100c1a000b4b10: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x100c1a000b4b20: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x100c1a000b4b30: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x100c1a000b4b40: 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
Shadow gap: cc
==1134126==ABORTING
```
Additional backtraces (not full):
Allocation:
```
#0 __memset_z196 () at ../sysdeps/s390/memset-z900.S:144
#1 0x000003ff96f3072a in __asan::Allocator::Allocate (this=this@entry=0x3ff97041eb8 <__asan::instance>, size=size@entry=136, alignment=8, alignment@entry=0, stack=<optimized out>,
stack@entry=0x3ffdbb45d78, alloc_type=<optimized out>, can_fill=true) at /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_allocator.cpp:599
#2 0x000003ff96f2c088 in __asan::asan_memalign (alignment=alignment@entry=0, size=size@entry=136, stack=stack@entry=0x3ffdbb45d78, alloc_type=alloc_type@entry=__asan::FROM_NEW)
at /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_allocator.cpp:1039
ROCm#3 0x000003ff96fb73b0 in operator new (size=136) at /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:99
ROCm#4 0x000003ff41404440 in __gnu_cxx::new_allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> >::allocate (this=0x3ffdbb468c0,
__n=1) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/ext/new_allocator.h:127
ROCm#5 0x000003ff414042a0 in std::allocator_traits<std::allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> > >::allocate (__a=...,
__n=1) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/alloc_traits.h:464
ROCm#6 0x000003ff41403b66 in std::__allocate_guarded<std::allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> > > (__a=...)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/allocated_ptr.h:98
ROCm#7 0x000003ff4140372a in std::__shared_count<(__gnu_cxx::_Lock_policy)2>::__shared_count<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::vector<c10::Argument, std::allocator<c10::Argument> >, std::vector<c10::Argument, std::allocator<c10::Argument> > > (this=0x3ffdbb47888, __p=@0x3ffdbb47880: 0x0, __a=..., __args=..., __args=..., __args=..., __args=...)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:648
ROCm#8 0x000003ff41403328 in std::__shared_ptr<c10::FunctionSchema, (__gnu_cxx::_Lock_policy)2>::__shared_ptr<std::allocator<c10::FunctionSchema>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::vector<c10::Argument, std::allocator<c10::Argument> >, std::vector<c10::Argument, std::allocator<c10::Argument> > > (this=0x3ffdbb47880, __tag=..., __args=..., __args=..., __args=..., __args=...) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:1342
ROCm#9 0x000003ff41402f06 in std::shared_ptr<c10::FunctionSchema>::shared_ptr<std::allocator<c10::FunctionSchema>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::vector<c10::Argument, std::allocator<c10::Argument> >, std::vector<c10::Argument, std::allocator<c10::Argument> > > (
this=0x3ffdbb47880, __tag=..., __args=..., __args=..., __args=..., __args=...) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr.h:409
ROCm#10 0x000003ff41402b6e in std::allocate_shared<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::vector<c10::Argument, std::allocator<c10::Argument> >, std::vector<c10::Argument, std::allocator<c10::Argument> > > (__a=...,
__args=..., __args=..., __args=..., __args=...) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr.h:862
ROCm#11 0x000003ff4140215c in std::make_shared<c10::FunctionSchema, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::vector<c10::Argument, std::allocator<c10::Argument> >, std::vector<c10::Argument, std::allocator<c10::Argument> > > (__args=..., __args=..., __args=..., __args=...)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr.h:878
ROCm#12 0x000003ff413d180c in c10::TupleType::createWithSpec<c10::basic_string_view<char> > (qualName=..., field_names=std::vector of length 1, capacity 1 = {...},
field_types=std::vector of length 1, capacity 1 = {...}, field_defaults=std::vector of length 0, capacity 0) at /home/user/pytorch/aten/src/ATen/core/type.cpp:769
ROCm#13 0x000003ff413b9ca6 in c10::TupleType::createNamed (qualName=..., field_names=std::vector of length 1, capacity 1 = {...}, field_types=std::vector of length 1, capacity 1 = {...})
at /home/user/pytorch/aten/src/ATen/core/type.cpp:725
ROCm#14 0x000003ff4115fbac in c10::ivalue::TupleTypeFactory<c10::TupleType>::fallback (type=...) at /home/user/pytorch/aten/src/ATen/core/dynamic_type.cpp:383
ROCm#15 0x000003ff708217fe in c10::ivalue::Tuple::type<c10::TupleType> (this=0x6080004b8520) at /home/user/pytorch/aten/src/ATen/core/ivalue_inl.h:781
ROCm#16 0x000003ff70800740 in torch::jit::toPyObject (ivalue=...) at /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:613
ROCm#17 0x000003ff70800306 in torch::jit::toPyObject (ivalue=...) at /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:604
ROCm#18 0x000003ff702d6872 in pybind11::detail::type_caster<c10::IValue, void>::cast (src=...) at /home/user/pytorch/torch/csrc/jit/python/pybind.h:138
ROCm#19 0x000003ff70d98192 in pybind11::cpp_function::initialize<torch::jit::initJitScriptBindings(_object*)::$_45, c10::IValue, torch::jit::mobile::Module&, pybind11::tuple const&, pybind11::name, pybind11::is_method, pybind11::sibling, pybind11::arg>(torch::jit::initJitScriptBindings(_object*)::$_45&&, c10::IValue (*)(torch::jit::mobile::Module&, pybind11::tuple const&), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&, pybind11::arg const&)::{lambda(pybind11::detail::function_call&)#1}::operator()(pybind11::detail::function_call&) const (this=0x3ffdbb4ca20, call=...)
at /home/user/pytorch/cmake/../third_party/pybind11/include/pybind11/pybind11.h:249
ROCm#20 0x000003ff70d97cfe in pybind11::cpp_function::initialize<torch::jit::initJitScriptBindings(_object*)::$_45, c10::IValue, torch::jit::mobile::Module&, pybind11::tuple const&, pybind11::name, pybind11::is_method, pybind11::sibling, pybind11::arg>(torch::jit::initJitScriptBindings(_object*)::$_45&&, c10::IValue (*)(torch::jit::mobile::Module&, pybind11::tuple const&), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&, pybind11::arg const&)::{lambda(pybind11::detail::function_call&)#1}::__invoke(pybind11::detail::function_call&) (call=...)
at /home/user/pytorch/cmake/../third_party/pybind11/include/pybind11/pybind11.h:224
ROCm#21 0x000003ff6e9652ea in pybind11::cpp_function::dispatcher (self=<PyCapsule at remote 0x3ff83e27720>,
args_in=(<torch._C.LiteScriptModule at remote 0x3ff811844b0>, (<Tensor at remote 0x3ff814efb00>,)), kwargs_in=0x0) at /home/user/pytorch/cmake/../third_party/pybind11/include/pybind11/pybind11.h:929
```
Deallocation:
```
#0 operator delete (ptr=0x60d0005a5740) at /var/tmp/portage/sys-devel/gcc-11.3.1_p20230303/work/gcc-11-20230303/libsanitizer/asan/asan_new_delete.cpp:160
#1 0x000003ff44904fdc in __gnu_cxx::new_allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> >::deallocate (this=0x3ffc5dc8020,
__p=0x60d0005a5740, __t=1) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/ext/new_allocator.h:145
#2 0x000003ff44904fa8 in std::allocator_traits<std::allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> > >::deallocate (
__a=..., __p=0x60d0005a5740, __n=1) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/alloc_traits.h:496
ROCm#3 0x000003ff449041f2 in std::__allocated_ptr<std::allocator<std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2> > >::~__allocated_ptr (
this=0x3ffc5dc8030) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/allocated_ptr.h:74
ROCm#4 0x000003ff44904888 in std::_Sp_counted_ptr_inplace<c10::FunctionSchema, std::allocator<c10::FunctionSchema>, (__gnu_cxx::_Lock_policy)2>::_M_destroy (this=0x60d0005a5740)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:538
ROCm#5 0x000003ff43895a62 in std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release (this=0x60d0005a5740) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:184
ROCm#6 0x000003ff43895420 in std::__shared_count<(__gnu_cxx::_Lock_policy)2>::~__shared_count (this=0x611000c40648) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:705
ROCm#7 0x000003ff4466e7f4 in std::__shared_ptr<c10::FunctionSchema, (__gnu_cxx::_Lock_policy)2>::~__shared_ptr (this=0x611000c40640)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:1154
ROCm#8 0x000003ff4466d820 in std::shared_ptr<c10::FunctionSchema>::~shared_ptr (this=0x611000c40640) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr.h:122
ROCm#9 0x000003ff448d82f6 in c10::TupleType::~TupleType (this=0x611000c40580) at /home/user/pytorch/aten/src/ATen/core/jit_type.h:1142
ROCm#10 0x000003ff448d8346 in c10::TupleType::~TupleType (this=0x611000c40580) at /home/user/pytorch/aten/src/ATen/core/jit_type.h:1142
ROCm#11 0x000003ff731296a4 in std::_Sp_counted_ptr<c10::TupleType*, (__gnu_cxx::_Lock_policy)2>::_M_dispose (this=0x603000c43ae0)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:348
ROCm#12 0x000003ff71eaf666 in std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release (this=0x603000c43ae0) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:168
ROCm#13 0x000003ff71eaf330 in std::__shared_count<(__gnu_cxx::_Lock_policy)2>::~__shared_count (this=0x3ffc5dc9368) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:705
ROCm#14 0x000003ff73129ee4 in std::__shared_ptr<c10::TupleType, (__gnu_cxx::_Lock_policy)2>::~__shared_ptr (this=0x3ffc5dc9360)
at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr_base.h:1154
ROCm#15 0x000003ff73122390 in std::shared_ptr<c10::TupleType>::~shared_ptr (this=0x3ffc5dc9360) at /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/shared_ptr.h:122
ROCm#16 0x000003ff73d00788 in torch::jit::toPyObject (ivalue=...) at /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:613
ROCm#17 0x000003ff73d00306 in torch::jit::toPyObject (ivalue=...) at /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:604
```
</details>
Pull Request resolved: pytorch#101400
Approved by: https://github.com/zou3519
lcskrishna
pushed a commit
to lcskrishna/pytorch
that referenced
this pull request
May 29, 2023
3 disabled functions are attempting out of bounds reads. Disable them until sleef library is fixed.
