Merge remote-tracking branch 'anton/abiv2' into next

This series adds support for building the powerpc 64-bit
LE kernel using the new ABI v2. We already supported
running ABI v2 userspace programs but this adds support
for building the kernel itself using the new ABI.

arch/powerpc/Makefile

@@ -113,8 +113,13 @@ else
 endif
 endif
 
-CFLAGS-$(CONFIG_PPC64)	:= -mtraceback=no -mcall-aixdesc
-CFLAGS-$(CONFIG_PPC64)	+= $(call cc-option,-mabi=elfv1)
+CFLAGS-$(CONFIG_PPC64)	:= -mtraceback=no
+ifeq ($(CONFIG_CPU_LITTLE_ENDIAN),y)
+CFLAGS-$(CONFIG_PPC64)	+= $(call cc-option,-mabi=elfv2,-mcall-aixdesc)
+AFLAGS-$(CONFIG_PPC64)	+= $(call cc-option,-mabi=elfv2)
+else
+CFLAGS-$(CONFIG_PPC64)	+= -mcall-aixdesc
+endif
 CFLAGS-$(CONFIG_PPC64)	+= $(call cc-option,-mcmodel=medium,-mminimal-toc)
 CFLAGS-$(CONFIG_PPC64)	+= $(call cc-option,-mno-pointers-to-nested-functions)
 CFLAGS-$(CONFIG_PPC32)	:= -ffixed-r2 $(MULTIPLEWORD)
@@ -151,7 +156,7 @@ endif
 CFLAGS-$(CONFIG_TUNE_CELL) += $(call cc-option,-mtune=cell)
 
 KBUILD_CPPFLAGS	+= -Iarch/$(ARCH)
-KBUILD_AFLAGS	+= -Iarch/$(ARCH)
+KBUILD_AFLAGS	+= -Iarch/$(ARCH) $(AFLAGS-y)
 KBUILD_CFLAGS	+= -msoft-float -pipe -Iarch/$(ARCH) $(CFLAGS-y)
 CPP		= $(CC) -E $(KBUILD_CFLAGS)
 

arch/powerpc/boot/util.S

@@ -45,7 +45,7 @@ udelay:
 	mfspr	r4,SPRN_PVR
 	srwi	r4,r4,16
 	cmpwi	0,r4,1		/* 601 ? */
-	bne	.udelay_not_601
+	bne	.Ludelay_not_601
 00:	li	r0,86	/* Instructions / microsecond? */
 	mtctr	r0
 10:	addi	r0,r0,0 /* NOP */
@@ -54,7 +54,7 @@ udelay:
 	bne	00b
 	blr
 
-.udelay_not_601:
+.Ludelay_not_601:
 	mulli	r4,r3,1000	/* nanoseconds */
 	/*  Change r4 to be the number of ticks using:
 	 *	(nanoseconds + (timebase_period_ns - 1 )) / timebase_period_ns

arch/powerpc/include/asm/code-patching.h

@@ -42,15 +42,47 @@ void __patch_exception(int exc, unsigned long addr);
 } while (0)
 #endif
 
+#define OP_RT_RA_MASK	0xffff0000UL
+#define LIS_R2		0x3c020000UL
+#define ADDIS_R2_R12	0x3c4c0000UL
+#define ADDI_R2_R2	0x38420000UL
+
 static inline unsigned long ppc_function_entry(void *func)
 {
-#ifdef CONFIG_PPC64
+#if defined(CONFIG_PPC64)
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+	u32 *insn = func;
+
 	/*
-	 * On PPC64 the function pointer actually points to the function's
-	 * descriptor. The first entry in the descriptor is the address
-	 * of the function text.
+	 * A PPC64 ABIv2 function may have a local and a global entry
+	 * point. We need to use the local entry point when patching
+	 * functions, so identify and step over the global entry point
+	 * sequence.
+	 *
+	 * The global entry point sequence is always of the form:
+	 *
+	 * addis r2,r12,XXXX
+	 * addi  r2,r2,XXXX
+	 *
+	 * A linker optimisation may convert the addis to lis:
+	 *
+	 * lis   r2,XXXX
+	 * addi  r2,r2,XXXX
+	 */
+	if ((((*insn & OP_RT_RA_MASK) == ADDIS_R2_R12) ||
+	     ((*insn & OP_RT_RA_MASK) == LIS_R2)) &&
+	    ((*(insn+1) & OP_RT_RA_MASK) == ADDI_R2_R2))
+		return (unsigned long)(insn + 2);
+	else
+		return (unsigned long)func;
+#else
+	/*
+	 * On PPC64 ABIv1 the function pointer actually points to the
+	 * function's descriptor. The first entry in the descriptor is the
+	 * address of the function text.
 	 */
 	return ((func_descr_t *)func)->entry;
+#endif
 #else
 	return (unsigned long)func;
 #endif

arch/powerpc/include/asm/context_tracking.h

@@ -2,9 +2,9 @@
 #define _ASM_POWERPC_CONTEXT_TRACKING_H
 
 #ifdef CONFIG_CONTEXT_TRACKING
-#define SCHEDULE_USER bl	.schedule_user
+#define SCHEDULE_USER bl	schedule_user
 #else
-#define SCHEDULE_USER bl	.schedule
+#define SCHEDULE_USER bl	schedule
 #endif
 
 #endif

arch/powerpc/include/asm/exception-64e.h

@@ -174,10 +174,10 @@ exc_##label##_book3e:
 	mtlr	r16;
 #define TLB_MISS_STATS_D(name)						    \
 	addi	r9,r13,MMSTAT_DSTATS+name;				    \
-	bl	.tlb_stat_inc;
+	bl	tlb_stat_inc;
 #define TLB_MISS_STATS_I(name)						    \
 	addi	r9,r13,MMSTAT_ISTATS+name;				    \
-	bl	.tlb_stat_inc;
+	bl	tlb_stat_inc;
 #define TLB_MISS_STATS_X(name)						    \
 	ld	r8,PACA_EXTLB+EX_TLB_ESR(r13);				    \
 	cmpdi	cr2,r8,-1;						    \
@@ -185,7 +185,7 @@ exc_##label##_book3e:
 	addi	r9,r13,MMSTAT_DSTATS+name;				    \
 	b	62f;							    \
 61:	addi	r9,r13,MMSTAT_ISTATS+name;				    \
-62:	bl	.tlb_stat_inc;
+62:	bl	tlb_stat_inc;
 #define TLB_MISS_STATS_SAVE_INFO					    \
 	std	r14,EX_TLB_ESR(r12);	/* save ESR */
 #define TLB_MISS_STATS_SAVE_INFO_BOLTED					    \

arch/powerpc/include/asm/exception-64s.h

@@ -517,7 +517,7 @@ label##_relon_hv:							\
 #define DISABLE_INTS	RECONCILE_IRQ_STATE(r10,r11)
 
