summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAndrzej Ostruszka <aostruszka@marvell.com>2019-11-07 16:03:09 +0100
committerThomas Monjalon <thomas@monjalon.net>2019-11-08 15:17:05 +0100
commit098cc0fea3be3094a5f0691d1ae501c83fe0d687 (patch)
tree4471e2bdd514b6e15aed3f23f209f0ee89ad6eb1
parent909dd291f07713998c77681aa653e41b13d06121 (diff)
downloaddpdk-098cc0fea3be3094a5f0691d1ae501c83fe0d687.zip
dpdk-098cc0fea3be3094a5f0691d1ae501c83fe0d687.tar.gz
dpdk-098cc0fea3be3094a5f0691d1ae501c83fe0d687.tar.xz
build: add option to enable LTO
This patch adds an option to enable link time optimization. In addition to LTO option itself (-flto) fat-lto-objects are being used. This is because during the build pmdinfogen scans the generated ELF objects to find this_pmd_name* symbol in symbol table. Without fat-lto-objects gcc produces ELF only with extra symbols for internal use during linking. Signed-off-by: Andrzej Ostruszka <aostruszka@marvell.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
-rw-r--r--.travis.yml9
-rw-r--r--config/common_base5
-rw-r--r--config/meson.build13
-rw-r--r--doc/guides/prog_guide/index.rst1
-rw-r--r--doc/guides/prog_guide/lto.rst43
-rw-r--r--doc/guides/rel_notes/release_19_11.rst9
-rw-r--r--mk/toolchain/gcc/rte.toolchain-compat.mk4
-rw-r--r--mk/toolchain/gcc/rte.vars.mk12
-rw-r--r--mk/toolchain/icc/rte.vars.mk8
9 files changed, 104 insertions, 0 deletions
diff --git a/.travis.yml b/.travis.yml
index 3d6ef29..a1c31d4 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -105,6 +105,15 @@ matrix:
apt:
packages:
- *extra_packages
+ - env: DEF_LIB="static" OPTS="-Db_lto=true" EXTRA_PACKAGES=1 CC=gcc-7
+ compiler: gcc
+ addons:
+ apt:
+ sources:
+ - ubuntu-toolchain-r-test
+ packages:
+ - *extra_packages
+ - gcc-7
script: ./.ci/${TRAVIS_OS_NAME}-build.sh
diff --git a/config/common_base b/config/common_base
index 1858598..9142778 100644
--- a/config/common_base
+++ b/config/common_base
@@ -50,6 +50,11 @@ CONFIG_RTE_FORCE_INTRINSICS=n
CONFIG_RTE_ARCH_STRICT_ALIGN=n
#
+# Enable link time optimization
+#
+CONFIG_RTE_ENABLE_LTO=n
+
+#
# Compile to share library
#
CONFIG_RTE_BUILD_SHARED_LIB=n
diff --git a/config/meson.build b/config/meson.build
index e1ebdad..2b1cb92 100644
--- a/config/meson.build
+++ b/config/meson.build
@@ -225,3 +225,16 @@ add_project_arguments('-D_GNU_SOURCE', language: 'c')
if is_freebsd
add_project_arguments('-D__BSD_VISIBLE', language: 'c')
endif
+
+if get_option('b_lto')
+ if cc.has_argument('-ffat-lto-objects')
+ add_project_arguments('-ffat-lto-objects', language: 'c')
+ else
+ error('compiler does not support fat LTO objects - please turn LTO off')
+ endif
+ # workaround for gcc bug 81440
+ if cc.get_id() == 'gcc' and cc.version().version_compare('<8.0')
+ add_project_arguments('-Wno-lto-type-mismatch', language: 'c')
+ add_project_link_arguments('-Wno-lto-type-mismatch', language: 'c')
+ endif
+endif
diff --git a/doc/guides/prog_guide/index.rst b/doc/guides/prog_guide/index.rst
index 692409a..dc4851c 100644
--- a/doc/guides/prog_guide/index.rst
+++ b/doc/guides/prog_guide/index.rst
@@ -65,5 +65,6 @@ Programmer's Guide
ext_app_lib_make_help
perf_opt_guidelines
writing_efficient_code
+ lto
profile_app
glossary
diff --git a/doc/guides/prog_guide/lto.rst b/doc/guides/prog_guide/lto.rst
new file mode 100644
index 0000000..43f4c63
--- /dev/null
+++ b/doc/guides/prog_guide/lto.rst
@@ -0,0 +1,43 @@
+.. SPDX-License-Identifier: BSD-3-Clause
+ Copyright(c) 2019 Marvell International Ltd.
+
+Link Time Optimization
+======================
+
+The DPDK supports compilation with link time optimization turned on.
+This depends obviously on the ability of the compiler to do "whole
+program" optimization at link time and is available only for compilers
+that support that feature.
