librbd/cache/pwl: fix memory leak in SyncPoint persist context cleanup by tchaikov · Pull Request #63295 · ceph/ceph

tchaikov · 2025-05-15T12:08:41Z

Previously, SyncPoint allocated two C_Gather instances tracked by raw
pointers but failed to properly clean them up when only a single sync
point existed, causing memory leaks detected by AddressSanitizer.

This change fixes the leak by modifying AbstractWriteLog::shut_down()
to check for prior sync points in the chain. When the current sync point
is the only one present, we now activate the m_prior_log_entries_persisted
context to ensure:

The onfinish callback executes and releases the captured strong
reference to the enclosing SyncPoint
The parent m_sync_point_persist context completes and gets properly
released

This ensures all allocated contexts are cleaned up correctly during
shutdown, eliminating the memory leak.

The ASan report:

Indirect leak of 2064 byte(s) in 1 object(s) allocated from:
    #0 0x56440919ae2d in operator new(unsigned long) (/home/jenkins-build/build/workspace/ceph-pull-requests/build/bin/unittest_librbd+0x2f3de2d) (BuildId: 6a04677c6ee5235f1a41815df807f97c5b96d4cd)
    https://github.com/tchaikov/ceph/pull/1 0x56440bd67751 in __gnu_cxx::new_allocator<Context*>::allocate(unsigned long, void const*) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/ext/new_allocator.h:127:27
    https://github.com/tchaikov/ceph/pull/2 0x56440bd676e0 in std::allocator<Context*>::allocate(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/allocator.h:185:32
    https://github.com/tchaikov/ceph/pull/3 0x56440bd676e0 in std::allocator_traits<std::allocator<Context*>>::allocate(std::allocator<Context*>&, unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/alloc_traits.h:464:20
    https://github.com/tchaikov/ceph/pull/4 0x56440bd6730b in std::_Vector_base<Context*, std::allocator<Context*>>::_M_allocate(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:346:20
    https://github.com/tchaikov/ceph/pull/5 0x7fd33e00e8d1 in std::vector<Context*, std::allocator<Context*>>::reserve(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/vector.tcc:78:22
    https://github.com/tchaikov/ceph/pull/6 0x7fd33e00c51c in librbd::cache::pwl::SyncPoint::SyncPoint(unsigned long, ceph::common::CephContext*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/SyncPoint.cc:20:27
    https://github.com/tchaikov/ceph/pull/7 0x56440bd65f26 in decltype(::new((void*)(0)) librbd::cache::pwl::SyncPoint(std::declval<unsigned long&>(), std::declval<ceph::common::CephContext*&>())) std::construct_at<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(librbd::cache::pwl::SyncPoint*, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_construct.h:97:39
    https://github.com/tchaikov/ceph/pull/8 0x56440bd65b98 in void std::allocator_traits<std::allocator<librbd::cache::pwl::SyncPoint>>::construct<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint>&, librbd::cache::pwl::SyncPoint*, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/alloc_traits.h:518:4
    https://github.com/tchaikov/ceph/pull/9 0x56440bd657d3 in std::_Sp_counted_ptr_inplace<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, (__gnu_cxx::_Lock_policy)2>::_Sp_counted_ptr_inplace<unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:519:4
    https://github.com/tchaikov/ceph/pull/10 0x56440bd65371 in std::__shared_count<(__gnu_cxx::_Lock_policy)2>::__shared_count<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(librbd::cache::pwl::SyncPoint*&, std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:651:6
    https://github.com/ceph/ceph/pull/11 0x56440bd65163 in std::__shared_ptr<librbd::cache::pwl::SyncPoint, (__gnu_cxx::_Lock_policy)2>::__shared_ptr<std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:1342:14
    https://github.com/ceph/ceph/pull/12 0x56440bd650e6 in std::shared_ptr<librbd::cache::pwl::SyncPoint>::shared_ptr<std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:409:4
    https://github.com/ceph/ceph/pull/13 0x56440bd65057 in std::shared_ptr<librbd::cache::pwl::SyncPoint> std::allocate_shared<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint> const&, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:862:14
    https://github.com/ceph/ceph/pull/14 0x56440bca97e7 in std::shared_ptr<librbd::cache::pwl::SyncPoint> std::make_shared<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:878:14
    https://github.com/ceph/ceph/pull/15 0x56440bd443c8 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::new_sync_point(librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1905:20
    https://github.com/ceph/ceph/pull/16 0x56440bd42e4c in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::flush_new_sync_point(librbd::cache::pwl::C_FlushRequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1951:3
    https://github.com/ceph/ceph/pull/17 0x56440bd9cbf2 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::flush_new_sync_point_if_needed(librbd::cache::pwl::C_FlushRequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1990:5
    https://github.com/ceph/ceph/pull/18 0x56440bd9c636 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*)::'lambda'(librbd::cache::pwl::GuardedRequestFunctionContext&)::operator()(librbd::cache::pwl::GuardedRequestFunctionContext&) const /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:2152:9
    https://github.com/ceph/ceph/pull/19 0x56440bd9b9b4 in boost::detail::function::void_function_obj_invoker<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*)::'lambda'(librbd::cache::pwl::GuardedRequestFunctionContext&), void, librbd::cache::pwl::GuardedRequestFunctionContext&>::invoke(boost::detail::function::function_buffer&, librbd::cache::pwl::GuardedRequestFunctionContext&) /opt/ceph/include/boost/function/function_template.hpp:100:11
    https://github.com/ceph/ceph/pull/20 0x56440bd29321 in boost::function_n<void, librbd::cache::pwl::GuardedRequestFunctionContext&>::operator()(librbd::cache::pwl::GuardedRequestFunctionContext&) const /opt/ceph/include/boost/function/function_template.hpp:789:14
    https://github.com/ceph/ceph/pull/21 0x56440bd28d85 in librbd::cache::pwl::GuardedRequestFunctionContext::finish(int) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/Request.h:335:5
    https://github.com/ceph/ceph/pull/22 0x5644091e0fe0 in Context::complete(int) /home/jenkins-build/build/workspace/ceph-pull-requests/src/include/Context.h:102:5
    https://github.com/ceph/ceph/pull/23 0x56440bd9b378 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::detain_guarded_request(librbd::cache::pwl::C_BlockIORequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::GuardedRequestFunctionContext*, bool) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1202:20
    https://github.com/ceph/ceph/pull/24 0x56440bd96c50 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:2154:3
    https://github.com/ceph/ceph/pull/25 0x56440bd1e4b5 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::shut_down(Context*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:703:3
    https://github.com/ceph/ceph/pull/26 0x56440bdb9022 in librbd::cache::pwl::TestMockCacheSSDWriteLog_compare_and_write_compare_matched_Test::TestBody() /home/jenkins-build/build/workspace/ceph-pull-requests/src/test/librbd/cache/pwl/test_mock_SSDWriteLog.cc:403:7