 #define ADD_NVGPRS				\
-	bl	.save_nvgprs
+	bl	save_nvgprs
 
 #define RUNLATCH_ON				\
 BEGIN_FTR_SECTION				\

arch/powerpc/include/asm/ftrace.h

@@ -61,6 +61,7 @@ struct dyn_arch_ftrace {
 #endif
 
 #if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_PPC64) && !defined(__ASSEMBLY__)
+#if !defined(_CALL_ELF) || _CALL_ELF != 2
 #define ARCH_HAS_SYSCALL_MATCH_SYM_NAME
 static inline bool arch_syscall_match_sym_name(const char *sym, const char *name)
 {
@@ -72,6 +73,7 @@ static inline bool arch_syscall_match_sym_name(const char *sym, const char *name
 	 */
 	return !strcmp(sym + 4, name + 3);
 }
+#endif
 #endif /* CONFIG_FTRACE_SYSCALLS && CONFIG_PPC64 && !__ASSEMBLY__ */
 
 #endif /* _ASM_POWERPC_FTRACE */

arch/powerpc/include/asm/irqflags.h

@@ -20,9 +20,9 @@
  */
 #define TRACE_WITH_FRAME_BUFFER(func)		\
 	mflr	r0;				\
-	stdu	r1, -32(r1);			\
+	stdu	r1, -STACK_FRAME_OVERHEAD(r1);	\
 	std	r0, 16(r1);			\
-	stdu	r1, -32(r1);			\
+	stdu	r1, -STACK_FRAME_OVERHEAD(r1);	\
 	bl func;				\
 	ld	r1, 0(r1);			\
 	ld	r1, 0(r1);
@@ -36,8 +36,8 @@
  * have to call a C function so call a wrapper that saves all the
  * C-clobbered registers.
  */
-#define TRACE_ENABLE_INTS	TRACE_WITH_FRAME_BUFFER(.trace_hardirqs_on)
-#define TRACE_DISABLE_INTS	TRACE_WITH_FRAME_BUFFER(.trace_hardirqs_off)
+#define TRACE_ENABLE_INTS	TRACE_WITH_FRAME_BUFFER(trace_hardirqs_on)
+#define TRACE_DISABLE_INTS	TRACE_WITH_FRAME_BUFFER(trace_hardirqs_off)
 
 /*
  * This is used by assembly code to soft-disable interrupts first and

arch/powerpc/include/asm/kprobes.h

@@ -30,6 +30,7 @@
 #include <linux/ptrace.h>
 #include <linux/percpu.h>
 #include <asm/probes.h>
+#include <asm/code-patching.h>
 
 #define  __ARCH_WANT_KPROBES_INSN_SLOT
 
@@ -56,9 +57,9 @@ typedef ppc_opcode_t kprobe_opcode_t;
 		if ((colon = strchr(name, ':')) != NULL) {		\
 			colon++;					\
 			if (*colon != '\0' && *colon != '.')		\
-				addr = *(kprobe_opcode_t **)addr;	\
+				addr = (kprobe_opcode_t *)ppc_function_entry(addr); \
 		} else if (name[0] != '.')				\
-			addr = *(kprobe_opcode_t **)addr;		\
+			addr = (kprobe_opcode_t *)ppc_function_entry(addr); \
 	} else {							\
 		char dot_name[KSYM_NAME_LEN];				\
 		dot_name[0] = '.';					\

arch/powerpc/include/asm/linkage.h

@@ -2,6 +2,7 @@
 #define _ASM_POWERPC_LINKAGE_H
 
 #ifdef CONFIG_PPC64
+#if !defined(_CALL_ELF) || _CALL_ELF != 2
 #define cond_syscall(x) \
 	asm ("\t.weak " #x "\n\t.set " #x ", sys_ni_syscall\n"		\
 	     "\t.weak ." #x "\n\t.set ." #x ", .sys_ni_syscall\n")
@@ -9,5 +10,6 @@
 	asm ("\t.globl " #alias "\n\t.set " #alias ", " #name "\n"	\
 	     "\t.globl ." #alias "\n\t.set ." #alias ", ." #name)
 #endif
+#endif
 
 #endif	/* _ASM_POWERPC_LINKAGE_H */

arch/powerpc/include/asm/module.h

@@ -35,6 +35,7 @@ struct mod_arch_specific {
 #ifdef __powerpc64__
 	unsigned int stubs_section;	/* Index of stubs section in module */
 	unsigned int toc_section;	/* What section is the TOC? */
+	bool toc_fixed;			/* Have we fixed up .TOC.? */
 #ifdef CONFIG_DYNAMIC_FTRACE
 	unsigned long toc;
 	unsigned long tramp;
@@ -77,6 +78,9 @@ struct mod_arch_specific {
 #    endif	/* MODULE */
 #endif
 
+bool is_module_trampoline(u32 *insns);
+int module_trampoline_target(struct module *mod, u32 *trampoline,
+			     unsigned long *target);
 
 struct exception_table_entry;
 void sort_ex_table(struct exception_table_entry *start,

arch/powerpc/include/asm/ppc_asm.h

@@ -57,7 +57,7 @@ BEGIN_FW_FTR_SECTION;							\
 	LDX_BE	r10,0,r10;		/* get log write index */	\
 	cmpd	cr1,r11,r10;						\
 	beq+	cr1,33f;						\
-	bl	.accumulate_stolen_time;				\
+	bl	accumulate_stolen_time;				\
 	ld	r12,_MSR(r1);						\
 	andi.	r10,r12,MSR_PR;		/* Restore cr0 (coming from user) */ \
 33:									\
@@ -189,9 +189,45 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_SPLPAR)
 #define __STK_REG(i)   (112 + ((i)-14)*8)
 #define STK_REG(i)     __STK_REG(__REG_##i)
 
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+#define STK_GOT		24
+#define __STK_PARAM(i)	(32 + ((i)-3)*8)
+#else
+#define STK_GOT		40
 #define __STK_PARAM(i)	(48 + ((i)-3)*8)
+#endif
 #define STK_PARAM(i)	__STK_PARAM(__REG_##i)
 
