US 2019/0102278 A1 · MEMORY LEAK PROFILING EVENTS

Abstract

Techniques for profiling memory leaks are described. In one or more embodiments, a memory profiling system identifies a set of one or more objects on the heap during application runtime. For each respective object in the subset of objects, the memory profiling system stores a set of sample information including timestamp that identifies a time associated with an allocation on the heap memory was performed for the respective object and a stack trace identifying at least one subroutine that triggered the allocation on the heap memory. Responsive to detecting a memory leak, the memory profiling system generates a memory leak profile for at least one object in the subset of objects that is causing the memory leak. The memory leak profile identifies when the allocation on the memory store for the at least one object was performed and information about object that remained live after the potential memory leak.

Claims20

Dependent claims

A method comprising:

for each respective object of a sample of objects, as space is allocated for the respective object:
storing a set of allocation-time information including a timestamp that identifies a time associated with an allocation on a memory store for the respective object;
performing a stack trace identifying at least one subroutine that triggered the allocation on the memory store;
and
storing, in the allocation-time information, information that identifies the at least one subroutine that triggered the allocation on the memory store;
responsive to automatically detecting a potential memory leak, generating memory leak profile data for at least one object in the sample objects;
wherein the memory leak profile data identifies:
based on the allocation-time information, when the allocation on the memory store for the at least one object was performed;
and
based on information obtained after the time of the potential memory leak, identifying information about objects that remained live after the potential memory leak is detected.

7 more elements

One or more non-transitory computer-readable media storing instructions, which, when executed by one or more hardware processors, cause operations comprising:

for each respective object of a sample of objects, as space is allocated for the respective object:
storing a set of allocation-time information including a timestamp that identifies a time associated with an allocation on a memory store for the respective object;
performing a stack trace identifying at least one subroutine that triggered the allocation on the memory store;
and
storing, in the allocation-time information, information that identifies the at least one subroutine that triggered the allocation on the memory store;
responsive to automatically detecting a potential memory leak, generating memory leak profile data for at least one object in the sample objects;
wherein the memory leak profile data identifies:
based on the allocation-time information, when the allocation on the memory store for the at least one object was performed;
and
based on information obtained after the time of the potential memory leak, identifying information about objects that remained live after the potential memory leak is detected

7 more elements

Dependencies are inferred from the claim text.

Memory leak profiling events

Life of this patent

Memory leak profiling events

Life of this patent

Abstract

Claims20