[clang] 3bf1aca - [CMake][ELF] Add -fno-semantic-interposition and -Bsymbolic-functions

[Fangrui Song via cfe-commits]

Author: Fangrui Song
Date: 2021-05-12T10:34:31-07:00
New Revision: 3bf1acab5b454ad7fb2074b34663108b53620695

URL: https://github.com/llvm/llvm-project/commit/3bf1acab5b454ad7fb2074b34663108b53620695
DIFF: https://github.com/llvm/llvm-project/commit/3bf1acab5b454ad7fb2074b34663108b53620695.diff

LOG: [CMake][ELF] Add -fno-semantic-interposition and -Bsymbolic-functions

llvm-dev message: https://lists.llvm.org/pipermail/llvm-dev/2021-May/150465.html

In an ELF shared object, a default visibility defined symbol is preemptible by default.
This creates some missed optimization opportunities. -fno-semantic-interposition can optimize -fPIC:

* in Clang: avoid GOT/PLT cost for variable access/function calls to external linkage definition in the same TU
* in GCC: enable interprocedural optimizations (including inlining) and avoid PLT

See https://gist.github.com/MaskRay/2d4dfcfc897341163f734afb59f689c6 for more information.

-Bsymbolic-functions is more aggressive than -fvisibility-inlines-hidden (present since 2012) as it applies
to all function definitions.  It can

* avoid PLT for cross-TU function calls && reduce dynamic symbol lookup
* reduce dynamic symbol lookup for taking function addresses and optimize out GOT/TOC on x86-64/ppc64

With both options, the libLLVM.so and libclang-cpp.so performance should
be closer to PIE binary linking against `libLLVM*.a` and `libclang*.a`

(In a -DLLVM_TARGETS_TO_BUILD=X86 build, the number of JUMP_SLOT decreases from 12716 to 1628, and the number of GLOB_DAT decreases from 1918 to 1313
The built clang with `-DLLVM_LINK_LLVM_DYLIB=on -DCLANG_LINK_CLANG_DYLIB=on` is significantly faster.
See the Linux kernel build result https://bugs.archlinux.org/task/70697
)

Some implication:

Interposing a subset of functions is no longer supported.
(This is fragile anyway and cannot really be supported. For Mach-O we don't use
`ld -interpose`, so interposition is not supported on Mach-O at all.)

Compiling a program which takes the address of any LLVM function with
`{gcc,clang} -fno-pic` and expects the address to equal to the address taken
from libLLVM.so or libclang-cpp.so is unsupported. I am fairly confident that
llvm-project shouldn't have different behaviors depending on such pointer
equality (as we've been using -fvisibility-inlines-hidden which applies to
inline functions for a long time), but if we accidentally do, users should be
aware that they should not make assumption on pointer equality in `-fno-pic`
mode.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D102090

Added: 
    

Modified: 
    clang/tools/clang-shlib/CMakeLists.txt
    llvm/cmake/modules/HandleLLVMOptions.cmake
    llvm/tools/llvm-shlib/CMakeLists.txt

Removed: 
    


################################################################################
diff  --git a/clang/tools/clang-shlib/CMakeLists.txt b/clang/tools/clang-shlib/CMakeLists.txt
index 5949223fc8e34..73e806c495556 100644
--- a/clang/tools/clang-shlib/CMakeLists.txt
+++ b/clang/tools/clang-shlib/CMakeLists.txt
@@ -48,3 +48,8 @@ add_clang_library(clang-cpp
                   ${_OBJECTS}
                   LINK_LIBS
                   ${_DEPS})
+# Optimize function calls and global variable access for default visibility
+# definitions to avoid PLT and reduce dynamic relocations.
+if (NOT APPLE)
+  target_link_options(clang-cpp PRIVATE LINKER:-Bsymbolic-functions)
+endif()

diff  --git a/llvm/cmake/modules/HandleLLVMOptions.cmake b/llvm/cmake/modules/HandleLLVMOptions.cmake
index a64644a08e599..81f46cf116cf6 100644
--- a/llvm/cmake/modules/HandleLLVMOptions.cmake
+++ b/llvm/cmake/modules/HandleLLVMOptions.cmake
@@ -305,6 +305,7 @@ if( LLVM_ENABLE_PIC )
     # On Windows all code is PIC. MinGW warns if -fPIC is used.
   else()
     add_flag_or_print_warning("-fPIC" FPIC)
+    add_flag_if_supported("-fno-semantic-interposition" FNO_SEMANTIC_INTERPOSITION)
   endif()
   # GCC for MIPS can miscompile LLVM due to PR37701.
   if(CMAKE_COMPILER_IS_GNUCXX AND LLVM_NATIVE_ARCH STREQUAL "Mips" AND

diff  --git a/llvm/tools/llvm-shlib/CMakeLists.txt b/llvm/tools/llvm-shlib/CMakeLists.txt
index b0ee19049e6fa..7259ba2636581 100644
--- a/llvm/tools/llvm-shlib/CMakeLists.txt
+++ b/llvm/tools/llvm-shlib/CMakeLists.txt
@@ -50,6 +50,9 @@ if(LLVM_BUILD_LLVM_DYLIB)
       # Solaris ld does not accept global: *; so there is no way to version *all* global symbols
       set(LIB_NAMES -Wl,--version-script,${LLVM_LIBRARY_DIR}/tools/llvm-shlib/simple_version_script.map ${LIB_NAMES})
     endif()
+    # Optimize function calls and global variable access for default visibility
+    # definitions to avoid PLT and reduce dynamic relocations.
+    target_link_options(LLVM PRIVATE LINKER:-Bsymbolic-functions)
   elseif("${CMAKE_SYSTEM_NAME}" STREQUAL "Darwin")
     set(LIB_NAMES -Wl,-all_load ${LIB_NAMES})
   endif()