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+
+#define _GLOBAL(name) \
+	.section ".text"; \
+	.align 2 ; \
+	.type name,@function; \
+	.globl name; \
+name:
+
+#define _GLOBAL_TOC(name) \
+	.section ".text"; \
+	.align 2 ; \
+	.type name,@function; \
+	.globl name; \
+name: \
+0:	addis r2,r12,(.TOC.-0b)@ha; \
+	addi r2,r2,(.TOC.-0b)@l; \
+	.localentry name,.-name
+
+#define _KPROBE(name) \
+	.section ".kprobes.text","a"; \
+	.align 2 ; \
+	.type name,@function; \
+	.globl name; \
+name:
+
+#define DOTSYM(a)	a
+
+#else
+
 #define XGLUE(a,b) a##b
 #define GLUE(a,b) XGLUE(a,b)
 
@@ -209,19 +245,7 @@ name: \
 	.type GLUE(.,name),@function; \
 GLUE(.,name):
 
-#define _INIT_GLOBAL(name) \
-	__REF; \
-	.align 2 ; \
-	.globl name; \
-	.globl GLUE(.,name); \
-	.section ".opd","aw"; \
-name: \
-	.quad GLUE(.,name); \
-	.quad .TOC.@tocbase; \
-	.quad 0; \
-	.previous; \
-	.type GLUE(.,name),@function; \
-GLUE(.,name):
+#define _GLOBAL_TOC(name) _GLOBAL(name)
 
 #define _KPROBE(name) \
 	.section ".kprobes.text","a"; \
@@ -237,29 +261,9 @@ name: \
 	.type GLUE(.,name),@function; \
 GLUE(.,name):
 
-#define _STATIC(name) \
-	.section ".text"; \
-	.align 2 ; \
-	.section ".opd","aw"; \
-name: \
-	.quad GLUE(.,name); \
-	.quad .TOC.@tocbase; \
-	.quad 0; \
-	.previous; \
-	.type GLUE(.,name),@function; \
-GLUE(.,name):
+#define DOTSYM(a)	GLUE(.,a)
 
-#define _INIT_STATIC(name) \
-	__REF; \
-	.align 2 ; \
-	.section ".opd","aw"; \
-name: \
-	.quad GLUE(.,name); \
-	.quad .TOC.@tocbase; \
-	.quad 0; \
-	.previous; \
-	.type GLUE(.,name),@function; \
-GLUE(.,name):
+#endif
 
 #else /* 32-bit */
 

arch/powerpc/include/asm/sections.h

@@ -39,6 +39,7 @@ static inline int overlaps_kernel_text(unsigned long start, unsigned long end)
 		(unsigned long)_stext < end;
 }
 
+#if !defined(_CALL_ELF) || _CALL_ELF != 2
 #undef dereference_function_descriptor
 static inline void *dereference_function_descriptor(void *ptr)
 {
@@ -49,6 +50,7 @@ static inline void *dereference_function_descriptor(void *ptr)
 		ptr = p;
 	return ptr;
 }
+#endif
 
 #endif
 

arch/powerpc/include/asm/systbl.h

@@ -62,7 +62,7 @@ COMPAT_SYS_SPU(fcntl)
 SYSCALL(ni_syscall)
 SYSCALL_SPU(setpgid)
 SYSCALL(ni_syscall)
-SYSX(sys_ni_syscall,sys_olduname, sys_olduname)
+SYSX(sys_ni_syscall,sys_olduname,sys_olduname)
 SYSCALL_SPU(umask)
 SYSCALL_SPU(chroot)
 COMPAT_SYS(ustat)
@@ -258,7 +258,7 @@ SYSCALL_SPU(tgkill)
 COMPAT_SYS_SPU(utimes)
 COMPAT_SYS_SPU(statfs64)
 COMPAT_SYS_SPU(fstatfs64)
-SYSX(sys_ni_syscall, ppc_fadvise64_64, ppc_fadvise64_64)
+SYSX(sys_ni_syscall,ppc_fadvise64_64,ppc_fadvise64_64)
 PPC_SYS_SPU(rtas)
 OLDSYS(debug_setcontext)
 SYSCALL(ni_syscall)
@@ -295,7 +295,7 @@ SYSCALL_SPU(mkdirat)
 SYSCALL_SPU(mknodat)
 SYSCALL_SPU(fchownat)
 COMPAT_SYS_SPU(futimesat)
-SYSX_SPU(sys_newfstatat, sys_fstatat64, sys_fstatat64)
+SYSX_SPU(sys_newfstatat,sys_fstatat64,sys_fstatat64)
 SYSCALL_SPU(unlinkat)
 SYSCALL_SPU(renameat)
 SYSCALL_SPU(linkat)

arch/powerpc/include/uapi/asm/elf.h

@@ -291,9 +291,17 @@ do {									\
 #define R_PPC64_DTPREL16_HIGHERA 104 /* half16	(sym+add)@dtprel@highera */
 #define R_PPC64_DTPREL16_HIGHEST 105 /* half16	(sym+add)@dtprel@highest */
 #define R_PPC64_DTPREL16_HIGHESTA 106 /* half16	(sym+add)@dtprel@highesta */
+#define R_PPC64_TLSGD		107
+#define R_PPC64_TLSLD		108
+#define R_PPC64_TOCSAVE		109
+
+#define R_PPC64_REL16		249
+#define R_PPC64_REL16_LO	250
+#define R_PPC64_REL16_HI	251
+#define R_PPC64_REL16_HA	252
 
 /* Keep this the last entry.  */
-#define R_PPC64_NUM		107
+#define R_PPC64_NUM		253
 
 /* There's actually a third entry here, but it's unused */
 struct ppc64_opd_entry

