Summary
- added atomic to eliminate if failing test deadlock is a staleness issue (details)
- Eliminating the busy wait spin at the end seems to resolve deadlock. Added further tests to try and narrow down failure case, and to see if the failure is waituntil specific (details)
- more updates for llheap paper (details)
The file was modified | libcfa/src/concurrency/channel.hfa |
The file was added | tests/concurrent/waituntil/.expect/one_chan_spin.txt |
The file was added | tests/concurrent/channels/handoff.cfa |
The file was added | tests/concurrent/waituntil/one_chan_spin.cfa |
The file was added | tests/concurrent/channels/.expect/handoff.txt |
The file was modified | tests/concurrent/waituntil/one_chan.cfa |
The file was modified | doc/papers/llheap/Paper.tex |