Fixes: https://tracker.ceph.com/issues/71335

Contribution Guidelines

To sign and title your commits, please refer to Submitting Patches to Ceph.
If you are submitting a fix for a stable branch (e.g. "quincy"), please refer to Submitting Patches to Ceph - Backports for the proper workflow.
When filling out the below checklist, you may click boxes directly in the GitHub web UI. When entering or editing the entire PR message in the GitHub web UI editor, you may also select a checklist item by adding an x between the brackets: [x]. Spaces and capitalization matter when checking off items this way.

Checklist

Tracker (select at least one)
- References tracker ticket
- Very recent bug; references commit where it was introduced
- New feature (ticket optional)
- Doc update (no ticket needed)
- Code cleanup (no ticket needed)
Component impact
- Affects Dashboard, opened tracker ticket
- Affects Orchestrator, opened tracker ticket
- No impact that needs to be tracked
Documentation (select at least one)
- Updates relevant documentation
- No doc update is appropriate
Tests (select at least one)
- Includes unit test(s)
- Includes integration test(s)
- Includes bug reproducer
- No tests

Show available Jenkins commands

jenkins test classic perf Jenkins Job | Jenkins Job Definition
jenkins test crimson perf Jenkins Job | Jenkins Job Definition
jenkins test signed Jenkins Job | Jenkins Job Definition
jenkins test make check Jenkins Job | Jenkins Job Definition
jenkins test make check arm64 Jenkins Job | Jenkins Job Definition
jenkins test submodules Jenkins Job | Jenkins Job Definition
jenkins test dashboard Jenkins Job | Jenkins Job Definition
jenkins test dashboard cephadm Jenkins Job | Jenkins Job Definition
jenkins test api Jenkins Job | Jenkins Job Definition
jenkins test docs ReadTheDocs | Github Workflow Definition
jenkins test ceph-volume all Jenkins Jobs | Jenkins Jobs Definition
jenkins test windows Jenkins Job | Jenkins Job Definition
jenkins test rook e2e Jenkins Job | Jenkins Job Definition

tchaikov · 2025-05-15T12:25:20Z

@idryomov hi Ilya, could you please help review this change?

tchaikov · 2025-05-15T22:14:27Z

this change breaks the librbd tests. looking ..

tchaikov · 2025-05-16T03:53:36Z

the lifecycle of C_GatherBase derived classes are managed by itself. i need to take a closer look to see how to free them if, for instance, they and their sub contexts are not completed before the test case finishes.

src/librbd/cache/pwl/SyncPoint.cc

idryomov · 2025-05-16T06:07:52Z

the lifecycle of C_GatherBase derived classes are managed by itself.