arch/powerpc/kernel/cpu_setup_fsl_booke.S

@@ -94,12 +94,12 @@ _GLOBAL(setup_altivec_idle)
 _GLOBAL(__setup_cpu_e6500)
 	mflr	r6
 #ifdef CONFIG_PPC64
-	bl	.setup_altivec_ivors
+	bl	setup_altivec_ivors
 	/* Touch IVOR42 only if the CPU supports E.HV category */
 	mfspr	r10,SPRN_MMUCFG
 	rlwinm.	r10,r10,0,MMUCFG_LPIDSIZE
 	beq	1f
-	bl	.setup_lrat_ivor
+	bl	setup_lrat_ivor
 1:
 #endif
 	bl	setup_pw20_idle
@@ -164,15 +164,15 @@ _GLOBAL(__setup_cpu_e5500)
 #ifdef CONFIG_PPC_BOOK3E_64
 _GLOBAL(__restore_cpu_e6500)
 	mflr	r5
-	bl	.setup_altivec_ivors
+	bl	setup_altivec_ivors
 	/* Touch IVOR42 only if the CPU supports E.HV category */
 	mfspr	r10,SPRN_MMUCFG
 	rlwinm.	r10,r10,0,MMUCFG_LPIDSIZE
 	beq	1f
-	bl	.setup_lrat_ivor
+	bl	setup_lrat_ivor
 1:
-	bl	.setup_pw20_idle
-	bl	.setup_altivec_idle
+	bl	setup_pw20_idle
+	bl	setup_altivec_idle
 	bl	__restore_cpu_e5500
 	mtlr	r5
 	blr
@@ -181,9 +181,9 @@ _GLOBAL(__restore_cpu_e5500)
 	mflr	r4
 	bl	__e500_icache_setup
 	bl	__e500_dcache_setup
-	bl	.__setup_base_ivors
-	bl	.setup_perfmon_ivor
-	bl	.setup_doorbell_ivors
+	bl	__setup_base_ivors
+	bl	setup_perfmon_ivor
+	bl	setup_doorbell_ivors
 	/*
 	 * We only want to touch IVOR38-41 if we're running on hardware
 	 * that supports category E.HV.  The architectural way to determine
@@ -192,7 +192,7 @@ _GLOBAL(__restore_cpu_e5500)
 	mfspr	r10,SPRN_MMUCFG
 	rlwinm.	r10,r10,0,MMUCFG_LPIDSIZE
 	beq	1f
-	bl	.setup_ehv_ivors
+	bl	setup_ehv_ivors
 1:
 	mtlr	r4
 	blr
@@ -201,9 +201,9 @@ _GLOBAL(__setup_cpu_e5500)
 	mflr	r5
 	bl	__e500_icache_setup
 	bl	__e500_dcache_setup
-	bl	.__setup_base_ivors
-	bl	.setup_perfmon_ivor
-	bl	.setup_doorbell_ivors
+	bl	__setup_base_ivors
+	bl	setup_perfmon_ivor
+	bl	setup_doorbell_ivors
 	/*
 	 * We only want to touch IVOR38-41 if we're running on hardware
 	 * that supports category E.HV.  The architectural way to determine
@@ -212,7 +212,7 @@ _GLOBAL(__setup_cpu_e5500)
 	mfspr	r10,SPRN_MMUCFG
 	rlwinm.	r10,r10,0,MMUCFG_LPIDSIZE
 	beq	1f
-	bl	.setup_ehv_ivors
+	bl	setup_ehv_ivors
 	b	2f
 1:
 	ld	r10,CPU_SPEC_FEATURES(r4)

arch/powerpc/kernel/entry_64.S

@@ -39,8 +39,8 @@
  * System calls.
  */
 	.section	".toc","aw"
-.SYS_CALL_TABLE:
-	.tc .sys_call_table[TC],.sys_call_table
+SYS_CALL_TABLE:
+	.tc sys_call_table[TC],sys_call_table
 
 /* This value is used to mark exception frames on the stack. */
 exception_marker:
@@ -106,7 +106,7 @@ BEGIN_FW_FTR_SECTION
 	LDX_BE	r10,0,r10		/* get log write index */
 	cmpd	cr1,r11,r10
 	beq+	cr1,33f
-	bl	.accumulate_stolen_time
+	bl	accumulate_stolen_time
 	REST_GPR(0,r1)
 	REST_4GPRS(3,r1)
 	REST_2GPRS(7,r1)
@@ -143,7 +143,7 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_SPLPAR)
 	std	r10,SOFTE(r1)
 
 #ifdef SHOW_SYSCALLS
-	bl	.do_show_syscall
+	bl	do_show_syscall
 	REST_GPR(0,r1)
 	REST_4GPRS(3,r1)
 	REST_2GPRS(7,r1)
@@ -162,7 +162,7 @@ system_call:			/* label this so stack traces look sane */
  * Need to vector to 32 Bit or default sys_call_table here,
  * based on caller's run-mode / personality.
  */
-	ld	r11,.SYS_CALL_TABLE@toc(2)
+	ld	r11,SYS_CALL_TABLE@toc(2)
 	andi.	r10,r10,_TIF_32BIT
 	beq	15f
 	addi	r11,r11,8	/* use 32-bit syscall entries */
@@ -174,14 +174,14 @@ system_call:			/* label this so stack traces look sane */
 	clrldi	r8,r8,32
 15:
 	slwi	r0,r0,4
-	ldx	r10,r11,r0	/* Fetch system call handler [ptr] */
-	mtctr   r10
+	ldx	r12,r11,r0	/* Fetch system call handler [ptr] */
+	mtctr   r12
 	bctrl			/* Call handler */
 
 syscall_exit:
 	std	r3,RESULT(r1)
 #ifdef SHOW_SYSCALLS
-	bl	.do_show_syscall_exit
+	bl	do_show_syscall_exit
 	ld	r3,RESULT(r1)
 #endif
 	CURRENT_THREAD_INFO(r12, r1)
@@ -248,9 +248,9 @@ syscall_error:
 	
 /* Traced system call support */
 syscall_dotrace:
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.do_syscall_trace_enter
+	bl	do_syscall_trace_enter
 	/*
 	 * Restore argument registers possibly just changed.
 	 * We use the return value of do_syscall_trace_enter
@@ -308,7 +308,7 @@ syscall_exit_work:
 4:	/* Anything else left to do? */
 	SET_DEFAULT_THREAD_PPR(r3, r10)		/* Set thread.ppr = 3 */
 	andi.	r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
-	beq	.ret_from_except_lite
+	beq	ret_from_except_lite
 
 	/* Re-enable interrupts */
 #ifdef CONFIG_PPC_BOOK3E
@@ -319,10 +319,10 @@ syscall_exit_work:
 	mtmsrd	r10,1
 #endif /* CONFIG_PPC_BOOK3E */
 
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.do_syscall_trace_leave
-	b	.ret_from_except
+	bl	do_syscall_trace_leave
+	b	ret_from_except
 
 /* Save non-volatile GPRs, if not already saved. */
 _GLOBAL(save_nvgprs)
@@ -345,42 +345,44 @@ _GLOBAL(save_nvgprs)
  */
 
 _GLOBAL(ppc_fork)
-	bl	.save_nvgprs
-	bl	.sys_fork
+	bl	save_nvgprs
+	bl	sys_fork
 	b	syscall_exit
 
 _GLOBAL(ppc_vfork)
-	bl	.save_nvgprs
-	bl	.sys_vfork
+	bl	save_nvgprs
+	bl	sys_vfork
 	b	syscall_exit
 
