Details
Diff Detail
- Build Status
Buildable 5658 Build 5658: arc lint + arc unit
Event Timeline
lib/sanitizer_common/sanitizer_stoptheworld_mac.cc | ||
---|---|---|
74–78 | Do you expect any performance issues from this O(n^2) behavior? What is the maximum reasonable thread count? Why do we even need to check for ContainsThread? Do you expect that threads contains duplicates? | |
87–89 | Can we iterate over suspended_thread_list instead and skip the check for mach_thread_self? | |
91–94 | clang-format please |
lib/sanitizer_common/sanitizer_stoptheworld_mac.cc | ||
---|---|---|
74–78 | I expect to be running this on programs with up to 250 threads or so, not sure what a normal maximum use case would be. I based this on the Linux implementation, which does have the nested loop. That implementation is more complex, so it may be necessary there, but it should be fine to remove it here. |
Do you expect any performance issues from this O(n^2) behavior? What is the maximum reasonable thread count? Why do we even need to check for ContainsThread? Do you expect that threads contains duplicates?