Yup -- as with most other Context-like types.

i need to take a closer look to see how to free them if, for instance, they and their sub contexts are not completed before the test case finishes.

In case C_Gather is activated, I think the test should just wait for it to complete. In case C_Gather is not activated, in theory it could be freed, but the design of SyncPoint class doesn't seem to allow that which means that the test would probably need to be changed to block until any SyncPoint object that it ended up instantiating retires on its own.

tchaikov · 2025-06-03T08:21:26Z

the lifecycle of C_GatherBase derived classes are managed by itself.

Yup -- as with most other Context-like types.

i need to take a closer look to see how to free them if, for instance, they and their sub contexts are not completed before the test case finishes.

In case C_Gather is activated, I think the test should just wait for it to complete. In case C_Gather is not activated, in theory it could be freed, but the design of SyncPoint class doesn't seem to allow that which means that the test would probably need to be changed to block until any SyncPoint object that it ended up instantiating retires on its own.

i think it's the 2nd case where C_Gather is not activated. fortunately, SyncPoint allows it to be activated manually. and the change uses it to address the memleak.

src/librbd/cache/pwl/AbstractWriteLog.cc

tchaikov · 2025-06-04T09:52:18Z

changelog

check m_current_sync_point with m_lock
also activate m_sync_point_persisit if there are no writes in current sync point's log entry.
do not check for number of writes separately, and trade the check for an ceph_assert().