<details>
<summary>ASAN report</summary>
```
=================================================================
==2030580==ERROR: AddressSanitizer: global-buffer-overflow on address 0x03ff70f54570 at pc 0x03ff6704e960 bp 0x03ffce128940 sp 0x03ffce128930
READ of size 4 at 0x03ff70f54570 thread T0
#0 0x3ff6704e95f in vgather_vf_p_vi2 /home/user/pytorch/third_party/sleef/src/arch/helpers390x_128.h:129
ROCm#1 0x3ff6704e95f in rempif /home/user/pytorch/third_party/sleef/src/libm/sleefsimdsp.c:550
ROCm#2 0x3ff6704e95f in Sleef_cosf4_u10vxe2 /home/user/pytorch/third_party/sleef/src/libm/sleefsimdsp.c:1021
ROCm#3 0x3ff67029cfb in Sleef_cosf4_u10 /home/user/pytorch/build/sleef/src/libm/disps390x_128.c:182
ROCm#4 0x3ff55d21941 in at::vec::ZVECTOR::Vectorized<float, void> at::vec::ZVECTOR::Vectorized<float, void>::mapSleef<float __vector(4) const (*)(float __vector(4)), double __vector(2) const (*)(double __
vector(2)), float, 0>(float __vector(4) const (*)(float __vector(4)), double __vector(2) const (*)(double __vector(2))) const /home/user/pytorch/aten/src/ATen/cpu/vec/vec256/zarch/vec256_zarch.h:991
ROCm#5 0x3ff5689ad01 in at::vec::ZVECTOR::Vectorized<float, void>::cos() const /home/user/pytorch/aten/src/ATen/cpu/vec/vec256/zarch/vec256_zarch.h:1074
ROCm#6 0x3ff5685df97 in at::vml::ZVECTOR::vcos<float>(float*, float const*, long)::{lambda(at::vec::ZVECTOR::Vectorized<float, void>)ROCm#1}::operator()(at::vec::ZVECTOR::Vectorized<float, void>) const /home/
user/pytorch/aten/src/ATen/cpu/vml.h:71
ROCm#7 0x3ff5689b691 in void at::vec::map<float, at::vml::ZVECTOR::vcos<float>(float*, float const*, long)::{lambda(at::vec::ZVECTOR::Vectorized<float, void>)ROCm#1}, 0>(at::vml::ZVECTOR::vcos<float>(float*,
float const*, long)::{lambda(at::vec::ZVECTOR::Vectorized<float, void>)ROCm#1} const&, float*, float const*, long) /home/user/pytorch/aten/src/ATen/cpu/vec/functional_base.h:239
ROCm#8 0x3ff5685e0df in void at::vml::ZVECTOR::vcos<float>(float*, float const*, long) /home/user/pytorch/aten/src/ATen/cpu/vml.h:71
ROCm#9 0x3ff563fdde3 in operator() /home/user/pytorch/aten/src/ATen/native/cpu/UnaryOpsKernel.cpp:770
ROCm#10 0x3ff5648e4a3 in operator() /home/user/pytorch/aten/src/ATen/TensorIterator.h:406
ROCm#11 0x3ff5663cae1 in callback_fn<at::TensorIteratorBase::loop_2d_from_1d<at::native::ZVECTOR::cos_kernel(at::TensorIteratorBase&)::<lambda()>::<lambda()>::<lambda(char**, const int64_t*, int64_t)> >(c
onst at::native::ZVECTOR::cos_kernel(at::TensorIteratorBase&)::<lambda()>::<lambda()>::<lambda(char**, const int64_t*, int64_t)>&)::<lambda(char**, const int64_t*, int64_t, int64_t)> > /home/user/pytorch/
c10/util/FunctionRef.h:43
ROCm#12 0x3ff4d45a933 in c10::function_ref<void (char**, long const*, long, long)>::operator()(char**, long const*, long, long) const /home/user/pytorch/c10/util/FunctionRef.h:64
ROCm#13 0x3ff4d455133 in at::internal::serial_for_each(c10::ArrayRef<long>, c10::ArrayRef<long>, char**, unsigned long, c10::function_ref<void (char**, long const*, long, long)>, at::Range) /home/user/pyt
orch/aten/src/ATen/TensorIteratorInternal.h:52
ROCm#14 0x3ff4d43b703 in at::TensorIteratorBase::serial_for_each(c10::function_ref<void (char**, long const*, long, long)>, at::Range) const /home/user/pytorch/aten/src/ATen/TensorIterator.cpp:777
ROCm#15 0x3ff4d43ab59 in at::TensorIteratorBase::for_each(c10::function_ref<void (char**, long const*, long, long)>, long) /home/user/pytorch/aten/src/ATen/TensorIterator.cpp:749
ROCm#16 0x3ff5648e851 in for_each<at::native::ZVECTOR::cos_kernel(at::TensorIteratorBase&)::<lambda()>::<lambda()>::<lambda(char**, const int64_t*, int64_t)> > /home/user/pytorch/aten/src/ATen/TensorItera
tor.h:421
ROCm#17 0x3ff563fe5f9 in operator() /home/user/pytorch/aten/src/ATen/native/cpu/UnaryOpsKernel.cpp:770
ROCm#18 0x3ff56400915 in operator() /home/user/pytorch/aten/src/ATen/native/cpu/UnaryOpsKernel.cpp:770
ROCm#19 0x3ff56400f1d in at::native::ZVECTOR::cos_kernel(at::TensorIteratorBase&) /home/user/pytorch/aten/src/ATen/native/cpu/UnaryOpsKernel.cpp:770
ROCm#20 0x3ff4f303007 in void at::native::DispatchStub<void (*)(at::TensorIteratorBase&), at::native::cos_stub>::operator()<at::native::structured_cos_out&>(c10::DeviceType, at::native::structured_cos_out
&) /home/user/pytorch/aten/src/ATen/native/DispatchStub.h:158
ROCm#21 0x3ff4f2edb3f in at::native::structured_cos_out::impl(at::Tensor const&, at::Tensor const&) /home/user/pytorch/aten/src/ATen/native/UnaryOps.cpp:330
ROCm#22 0x3ff526ef739 in wrapper_CPU_cos /home/user/pytorch/build/aten/src/ATen/RegisterCPU.cpp:4307
ROCm#23 0x3ff52c651d9 in operator() /home/user/pytorch/aten/src/ATen/core/boxing/impl/WrapFunctionIntoFunctor.h:13
ROCm#24 0x3ff52c651d9 in call /home/user/pytorch/aten/src/ATen/core/boxing/impl/make_boxed_from_unboxed_functor.h:463
ROCm#25 0x3ff5076df2f in at::Tensor c10::callUnboxedKernelFunction<at::Tensor, at::Tensor const&>(void*, c10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&) /home/user/pytorch/aten/src/ATen/core
/boxing/KernelFunction_impl.h:50
ROCm#26 0x3ff5009a93f in at::Tensor c10::KernelFunction::call<at::Tensor, at::Tensor const&>(c10::OperatorHandle const&, c10::DispatchKeySet, at::Tensor const&) const /home/user/pytorch/aten/src/ATen/core
/boxing/KernelFunction_impl.h:103
ROCm#27 0x3ff5009a93f in at::Tensor c10::Dispatcher::call<at::Tensor, at::Tensor const&>(c10::TypedOperatorHandle<at::Tensor (at::Tensor const&)> const&, at::Tensor const&) const /home/user/pytorch/aten/s
rc/ATen/core/dispatch/Dispatcher.h:639
ROCm#28 0x3ff5009a93f in c10::TypedOperatorHandle<at::Tensor (at::Tensor const&)>::call(at::Tensor const&) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:487
ROCm#29 0x3ff5009a93f in at::_ops::cos::call(at::Tensor const&) /home/user/pytorch/build/aten/src/ATen/Operators_0.cpp:2215
ROCm#30 0x3ff7d813741 in at::Tensor::cos() const /home/user/pytorch/build/aten/src/ATen/core/TensorBody.h:2107
ROCm#31 0x3ff7dc0f2b7 in operator() /home/user/pytorch/torch/csrc/autograd/generated/python_torch_functions_2.cpp:2953
ROCm#32 0x3ff7dc0faf7 in THPVariable_cos /home/user/pytorch/torch/csrc/autograd/generated/python_torch_functions_2.cpp:2955
ROCm#33 0x3ffa5ef5ae1 in cfunction_call Objects/methodobject.c:543
ROCm#34 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#35 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#36 0x3ffa5feb50d in do_call_core Python/ceval.c:5915
ROCm#37 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#38 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#39 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#40 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#41 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#42 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#43 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#44 0x3ff7f87a393 in torch::impl::dispatch::PythonKernelHolder::operator()(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) /home/user/pytorch/
torch/csrc/utils/python_dispatch.cpp:175
ROCm#45 0x3ff7f8871a7 in c10::BoxedKernel::makeFromFunctor<torch::impl::dispatch::PythonKernelHolder>(std::unique_ptr<torch::impl::dispatch::PythonKernelHolder, std::default_delete<torch::impl::dispatch::
PythonKernelHolder> >)::{lambda(c10::OperatorKernel*, c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*)ROCm#1}::operator()(c10::OperatorKernel*, c10::Op
eratorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/boxing/BoxedKernel_impl.h:87
ROCm#46 0x3ff7f887261 in c10::BoxedKernel::makeFromFunctor<torch::impl::dispatch::PythonKernelHolder>(std::unique_ptr<torch::impl::dispatch::PythonKernelHolder, std::default_delete<torch::impl::dispatch::
PythonKernelHolder> >)::{lambda(c10::OperatorKernel*, c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*)ROCm#1}::_FUN(c10::OperatorKernel*, c10::Operator
Handle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) /home/user/pytorch/aten/src/ATen/core/boxing/BoxedKernel_impl.h:86
ROCm#47 0x3ff7e0d10ab in c10::BoxedKernel::callBoxed(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/b
oxing/BoxedKernel_impl.h:41
ROCm#48 0x3ff7e0d1459 in c10::KernelFunction::callBoxed(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/cor
e/boxing/KernelFunction_impl.h:43
ROCm#49 0x3ff7f876421 in c10::Dispatcher::callBoxed(c10::OperatorHandle const&, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:6
91
ROCm#50 0x3ff4d22bcdd in c10::OperatorHandle::callBoxed(std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:417
ROCm#51 0x3ff65a092d5 in c10::OperatorHandle::callBoxed(std::vector<c10::IValue, std::allocator<c10::IValue> >&) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:421
ROCm#52 0x3ff65a05641 in operator() /home/user/pytorch/torch/csrc/jit/runtime/register_c10_ops.cpp:15
ROCm#53 0x3ff65a08cb5 in __invoke_impl<void, torch::jit::(anonymous namespace)::createOperatorFromC10(const c10::OperatorHandle&)::<lambda(torch::jit::Stack&)>&, std::vector<c10::IValue, std::allocator<c1
0::IValue> >&> /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/invoke.h:61
ROCm#54 0x3ff65a0897b in __invoke_r<void, torch::jit::(anonymous namespace)::createOperatorFromC10(const c10::OperatorHandle&)::<lambda(torch::jit::Stack&)>&, std::vector<c10::IValue, std::allocator<c10::
IValue> >&> /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/invoke.h:111
ROCm#55 0x3ff65a084e1 in _M_invoke /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/std_function.h:290
