Fixes order fiasco of device_config_parser_hook_ by ddelgadovargas-cyber · Pull Request #172581 · pytorch/pytorch

ddelgadovargas-cyber · 2026-01-15T20:31:26Z

If another static object (like g_device_config_parse_hook_registry_instance created by the REGISTER_ALLOCATOR_CONFIG_PARSE_HOOK macro) tries to call registerDeviceConfigParserHook before device_config_parser_hook_ is initialized, assigning to it (operator=) can fail, which leads to a runtime error.

When I use a compilation optimization of -O1 I see this issue:

[src/libcxx/include/__functional/function.h:496]:14: runtime error: member access within null pointer of type 'const __policy'
    #0 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:496]:14
    #1 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:483]:19
    #2 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:727]:8
    #3 0x563224e28b78 in c10::CachingAllocator::AcceleratorAllocatorConfig::registerDeviceConfigParserHook(std::__u::function<void (std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>> const&)>&&, std::__u::unordered_set<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>, std::__u::hash<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>, std::__u::equal_to<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>, std::__u::allocator<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>> const&) [torch/c10/core/AllocatorConfig.h:263]:32
    #4 0x563224e28e9d in DeviceConfigParserHookRegistry [torch/c10/core/AllocatorConfig.h:369]:5
    #5 0x563224e28e9d in __cxx_global_var_init.34 [torch/c10/cuda/CUDAAllocatorConfig.cpp:195]:1
    #6 0x563224e28e9d in _GLOBAL__sub_I_CUDAAllocatorConfig.cpp torch/c10/cuda/CUDAAllocatorConfig.cpp
    #7 0x5632459709ac in __libc_csu_init /[usr/grte/v5/debug-src/src/csu/elf-init.c:88]:7
    #8 0x7f748b9562e7 in __libc_start_main (/usr/grte/v5/lib64/libc.so.6+0x612e7) (BuildId: ca23ec6d935352118622ce674a8bb52d)
    #9 0x5632018f3729 in _start /usr/grte/v5/debug-src/src/csu/../sysdeps/x86_64/start.S:120

pytorch-bot · 2026-01-15T20:31:30Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/172581

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 2 Unrelated Failures

As of commit d0c9f88 with merge base d74f9ec ():

NEW FAILURE - The following job has failed:

trunk / linux-jammy-rocm-py3.10 / test (default, 5, 6, linux.rocm.gpu.gfx942.1) (gh)
test/dynamo/test_fx_graph_runnable.py::FxGraphRunnableTest::test_reduce_scatter_collective

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

trunk / linux-jammy-cuda12.8-py3.10-gcc11 / test (default, 4, 5, linux.g6.4xlarge.experimental.nvidia.gpu) (gh) (disabled by #118346 but the issue was closed recently and a rebase is needed to make it pass)
test/inductor/test_kernel_benchmark.py::TestKernelBenchmark::test_mm_triton_kernel_benchmark
trunk / linux-jammy-cuda13.0-py3.10-gcc11 / test (default, 3, 5, linux.g6.4xlarge.experimental.nvidia.gpu) (gh) (disabled by #118346 but the issue was closed recently and a rebase is needed to make it pass)
test/inductor/test_kernel_benchmark.py::TestKernelBenchmark::test_mm_triton_kernel_benchmark

This comment was automatically generated by Dr. CI and updates every 15 minutes.

linux-foundation-easycla · 2026-01-15T20:31:34Z

The committers listed above are authorized under a signed CLA.

✅ login: ddelgadovargas-cyber / name: Daniel Delgado Vargas (13bcabd, d0c9f88)

ddelgadovargas-cyber · 2026-01-15T20:52:55Z

@pytorchbot label "topic: not user facing"

guangyey · 2026-01-21T03:54:59Z

c10/core/AllocatorConfig.h

  // their own environment configuration extensions.
-  inline static std::function<void(const std::string&)>
-      device_config_parser_hook_{nullptr};
+  static std::function<void(const std::string&)>& device_config_parser_hook() {


Suggested change

static std::function<void(const std::string&)>& device_config_parser_hook() {

static std::function<void(const std::string&)>& getConfigParserHook() {

And move this function and its comments below static const std::unordered_set<std::string>& getKeys() { in line 252

guangyey

Thanks for your update!

guangyey · 2026-01-22T02:51:20Z

This PR still needs the final stamp from Meta.

guangyey · 2026-01-30T09:03:54Z

Hi @albanD Could you please help review this PR.

albanD

Thanks!

guangyey · 2026-02-03T01:43:54Z

Thanks~
@pytorchbot merge

pytorchmergebot · 2026-02-03T01:46:00Z

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

pytorchmergebot · 2026-02-03T01:46:18Z

Merge failed

Reason: 1 jobs have failed, first few of them are: Claude Code

Details for Dev Infra team

Raised by workflow job

guangyey · 2026-02-03T01:58:57Z

@pytorchbot merge

pytorchmergebot · 2026-02-03T02:01:00Z

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

pytorchmergebot · 2026-02-03T05:01:32Z

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / linux-jammy-rocm-py3.10 / test (default, 5, 6, linux.rocm.gpu.gfx942.1)