+To be more specific, compiler (in addition to performing LTO) have to
+support creation of ELF objects containing both normal code and internal
+representation (called fat-lto-objects in gcc and icc).
+This is required since during build some code is generated by parsing
+produced ELF objects (pmdinfogen).
+
+The amount of performance gain that one can get from LTO depends on the
+compiler and the code that is being compiled.
+However LTO is also useful for additional code analysis done by the
+compiler.
+In particular due to interprocedural analysis compiler can produce
+additional warnings about variables that might be used uninitialized.
+Some of these warnings might be "false positives" though and you might
+need to explicitly initialize variable in order to silence the compiler.
+
+Please note that turning LTO on causes considerable extension of
+build time.
+
+When using make based build, link time optimization can be enabled for
+the whole DPDK by setting:
+
+.. code-block:: console
+
+ CONFIG_ENABLE_LTO=y
+
+in config file.
+
+For the meson based build it can be enabled by setting meson built-in
+'b_lto' option:
+
+.. code-block:: console
+
+ meson build -Db_lto=true
diff --git a/doc/guides/rel_notes/release_19_11.rst b/doc/guides/rel_notes/release_19_11.rst
index fe11b4b..12ab110 100644
--- a/doc/guides/rel_notes/release_19_11.rst
+++ b/doc/guides/rel_notes/release_19_11.rst
@@ -248,6 +248,15 @@ New Features
traditional l2fwd example. It demonstrates usage of poll and event mode IO
mechanism under a single application.
+* **Added build support for Link Time Optimization.**
+
+ LTO is an optimization technique used by the compiler to perform whole
+ program analysis and optimization at link time. In order to do that
+ compilers store their internal representation of the source code that
+ the linker uses at the final stage of compilation process.
+
+ See :doc:`../prog_guide/lto` for more information:
+
Removed Items
-------------
diff --git a/mk/toolchain/gcc/rte.toolchain-compat.mk b/mk/toolchain/gcc/rte.toolchain-compat.mk
index ea40a11..69a53e5 100644
--- a/mk/toolchain/gcc/rte.toolchain-compat.mk
+++ b/mk/toolchain/gcc/rte.toolchain-compat.mk
@@ -88,6 +88,10 @@ else
MACHINE_CFLAGS := $(filter-out -march% -mtune% -msse%,$(MACHINE_CFLAGS))
endif
+ ifeq ($(shell test $(GCC_VERSION) -lt 74 && echo 1), 1)
+ CONFIG_RTE_ENABLE_LTO=n
+ endif
+
# Disable thunderx PMD for gcc < 4.7
ifeq ($(shell test $(GCC_VERSION) -lt 47 && echo 1), 1)
CONFIG_RTE_LIBRTE_THUNDERX_NICVF_PMD=d
diff --git a/mk/toolchain/gcc/rte.vars.mk b/mk/toolchain/gcc/rte.vars.mk
index b852fcf..9fc7041 100644
--- a/mk/toolchain/gcc/rte.vars.mk
+++ b/mk/toolchain/gcc/rte.vars.mk
@@ -62,6 +62,18 @@ endif
# process cpu flags
include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.toolchain-compat.mk
+ifeq ($(CONFIG_RTE_ENABLE_LTO),y)
+# 'fat-lto' is used since pmdinfogen needs to have 'this_pmd_nameX'
+# exported in symbol table and without this option only internal
+# representation is present.
+TOOLCHAIN_CFLAGS += -flto -ffat-lto-objects
+TOOLCHAIN_LDFLAGS += -flto
+# workaround for GCC bug 81440
+ifeq ($(shell test $(GCC_VERSION) -lt 80 && echo 1), 1)
+WERROR_FLAGS += -Wno-lto-type-mismatch
+endif
+endif
+
# workaround GCC bug with warning "missing initializer" for "= {0}"
ifeq ($(shell test $(GCC_VERSION) -lt 47 && echo 1), 1)
WERROR_FLAGS += -Wno-missing-field-initializers
diff --git a/mk/toolchain/icc/rte.vars.mk b/mk/toolchain/icc/rte.vars.mk
index aa1422b..8aa87aa 100644
--- a/mk/toolchain/icc/rte.vars.mk
+++ b/mk/toolchain/icc/rte.vars.mk
@@ -54,5 +54,13 @@ endif
# process cpu flags
include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.toolchain-compat.mk
+ifeq ($(CONFIG_RTE_ENABLE_LTO),y)
+# 'fat-lto' is used since pmdinfogen needs to have 'this_pmd_nameX'
+# exported in symbol table and without this option only internal
+# representation is present.
+TOOLCHAIN_CFLAGS += -flto -ffat-lto-objects
+TOOLCHAIN_LDFLAGS += -flto
+endif
+
export CC AS AR LD OBJCOPY OBJDUMP STRIP READELF
export TOOLCHAIN_CFLAGS TOOLCHAIN_LDFLAGS TOOLCHAIN_ASFLAGS