[SPARC64]: Stop putting -finline-limit=XXX into CFLAGS

It was a stupid workaround for the "static inline" vs.
"extern inline" issues of long ago, and it is what causes
schedule() to be inlined like crazy into kernel/sched.c
when -Os is specified.

MIPS and S390 should probably do the same.

Now CC_OPTIMIZE_FOR_SIZE can be safely used on sparc64
once more.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

arch/sparc64/Makefile

@@ -17,7 +17,6 @@ CC		:= $(shell if $(CC) -m64 -S -o /dev/null -xc /dev/null >/dev/null 2>&1; then
 NEW_GCC := $(call cc-option-yn, -m64 -mcmodel=medlow)
 NEW_GAS := $(shell if $(LD) -V 2>&1 | grep 'elf64_sparc' > /dev/null; then echo y; else echo n; fi)
 UNDECLARED_REGS := $(shell if $(CC) -c -x assembler /dev/null -Wa,--help | grep undeclared-regs > /dev/null; then echo y; else echo n; fi; )
-INLINE_LIMIT := $(call cc-option-yn, -m64 -finline-limit=100000)
 
 export NEW_GCC
 
@@ -49,10 +48,6 @@ else
   AFLAGS += -m64 -mcpu=ultrasparc $(CC_UNDECL)
 endif
 
-ifeq ($(INLINE_LIMIT),y)
-  CFLAGS := $(CFLAGS) -finline-limit=100000
-endif
-
 ifeq ($(CONFIG_MCOUNT),y)
   CFLAGS := $(CFLAGS) -pg
 endif

init/Kconfig

@@ -260,7 +260,6 @@ config CC_OPTIMIZE_FOR_SIZE
 	bool "Optimize for size (Look out for broken compilers!)"
 	default y
 	depends on ARM || H8300 || EXPERIMENTAL
-	depends on !SPARC64
 	help
 	  Enabling this option will pass "-Os" instead of "-O2" to gcc
 	  resulting in a smaller kernel.