KellenSunderland commented on a change in pull request #12631: [MXNET-953] Fix oob memory read URL: https://github.com/apache/incubator-mxnet/pull/12631#discussion_r220333817
########## File path: src/operator/tensor/elemwise_unary_op_basic.cc ########## @@ -542,8 +542,7 @@ void SizeComputeCPU(const nnvm::NodeAttrs& attrs, CHECK_EQ(req.size(), 1U); const TBlob& in_data = inputs[0]; const TBlob& out_data = outputs[0]; - const index_t size_var = in_data.Size(); - memcpy(out_data.dptr_, &size_var, 1U * sizeof(int64_t)); + *static_cast<int64_t*>(out_data.dptr_) = in_data.Size(); Review comment: Let me add some logging code to see what types are used here. My assumption is that uint64 is used for all size operations. It's large enough to handle large indexes and it's integer type, which makes sense for floats. Let me do a few test runs here in CI and validate those assumptions. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
