fix storage use_count #157694

guangyey · 2025-07-07T06:06:43Z

Stack from ghstack (oldest at bottom):

-> fix storage use_count #157694

Motivation

#155451 decoupled torch._C._storage_Use_Count from CUDA and introduced a corresponding unit test:

Lines 257 to 262 in 815545f

    
           def test_storage_use_count(self, device): 
        
               a = torch.randn(10, device=device) 
        
               prev_cf = torch._C._storage_Use_Count(a.untyped_storage()._cdata) 
        
               self.assertEqual(prev_cf, 1) 
        
               b = a.view(2, 5) 
        
               self.assertEqual(torch._C._storage_Use_Count(b.untyped_storage()._cdata), prev_cf + 1)

However, this test fails when PyTorch is built with debug assertions enabled. @clee2000 disabled this UT in #156731. The root cause is that _cdata is obtained from an intrusive_ptr, not a weak_intrusive_ptr. As a result, calling c10::weak_intrusive_ptr::use_count on it triggers the internal assertion:

pytorch/c10/util/intrusive_ptr.h

Lines 912 to 917 in 815545f

    
           TORCH_INTERNAL_ASSERT_DEBUG_ONLY( 
        
               owning_weak_ptr == NullType::singleton() || 
        
                   owning_weak_ptr->weakcount_.load() > 1 || 
        
                   (owning_weak_ptr->refcount_.load() == 0 && 
        
                    owning_weak_ptr->weakcount_.load() > 0), 
        
               "weak_intrusive_ptr: Can only weak_intrusive_ptr::reclaim() owning pointers that were created using weak_intrusive_ptr::release().");

For example:

a = torch.randn(10, device=device) # refcount=1, weakcount=1
prev_cf = torch._C._storage_Use_Count(a.untyped_storage()._cdata) # violate the assertation

This violates the expected invariant inside weak_intrusive_ptr::use_count, which assumes the pointer was originally constructed from a valid weak_intrusive_ptr. Actually, storage_impl is obtained from an intrusive_ptr.

pytorch/torch/csrc/Module.cpp

Lines 2105 to 2109 in 815545f

    
           py_module.def("_storage_Use_Count", [](size_t storage_impl_ptr) { 
        
             // NOLINTNEXTLINE(performance-no-int-to-ptr) 
        
             c10::StorageImpl* storage_impl = (c10::StorageImpl*)storage_impl_ptr; 
        
             return c10::raw::weak_intrusive_ptr::use_count(storage_impl); 
        
           });

Solution

Use c10::intrusive_ptr::use_count instead.

pytorch-bot · 2025-07-07T06:06:47Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/157694

📄 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 or our office hours

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

❌ 1 Cancelled Job, 1 Unrelated Failure

As of commit 25db3ac with merge base f8c0a4b ():

CANCELLED JOB - The following job was cancelled. Please retry:

trunk / linux-jammy-rocm-py3.10 / test (default, 1, 2, linux.rocm.gpu.2) (gh)
##[error]The operation was canceled.

FLAKY - The following job failed but was likely due to flakiness present on trunk:

trunk / macos-py3-arm64 / test (default, 1, 3, macos-m1-stable) (gh) (detected as infra flaky with no log or failing log classifier)

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

ghstack-source-id: ca0e9ef Pull Request resolved: #157694

albanD

Thanks!

[ghstack-poisoned]

guangyey · 2025-07-08T01:59:12Z

@pytorchbot merge

pytorchmergebot · 2025-07-08T02:01:01Z

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 · 2025-07-08T02:43:23Z

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / linux-jammy-rocm-py3.10 / test (default, 1, 2, linux.rocm.gpu.2)

Details for Dev Infra team

Raised by workflow job

guangyey · 2025-07-08T05:47:17Z

@pytorchbot merge

pytorchmergebot · 2025-07-08T05:49:06Z

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 · 2025-07-08T05:49:25Z

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / linux-jammy-rocm-py3.10 / test (default, 1, 2, linux.rocm.gpu.2)

Details for Dev Infra team

Raised by workflow job

guangyey · 2025-07-08T05:51:24Z

@pytorchbot merge -f "skip rocm queued CI"

pytorchmergebot · 2025-07-08T05:52:57Z

Merge started

Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use -f as last resort and instead consider -i/--ignore-current to continue the merge ignoring current failures. This will allow currently pending tests to finish and report signal before the merge.