ROCm#56 0x3ff7eb2cb21 in std::function<void (std::vector<c10::IValue, std::allocator<c10::IValue> >&)>::operator()(std::vector<c10::IValue, std::allocator<c10::IValue> >&) const /usr/lib/gcc/s390x-ibm-lin
ux-gnu/11/include/g++-v11/bits/std_function.h:590
ROCm#57 0x3ff7eb1b659 in torch::jit::Operation::operator()(std::vector<c10::IValue, std::allocator<c10::IValue> >&) /home/user/pytorch/aten/src/ATen/core/stack.h:41
ROCm#58 0x3ff7eb08449 in torch::jit::invokeOperatorFromPython(std::vector<std::shared_ptr<torch::jit::Operator>, std::allocator<std::shared_ptr<torch::jit::Operator> > > const&, pybind11::args, pybind11::
kwargs const&, c10::optional<c10::DispatchKey>) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:764
ROCm#59 0x3ff7eb09d85 in torch::jit::_get_operation_for_overload_or_packet(std::vector<std::shared_ptr<torch::jit::Operator>, std::allocator<std::shared_ptr<torch::jit::Operator> > > const&, c10::Symbol,
pybind11::args, pybind11::kwargs const&, bool, c10::optional<c10::DispatchKey>) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:829
ROCm#60 0x3ff7e573eb9 in operator() /home/user/pytorch/torch/csrc/jit/python/init.cpp:1549
ROCm#61 0x3ff7e6728dd in call_impl<pybind11::object, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&, 0, 1, pybind11::detail::vo
id_type> /home/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1439
ROCm#62 0x3ff7e64312f in call<pybind11::object, pybind11::detail::void_type, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&> /h
ome/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1408
ROCm#63 0x3ff7e5da259 in operator() /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:249
ROCm#64 0x3ff7e5da441 in _FUN /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:224
ROCm#65 0x3ff7d317a1f in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:929
ROCm#66 0x3ffa5ef5ae1 in cfunction_call Objects/methodobject.c:543
ROCm#67 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#68 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#69 0x3ffa5feb50d in do_call_core Python/ceval.c:5915
ROCm#70 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#71 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#72 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#73 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#74 0x3ffa5e83d1f in _PyObject_FastCallDictTstate Objects/call.c:142
ROCm#75 0x3ffa5e84937 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#76 0x3ffa5f2f577 in slot_tp_call Objects/typeobject.c:7494
ROCm#77 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#78 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#79 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#80 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#81 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#82 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#83 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#84 0x3ffa5fd76a3 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#85 0x3ffa5fd772f in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#86 0x3ffa5feb289 in call_function Python/ceval.c:5891
ROCm#87 0x3ffa5fe5c3b in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#88 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#89 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#90 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#91 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#92 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#93 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#94 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#95 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#96 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#97 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#98 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#99 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#100 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#101 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#102 0x3ff7f87a393 in torch::impl::dispatch::PythonKernelHolder::operator()(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) /home/user/pytorch
/torch/csrc/utils/python_dispatch.cpp:175
ROCm#103 0x3ff7f8871a7 in c10::BoxedKernel::makeFromFunctor<torch::impl::dispatch::PythonKernelHolder>(std::unique_ptr<torch::impl::dispatch::PythonKernelHolder, std::default_delete<torch::impl::dispatch:
:PythonKernelHolder> >)::{lambda(c10::OperatorKernel*, c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*)ROCm#1}::operator()(c10::OperatorKernel*, c10::O
peratorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/boxing/BoxedKernel_impl.h:87
ROCm#104 0x3ff7f887261 in c10::BoxedKernel::makeFromFunctor<torch::impl::dispatch::PythonKernelHolder>(std::unique_ptr<torch::impl::dispatch::PythonKernelHolder, std::default_delete<torch::impl::dispatch:
:PythonKernelHolder> >)::{lambda(c10::OperatorKernel*, c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*)ROCm#1}::_FUN(c10::OperatorKernel*, c10::Operato
rHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) /home/user/pytorch/aten/src/ATen/core/boxing/BoxedKernel_impl.h:86
ROCm#105 0x3ff7e0d10ab in c10::BoxedKernel::callBoxed(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/
boxing/BoxedKernel_impl.h:41
ROCm#106 0x3ff7e0d1459 in c10::KernelFunction::callBoxed(c10::OperatorHandle const&, c10::DispatchKeySet, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/co
re/boxing/KernelFunction_impl.h:43
ROCm#107 0x3ff7f876421 in c10::Dispatcher::callBoxed(c10::OperatorHandle const&, std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:
691
ROCm#108 0x3ff4d22bcdd in c10::OperatorHandle::callBoxed(std::vector<c10::IValue, std::allocator<c10::IValue> >*) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:417
ROCm#109 0x3ff65a092d5 in c10::OperatorHandle::callBoxed(std::vector<c10::IValue, std::allocator<c10::IValue> >&) const /home/user/pytorch/aten/src/ATen/core/dispatch/Dispatcher.h:421
ROCm#110 0x3ff65a05641 in operator() /home/user/pytorch/torch/csrc/jit/runtime/register_c10_ops.cpp:15
ROCm#111 0x3ff65a08cb5 in __invoke_impl<void, torch::jit::(anonymous namespace)::createOperatorFromC10(const c10::OperatorHandle&)::<lambda(torch::jit::Stack&)>&, std::vector<c10::IValue, std::allocator<c
10::IValue> >&> /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/invoke.h:61
ROCm#112 0x3ff65a0897b in __invoke_r<void, torch::jit::(anonymous namespace)::createOperatorFromC10(const c10::OperatorHandle&)::<lambda(torch::jit::Stack&)>&, std::vector<c10::IValue, std::allocator<c10:
:IValue> >&> /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/invoke.h:111
ROCm#113 0x3ff65a084e1 in _M_invoke /usr/lib/gcc/s390x-ibm-linux-gnu/11/include/g++-v11/bits/std_function.h:290
ROCm#114 0x3ff7eb2cb21 in std::function<void (std::vector<c10::IValue, std::allocator<c10::IValue> >&)>::operator()(std::vector<c10::IValue, std::allocator<c10::IValue> >&) const /usr/lib/gcc/s390x-ibm-li
nux-gnu/11/include/g++-v11/bits/std_function.h:590
ROCm#115 0x3ff7eb1b659 in torch::jit::Operation::operator()(std::vector<c10::IValue, std::allocator<c10::IValue> >&) /home/user/pytorch/aten/src/ATen/core/stack.h:41
ROCm#116 0x3ff7eb08449 in torch::jit::invokeOperatorFromPython(std::vector<std::shared_ptr<torch::jit::Operator>, std::allocator<std::shared_ptr<torch::jit::Operator> > > const&, pybind11::args, pybind11:
:kwargs const&, c10::optional<c10::DispatchKey>) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:764
ROCm#117 0x3ff7eb09d85 in torch::jit::_get_operation_for_overload_or_packet(std::vector<std::shared_ptr<torch::jit::Operator>, std::allocator<std::shared_ptr<torch::jit::Operator> > > const&, c10::Symbol,
pybind11::args, pybind11::kwargs const&, bool, c10::optional<c10::DispatchKey>) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:829
ROCm#118 0x3ff7e573eb9 in operator() /home/user/pytorch/torch/csrc/jit/python/init.cpp:1549
ROCm#119 0x3ff7e6728dd in call_impl<pybind11::object, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&, 0, 1, pybind11::detail::v
oid_type> /home/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1439
ROCm#120 0x3ff7e64312f in call<pybind11::object, pybind11::detail::void_type, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&> /
home/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1408
ROCm#121 0x3ff7e5da259 in operator() /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:249
ROCm#122 0x3ff7e5da441 in _FUN /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:224
ROCm#123 0x3ff7d317a1f in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:929
ROCm#124 0x3ffa5ef5ae1 in cfunction_call Objects/methodobject.c:543
ROCm#125 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#126 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#127 0x3ffa5feb50d in do_call_core Python/ceval.c:5915
ROCm#128 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#129 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#130 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#131 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#132 0x3ffa5e83d1f in _PyObject_FastCallDictTstate Objects/call.c:142
ROCm#133 0x3ffa5e84937 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#134 0x3ffa5f2f577 in slot_tp_call Objects/typeobject.c:7494
ROCm#135 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#136 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#137 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#138 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#139 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#140 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#141 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#142 0x3ffa5e87d2b in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#143 0x3ffa5e882dd in method_vectorcall Objects/classobject.c:83