 _GLOBAL(ppc_clone)
-	bl	.save_nvgprs
-	bl	.sys_clone
+	bl	save_nvgprs
+	bl	sys_clone
 	b	syscall_exit
 
 _GLOBAL(ppc32_swapcontext)
-	bl	.save_nvgprs
-	bl	.compat_sys_swapcontext
+	bl	save_nvgprs
+	bl	compat_sys_swapcontext
 	b	syscall_exit
 
 _GLOBAL(ppc64_swapcontext)
-	bl	.save_nvgprs
-	bl	.sys_swapcontext
+	bl	save_nvgprs
+	bl	sys_swapcontext
 	b	syscall_exit
 
 _GLOBAL(ret_from_fork)
-	bl	.schedule_tail
+	bl	schedule_tail
 	REST_NVGPRS(r1)
 	li	r3,0
 	b	syscall_exit
 
 _GLOBAL(ret_from_kernel_thread)
-	bl	.schedule_tail
+	bl	schedule_tail
 	REST_NVGPRS(r1)
-	ld	r14, 0(r14)
 	mtlr	r14
 	mr	r3,r15
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+	mr	r12,r14
+#endif
 	blrl
 	li	r3,0
 	b	syscall_exit
@@ -611,7 +613,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_DSCR)
 _GLOBAL(ret_from_except)
 	ld	r11,_TRAP(r1)
 	andi.	r0,r11,1
-	bne	.ret_from_except_lite
+	bne	ret_from_except_lite
 	REST_NVGPRS(r1)
 
 _GLOBAL(ret_from_except_lite)
@@ -661,23 +663,23 @@ _GLOBAL(ret_from_except_lite)
 #endif
 1:	andi.	r0,r4,_TIF_NEED_RESCHED
 	beq	2f
-	bl	.restore_interrupts
+	bl	restore_interrupts
 	SCHEDULE_USER
-	b	.ret_from_except_lite
+	b	ret_from_except_lite
 2:
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 	andi.	r0,r4,_TIF_USER_WORK_MASK & ~_TIF_RESTORE_TM
 	bne	3f		/* only restore TM if nothing else to do */
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.restore_tm_state
+	bl	restore_tm_state
 	b	restore
 3:
 #endif
-	bl	.save_nvgprs
-	bl	.restore_interrupts
+	bl	save_nvgprs
+	bl	restore_interrupts
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.do_notify_resume
-	b	.ret_from_except
+	bl	do_notify_resume
+	b	ret_from_except
 
 resume_kernel:
 	/* check current_thread_info, _TIF_EMULATE_STACK_STORE */
@@ -730,7 +732,7 @@ resume_kernel:
 	 * sure we are soft-disabled first and reconcile irq state.
 	 */
 	RECONCILE_IRQ_STATE(r3,r4)
-1:	bl	.preempt_schedule_irq
+1:	bl	preempt_schedule_irq
 
 	/* Re-test flags and eventually loop */
 	CURRENT_THREAD_INFO(r9, r1)
@@ -792,7 +794,7 @@ restore_no_replay:
 	 */
 do_restore:
 #ifdef CONFIG_PPC_BOOK3E
-	b	.exception_return_book3e
+	b	exception_return_book3e
 #else
 	/*
 	 * Clear the reservation. If we know the CPU tracks the address of
@@ -907,7 +909,7 @@ restore_check_irq_replay:
 	 *
 	 * Still, this might be useful for things like hash_page
 	 */
-	bl	.__check_irq_replay
+	bl	__check_irq_replay
 	cmpwi	cr0,r3,0
  	beq	restore_no_replay
  
@@ -928,13 +930,13 @@ restore_check_irq_replay:
 	cmpwi	cr0,r3,0x500
 	bne	1f
 	addi	r3,r1,STACK_FRAME_OVERHEAD;
- 	bl	.do_IRQ
-	b	.ret_from_except
+ 	bl	do_IRQ
+	b	ret_from_except
 1:	cmpwi	cr0,r3,0x900
 	bne	1f
 	addi	r3,r1,STACK_FRAME_OVERHEAD;
-	bl	.timer_interrupt
-	b	.ret_from_except
+	bl	timer_interrupt
+	b	ret_from_except
 #ifdef CONFIG_PPC_DOORBELL
 1:
 #ifdef CONFIG_PPC_BOOK3E
@@ -948,14 +950,14 @@ restore_check_irq_replay:
 #endif /* CONFIG_PPC_BOOK3E */
 	bne	1f
 	addi	r3,r1,STACK_FRAME_OVERHEAD;
-	bl	.doorbell_exception
-	b	.ret_from_except
+	bl	doorbell_exception
+	b	ret_from_except
 #endif /* CONFIG_PPC_DOORBELL */
-1:	b	.ret_from_except /* What else to do here ? */
+1:	b	ret_from_except /* What else to do here ? */
  
 unrecov_restore:
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.unrecoverable_exception
+	bl	unrecoverable_exception
 	b	unrecov_restore
 
 #ifdef CONFIG_PPC_RTAS
@@ -1021,7 +1023,7 @@ _GLOBAL(enter_rtas)
         std	r6,PACASAVEDMSR(r13)
 
 	/* Setup our real return addr */	
-	LOAD_REG_ADDR(r4,.rtas_return_loc)
+	LOAD_REG_ADDR(r4,rtas_return_loc)
 	clrldi	r4,r4,2			/* convert to realmode address */
        	mtlr	r4
 
@@ -1045,7 +1047,7 @@ _GLOBAL(enter_rtas)
 	rfid
 	b	.	/* prevent speculative execution */
 
-_STATIC(rtas_return_loc)
+rtas_return_loc:
 	FIXUP_ENDIAN
 
 	/* relocation is off at this point */
@@ -1054,7 +1056,7 @@ _STATIC(rtas_return_loc)
 
 	bcl	20,31,$+4
 0:	mflr	r3
-	ld	r3,(1f-0b)(r3)		/* get &.rtas_restore_regs */
+	ld	r3,(1f-0b)(r3)		/* get &rtas_restore_regs */
 
 	mfmsr   r6
 	li	r0,MSR_RI
@@ -1071,9 +1073,9 @@ _STATIC(rtas_return_loc)
 	b	.	/* prevent speculative execution */
 
 	.align	3
-1:	.llong	.rtas_restore_regs
+1:	.llong	rtas_restore_regs
 
-_STATIC(rtas_restore_regs)
+rtas_restore_regs:
 	/* relocation is on at this point */
 	REST_GPR(2, r1)			/* Restore the TOC */
 	REST_GPR(13, r1)		/* Restore paca */
@@ -1173,7 +1175,7 @@ _GLOBAL(mcount)
 _GLOBAL(_mcount)
 	blr
 