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

guangyey added a commit that referenced this pull request Jul 7, 2025

fix storage use_count

6422ec8

ghstack-source-id: ca0e9ef Pull Request resolved: #157694

guangyey requested review from albanD and clee2000 July 7, 2025 06:10

guangyey added the topic: bug fixes topic category label Jul 7, 2025

pytorchbot added the open source label Jul 7, 2025

guangyey added the topic: not user facing topic category label Jul 7, 2025

guangyey requested review from EikanWang and gujinghui July 7, 2025 07:03

albanD approved these changes Jul 7, 2025

View reviewed changes

Update

25db3ac

[ghstack-poisoned]

pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Jul 8, 2025

pytorchmergebot added the merging label Jul 8, 2025

pytorchmergebot removed the merging label Jul 8, 2025

pytorchmergebot added the merging label Jul 8, 2025

pytorchmergebot removed the merging label Jul 8, 2025

pytorchmergebot added the merging label Jul 8, 2025

pytorchmergebot added the Merged label Jul 8, 2025

pytorchmergebot closed this in 1b58e7a Jul 8, 2025

pytorchmergebot removed the merging label Jul 8, 2025

github-actions bot deleted the gh/guangyey/164/head branch August 8, 2025 02:21

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix storage use_count #157694

fix storage use_count #157694

Uh oh!

guangyey commented Jul 7, 2025 •

edited

Loading

Uh oh!

pytorch-bot bot commented Jul 7, 2025 •

edited

Loading

Uh oh!

albanD left a comment

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

	def test_storage_use_count(self, device):
	a = torch.randn(10, device=device)
	prev_cf = torch._C._storage_Use_Count(a.untyped_storage()._cdata)
	self.assertEqual(prev_cf, 1)
	b = a.view(2, 5)
	self.assertEqual(torch._C._storage_Use_Count(b.untyped_storage()._cdata), prev_cf + 1)

	TORCH_INTERNAL_ASSERT_DEBUG_ONLY(
	owning_weak_ptr == NullType::singleton() \|\|
	owning_weak_ptr->weakcount_.load() > 1 \|\|
	(owning_weak_ptr->refcount_.load() == 0 &&
	owning_weak_ptr->weakcount_.load() > 0),
	"weak_intrusive_ptr: Can only weak_intrusive_ptr::reclaim() owning pointers that were created using weak_intrusive_ptr::release().");

	py_module.def("_storage_Use_Count", [](size_t storage_impl_ptr) {
	// NOLINTNEXTLINE(performance-no-int-to-ptr)
	c10::StorageImpl* storage_impl = (c10::StorageImpl*)storage_impl_ptr;
	return c10::raw::weak_intrusive_ptr::use_count(storage_impl);
	});

fix storage use_count #157694

fix storage use_count #157694

Uh oh!

Conversation

guangyey commented Jul 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Solution

Uh oh!

pytorch-bot bot commented Jul 7, 2025 • 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/157694

❌ 1 Cancelled Job, 1 Unrelated Failure

Uh oh!

albanD left a comment

Choose a reason for hiding this comment

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Merge started

Uh oh!

pytorchmergebot commented Jul 8, 2025

Merge failed

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Merge started

Uh oh!

pytorchmergebot commented Jul 8, 2025

Merge failed

Uh oh!

guangyey commented Jul 8, 2025

Uh oh!

pytorchmergebot commented Jul 8, 2025

Merge started

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

guangyey commented Jul 7, 2025 •

edited

Loading

pytorch-bot bot commented Jul 7, 2025 •

edited

Loading