enhance build process to allow selective -O3 optimization

the motivation for this patch is that the vast majority of libc is
code that does not benefit at all from optimizations, but that certain
components like string/memory operations can be major performance
bottlenecks.

at the same time, the old -falign-*=1 options are removed, since they
were only beneficial for avoiding bloat when global -O3 was used, and
in that case, they may have prevented some of the performance gains.

to be the most useful, this patch will need further tuning. in
particular, research is needed to determine which components should be
built with -O3 by default, and it may be desirable to remove the
hard-coded -O3 and instead allow more customization of the
optimization level used for selected modules.

2 files changed, 57 insertions, 12 deletions

diff --git a/Makefile b/Makefile
index b9b78078..2cd2342a 100644
--- a/Makefile
+++ b/Makefile
@@ -80,6 +80,9 @@ include/bits/alltypes.h: include/bits/alltypes.h.in include/alltypes.h.in tools/
 
 src/ldso/dynlink.lo: arch/$(ARCH)/reloc.h
 
+OPTIMIZE_SRCS = $(wildcard $(OPTIMIZE_GLOBS:%=src/%))
+$(OPTIMIZE_SRCS:%.c=%.o) $(OPTIMIZE_SRCS:%.c=%.lo): CFLAGS += -O3
+
 %.o: $(ARCH)/%.s
 	$(CC) $(CFLAGS_ALL_STATIC) -c -o $@ $<
 
diff --git a/configure b/configure
index b679986f..327863c8 100755
--- a/configure
+++ b/configure
@@ -24,6 +24,7 @@ System types:
   --host=HOST             same as --target
 
 Optional features:
+  --enable-optimize=...   optimize listed components for speed over size [auto]
   --enable-debug          build with debugging information [disabled]
   --enable-warnings       build with recommended warnings flags [disabled]
   --enable-gcc-wrapper    build musl-gcc toolchain wrapper [auto]
@@ -104,6 +105,7 @@ fi
 CFLAGS_C99FSE=
 CFLAGS_AUTO=
 LDFLAGS_AUTO=
+OPTIMIZE_GLOBS=
 prefix=/usr/local/musl
 exec_prefix='$(prefix)'
 bindir='$(exec_prefix)/bin'
@@ -111,6 +113,7 @@ libdir='$(prefix)/lib'
 includedir='$(prefix)/include'
 syslibdir='/lib'
 target=
+optimize=auto
 debug=no
 warnings=no
 shared=yes
@@ -129,6 +132,9 @@ case "$arg" in
 --disable-shared|--enable-shared=no) shared=no ;;
 --enable-static|--enable-static=yes) static=yes ;;
 --disable-static|--enable-static=no) static=no ;;
+--enable-optimize) optimize=yes ;;
+--enable-optimize=*) optimize=${arg#*=} ;;
+--disable-optimize) optimize=no ;;
 --enable-debug|--enable-debug=yes) debug=yes ;;
 --disable-debug|--enable-debug=no) debug=no ;;
 --enable-warnings|--enable-warnings=yes) warnings=yes ;;
@@ -230,14 +236,57 @@ tryflag CFLAGS_C99FSE -fexcess-precision=standard \
 || { test "$ARCH" = i386 && tryflag CFLAGS_C99FSE -ffloat-store ; }
 tryflag CFLAGS_C99FSE -frounding-math
 
+
+#
+# If debugging is explicitly enabled, don't auto-enable optimizations
+#
+if test "$debug" = yes ; then
+CFLAGS_AUTO=-g
+test "$optimize" = auto && optimize=no
+fi
+
 #
-# Setup basic default CFLAGS: debug, optimization, and -pipe
+# Possibly add a -O option to CFLAGS and select modules to optimize with
+# -O3 based on the status of --enable-optimize and provided CFLAGS.
 #
-if fnmatch '-O*|*\ -O*' "$CFLAGS_AUTO $CFLAGS" ; then :
+printf "checking for optimization settings... "
+case "x$optimize" in
+xauto)
+if fnmatch '-O*|*\ -O*' "$CFLAGS_AUTO $CFLAGS" ; then
+printf "using provided CFLAGS\n" ;optimize=no
 else
-tryflag CFLAGS_AUTO -Os || tryflag CFLAGS_AUTO -O2
+printf "using defaults\n" ; optimize=yes
 fi
-test "x$debug" = xyes && CFLAGS_AUTO="-g"
+;;
+xsize|xnone) printf "minimize size\n" ; optimize=size ;;
+xno|x) printf "disabled\n" ; optimize=no ;;
+*) printf "custom\n" ;;
+esac
+
+test "$optimize" = no || tryflag CFLAGS_AUTO -Os || tryflag CFLAGS_AUTO -O2
+test "$optimize" = yes && optimize="internal,malloc,math,string"
+
+if fnmatch 'no|size' "$optimize" ; then :
+else
+printf "components to be optimized for speed:"
+while test "$optimize" ; do
+case "$optimize" in
+*,*) this=${optimize%%,*} optimize=${optimize#*,} ;;
+*) this=$optimize optimize=
+esac
+printf " $this"
+case "$this" in
+*/*.c) ;;
+*/*) this=$this*.c ;;
+*) this=$this/*.c ;;
+esac
+OPTIMIZE_GLOBS="$OPTIMIZE_GLOBS $this"
+done
+OPTIMIZE_GLOBS=${OPTIMIZE_GLOBS# }
+printf "\n"
+fi
+
+# Always try -pipe
 tryflag CFLAGS_AUTO -pipe
 
 #
@@ -267,14 +316,6 @@ tryflag CFLAGS_AUTO -fno-asynchronous-unwind-tables
 tryflag CFLAGS_AUTO -Wa,--noexecstack
 
 #
-# Some optimization levels add bloated alignment that hurt performance
-#
-tryflag CFLAGS_AUTO -falign-functions=1
-tryflag CFLAGS_AUTO -falign-labels=1
-tryflag CFLAGS_AUTO -falign-loops=1
-tryflag CFLAGS_AUTO -falign-jumps=1
-
-#
 # On x86, make sure we don't have incompatible instruction set
 # extensions enabled by default. This is bad for making static binaries.
 # We cheat and use i486 rather than i386 because i386 really does not
@@ -366,6 +407,7 @@ CPPFLAGS = $CPPFLAGS
 LDFLAGS = $LDFLAGS_AUTO $LDFLAGS
 CROSS_COMPILE = $CROSS_COMPILE
 LIBCC = $LIBCC
+OPTIMIZE_GLOBS = $OPTIMIZE_GLOBS
 EOF
 test "x$static" = xno && echo "STATIC_LIBS ="
 test "x$shared" = xno && echo "SHARED_LIBS ="