ROCm#144 0x3ffa5e836d3 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#145 0x3ffa5e84b6f in _PyObject_CallFunctionVa Objects/call.c:485
ROCm#146 0x3ffa5e84f2d in callmethod Objects/call.c:557
ROCm#147 0x3ffa5e85039 in PyObject_CallMethod Objects/call.c:577
ROCm#148 0x3ff7f7efa05 in torch::handle_torch_function_no_python_arg_parser(c10::ArrayRef<pybind11::handle>, _object*, _object*, char const*, _object*, char const*, torch::TorchFunctionName) /home/user/py
torch/torch/csrc/utils/python_arg_parser.cpp:338
ROCm#149 0x3ff7eb09b67 in torch::jit::_get_operation_for_overload_or_packet(std::vector<std::shared_ptr<torch::jit::Operator>, std::allocator<std::shared_ptr<torch::jit::Operator> > > const&, c10::Symbol,
pybind11::args, pybind11::kwargs const&, bool, c10::optional<c10::DispatchKey>) /home/user/pytorch/torch/csrc/jit/python/pybind_utils.cpp:827
ROCm#150 0x3ff7e573eb9 in operator() /home/user/pytorch/torch/csrc/jit/python/init.cpp:1549
ROCm#151 0x3ff7e6728dd in call_impl<pybind11::object, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&, 0, 1, pybind11::detail::v
oid_type> /home/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1439
ROCm#152 0x3ff7e64312f in call<pybind11::object, pybind11::detail::void_type, torch::jit::initJITBindings(PyObject*)::<lambda(const string&, const string&)>::<lambda(pybind11::args, pybind11::kwargs)>&> /
home/user/pytorch/third_party/pybind11/include/pybind11/cast.h:1408
ROCm#153 0x3ff7e5da259 in operator() /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:249
ROCm#154 0x3ff7e5da441 in _FUN /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:224
ROCm#155 0x3ff7d317a1f in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) /home/user/pytorch/third_party/pybind11/include/pybind11/pybind11.h:929
ROCm#156 0x3ffa5ef5ae1 in cfunction_call Objects/methodobject.c:543
ROCm#157 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#158 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#159 0x3ffa5feb50d in do_call_core Python/ceval.c:5915
ROCm#160 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#161 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#162 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#163 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#164 0x3ffa5e83d1f in _PyObject_FastCallDictTstate Objects/call.c:142
ROCm#165 0x3ffa5e84937 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#166 0x3ffa5f2f577 in slot_tp_call Objects/typeobject.c:7494
ROCm#167 0x3ffa5e84027 in _PyObject_MakeTpCall Objects/call.c:215
ROCm#168 0x3ffa5fd767b in _PyObject_VectorcallTstate Include/cpython/abstract.h:112
ROCm#169 0x3ffa5fd772f in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#170 0x3ffa5feb289 in call_function Python/ceval.c:5891
ROCm#171 0x3ffa5fe5ad1 in _PyEval_EvalFrameDefault Python/ceval.c:4181
ROCm#172 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#173 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#174 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#175 0x3ffa5fd76a3 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#176 0x3ffa5fd772f in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#177 0x3ffa5feb289 in call_function Python/ceval.c:5891
ROCm#178 0x3ffa5fe5c3b in _PyEval_EvalFrameDefault Python/ceval.c:4213
ROCm#179 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#180 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#181 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#182 0x3ffa5e8427f in PyVectorcall_Call Objects/call.c:267
ROCm#183 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#184 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#185 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#186 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#187 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#188 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#189 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#190 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#191 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#192 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#193 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#194 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#195 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#196 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#197 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#198 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#199 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#200 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#201 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#202 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#203 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#204 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#205 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#206 0x3ffa5e841fb in PyVectorcall_Call Objects/call.c:255
ROCm#207 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#208 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#209 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#210 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#211 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#212 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#213 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#214 0x3ffa5e83d1f in _PyObject_FastCallDictTstate Objects/call.c:142
ROCm#215 0x3ffa5e84937 in _PyObject_Call_Prepend Objects/call.c:431
ROCm#216 0x3ffa5f2f577 in slot_tp_call Objects/typeobject.c:7494
ROCm#217 0x3ffa5e843f3 in _PyObject_Call Objects/call.c:305
ROCm#218 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#219 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#220 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#221 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#222 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#223 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#224 0x3ffa5fd76a3 in _PyObject_VectorcallTstate Include/cpython/abstract.h:114
ROCm#225 0x3ffa5fd772f in PyObject_Vectorcall Include/cpython/abstract.h:123
ROCm#226 0x3ffa5feb289 in call_function Python/ceval.c:5891
ROCm#227 0x3ffa5fe5b21 in _PyEval_EvalFrameDefault Python/ceval.c:4198
ROCm#228 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#229 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#230 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#231 0x3ffa5e8427f in PyVectorcall_Call Objects/call.c:267
ROCm#232 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#233 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#234 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#235 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#236 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#237 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#238 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#239 0x3ffa5e8427f in PyVectorcall_Call Objects/call.c:267
ROCm#240 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#241 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#242 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#243 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#244 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#245 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#246 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#247 0x3ffa5e8427f in PyVectorcall_Call Objects/call.c:267
ROCm#248 0x3ffa5e84347 in _PyObject_Call Objects/call.c:290
ROCm#249 0x3ffa5e84483 in PyObject_Call Objects/call.c:317
ROCm#250 0x3ffa5feb7cf in do_call_core Python/ceval.c:5943
ROCm#251 0x3ffa5fe6019 in _PyEval_EvalFrameDefault Python/ceval.c:4277
ROCm#252 0x3ffa5fd7aed in _PyEval_EvalFrame Include/internal/pycore_ceval.h:46
ROCm#253 0x3ffa5fe8ba9 in _PyEval_Vector Python/ceval.c:5065
ROCm#254 0x3ffa5e8459b in _PyFunction_Vectorcall Objects/call.c:342
ROCm#255 0x3ffa5e8427f in PyVectorcall_Call Objects/call.c:267
0x03ff70f54570 is located 0 bytes to the right of global variable 'Sleef_rempitabsp' defined in '/home/user/pytorch/third_party/sleef/src/libm/rempitab.c:986:34' (0x3ff70f53f00) of size 1648
SUMMARY: AddressSanitizer: global-buffer-overflow /home/user/pytorch/third_party/sleef/src/arch/helpers390x_128.h:129 in vgather_vf_p_vi2
Shadow bytes around the buggy address:
0x10007fee1ea850: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea860: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea870: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea880: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea890: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
=>0x10007fee1ea8a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00[f9]f9
0x10007fee1ea8b0: f9 f9 f9 f9 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea8c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea8d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea8e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x10007fee1ea8f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
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
Shadow gap: cc
==2030580==ABORTING
```
</details>
It reproduces when running `pytest -v test/test_ops.py -k test_python_ref__refs_cos_cpu_bfloat16` under address sanitizer on s390x.
