projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5f46c48) | patch
Change to JITDefault code model for ELF targets
authorAndrew Kaylor <andrew.kaylor@intel.com>
Thu, 21 Feb 2013 23:45:19 +0000 (23:45 +0000)
committerAndrew Kaylor <andrew.kaylor@intel.com>
Thu, 21 Feb 2013 23:45:19 +0000 (23:45 +0000)
commitcd5c7247ab895dfe4197e08736d025e807ce9bca
tree2a75e1650cde024a09b08babf1531b53d12b5691tree | snapshot
parent5f46c485148917307d544f943269b13f0c80dd32commit | diff
Change to JITDefault code model for ELF targets

On x86-64 platforms, the small code model assumes that code will be loaded below the 2GB boundary.  With the static relocation model, the fact that the expression code is initially loaded (in the LLDB debugger address space) above that boundary causes problems.  Switching to the JITDefault code model causes the large code model to be used for 64-bit targets and small code model of 32-bit targets.

llvm-svn: 175828
lldb/source/Expression/ClangExpressionParser.cpp diff | blob | history
Domain: System / Toolchain;