-_GLOBAL(ftrace_caller)
+_GLOBAL_TOC(ftrace_caller)
 	/* Taken from output of objdump from lib64/glibc */
 	mflr	r3
 	ld	r11, 0(r1)
@@ -1197,10 +1199,7 @@ _GLOBAL(ftrace_graph_stub)
 _GLOBAL(ftrace_stub)
 	blr
 #else
-_GLOBAL(mcount)
-	blr
-
-_GLOBAL(_mcount)
+_GLOBAL_TOC(_mcount)
 	/* Taken from output of objdump from lib64/glibc */
 	mflr	r3
 	ld	r11, 0(r1)
@@ -1238,7 +1237,7 @@ _GLOBAL(ftrace_graph_caller)
 	ld	r11, 112(r1)
 	addi	r3, r11, 16
 
-	bl	.prepare_ftrace_return
+	bl	prepare_ftrace_return
 	nop
 
 	ld	r0, 128(r1)
@@ -1254,7 +1253,7 @@ _GLOBAL(return_to_handler)
 	mr	r31, r1
 	stdu	r1, -112(r1)
 
-	bl	.ftrace_return_to_handler
+	bl	ftrace_return_to_handler
 	nop
 
 	/* return value has real return address */
@@ -1284,7 +1283,7 @@ _GLOBAL(mod_return_to_handler)
 	 */
 	ld	r2, PACATOC(r13)
 
-	bl	.ftrace_return_to_handler
+	bl	ftrace_return_to_handler
 	nop
 
 	/* return value has real return address */

arch/powerpc/kernel/exceptions-64e.S

@@ -499,7 +499,7 @@ exc_##n##_bad_stack:							    \
 	CHECK_NAPPING();						\
 	addi	r3,r1,STACK_FRAME_OVERHEAD;				\
 	bl	hdlr;							\
-	b	.ret_from_except_lite;
+	b	ret_from_except_lite;
 
 /* This value is used to mark exception frames on the stack. */
 	.section	".toc","aw"
@@ -550,11 +550,11 @@ interrupt_end_book3e:
 	CRIT_EXCEPTION_PROLOG(0x100, BOOKE_INTERRUPT_CRITICAL,
 			      PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON_CRIT(0x100)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	bl	special_reg_save
 	CHECK_NAPPING();
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.unknown_exception
+	bl	unknown_exception
 	b	ret_from_crit_except
 
 /* Machine Check Interrupt */
@@ -562,11 +562,11 @@ interrupt_end_book3e:
 	MC_EXCEPTION_PROLOG(0x000, BOOKE_INTERRUPT_MACHINE_CHECK,
 			    PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON_MC(0x000)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	bl	special_reg_save
 	CHECK_NAPPING();
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.machine_check_exception
+	bl	machine_check_exception
 	b	ret_from_mc_except
 
 /* Data Storage Interrupt */
@@ -591,7 +591,7 @@ interrupt_end_book3e:
 
 /* External Input Interrupt */
 	MASKABLE_EXCEPTION(0x500, BOOKE_INTERRUPT_EXTERNAL,
-			   external_input, .do_IRQ, ACK_NONE)
+			   external_input, do_IRQ, ACK_NONE)
 
 /* Alignment */
 	START_EXCEPTION(alignment);
@@ -612,9 +612,9 @@ interrupt_end_book3e:
 	std	r14,_DSISR(r1)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	ld	r14,PACA_EXGEN+EX_R14(r13)
-	bl	.save_nvgprs
-	bl	.program_check_exception
-	b	.ret_from_except
+	bl	save_nvgprs
+	bl	program_check_exception
+	b	ret_from_except
 
 /* Floating Point Unavailable Interrupt */
 	START_EXCEPTION(fp_unavailable);
@@ -625,13 +625,13 @@ interrupt_end_book3e:
 	ld	r12,_MSR(r1)
 	andi.	r0,r12,MSR_PR;
 	beq-	1f
-	bl	.load_up_fpu
+	bl	load_up_fpu
 	b	fast_exception_return
 1:	INTS_DISABLE
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.kernel_fp_unavailable_exception
-	b	.ret_from_except
+	bl	kernel_fp_unavailable_exception
+	b	ret_from_except
 
 /* Altivec Unavailable Interrupt */
 	START_EXCEPTION(altivec_unavailable);
@@ -644,16 +644,16 @@ BEGIN_FTR_SECTION
 	ld	r12,_MSR(r1)
 	andi.	r0,r12,MSR_PR;
 	beq-	1f
-	bl	.load_up_altivec
+	bl	load_up_altivec
 	b	fast_exception_return
 1:
 END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 #endif
 	INTS_DISABLE
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.altivec_unavailable_exception
-	b	.ret_from_except
+	bl	altivec_unavailable_exception
+	b	ret_from_except
 
 /* AltiVec Assist */
 	START_EXCEPTION(altivec_assist);
@@ -662,39 +662,39 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 				PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x220)
 	INTS_DISABLE
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 #ifdef CONFIG_ALTIVEC
 BEGIN_FTR_SECTION
-	bl	.altivec_assist_exception
+	bl	altivec_assist_exception
 END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 #else
-	bl	.unknown_exception
+	bl	unknown_exception
 #endif
-	b	.ret_from_except
+	b	ret_from_except
 
 
 /* Decrementer Interrupt */
 	MASKABLE_EXCEPTION(0x900, BOOKE_INTERRUPT_DECREMENTER,
-			   decrementer, .timer_interrupt, ACK_DEC)
+			   decrementer, timer_interrupt, ACK_DEC)
 
 /* Fixed Interval Timer Interrupt */
 	MASKABLE_EXCEPTION(0x980, BOOKE_INTERRUPT_FIT,
-			   fixed_interval, .unknown_exception, ACK_FIT)
+			   fixed_interval, unknown_exception, ACK_FIT)
 
 /* Watchdog Timer Interrupt */
 	START_EXCEPTION(watchdog);
 	CRIT_EXCEPTION_PROLOG(0x9f0, BOOKE_INTERRUPT_WATCHDOG,
 			      PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON_CRIT(0x9f0)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	bl	special_reg_save
 	CHECK_NAPPING();
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 #ifdef CONFIG_BOOKE_WDT
-	bl	.WatchdogException
+	bl	WatchdogException
 #else
-	bl	.unknown_exception
+	bl	unknown_exception
 #endif
 	b	ret_from_crit_except
 