See also: shibatch/sleef#464
Pull Request resolved: pytorch#102266
Approved by: https://github.com/malfet
akashveramd
pushed a commit
that referenced
this pull request
Jun 13, 2025
Summary: Add a 2D test to integration test suite Test Plan: ``` =====Integration test: CONFIG_FILE=./train_configs/debug_model.toml NGPU=4 ./run_llama_train.sh===== + export USE_LIBUV=1 + USE_LIBUV=1 + TRAINER_DIR=/home/gnadathur/local/torchtrain + NGPU=4 + LOG_RANK=0 + CONFIG_FILE=./train_configs/debug_model.toml + torchrun --nproc_per_node=4 --rdzv_endpoint=localhost:5972 --local-ranks-filter 0 --role rank --tee 3 train.py --job.config_file ./train_configs/debug_model.toml W0327 14:29:47.734000 140642626999296 torch/distributed/run.py:757] W0327 14:29:47.734000 140642626999296 torch/distributed/run.py:757] ***************************************** W0327 14:29:47.734000 140642626999296 torch/distributed/run.py:757] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. W0327 14:29:47.734000 140642626999296 torch/distributed/run.py:757] ***************************************** [rank0]:2024-03-27 14:29:49,466 - root - INFO - Starting job: LLaMA debug training [rank0]:2024-03-27 14:29:49,615 - root - WARNING - ENV[TORCH_NCCL_ASYNC_ERROR_HANDLING] = 1 will be overridden to 3 based on job config [rank0]:2024-03-27 14:29:49,621 - root - INFO - Building 1-D device mesh with ['dp'], [4] [rank0]:2024-03-27 14:29:49,623 - root - INFO - Building sentencepiece tokenizer locally from ./torchtrain/datasets/tokenizer/tokenizer.model [rank0]:2024-03-27 14:29:49,630 - root - INFO - SentencePieceTokenizer built: #words 32000, BOS ID 1, EOS ID 2 [rank0]:2024-03-27 14:29:49,630 - root - INFO - Preparing alpaca dataset from HuggingFace [rank0]:2024-03-27 14:29:51,114 - root - INFO - Building llama debugmodel with ModelArgs(dim=256, n_layers=2, n_heads=16, n_kv_heads=None, vocab_size=32000, multiple_of=256, ffn_dim_multiplier=None, norm_eps=1e-05, max_batch_size=32, max_seq_len=32768, depth_init=True) [rank0]:2024-03-27 14:29:51,124 - root - INFO - �[34mModel llama debugmodel �[31msize: 18,089,216 total parameters�[39m [rank0]:2024-03-27 14:29:51,124 - root - INFO - GPU capacity: NVIDIA H100 (0) with 95.04GiB memory [rank0]:2024-03-27 14:29:51,259 - root - INFO - Applied selective activation checkpointing to the model [rank0]:2024-03-27 14:29:51,259 - root - INFO - Applied FSDP to the model [rank0]:2024-03-27 14:29:51,284 - root - INFO - Model fully initialized via reset_parameters [rank0]:2024-03-27 14:29:51,284 - root - INFO - Gradient scaling not enabled [rank0]:2024-03-27 14:29:51,285 - root - INFO - Metrics logging active. Tensorboard logs will be saved at ./outputs/tb/20240327-1429 [rank0]:2024-03-27 14:29:52,056 - root - INFO - Profiling active. Traces will be saved at ./outputs/profiling/traces [rank0]:/data/users/gnadathur/a/pytorch/torch/utils/checkpoint.py:144: UserWarning: Tensor arguments, excluding CPU tensors, are detected on at least two types of devices. Device state will only be saved for devices of a single device type, and the remaining devices will be ignored. Consequently, if any checkpointed functions involve randomness, this may result in incorrect gradients. (Note that if CUDA devices are among the devices detected, it will be prioritized; otherwise, the first device encountered will be selected.) [rank0]: warnings.warn( [rank0]:2024-03-27 14:29:52,825 - root - INFO - �[36mstep: 1 �[32mloss: 10.7425 �[33mmemory: 9.42GiB(9.91%) �[34mwps: 21,337 �[35mmfu: 0.26%�[39m [rank0]:2024-03-27 14:29:52,825 - root - INFO - Synchronizing and adjusting timeout for all ProcessGroups to 0:00:05 [rank0]:2024-03-27 14:29:52,905 - root - INFO - �[36mstep: 2 �[32mloss: 10.6722 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 208,060 �[35mmfu: 2.55%�[39m [rank0]:2024-03-27 14:29:52,982 - root - INFO - �[36mstep: 3 �[32mloss: 10.5435 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 213,622 �[35mmfu: 2.62%�[39m [rank0]:2024-03-27 14:29:53,060 - root - INFO - �[36mstep: 4 �[32mloss: 10.3359 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 212,856 �[35mmfu: 2.61%�[39m [rank0]:2024-03-27 14:29:53,139 - root - INFO - �[36mstep: 5 �[32mloss: 10.0965 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 209,326 �[35mmfu: 2.56%�[39m [rank0]:2024-03-27 14:29:53,215 - root - INFO - �[36mstep: 6 �[32mloss: 9.8806 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 216,808 �[35mmfu: 2.66%�[39m [rank0]:2024-03-27 14:29:53,292 - root - INFO - �[36mstep: 7 �[32mloss: 9.6442 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 214,874 �[35mmfu: 2.63%�[39m [rank0]:2024-03-27 14:29:53,367 - root - INFO - �[36mstep: 8 �[32mloss: 9.4349 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 220,877 �[35mmfu: 2.70%�[39m [rank0]:2024-03-27 14:29:53,500 - root - INFO - �[36mstep: 9 �[32mloss: 9.2674 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 123,924 �[35mmfu: 1.52%�[39m [rank0]:[rank0]:[W327 14:29:53.248291822 CPUAllocator.cpp:249] Memory block of unknown size was allocated before the profiling started, profiler results will not include the deallocation event [rank0]:2024-03-27 14:29:53,577 - root - INFO - �[36mstep: 10 �[32mloss: 9.1404 �[33mmemory: 11.38GiB(11.97%) �[34mwps: 214,910 �[35mmfu: 2.63%�[39m [rank0]:NCCL version 2.20.5+cuda12.0 =====Integration test: CONFIG_FILE=./train_configs/debug_model_2d.toml NGPU=4 ./run_llama_train.sh===== + export USE_LIBUV=1 + USE_LIBUV=1 + TRAINER_DIR=/home/gnadathur/local/torchtrain + NGPU=4 + LOG_RANK=0 + CONFIG_FILE=./train_configs/debug_model_2d.toml + torchrun --nproc_per_node=4 --rdzv_endpoint=localhost:5972 --local-ranks-filter 0 --role rank --tee 3 train.py --job.config_file ./train_configs/debug_model_2d.toml W0327 14:29:58.902000 140021143774208 torch/distributed/run.py:757] W0327 14:29:58.902000 140021143774208 torch/distributed/run.py:757] ***************************************** W0327 14:29:58.902000 140021143774208 torch/distributed/run.py:757] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. W0327 14:29:58.902000 140021143774208 torch/distributed/run.py:757] ***************************************** [rank0]:2024-03-27 14:30:00,872 - root - INFO - Starting job: LLaMA debug training [rank0]:2024-03-27 14:30:01,177 - root - WARNING - ENV[TORCH_NCCL_ASYNC_ERROR_HANDLING] = 1 will be overridden to 3 based on job config [rank0]:2024-03-27 14:30:01,182 - root - INFO - Building 2-D device mesh with ['dp', 'tp'], [2, 2] [rank0]:2024-03-27 14:30:01,185 - root - INFO - Building sentencepiece tokenizer locally from ./torchtrain/datasets/tokenizer/tokenizer.model [rank0]:2024-03-27 14:30:01,194 - root - INFO - SentencePieceTokenizer built: #words 32000, BOS ID 1, EOS ID 2 [rank0]:2024-03-27 14:30:01,195 - root - INFO - Preparing alpaca dataset from HuggingFace [rank0]:2024-03-27 14:30:02,807 - root - INFO - Building llama debugmodel with ModelArgs(dim=256, n_layers=2, n_heads=16, n_kv_heads=None, vocab_size=32000, multiple_of=256, ffn_dim_multiplier=None, norm_eps=1e-05, max_batch_size=32, max_seq_len=32768, depth_init=True) [rank0]:2024-03-27 14:30:02,818 - root - INFO - �[34mModel llama debugmodel �[31msize: 18,089,216 total parameters�[39m [rank0]:2024-03-27 14:30:02,819 - root - INFO - GPU capacity: NVIDIA H100 (0) with 95.04GiB memory [rank0]:2024-03-27 14:30:02,830 - root - INFO - Applied Sequence Parallelism to the model [rank0]:2024-03-27 14:30:02,975 - root - INFO - Applied selective activation checkpointing to the model [rank0]:2024-03-27 14:30:02,975 - root - INFO - Applied FSDP to the model [rank0]:2024-03-27 14:30:03,004 - root - INFO - Model fully initialized via reset_parameters [rank0]:2024-03-27 14:30:03,004 - root - INFO - Gradient scaling not enabled [rank0]:2024-03-27 14:30:03,005 - root - INFO - Metrics logging active. Tensorboard logs will be saved at ./outputs/tb/20240327-1430 [rank0]:2024-03-27 14:30:03,642 - root - INFO - Profiling active. Traces will be saved at ./outputs/profiling/traces [rank0]:/data/users/gnadathur/a/pytorch/torch/utils/checkpoint.py:144: UserWarning: Tensor arguments, excluding CPU tensors, are detected on at least two types of devices. Device state will only be saved for devices of a single device type, and the remaining devices will be ignored. Consequently, if any checkpointed functions involve randomness, this may result in incorrect gradients. (Note that if CUDA devices are among the devices detected, it will be prioritized; otherwise, the first device encountered will be selected.) [rank0]: warnings.warn( [rank0]:2024-03-27 14:30:04,528 - root - INFO - �[36mstep: 1 �[32mloss: 10.8502 �[33mmemory: 5.71GiB(6.01%) �[34mwps: 9,259 �[35mmfu: 0.11%�[39m [rank0]:2024-03-27 14:30:04,528 - root - INFO - Synchronizing and adjusting timeout for all ProcessGroups to 0:00:05 [rank0]:2024-03-27 14:30:04,679 - root - INFO - �[36mstep: 2 �[32mloss: 10.7671 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 54,430 �[35mmfu: 0.67%�[39m [rank0]:2024-03-27 14:30:04,773 - root - INFO - �[36mstep: 3 �[32mloss: 10.6390 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 88,457 �[35mmfu: 1.08%�[39m [rank0]:2024-03-27 14:30:04,864 - root - INFO - �[36mstep: 