On Tue, February 12, 2013 10:32 am, Mitch Hayenga wrote: > This changeset breaks compilation for me (gcc 4.7.2), due to the change in > the MachineID struct I believe (src/mem/ruby/system/MachineID.hh). > > The first bad revision is: > changeset: 9496:28d88a0fda74 > user: Nilay Vaish <[email protected]> > date: Sun Feb 10 21:26:22 2013 -0600 > summary: ruby: record fully busy cycle with in the controller > > In file included from build/ARM/mem/ruby/system/CacheMemory.hh:45:0, > from build/ARM/mem/ruby/system/Sequencer.hh:39, > from build/ARM/cpu/testers/rubytest/Check.cc:33: > build/ARM/mem/ruby/slicc_interface/RubySlicc_ComponentMapping.hh: In > function 'MachineID map_Address_to_Directory(const Address&)': > build/ARM/mem/ruby/slicc_interface/RubySlicc_ComponentMapping.hh:53:66: > error: narrowing conversion of 'map_Address_to_DirectoryNode((* & addr))' > from 'NodeID {aka int}' to 'uint32_t {aka unsigned int}' inside { } > [-Werror=narrowing] > build/ARM/mem/ruby/slicc_interface/RubySlicc_ComponentMapping.hh: In > function 'NetDest broadcast(MachineType)': > build/ARM/mem/ruby/slicc_interface/RubySlicc_ComponentMapping.hh:62:34: > error: narrowing conversion of 'i' from 'int' to 'uint32_t {aka unsigned > int}' inside { } [-Werror=narrowing] > cc1plus: all warnings being treated as errors > scons: *** [build/ARM/cpu/testers/rubytest/Check.o] Error 1 > scons: building terminated because of errors. > >
Mitch, you will have to fix this one. I see other compilation errors when I try to compile with gcc-4.7.2. I can change the type of variable 'i' from int to uint32_t, but I would not be able to compile to ensure that no other problems arise with gcc-4.7.2. -- Nilay _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