Previously, SyncPoint allocated two C_Gather instances tracked by raw pointers but failed to properly clean them up when only a single sync point existed, causing memory leaks detected by AddressSanitizer. This change fixes the leak by modifying AbstractWriteLog::shut_down() to check for prior sync points in the chain. When the current sync point is the only one present, we now activate the m_prior_log_entries_persisted context to ensure: - The onfinish callback executes and releases the captured strong reference to the enclosing SyncPoint - The parent m_sync_point_persist context completes and gets properly released This ensures all allocated contexts are cleaned up correctly during shutdown, eliminating the memory leak. The ASan report: ``` Indirect leak of 2064 byte(s) in 1 object(s) allocated from: #0 0x56440919ae2d in operator new(unsigned long) (/home/jenkins-build/build/workspace/ceph-pull-requests/build/bin/unittest_librbd+0x2f3de2d) (BuildId: 6a04677c6ee5235f1a41815df807f97c5b96d4cd) #1 0x56440bd67751 in __gnu_cxx::new_allocator<Context*>::allocate(unsigned long, void const*) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/ext/new_allocator.h:127:27 #2 0x56440bd676e0 in std::allocator<Context*>::allocate(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/allocator.h:185:32 #3 0x56440bd676e0 in std::allocator_traits<std::allocator<Context*>>::allocate(std::allocator<Context*>&, unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/alloc_traits.h:464:20 #4 0x56440bd6730b in std::_Vector_base<Context*, std::allocator<Context*>>::_M_allocate(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_vector.h:346:20 #5 0x7fd33e00e8d1 in std::vector<Context*, std::allocator<Context*>>::reserve(unsigned long) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/vector.tcc:78:22 #6 0x7fd33e00c51c in librbd::cache::pwl::SyncPoint::SyncPoint(unsigned long, ceph::common::CephContext*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/SyncPoint.cc:20:27 #7 0x56440bd65f26 in decltype(::new((void*)(0)) librbd::cache::pwl::SyncPoint(std::declval<unsigned long&>(), std::declval<ceph::common::CephContext*&>())) std::construct_at<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(librbd::cache::pwl::SyncPoint*, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_construct.h:97:39 #8 0x56440bd65b98 in void std::allocator_traits<std::allocator<librbd::cache::pwl::SyncPoint>>::construct<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint>&, librbd::cache::pwl::SyncPoint*, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/alloc_traits.h:518:4 #9 0x56440bd657d3 in std::_Sp_counted_ptr_inplace<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, (__gnu_cxx::_Lock_policy)2>::_Sp_counted_ptr_inplace<unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:519:4 #10 0x56440bd65371 in std::__shared_count<(__gnu_cxx::_Lock_policy)2>::__shared_count<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(librbd::cache::pwl::SyncPoint*&, std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:651:6 #11 0x56440bd65163 in std::__shared_ptr<librbd::cache::pwl::SyncPoint, (__gnu_cxx::_Lock_policy)2>::__shared_ptr<std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr_base.h:1342:14 #12 0x56440bd650e6 in std::shared_ptr<librbd::cache::pwl::SyncPoint>::shared_ptr<std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::_Sp_alloc_shared_tag<std::allocator<librbd::cache::pwl::SyncPoint>>, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:409:4 ceph#13 0x56440bd65057 in std::shared_ptr<librbd::cache::pwl::SyncPoint> std::allocate_shared<librbd::cache::pwl::SyncPoint, std::allocator<librbd::cache::pwl::SyncPoint>, unsigned long&, ceph::common::CephContext*&>(std::allocator<librbd::cache::pwl::SyncPoint> const&, unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:862:14 ceph#14 0x56440bca97e7 in std::shared_ptr<librbd::cache::pwl::SyncPoint> std::make_shared<librbd::cache::pwl::SyncPoint, unsigned long&, ceph::common::CephContext*&>(unsigned long&, ceph::common::CephContext*&) /usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/shared_ptr.h:878:14 ceph#15 0x56440bd443c8 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::new_sync_point(librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1905:20 ceph#16 0x56440bd42e4c in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::flush_new_sync_point(librbd::cache::pwl::C_FlushRequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1951:3 ceph#17 0x56440bd9cbf2 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::flush_new_sync_point_if_needed(librbd::cache::pwl::C_FlushRequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::DeferredContexts&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1990:5 ceph#18 0x56440bd9c636 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*)::'lambda'(librbd::cache::pwl::GuardedRequestFunctionContext&)::operator()(librbd::cache::pwl::GuardedRequestFunctionContext&) const /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:2152:9 ceph#19 0x56440bd9b9b4 in boost::detail::function::void_function_obj_invoker<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*)::'lambda'(librbd::cache::pwl::GuardedRequestFunctionContext&), void, librbd::cache::pwl::GuardedRequestFunctionContext&>::invoke(boost::detail::function::function_buffer&, librbd::cache::pwl::GuardedRequestFunctionContext&) /opt/ceph/include/boost/function/function_template.hpp:100:11 ceph#20 0x56440bd29321 in boost::function_n<void, librbd::cache::pwl::GuardedRequestFunctionContext&>::operator()(librbd::cache::pwl::GuardedRequestFunctionContext&) const /opt/ceph/include/boost/function/function_template.hpp:789:14 ceph#21 0x56440bd28d85 in librbd::cache::pwl::GuardedRequestFunctionContext::finish(int) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/Request.h:335:5 ceph#22 0x5644091e0fe0 in Context::complete(int) /home/jenkins-build/build/workspace/ceph-pull-requests/src/include/Context.h:102:5 ceph#23 0x56440bd9b378 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::detain_guarded_request(librbd::cache::pwl::C_BlockIORequest<librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>>*, librbd::cache::pwl::GuardedRequestFunctionContext*, bool) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:1202:20 ceph#24 0x56440bd96c50 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::internal_flush(bool, Context*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:2154:3 ceph#25 0x56440bd1e4b5 in librbd::cache::pwl::AbstractWriteLog<librbd::MockImageCtx>::shut_down(Context*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/librbd/cache/pwl/AbstractWriteLog.cc:703:3 ceph#26 0x56440bdb9022 in librbd::cache::pwl::TestMockCacheSSDWriteLog_compare_and_write_compare_matched_Test::TestBody() /home/jenkins-build/build/workspace/ceph-pull-requests/src/test/librbd/cache/pwl/test_mock_SSDWriteLog.cc:403:7 ``` Fixes: https://tracker.ceph.com/issues/71335 Signed-off-by: Kefu Chai <tchaikov@gmail.com>

tchaikov · 2025-06-12T13:19:48Z

courtesy of @CongMinYin ,

@CongMinYin discussed with me offline, we came to the consensus that this is the right fix.

@idryomov could you help take another look?

CongMinYin · 2025-06-12T14:51:52Z

I reviewed the code flow. Then, I think the fix was reasonable. We discussed this fix offline and reached a consensus. For better explanation, I have drawn this simple flowchart.

idryomov · 2025-06-12T16:29:56Z

I reviewed the code flow. Then, I think the fix was reasonable.

