Some possible approaches: - update the tests to check that we don't memcpy in this case - change FieldMemcpyizer (in CGClass) to allow volatile fields and create a volatile memcpy call
The test was added for PR9027, which is just trying to make sure we emit a volatile load/store in this case, and either approach satisfies its requirements, but checking for the absence of a memcpy seems like a somewhat better approach to me. http://reviews.llvm.org/D7060 EMAIL PREFERENCES http://reviews.llvm.org/settings/panel/emailpreferences/ _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
