diff options
| -rw-r--r-- | Rules.mak | 13 |
1 files changed, 12 insertions, 1 deletions
@@ -89,6 +89,10 @@ ARFLAGS:=r OPTIMIZATION:= PICFLAG:=-fPIC +PIEFLAG:=$(call check_gcc,-fPIE,) +ifeq ($(strip $(PIEFLAG)),-fPIE) +LDPIEFLAG:=-Wl,-pie +endif # Some nice CPU specific optimizations ifeq ($(strip $(TARGET_ARCH)),i386) @@ -172,6 +176,7 @@ ifeq ($(strip $(TARGET_ARCH)),cris) CPU_LDFLAGS-$(CONFIG_CRIS)+=-mcrislinux CPU_CFLAGS-$(CONFIG_CRIS)+=-mlinux PICFLAG:=-fpic + PIEFLAG:=$(call check_gcc,-fpie,) endif ifeq ($(strip $(TARGET_ARCH)),powerpc) @@ -179,13 +184,14 @@ ifeq ($(strip $(TARGET_ARCH)),powerpc) # enough. Therefore use -fpic which will reduce code size and generates # faster code. PICFLAG:=-fpic + PIEFLAG=$(call check_gcc,-fpie,) endif ifeq ($(strip $(TARGET_ARCH)),frv) CPU_LDFLAGS-$(CONFIG_FRV)+=-melf32frvfd CPU_CFLAGS-$(CONFIG_FRV)+=-mfdpic PICFLAG=-fPIC -DPIC - PIEFLAG=-fpie + PIEFLAG=$(call check_gcc,-fpie,) # Using -pie causes the program to have an interpreter, which is # forbidden, so we must make do with -shared. Unfortunately, # -shared by itself would get us global function descriptors @@ -220,6 +226,11 @@ ifeq ($(strip $(TARGET_ARCH)),arm) endif endif +ifneq ($(strip $(UCLIBC_PIE_SUPPORT)),y) +PIEFLAG= +LDPIEFLAG= +endif + ifeq ($(SSP_CFLAGS),) SSP_CFLAGS=$(call check_gcc,-fno-stack-protector-all,) SSP_CFLAGS+=$(call check_gcc,-fstack-protector,) |