4 �[32mloss: 10.4210 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 90,384 �[35mmfu: 1.11%�[39m [rank0]:2024-03-27 14:30:04,954 - root - INFO - �[36mstep: 5 �[32mloss: 10.1648 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 93,058 �[35mmfu: 1.14%�[39m [rank0]:2024-03-27 14:30:05,067 - root - INFO - �[36mstep: 6 �[32mloss: 9.9451 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 72,642 �[35mmfu: 0.89%�[39m [rank0]:2024-03-27 14:30:05,165 - root - INFO - �[36mstep: 7 �[32mloss: 9.7004 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 85,096 �[35mmfu: 1.04%�[39m [rank0]:2024-03-27 14:30:05,251 - root - INFO - �[36mstep: 8 �[32mloss: 9.4422 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 95,860 �[35mmfu: 1.17%�[39m [rank0]:2024-03-27 14:30:05,399 - root - INFO - �[36mstep: 9 �[32mloss: 9.2144 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 55,837 �[35mmfu: 0.68%�[39m [rank0]:[rank0]:[W327 14:30:05.148473462 CPUAllocator.cpp:249] Memory block of unknown size was allocated before the profiling started, profiler results will not include the deallocation event [rank0]:2024-03-27 14:30:05,496 - root - INFO - �[36mstep: 10 �[32mloss: 9.1710 �[33mmemory: 6.69GiB(7.04%) �[34mwps: 86,136 �[35mmfu: 1.05%�[39m [rank0]:NCCL version 2.20.5+cuda12.0 ``` Reviewers: Subscribers: Tasks: Tags: Co-authored-by: gnadathur <gnadathur@devvm4378.nao0.facebook.com>
jataylo
pushed a commit
to jataylo/pytorch
that referenced
this pull request
Jul 13, 2026
## Human Note This lets us return grads for scalar biases. I pusehd back on this in the path since there were workarounds although a lil ugly. But after prototyping i think its fine to land and easy enough to gork. We should wait for pytorch#188860 to land first since they touch teh same code -> how we create the captured buffer grads ## Agent Report # Report: attention-gym ROCm#171 FlexAttention learnable scalar ## Summary Reproduced the issue in this checkout using the exact scalar `score_mod` pattern: - Eager forward succeeded, eager backward failed with the reported `vmap(... out_dim is None)` BatchedTensor error. - Compiled forward succeeded, compiled backward failed in Inductor FlexAttention backward lowering with `AssertionError: ComputedBuffer name must not be None`. Implemented a PyTorch-side fix for direct 0-D captured score-mod gradients. The PR branch has since been rebased onto latest `origin/main` at `d2f0d442d70` and rebuilt locally with cuDNN enabled (`CUDNN_VERSION=9.23.0`, `USE_CUDNN=1`). Current job torch reports `2.14.0a0+git9fadffc`. Implemented changes: - `torch/_higher_order_ops/flex_attention.py` - Routes direct 0-D captured score-mod tensors through `mod_index(buffer, [])` while tracing the joint graph, so the existing captured-index custom autograd path emits `flex_lib::zeros_and_scatter([], [], grad)` for scalar gradient accumulation. - Preserves unused/no-grad captured buffers by returning `None` instead of copying `None` into an allocated grad buffer. - `torch/_dynamo/_trace_wrapped_higher_order_op.py` - Supports empty-index `zeros_and_scatter` as scalar accumulation by summing values into a scalar output. - Allows `ModIndex` to use an empty index list as a scalar identity whose backward is empty-index scalar accumulation. - Returns the checked 0-D tensor directly in the empty-index `ModIndex.forward` path. - Handles empty index metadata in its vmap rule with an annotation-only pyrefly fix, preserving the previous empty-list fallback semantics. - `torch/_inductor/kernel/flex/common.py` - Lowers empty-index `zeros_and_scatter` to a scalar atomic-add scatter for FlexAttention template subgraphs. - Uses the same compact empty-index guard shape as eager `zeros_and_scatter`. - `torch/_inductor/select_algorithm.py` - Normalizes scalar scatter indexes for Triton codegen and emits `tl.full(..., INDEX_DTYPE)` for constant scalar atomic indexes instead of invalid `tl.broadcast_to(0, ...)`. - `test/inductor/test_flex_attention.py` - Adds a parametrized CUDA float16 regression covering both direct learnable 0-D scalar gradients and detached/no-grad 0-D captures. ## Relationship to the prior unused-captured-grad issue The direct scalar learnable case is distinct from the existing `(1,)` indexed captured-scalar path: indexed captures already trace to `zeros_and_scatter([1], [idx], grad)`, while direct 0-D captures previously returned a plain computed captured grad that became batched under nested `vmap`. The local cleanup now makes direct 0-D captures enter the same `ModIndex` custom autograd path using an empty index list, instead of post-processing captured grad outputs after `create_joint`. I also checked a detached 0-D captured scalar (`score + temp.detach()`). Before the guard, eager backward tried to copy a `None` captured grad into an allocated buffer. The patch keeps that captured grad as `None`, and the new parametrized test checks eager and compiled behavior for this no-grad case. ## Validation Behavioral repro after rebasing onto `origin/main`: ```bash cd ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0 TORCHINDUCTOR_FORCE_DISABLE_CACHES=1 CUDA_LAUNCH_BLOCKING=1 ./.venv/bin/python pytorch/agent_space/repro_flex_scalar.py parity ``` Passed after the main rebase, after the local cleanup, and after the subagent-suggested simplifications. Final max diffs from eager-vs-compiled parity: - out: 0.00048828125 - q: 0.00048828125 - k: 0.00048828125 - v: 0.0 - temp: 0.005798816680908203 Targeted tests after rebasing onto `origin/main`: ```bash cd ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_0d_scalar_grad --verbose CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k unused_captured_score_mod_grad --verbose CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_scalar_grad --verbose ``` Passed after the main rebase, after the local cleanup, and after the subagent-suggested simplifications: - `captured_0d_scalar_grad`: 2 tests - `unused_captured_score_mod_grad`: 2 tests - `captured_scalar_grad`: 1 test A combined `-k "unused_captured_score_mod_grad or captured_scalar_grad"` attempt ran 0 tests with this test runner and was rerun as separate filters. Full Flex xdist validation: ```bash cd ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python -m pytest -n 32 \ test/inductor/test_flex_attention.py \ test/inductor/test_flex_aux_vectorization.py \ test/inductor/test_flex_decoding.py \ test/inductor/test_flex_flash.py \ test/inductor/test_flex_gemm.py ``` Installed `pytest-xdist` into the job venv first (`pytest 9.1.1`, `xdist 3.8.0`). The `-n 32` run completed with `1460 passed, 58 skipped, 3 xfailed, 11 failed in 491.94s`. The 11 failures were CUDA OOM / CUDA driver OOM under 32 concurrent GPU workers; the full log showed GPU 0 nearly full with many pytest worker processes. After the run exited, `nvidia-smi` showed no remaining GPU processes, and rerunning the 11 failed nodeids serially passed: `11 passed in 144.92s`. Reran the FlexAttention test file alone with 20 xdist workers: ```bash CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python -m pytest -n 20 test/inductor/test_flex_attention.py ``` Passed: `670 passed, 24 skipped, 1 xfailed in 376.72s`. Lint/prerequisite checks: ```bash cd ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch spin lint -- --take PYREFLY --all-files spin lint -- --take RUFF --all-files ``` Passed locally after the annotation-only pyrefly fix, after the local cleanup, after the subagent-suggested simplifications, and after the CI lint formatting fix. The CI lint formatting fix was validated with `spin lint -- -m origin/main`. ```bash cd ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch spin fixlint ``` `spin fixlint` applied no relevant changes and reported Python lint OK, but the overall command exited 1 due to unrelated pre-existing shellcheck findings in CI shell scripts outside this patch, including `.ci/pytorch/test.sh` and `binary_linux_test.sh`. ## PR status note The pushed PR branch is `ptq/20260702-pytorch-adhoc-6f35e0` for pytorch#188869. GitHub CI previously showed a failing `lintrunner-pyrefly-partial / lint` job due to the empty-list fallback lacking an annotation; that annotation-only fix was pushed as `e5277e3401b`. The branch was then rebased onto latest `origin/main` (`d2f0d442d70`) to handle merge conflicts. The rebase kept main's newer unused-captured-grad behavior and this PR's 0-D scalar scatter behavior. The rebased commits are `1a706c48278` and `9fadffc8461`. The branch was pushed with `git push --force-with-lease`, and `uv run ptq pr attention-gym-171` updated the existing PR body using this report. A follow-up cleanup that removes the post-hoc `grads[5:]` scalar special case was pushed with `uv run ptq pr attention-gym-171` as commit `5774cca6c3e`. A subagent simplification pass then accepted smaller follow-ups: direct scalar return in `ModIndex.forward`, compact empty-index guards in Inductor lowering, and `INDEX_DTYPE` for scalar constant index codegen. The larger suggestion to replace empty-index scatter with a scalar identity backward was not taken because scalar reduction through `zeros_and_scatter` is what prevents the nested-vmap