Please approve the PR then ;) I'll include this in the next integration branch for teuthology.

idryomov · 2025-06-19T18:07:03Z

No related failures:

https://pulpito.ceph.com/dis-2025-06-17_15:02:32-rbd-wip-dis-testing-distro-default-smithi/
https://pulpito.ceph.com/dis-2025-06-18_05:42:00-rbd-wip-dis-testing-distro-default-smithi/
https://pulpito.ceph.com/dis-2025-06-18_21:18:46-rbd-wip-dis-testing-distro-default-smithi/

tchaikov requested a review from a team as a code owner May 15, 2025 12:08

github-actions bot added the rbd label May 15, 2025

tchaikov force-pushed the rbd-cache-pwl-fix-leaks branch from 2a6483e to 0ffe250 Compare May 15, 2025 12:22

tchaikov requested a review from idryomov May 15, 2025 12:24

tchaikov mentioned this pull request May 15, 2025

script/run-make: enable ASan #56537

Draft

14 tasks

tchaikov removed request for a team and idryomov May 15, 2025 22:13

tchaikov marked this pull request as draft May 15, 2025 22:14

tchaikov closed this May 16, 2025

idryomov reviewed May 16, 2025

View reviewed changes

src/librbd/cache/pwl/SyncPoint.cc Outdated Show resolved Hide resolved

tchaikov reopened this Jun 3, 2025

tchaikov force-pushed the rbd-cache-pwl-fix-leaks branch from 0ffe250 to c7ac59d Compare June 3, 2025 08:18

tchaikov marked this pull request as ready for review June 3, 2025 08:18

tchaikov requested a review from idryomov June 3, 2025 08:22

tchaikov changed the title ~~librbd/cache/pwl: Use smart pointers to fix memory leak in SyncPoint~~ librbd/cache/pwl: fix memory leak in SyncPoint persist context cleanup Jun 3, 2025

idryomov reviewed Jun 3, 2025

View reviewed changes

src/librbd/cache/pwl/AbstractWriteLog.cc Show resolved Hide resolved

tchaikov force-pushed the rbd-cache-pwl-fix-leaks branch from c7ac59d to 9f07e29 Compare June 4, 2025 09:49

tchaikov force-pushed the rbd-cache-pwl-fix-leaks branch from 9f07e29 to 05fd6f9 Compare June 12, 2025 09:17

idryomov added bug-fix wip-dis-testing labels Jun 12, 2025

CongMinYin approved these changes Jun 13, 2025

View reviewed changes

idryomov approved these changes Jun 19, 2025

View reviewed changes

idryomov merged commit 18e473b into ceph:main Jun 19, 2025
13 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

librbd/cache/pwl: fix memory leak in SyncPoint persist context cleanup#63295

librbd/cache/pwl: fix memory leak in SyncPoint persist context cleanup#63295
idryomov merged 1 commit intoceph:mainfrom
tchaikov:rbd-cache-pwl-fix-leaks

tchaikov commented May 15, 2025 •

edited

Loading

Uh oh!

tchaikov commented May 15, 2025

Uh oh!

tchaikov commented May 15, 2025

Uh oh!

tchaikov commented May 16, 2025

Uh oh!

Uh oh!

idryomov commented May 16, 2025

Uh oh!

tchaikov commented Jun 3, 2025 •

edited

Loading

Uh oh!

Uh oh!

tchaikov commented Jun 4, 2025 •

edited

Loading

Uh oh!

tchaikov commented Jun 12, 2025

Uh oh!

CongMinYin commented Jun 12, 2025

Uh oh!

idryomov commented Jun 12, 2025

Uh oh!

idryomov commented Jun 19, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

tchaikov commented May 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Contribution Guidelines

Checklist

Uh oh!

tchaikov commented May 15, 2025

Uh oh!

tchaikov commented May 15, 2025

Uh oh!

tchaikov commented May 16, 2025

Uh oh!

Uh oh!

idryomov commented May 16, 2025

Uh oh!

tchaikov commented Jun 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

tchaikov commented Jun 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tchaikov commented Jun 12, 2025

Uh oh!

CongMinYin commented Jun 12, 2025

Uh oh!

idryomov commented Jun 12, 2025

Uh oh!

idryomov commented Jun 19, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

tchaikov commented May 15, 2025 •

edited

Loading

tchaikov commented Jun 3, 2025 •

edited

Loading

tchaikov commented Jun 4, 2025 •

edited

Loading