Bug report
Bug description:
This is a sub-issue of #146102 with the gist here.
I think that this might be a false positive in the umbrella. When I ran the reproducer, I got no segfault.
However, when I ran this modified reproducer, ASan detected a memory leak:
Modified Reproducer
import sys
import _testcapi
for n in range(500):
sys.modules.pop("readline", None)
try:
_testcapi.set_nomemory(n, 0)
import readline
except MemoryError:
pass
finally:
_testcapi.remove_mem_hooks()ASan Report
=================================================================
==1644389==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 10370 byte(s) in 305 object(s) allocated from:
#0 0x7ff1ebd2506e in strdup (/usr/lib/libasan.so.8+0x12506e) (BuildId: ee5fbab73143ab257a66a33afe0f038a4af7a74e)
#1 0x7ff1eb6dbc96 in setup_readline Modules/readline.c:1407
#2 0x7ff1eb6de59d in PyInit_readline Modules/readline.c:1689
#3 0x5620e9ae1f76 in _PyImport_RunModInitFunc Python/importdl.c:436
#4 0x5620e9ad934a in reload_singlephase_extension Python/import.c:1982
#5 0x5620e9ad9b31 in import_find_extension Python/import.c:2043
#6 0x5620e9adb595 in _imp_create_dynamic_impl Python/import.c:5468
#7 0x5620e9adbb3c in _imp_create_dynamic Python/clinic/import.c.h:489
#8 0x5620e9825587 in cfunction_vectorcall_FASTCALL Objects/methodobject.c:449
#9 0x5620e97704ed in _PyVectorcall_Call Objects/call.c:273
#10 0x5620e97709bd in _PyObject_Call Objects/call.c:348
#11 0x5620e97709ff in PyObject_Call Objects/call.c:373
#12 0x5620e9a09460 in _PyEval_EvalFrameDefault Python/generated_cases.c.h:2831
#13 0x5620e9a410c9 in _PyEval_EvalFrame Include/internal/pycore_ceval.h:122
#14 0x5620e9a41436 in _PyEval_Vector Python/ceval.c:2141
#15 0x5620e976c905 in _PyFunction_Vectorcall Objects/call.c:413
#16 0x5620e976cdca in _PyObject_VectorcallTstate Include/internal/pycore_call.h:144
#17 0x5620e976f058 in object_vacall Objects/call.c:823
#18 0x5620e976f333 in PyObject_CallMethodObjArgs Objects/call.c:960
#19 0x5620e9ad5c3c in import_find_and_load Python/import.c:4133
#20 0x5620e9ade62a in PyImport_ImportModuleLevelObject Python/import.c:4249
#21 0x5620e99f8305 in _PyEval_ImportNameWithImport Python/ceval.c:2996
#22 0x5620e99f858d in _PyEval_ImportName Python/ceval.c:2975
#23 0x5620e9a1dd40 in _PyEval_EvalFrameDefault Python/generated_cases.c.h:7128
#24 0x5620e9a410c9 in _PyEval_EvalFrame Include/internal/pycore_ceval.h:122
#25 0x5620e9a41436 in _PyEval_Vector Python/ceval.c:2141
#26 0x5620e9a416ef in PyEval_EvalCode Python/ceval.c:679
#27 0x5620e9b4929e in run_eval_code_obj Python/pythonrun.c:1369
#28 0x5620e9b4961c in run_mod Python/pythonrun.c:1472
#29 0x5620e9b4a51c in pyrun_file Python/pythonrun.c:1296
Direct leak of 34 byte(s) in 1 object(s) allocated from:
#0 0x7ff1ebd2506e in strdup (/usr/lib/libasan.so.8+0x12506e) (BuildId: ee5fbab73143ab257a66a33afe0f038a4af7a74e)
#1 0x7ff1eb6dbc96 in setup_readline Modules/readline.c:1407
#2 0x7ff1eb6de59d in PyInit_readline Modules/readline.c:1689
#3 0x5620e9ae1f76 in _PyImport_RunModInitFunc Python/importdl.c:436
#4 0x5620e9ad9e0a in import_run_extension Python/import.c:2167
#5 0x5620e9adb9ff in _imp_create_dynamic_impl Python/import.c:5529
#6 0x5620e9adbb3c in _imp_create_dynamic Python/clinic/import.c.h:489
#7 0x5620e9825587 in cfunction_vectorcall_FASTCALL Objects/methodobject.c:449
#8 0x5620e97704ed in _PyVectorcall_Call Objects/call.c:273
#9 0x5620e97709bd in _PyObject_Call Objects/call.c:348
#10 0x5620e97709ff in PyObject_Call Objects/call.c:373
#11 0x5620e9a09460 in _PyEval_EvalFrameDefault Python/generated_cases.c.h:2831
#12 0x5620e9a410c9 in _PyEval_EvalFrame Include/internal/pycore_ceval.h:122
#13 0x5620e9a41436 in _PyEval_Vector Python/ceval.c:2141
#14 0x5620e976c905 in _PyFunction_Vectorcall Objects/call.c:413
#15 0x5620e976cdca in _PyObject_VectorcallTstate Include/internal/pycore_call.h:144
#16 0x5620e976f058 in object_vacall Objects/call.c:823
#17 0x5620e976f333 in PyObject_CallMethodObjArgs Objects/call.c:960
#18 0x5620e9ad5c3c in import_find_and_load Python/import.c:4133
#19 0x5620e9ade62a in PyImport_ImportModuleLevelObject Python/import.c:4249
#20 0x5620e99f8305 in _PyEval_ImportNameWithImport Python/ceval.c:2996
#21 0x5620e99f858d in _PyEval_ImportName Python/ceval.c:2975
#22 0x5620e9a1dd40 in _PyEval_EvalFrameDefault Python/generated_cases.c.h:7128
#23 0x5620e9a410c9 in _PyEval_EvalFrame Include/internal/pycore_ceval.h:122
#24 0x5620e9a41436 in _PyEval_Vector Python/ceval.c:2141
#25 0x5620e9a416ef in PyEval_EvalCode Python/ceval.c:679
#26 0x5620e9b4929e in run_eval_code_obj Python/pythonrun.c:1369
#27 0x5620e9b4961c in run_mod Python/pythonrun.c:1472
#28 0x5620e9b4a51c in pyrun_file Python/pythonrun.c:1296
#29 0x5620e9b4d366 in _PyRun_SimpleFileObject Python/pythonrun.c:518
SUMMARY: AddressSanitizer: 10404 byte(s) leaked in 306 allocation(s).
I'm not sure how important this is given that it only happens during OOM.
CPython versions tested on:
CPython main branch
Operating systems tested on:
Linux
Bug report
Bug description:
This is a sub-issue of #146102 with the gist here.
I think that this might be a false positive in the umbrella. When I ran the reproducer, I got no segfault.
However, when I ran this modified reproducer, ASan detected a memory leak:
Modified Reproducer
ASan Report
I'm not sure how important this is given that it only happens during OOM.
CPython versions tested on:
CPython main branch
Operating systems tested on:
Linux