Details for Dev Infra team

Raised by workflow job

guangyey · 2026-02-03T05:30:29Z

@pytorchbot merge -i

pytorchmergebot · 2026-02-03T05:32:32Z

Merge started

Your change will be merged while ignoring the following 3 checks: trunk / linux-jammy-cuda13.0-py3.10-gcc11 / test (default, 3, 5, linux.g6.4xlarge.experimental.nvidia.gpu), trunk / linux-jammy-cuda12.8-py3.10-gcc11 / test (default, 4, 5, linux.g6.4xlarge.experimental.nvidia.gpu), trunk / linux-jammy-rocm-py3.10 / test (default, 5, 6, linux.rocm.gpu.gfx942.1)

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

If another static object (like `g_device_config_parse_hook_registry_instance` created by the `REGISTER_ALLOCATOR_CONFIG_PARSE_HOOK` macro) tries to call `registerDeviceConfigParserHook` before `device_config_parser_hook_` is initialized, assigning to it (operator=) can fail, which leads to a runtime error. When I use a compilation optimization of ` -O1` I see this issue: ``` [src/libcxx/include/__functional/function.h:496]:14: runtime error: member access within null pointer of type 'const __policy' #0 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:496]:14 pytorch#1 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:483]:19 pytorch#2 0x563224e28b78 in operator= [crosstool/v18/stable/src/libcxx/include/__functional/function.h:727]:8 pytorch#3 0x563224e28b78 in c10::CachingAllocator::AcceleratorAllocatorConfig::registerDeviceConfigParserHook(std::__u::function<void (std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>> const&)>&&, std::__u::unordered_set<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>, std::__u::hash<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>, std::__u::equal_to<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>, std::__u::allocator<std::__u::basic_string<char, std::__u::char_traits<char>, std::__u::allocator<char>>>> const&) [torch/c10/core/AllocatorConfig.h:263]:32 pytorch#4 0x563224e28e9d in DeviceConfigParserHookRegistry [torch/c10/core/AllocatorConfig.h:369]:5 pytorch#5 0x563224e28e9d in __cxx_global_var_init.34 [torch/c10/cuda/CUDAAllocatorConfig.cpp:195]:1 pytorch#6 0x563224e28e9d in _GLOBAL__sub_I_CUDAAllocatorConfig.cpp torch/c10/cuda/CUDAAllocatorConfig.cpp pytorch#7 0x5632459709ac in __libc_csu_init /[usr/grte/v5/debug-src/src/csu/elf-init.c:88]:7 pytorch#8 0x7f748b9562e7 in __libc_start_main (/usr/grte/v5/lib64/libc.so.6+0x612e7) (BuildId: ca23ec6d935352118622ce674a8bb52d) pytorch#9 0x5632018f3729 in _start /usr/grte/v5/debug-src/src/csu/../sysdeps/x86_64/start.S:120 ``` Pull Request resolved: pytorch#172581 Approved by: https://github.com/guangyey, https://github.com/albanD

Fixes order fiasco of device_config_parser_hook_

13bcabd

pytorchbot added the open source label Jan 15, 2026

pytorch-bot bot added the topic: not user facing topic category label Jan 15, 2026

jcaip requested a review from guangyey January 20, 2026 13:08

jcaip added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jan 20, 2026

guangyey reviewed Jan 21, 2026

View reviewed changes

Adress review comments

d0c9f88

ddelgadovargas-cyber requested a review from guangyey January 21, 2026 20:18

guangyey approved these changes Jan 22, 2026

View reviewed changes

guangyey requested review from albanD and ezyang January 22, 2026 02:50

guangyey requested a review from jcaip January 22, 2026 02:53

albanD approved these changes Feb 2, 2026

View reviewed changes

pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Feb 3, 2026

pytorchmergebot added the merging label Feb 3, 2026

pytorchmergebot removed the merging label Feb 3, 2026

pytorchmergebot added the merging label Feb 3, 2026

pytorchmergebot removed the merging label Feb 3, 2026

pytorchmergebot added the merging label Feb 3, 2026

pytorchmergebot added the Merged label Feb 3, 2026

pytorchmergebot closed this in 81fa7ec Feb 3, 2026

pytorchmergebot removed the merging label Feb 3, 2026

	static std::function<void(const std::string&)>& device_config_parser_hook() {
	static std::function<void(const std::string&)>& getConfigParserHook() {

Conversation

ddelgadovargas-cyber commented Jan 15, 2026

Uh oh!

pytorch-bot bot commented Jan 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/172581

❌ 1 New Failure, 2 Unrelated Failures

Uh oh!

linux-foundation-easycla bot commented Jan 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ddelgadovargas-cyber commented Jan 15, 2026

Uh oh!

guangyey Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

ddelgadovargas-cyber Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

guangyey left a comment

Choose a reason for hiding this comment

Uh oh!

guangyey commented Jan 22, 2026

Uh oh!

guangyey commented Jan 30, 2026

Uh oh!

albanD left a comment

Choose a reason for hiding this comment

Uh oh!

guangyey commented Feb 3, 2026

Uh oh!

pytorchmergebot commented Feb 3, 2026

Merge started

Uh oh!

pytorchmergebot commented Feb 3, 2026

Merge failed

Uh oh!

guangyey commented Feb 3, 2026

Uh oh!

pytorchmergebot commented Feb 3, 2026

Merge started

Uh oh!

pytorchmergebot commented Feb 3, 2026

Merge failed

Uh oh!

guangyey commented Feb 3, 2026

Uh oh!

pytorchmergebot commented Feb 3, 2026

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

pytorch-bot bot commented Jan 15, 2026 •

edited

Loading

linux-foundation-easycla bot commented Jan 15, 2026 •

edited

Loading