@@ -712,10 +712,10 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 				PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0xf20)
 	INTS_DISABLE
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.unknown_exception
-	b	.ret_from_except
+	bl	unknown_exception
+	b	ret_from_except
 
 /* Debug exception as a critical interrupt*/
 	START_EXCEPTION(debug_crit);
@@ -774,9 +774,9 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 	mr	r4,r14
 	ld	r14,PACA_EXCRIT+EX_R14(r13)
 	ld	r15,PACA_EXCRIT+EX_R15(r13)
-	bl	.save_nvgprs
-	bl	.DebugException
-	b	.ret_from_except
+	bl	save_nvgprs
+	bl	DebugException
+	b	ret_from_except
 
 kernel_dbg_exc:
 	b	.	/* NYI */
@@ -839,9 +839,9 @@ kernel_dbg_exc:
 	mr	r4,r14
 	ld	r14,PACA_EXDBG+EX_R14(r13)
 	ld	r15,PACA_EXDBG+EX_R15(r13)
-	bl	.save_nvgprs
-	bl	.DebugException
-	b	.ret_from_except
+	bl	save_nvgprs
+	bl	DebugException
+	b	ret_from_except
 
 	START_EXCEPTION(perfmon);
 	NORMAL_EXCEPTION_PROLOG(0x260, BOOKE_INTERRUPT_PERFORMANCE_MONITOR,
@@ -850,23 +850,23 @@ kernel_dbg_exc:
 	INTS_DISABLE
 	CHECK_NAPPING()
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.performance_monitor_exception
-	b	.ret_from_except_lite
+	bl	performance_monitor_exception
+	b	ret_from_except_lite
 
 /* Doorbell interrupt */
 	MASKABLE_EXCEPTION(0x280, BOOKE_INTERRUPT_DOORBELL,
-			   doorbell, .doorbell_exception, ACK_NONE)
+			   doorbell, doorbell_exception, ACK_NONE)
 
 /* Doorbell critical Interrupt */
 	START_EXCEPTION(doorbell_crit);
 	CRIT_EXCEPTION_PROLOG(0x2a0, BOOKE_INTERRUPT_DOORBELL_CRITICAL,
 			      PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON_CRIT(0x2a0)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	bl	special_reg_save
 	CHECK_NAPPING();
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.unknown_exception
+	bl	unknown_exception
 	b	ret_from_crit_except
 
 /*
@@ -878,21 +878,21 @@ kernel_dbg_exc:
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x2c0)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	INTS_RESTORE_HARD
-	bl	.unknown_exception
-	b	.ret_from_except
+	bl	unknown_exception
+	b	ret_from_except
 
 /* Guest Doorbell critical Interrupt */
 	START_EXCEPTION(guest_doorbell_crit);
 	CRIT_EXCEPTION_PROLOG(0x2e0, BOOKE_INTERRUPT_GUEST_DBELL_CRIT,
 			      PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON_CRIT(0x2e0)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	bl	special_reg_save
 	CHECK_NAPPING();
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.unknown_exception
+	bl	unknown_exception
 	b	ret_from_crit_except
 
 /* Hypervisor call */
@@ -901,10 +901,10 @@ kernel_dbg_exc:
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x310)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	INTS_RESTORE_HARD
-	bl	.unknown_exception
-	b	.ret_from_except
+	bl	unknown_exception
+	b	ret_from_except
 
 /* Embedded Hypervisor priviledged  */
 	START_EXCEPTION(ehpriv);
@@ -912,10 +912,10 @@ kernel_dbg_exc:
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x320)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	INTS_RESTORE_HARD
-	bl	.unknown_exception
-	b	.ret_from_except
+	bl	unknown_exception
+	b	ret_from_except
 
 /* LRAT Error interrupt */
 	START_EXCEPTION(lrat_error);
@@ -1014,16 +1014,16 @@ storage_fault_common:
 	mr	r5,r15
 	ld	r14,PACA_EXGEN+EX_R14(r13)
 	ld	r15,PACA_EXGEN+EX_R15(r13)
-	bl	.do_page_fault
+	bl	do_page_fault
 	cmpdi	r3,0
 	bne-	1f
-	b	.ret_from_except_lite
-1:	bl	.save_nvgprs
+	b	ret_from_except_lite
+1:	bl	save_nvgprs
 	mr	r5,r3
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	ld	r4,_DAR(r1)
-	bl	.bad_page_fault
-	b	.ret_from_except
+	bl	bad_page_fault
+	b	ret_from_except
 
 /*
  * Alignment exception doesn't fit entirely in the 0x100 bytes so it
@@ -1035,10 +1035,10 @@ alignment_more:
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	ld	r14,PACA_EXGEN+EX_R14(r13)
 	ld	r15,PACA_EXGEN+EX_R15(r13)
-	bl	.save_nvgprs
+	bl	save_nvgprs
 	INTS_RESTORE_HARD
-	bl	.alignment_exception
-	b	.ret_from_except
+	bl	alignment_exception
+	b	ret_from_except
 
 /*
  * We branch here from entry_64.S for the last stage of the exception
@@ -1172,7 +1172,7 @@ bad_stack_book3e:
 	std	r12,0(r11)
 	ld	r2,PACATOC(r13)
 1:	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.kernel_bad_stack
+	bl	kernel_bad_stack
 	b	1b
 
 /*
@@ -1521,13 +1521,13 @@ _GLOBAL(start_initialization_book3e)
 	 * and always use AS 0, so we just set it up to match our link
 	 * address and never use 0 based addresses.
 	 */
-	bl	.initial_tlb_book3e
+	bl	initial_tlb_book3e
 
 	/* Init global core bits */
-	bl	.init_core_book3e
+	bl	init_core_book3e
 
 	/* Init per-thread bits */
-	bl	.init_thread_book3e
+	bl	init_thread_book3e
 
 	/* Return to common init code */
 	tovirt(r28,r28)
@@ -1548,7 +1548,7 @@ _GLOBAL(start_initialization_book3e)
  */
 _GLOBAL(book3e_secondary_core_init_tlb_set)
 	li	r4,1
-	b	.generic_secondary_smp_init
+	b	generic_secondary_smp_init
 
 _GLOBAL(book3e_secondary_core_init)
 	mflr	r28
@@ -1558,18 +1558,18 @@ _GLOBAL(book3e_secondary_core_init)
 	bne	2f
 
