https://community.intel.com/t5/Intel-Moderncode-for-Parallel/AppCore-bug-found-and-fixed/m-p/990947#M6247 <DIV>I found and fixed the last bug in AppCore wrt SMR and the node cache. <BR />AppCore should now be basically bug free, I am currently verifying that. You <BR />should probably re-download. Again, the bug is very subtle hard to trip. It <BR />had to do with the way the lock-free single-producer/consumer queue <BR />allocated its very first dummy node. The broken code simple did not set the <BR />fp_dtor member of the dummy node to null. Stupid me!!!</DIV> <DIV><BR />Here is how you would reproduce it:</DIV> <DIV>1. create a new queue object</DIV> <DIV>2. enqueue/dequeue an object with Thread A</DIV> <DIV>3. pop the bad node from the cache in Thread A</DIV> <DIV>4. collect the bad node into Thread A SMR cache</DIV> <DIV>5. join and/or scan Thread A SMR cache</DIV> <DIV>6. SMR scanner can sometimes crash when it iterates over the bad node by <BR />calling a corrupted deferred callback function pointer, if it does goto step <BR />7</DIV> <DIV>7. BOOM!</DIV> <DIV><BR />:(...</DIV> <DIV></DIV> <DIV>Sorry! </DIV> <DIV></DIV> Wed, 11 May 2005 04:33:15 GMT Chris_M__Thomasson 2005-05-11T04:33:15Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/AppCore-bug-found-and-fixed/m-p/990947#M6247 <DIV>I found and fixed the last bug in AppCore wrt SMR and the node cache. <BR />AppCore should now be basically bug free, I am currently verifying that. You <BR />should probably re-download. Again, the bug is very subtle hard to trip. It <BR />had to do with the way the lock-free single-producer/consumer queue <BR />allocated its very first dummy node. The broken code simple did not set the <BR />fp_dtor member of the dummy node to null. Stupid me!!!</DIV> <DIV><BR />Here is how you would reproduce it:</DIV> <DIV>1. create a new queue object</DIV> <DIV>2. enqueue/dequeue an object with Thread A</DIV> <DIV>3. pop the bad node from the cache in Thread A</DIV> <DIV>4. collect the bad node into Thread A SMR cache</DIV> <DIV>5. join and/or scan Thread A SMR cache</DIV> <DIV>6. SMR scanner can sometimes crash when it iterates over the bad node by <BR />calling a corrupted deferred callback function pointer, if it does goto step <BR />7</DIV> <DIV>7. BOOM!</DIV> <DIV><BR />:(...</DIV> <DIV></DIV> <DIV>Sorry! </DIV> <DIV></DIV> Wed, 11 May 2005 04:33:15 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/AppCore-bug-found-and-fixed/m-p/990947#M6247 Chris_M__Thomasson 2005-05-11T04:33:15Z