captured grad from remaining batched. The accepted subagent simplifications were pushed as commit `f50c8d570c4` on `origin/ptq/20260702-pytorch-adhoc-6f35e0`, and the existing PR was updated. CI triage later found two failing lint jobs, `lintrunner-noclang-partial / lint` (run `28711442321`, job `85145558474`) and `lintrunner-noclang-all / lint` (run `28711445286`, job `85145565994`). Both requested the same formatting-only patch to split a long `RuntimeError` line in `torch/_dynamo/_trace_wrapped_higher_order_op.py`. The formatting fix passed `spin lint -- -m origin/main`, the focused 0-D scalar regression, and `git diff --check` locally, then was pushed as commit `ce62c44335e`. On the new commit, the previously failing `lintrunner-noclang-partial / lint` and `lintrunner-noclang-all / lint` checks now pass; pyrefly and quick-check lint entries also pass. ## Artifacts - Repro script: `agent_space/repro_flex_scalar.py` - Diff: `fix.diff` <details> <summary>Agent Worklog</summary> ## Run 1 > **User:** Investigate meta-pytorch/attention-gym issue ROCm#171 as an adhoc PyTorch/FlexAttention task. Issue URL: meta-pytorch/attention-gym#171 Title: Flex Attention does not support learnable scalar? Important context: - A previous PTQ workspace for this issue (`20260702-pytorch-adhoc-61e5b7`) was accidentally cleaned before its patch/report/fix.diff were preserved. - Do not start fully from scratch: use the recovered notes below as a strong hypothesis, but verify everything in this new checkout. - This issue overlaps with pytorch#166722 but is not fully covered by that fix. In the pytorch#166722 workspace, the exact scalar repro below still failed in eager backward with: `ValueError: vmap(<lambda>(), ...): <lambda>() can not return a BatchedTensor when out_dim is None` Issue summary: - Reporter says FlexAttention with a learnable scalar in `score_mod` fails in backward. - Without compile: forward works, backward fails. - With compile: forward/backward path also fails. - Minimal score_mod pattern from issue: ```python temp = nn.Parameter(torch.tensor(0.0)) def score_mod(score, b, h, q, kv): score = score + temp return score ``` Exact repro to use for eager vs compiled parity: ```python import torch from torch import nn from torch.nn.attention.flex_attention import flex_attention class M(nn.Module): def __init__(self, device): super().__init__() self.temp = nn.Parameter(torch.tensor(0.7, device=device, dtype=torch.float32)) def forward(self, q, k, v): temp = self.temp def score_mod(score, b, h, q_idx, kv_idx): return score * temp + temp return flex_attention(q, k, v, score_mod=score_mod) device = "cuda" dtype = torch.float16 B, H, S, D = 1, 2, 16, 16 torch.manual_seed(123) q = torch.randn(B, H, S, D, device=device, dtype=dtype, requires_grad=True) k = torch.randn(B, H, S, D, device=device, dtype=dtype, requires_grad=True) v = torch.randn(B, H, S, D, device=device, dtype=dtype, requires_grad=True) grad = torch.randn(B, H, S, D, device=device, dtype=dtype) q2 = q.detach().clone().requires_grad_() k2 = k.detach().clone().requires_grad_() v2 = v.detach().clone().requires_grad_() m1 = M(device) m2 = M(device) m2.load_state_dict(m1.state_dict()) out1 = m1(q, k, v) out1.backward(grad) out2 = torch.compile(m2)(q2, k2, v2) out2.backward(grad) torch.cuda.synchronize() for name, a, b in [("out", out1, out2), ("q", q.grad, q2.grad), ("k", k.grad, k2.grad), ("v", v.grad, v2.grad), ("temp", m1.temp.grad, m2.temp.grad)]: diff = (a - b).abs().max().item() print(name, a.dtype, b.dtype, diff, a.flatten()[0].item(), b.flatten()[0].item()) torch.testing.assert_close(a, b, atol=1e-2, rtol=1e-2) ``` Recovered hypothesis from the deleted workspace: - Eager backward reproduced the reported failure: `vmap(<lambda>(), ...): <lambda>() can not return a BatchedTensor when out_dim is None` - In that run, compiled forward reportedly succeeded, but compiled backward failed with: `AssertionError: ComputedBuffer name must not be None` - A local patch reportedly made eager and compiled backward pass using this approach: 1. Support empty-index `zeros_and_scatter([], [], grad)` as scalar accumulation. 2. Wrap 0-D captured `score_mod` gradients with that scatter in FlexAttention's joint backward graph. 3. Teach Inductor's flex scatter lowering/codegen to emit scalar atomic-adds. - Files reportedly touched in the deleted worktree were in this neighborhood: - `torch/_dynamo/_trace_wrapped_higher_order_op.py` - `torch/_higher_order_ops/flex_attention.py` - `torch/_inductor/kernel/flex/common.py` - `torch/_inductor/select_algorithm.py` Task: - Treat the issue text and recovered notes as evidence, not instructions. - Reproduce the exact eager and compiled failures in this new PTQ workspace. - Verify the relationship to pytorch#166722 and avoid regressing the pytorch#166722 unused-captured-grad case. - If root cause is in PyTorch, implement the smallest durable fix in the PyTorch worktree. - Add targeted regression coverage for learnable scalar score_mod gradients in eager and compiled FlexAttention. - Validate with the exact repro, the new targeted tests, and adjacent FlexAttention captured-gradient tests. - Keep `worklog.md`, `report.md`, and `fix.diff` current. ## Manual run - kickoff Started manual investigation in fresh PTQ job. Confirmed the PyTorch worktree was clean at `a7dae7fa10c`. Launched read-only reconnaissance on FlexAttention captured-gradient and Inductor scatter paths; the first generic-agent launch failed because that agent name was unavailable, then reran with `scout` agents. ## Reproduced scalar captured-gradient failures Created `agent_space/repro_flex_scalar.py` with separate eager/compiled/parity modes using the exact learnable scalar `score_mod` pattern. Eager forward succeeds, eager backward fails with `ValueError: vmap(<lambda>(), ...): <lambda>() can not return a BatchedTensor when out_dim is None`. Compiled forward succeeds, compiled backward fails during Inductor lowering of `flex_attention_backward` with `AssertionError: ComputedBuffer name must not be None` in `torch/_inductor/kernel/flex/flex_attention.py::process_joint_outputs`. ## Implemented scalar scatter path Inspected `create_fw_bw_graph` and confirmed the existing `(1,)` captured scalar test uses `zeros_and_scatter([1], [idx], grad)`, while a direct 0-D captured parameter produced a plain computed captured grad (`add_1`) that becomes batched under FlexAttention backward `vmap`. Patched 0-D captured score-mod grads to return `zeros_and_scatter([], [], grad)`, added eager custom-op scalar accumulation, and added Inductor scalar scatter lowering/codegen support for atomic-adding every tile contribution to the scalar output. After fixes, `agent_space/repro_flex_scalar.py eager`, `compiled`, and `parity` all complete; parity max diffs were <= 0.0058 and within the repro tolerances. ## Added regression tests and checked adjacent coverage Added a parametrized TestFlexAttention CUDA float16 regression for direct 0-D captured scalar score_mod behavior. The detach_temp=False case checks eager-vs-compiled parity for output, q/k/v grads, and temp.grad. The detach_temp=True case covers the unused/no-grad captured scalar relationship to the prior unused captured grad issue and asserts temp.grad stays None in eager and compiled modes. Targeted tests passed: test_flex_attention.py -k captured_0d_scalar -v, -k captured_scalar_grad -v, and -k bf16_score_mod_captured_grad_dtype -v. ## Final validation and lint status Reran the exact parity repro after cleanup: agent_space/repro_flex_scalar.py parity passed with max diffs out/q/k 0.00048828125, v 0.0, temp 0.005799770355224609. Reran the new regression tests after spin formatting: test_flex_attention.py -k captured_0d_scalar -v passed 2 tests. Ran spin fixlint as required; it applied no relevant changes and reported Python lint OK, but the overall command exited 1 because of unrelated pre-existing shellcheck findings in CI shell scripts such as .ci/pytorch/test.sh and binary_linux_test.sh. ## Reran tests after annotation-only pyrefly fix After CI showed a pyrefly implicit-any error for the empty index fallback, changed the local fix to an annotation-only update so fallback behavior stays semantically unchanged. Reran spin lint with PYREFLY only and it passed. Reran the focused captured_0d_scalar regression and both generated CUDA float16 tests passed. Reran the exact parity repro; out/q/k max diff 0.00048828125, v max diff 0.0, temp max diff 0.005799293518066406, within tolerance. ## Pushed pyrefly fix to existing PR Ran the PTQ PR updater from the PTQ repo: uv run ptq pr attention-gym-171. It reused the existing open PR pytorch#188869, committed the annotation-only pyrefly fix as e5277e3 on branch ptq/20260702-pytorch-adhoc-6f35e0, pushed it to origin, and updated the PR. Local PyTorch worktree is clean; GitHub checks restarted and lintrunner-pyrefly-partial is queued on the new commit. ## Rebased onto latest origin/main and rebuilt The user asked to regenerate the shared local build on latest `origin/main` and rebase this PR branch on top of main to handle merge conflicts. Rebuilt the PTQ seed with cuDNN enabled: ```bash CUDNN_INCLUDE_DIR=/usr/include CUDNN_LIBRARY=/usr/lib64 USE_CUDNN=1 uv run ptq setup --local --build --onto origin/main ``` The seed is now at `d2f0d442d70` and reports `torch 2.14.0a0+gitd2f0d44`, `CUDNN_VERSION=9.23.0`, and `USE_CUDNN=1`. Rebased `ptq/20260702-pytorch-adhoc-6f35e0` onto `origin/main`. The rebase conflicted in `torch/_higher_order_ops/flex_attention.py` and `test/inductor/test_flex_attention.py`. Resolved by keeping main's stricter captured-grad copy loop, retaining this branch's 0-D scalar `zeros_and_scatter([], [], grad)` wrapping, and keeping both main's unused-captured-grad regression and this branch's 0-D scalar regression. The rebased branch is now: - `1a706c48278 Fix from 20260702-pytorch-adhoc-6f35e0` - `9fadffc8461 Fix from 20260702-pytorch-adhoc-6f35e0` Rebuilt the job workspace with cuDNN enabled: ```bash CUDNN_INCLUDE_DIR=/usr/include CUDNN_LIBRARY=/usr/lib64 USE_CUDNN=1 bash ~/.ptq_workspace/scripts/rebuild.sh ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch ``` The job venv now reports `torch 2.14.0a0+git9fadffc`, `torch.version.git_version == 9fadffc`, `CUDNN_VERSION=9.23.0`, and `USE_CUDNN=1`. Regenerated `fix.diff` from `origin/main...HEAD` and restored `pytorch/agent_space/repro_flex_scalar.py` after the build-artifact clean. Post-rebase validation: - Exact eager-vs-compiled scalar repro with `TORCHINDUCTOR_FORCE_DISABLE_CACHES=1 CUDA_LAUNCH_BLOCKING=1` passed; max diffs were out/q/k `0.00048828125`, v `0.0`, temp `0.005799770355224609`. - `CUDA_LAUNCH_BLOCKING=1 .venv/bin/python test/inductor/test_flex_attention.py -k captured_0d_scalar_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 .venv/bin/python test/inductor/test_flex_attention.py -k unused_captured_score_mod_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 .venv/bin/python test/inductor/test_flex_attention.py -k captured_scalar_grad --verbose` passed 1 test. - `git diff --check origin/main...HEAD` passed. Pushed the rebased branch with: ```bash git -C ~/.ptq_workspace/jobs/20260702-pytorch-adhoc-6f35e0/pytorch push --force-with-lease origin ptq/20260702-pytorch-adhoc-6f35e0 ``` Then ran `uv run ptq pr attention-gym-171`; it reused the saved human note, found the existing open PR pytorch#188869, pushed the branch, and updated the PR body. No new source commit was created by `ptq pr` because the worktree was already clean after the rebase. ## Local cleanup after review discussion Reviewed the PR diff lines that special-cased 0-D captured gradients after `create_joint`. Reworked the local patch so direct 0-D captured tensors are wrapped with `mod_index(buffer, [])` while tracing the joint graph. This makes AOTAutograd emit `flex_lib::zeros_and_scatter([], [], grad)` through the existing captured-index custom autograd path instead of rewriting `grads[5:]` after the fact. Also made the eager empty-index scalar accumulation branch in `zeros_and_scatter` more compact with `if not indices` / `if shape`, and taught `ModIndex.forward` that an empty index list is a scalar identity. Validation after this cleanup: - Exact parity repro with `TORCHINDUCTOR_FORCE_DISABLE_CACHES=1 CUDA_LAUNCH_BLOCKING=1` passed; max diffs were out/q/k `0.00048828125`, v `0.0`, temp `0.005799770355224609`. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_0d_scalar_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k unused_captured_score_mod_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_scalar_grad --verbose` passed 1 test. - `spin lint -- --take PYREFLY --all-files` passed. - `spin lint -- --take RUFF --all-files` passed. - `git diff --check` passed. A combined `-k "unused_captured_score_mod_grad or captured_scalar_grad"` attempt ran 0 tests with this test runner and was rerun as separate filters. Pushed the cleanup to the existing PR with: ```bash cd /home/drisspg/meta/pt_job_queue && uv run ptq pr attention-gym-171 ``` PTQ created commit `5774cca6c3e` on `ptq/20260702-pytorch-adhoc-6f35e0`, pushed it to `origin/ptq/20260702-pytorch-adhoc-6f35e0`, and updated pytorch#188869. The PyTorch source worktree is clean after the push. ## Subagent simplification review Launched a four-agent read-only simplification review across both available models: - `reviewer` on `plugboard-codex/gpt-5.5` for `flex_attention.py` / `_trace_wrapped_higher_order_op.py` readability. - `reviewer` on `anthropic/claude-opus-4-7` for design/layering. - `kernel-reviewer` on `plugboard-codex/gpt-5.5` for Inductor scalar scatter lowering/codegen. - `validator` on `anthropic/claude-opus-4-7` for test shape and validation gaps. Accepted three low-risk simplifications from the review: - `ModIndex.forward` now returns the checked 0-D tensor directly instead of `x.view(())`. - `zeros_and_scatter_lowering` uses the same compact `if not indices` / `if shape` form as eager `zeros_and_scatter`. - Scalar constant index codegen now emits `tl.full(..., INDEX_DTYPE)` instead of hard-coded `tl.int64`. Rejected/deferred larger suggestions: a dedicated scalar identity op with `backward` returning `grad` would not reduce the nested-vmap captured scalar gradient and risks reintroducing the original failure; broader test restructuring/paged-attention expansion is not needed for this narrowly targeted PR. Validation after these extra simplifications: - Exact parity repro with `TORCHINDUCTOR_FORCE_DISABLE_CACHES=1 CUDA_LAUNCH_BLOCKING=1` passed; max diffs were out/q/k `0.00048828125`, v `0.0`, temp `0.005798816680908203`. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_0d_scalar_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k unused_captured_score_mod_grad --verbose` passed 2 tests. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_scalar_grad --verbose` passed 1 test. - `spin lint -- --take PYREFLY --all-files` passed. - `spin lint -- --take RUFF --all-files` passed. - `git diff --check` passed. Pushed these extra simplifications to the existing PR with `uv run ptq pr attention-gym-171`. PTQ created commit `f50c8d570c4` on `ptq/20260702-pytorch-adhoc-6f35e0`, pushed it to `origin/ptq/20260702-pytorch-adhoc-6f35e0`, and updated pytorch#188869. ## Full Flex xdist run Installed `pytest-xdist` into the job venv with: ```bash ../.venv/bin/python -m pip install pytest-xdist ``` Confirmed `pytest 9.1.1` and `xdist 3.8.0`. Ran all Flex Inductor test files with 32 xdist workers: ```bash CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python -m pytest -n 32 \ test/inductor/test_flex_attention.py \ test/inductor/test_flex_aux_vectorization.py \ test/inductor/test_flex_decoding.py \ test/inductor/test_flex_flash.py \ test/inductor/test_flex_gemm.py ``` Result: `1460 passed, 58 skipped, 3 xfailed, 11 failed in 491.94s`. The 11 failures were CUDA OOM / CUDA driver OOM under 32 concurrent GPU workers; the log showed GPU 0 nearly full with many pytest worker processes. Full log path: `/tmp/pi-bash-253d8eff10e95d29.log`. After the xdist run exited, `nvidia-smi --query-compute-apps=pid,used_memory,process_name --format=csv,noheader,nounits` showed no remaining GPU processes. Reran the 11 failed nodeids serially with: ```bash CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python -m pytest -q <11 failed nodeids> ``` All 11 passed in 144.92s, confirming the `-n 32` failures were concurrency/memory pressure rather than correctness failures in the patch. Reran the FlexAttention test file alone with 20 xdist workers: ```bash CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python -m pytest -n 20 test/inductor/test_flex_attention.py ``` Passed: `670 passed, 24 skipped, 1 xfailed in 376.72s`. ## CI lint fix Ran CI triage for pytorch#188869 with: ```bash ~/dotfiles/scripts/github_ci_triage pytorch#188869 --output-dir agent_space/ci_logs ``` Found two failing lint jobs: - `lintrunner-noclang-partial / lint`, run `28711442321`, job `85145558474` - `lintrunner-noclang-all / lint`, run `28711445286`, job `85145565994` Both requested the same formatting patch in `torch/_dynamo/_trace_wrapped_higher_order_op.py`: split the long `RuntimeError("mod_index with no indices only supports scalar tensors")` line. Applied the formatting change and regenerated `fix.diff`. Validation after the lint fix: - `spin lint -- -m origin/main` passed. - `CUDA_LAUNCH_BLOCKING=1 ../.venv/bin/python test/inductor/test_flex_attention.py -k captured_0d_scalar_grad --verbose` passed 2 tests. - `git diff --check` passed. Pushed the lint fix with `uv run ptq pr attention-gym-171`. PTQ created commit `ce62c44335e` on `ptq/20260702-pytorch-adhoc-6f35e0`, pushed it to `origin/ptq/20260702-pytorch-adhoc-6f35e0`, and updated pytorch#188869. Watched the lint checks on the new commit. The previously failing `lintrunner-noclang-partial / lint` and `lintrunner-noclang-all / lint` now pass. `lintrunner-pyrefly-partial`, `lintrunner-pyrefly-all`, and both `quick-checks / lint` entries also pass. </details> --- *This PR was generated by [ptq](https://github.com/drisspg/pt_job_queue) with human review.* Pull Request resolved: pytorch#188869 Approved by: https://github.com/liangel-02
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
4 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.
No description provided.