 	/* Setup TLB for this core */
-	bl	.initial_tlb_book3e
+	bl	initial_tlb_book3e
 
 	/* We can return from the above running at a different
 	 * address, so recalculate r2 (TOC)
 	 */
-	bl	.relative_toc
+	bl	relative_toc
 
 	/* Init global core bits */
-2:	bl	.init_core_book3e
+2:	bl	init_core_book3e
 
 	/* Init per-thread bits */
-3:	bl	.init_thread_book3e
+3:	bl	init_thread_book3e
 
 	/* Return to common init code at proper virtual address.
 	 *
@@ -1596,14 +1596,14 @@ _GLOBAL(book3e_secondary_thread_init)
 	mflr	r28
 	b	3b
 
-_STATIC(init_core_book3e)
+init_core_book3e:
 	/* Establish the interrupt vector base */
 	LOAD_REG_IMMEDIATE(r3, interrupt_base_book3e)
 	mtspr	SPRN_IVPR,r3
 	sync
 	blr
 
-_STATIC(init_thread_book3e)
+init_thread_book3e:
 	lis	r3,(SPRN_EPCR_ICM | SPRN_EPCR_GICM)@h
 	mtspr	SPRN_EPCR,r3
 

arch/powerpc/kernel/ftrace.c

@@ -105,11 +105,9 @@ __ftrace_make_nop(struct module *mod,
 		  struct dyn_ftrace *rec, unsigned long addr)
 {
 	unsigned int op;
-	unsigned int jmp[5];
 	unsigned long ptr;
 	unsigned long ip = rec->ip;
-	unsigned long tramp;
-	int offset;
+	void *tramp;
 
 	/* read where this goes */
 	if (probe_kernel_read(&op, (void *)ip, sizeof(int)))
@@ -122,96 +120,41 @@ __ftrace_make_nop(struct module *mod,
 	}
 
 	/* lets find where the pointer goes */
-	tramp = find_bl_target(ip, op);
+	tramp = (void *)find_bl_target(ip, op);
 
-	/*
-	 * On PPC64 the trampoline looks like:
-	 * 0x3d, 0x82, 0x00, 0x00,    addis   r12,r2, <high>
-	 * 0x39, 0x8c, 0x00, 0x00,    addi    r12,r12, <low>
-	 *   Where the bytes 2,3,6 and 7 make up the 32bit offset
-	 *   to the TOC that holds the pointer.
-	 *   to jump to.
-	 * 0xf8, 0x41, 0x00, 0x28,    std     r2,40(r1)
-	 * 0xe9, 0x6c, 0x00, 0x20,    ld      r11,32(r12)
-	 *   The actually address is 32 bytes from the offset
-	 *   into the TOC.
-	 * 0xe8, 0x4c, 0x00, 0x28,    ld      r2,40(r12)
-	 */
+	pr_devel("ip:%lx jumps to %p", ip, tramp);
 
-	pr_devel("ip:%lx jumps to %lx r2: %lx", ip, tramp, mod->arch.toc);
-
-	/* Find where the trampoline jumps to */
-	if (probe_kernel_read(jmp, (void *)tramp, sizeof(jmp))) {
-		printk(KERN_ERR "Failed to read %lx\n", tramp);
-		return -EFAULT;
-	}
-
-	pr_devel(" %08x %08x", jmp[0], jmp[1]);
-
-	/* verify that this is what we expect it to be */
-	if (((jmp[0] & 0xffff0000) != 0x3d820000) ||
-	    ((jmp[1] & 0xffff0000) != 0x398c0000) ||
-	    (jmp[2] != 0xf8410028) ||
-	    (jmp[3] != 0xe96c0020) ||
-	    (jmp[4] != 0xe84c0028)) {
+	if (!is_module_trampoline(tramp)) {
 		printk(KERN_ERR "Not a trampoline\n");
 		return -EINVAL;
 	}
 
-	/* The bottom half is signed extended */
-	offset = ((unsigned)((unsigned short)jmp[0]) << 16) +
-		(int)((short)jmp[1]);
-
-	pr_devel(" %x ", offset);
-
-	/* get the address this jumps too */
-	tramp = mod->arch.toc + offset + 32;
-	pr_devel("toc: %lx", tramp);
-
-	if (probe_kernel_read(jmp, (void *)tramp, 8)) {
-		printk(KERN_ERR "Failed to read %lx\n", tramp);
+	if (module_trampoline_target(mod, tramp, &ptr)) {
+		printk(KERN_ERR "Failed to get trampoline target\n");
 		return -EFAULT;
 	}
 
-	pr_devel(" %08x %08x\n", jmp[0], jmp[1]);
-
-#ifdef __LITTLE_ENDIAN__
-	ptr = ((unsigned long)jmp[1] << 32) + jmp[0];
-#else
-	ptr = ((unsigned long)jmp[0] << 32) + jmp[1];
-#endif
+	pr_devel("trampoline target %lx", ptr);
 
 	/* This should match what was called */
 	if (ptr != ppc_function_entry((void *)addr)) {
-		printk(KERN_ERR "addr does not match %lx\n", ptr);
+		printk(KERN_ERR "addr %lx does not match expected %lx\n",
+			ptr, ppc_function_entry((void *)addr));
 		return -EINVAL;
 	}
 
 	/*
-	 * We want to nop the line, but the next line is
-	 *  0xe8, 0x41, 0x00, 0x28   ld r2,40(r1)
-	 * This needs to be turned to a nop too.
-	 */
-	if (probe_kernel_read(&op, (void *)(ip+4), MCOUNT_INSN_SIZE))
-		return -EFAULT;
-
-	if (op != 0xe8410028) {
-		printk(KERN_ERR "Next line is not ld! (%08x)\n", op);
-		return -EINVAL;
-	}
-
-	/*
-	 * Milton Miller pointed out that we can not blindly do nops.
-	 * If a task was preempted when calling a trace function,
-	 * the nops will remove the way to restore the TOC in r2
-	 * and the r2 TOC will get corrupted.
-	 */
-
-	/*
-	 * Replace:
-	 *   bl <tramp>  <==== will be replaced with "b 1f"
-	 *   ld r2,40(r1)
-	 *  1:
+	 * Our original call site looks like:
+	 *
+	 * bl <tramp>
+	 * ld r2,XX(r1)
+	 *
+	 * Milton Miller pointed out that we can not simply nop the branch.
+	 * If a task was preempted when calling a trace function, the nops
+	 * will remove the way to restore the TOC in r2 and the r2 TOC will
+	 * get corrupted.
+	 *
+	 * Use a b +8 to jump over the load.
 	 */
 	op = 0x48000008;	/* b +8 */
 
@@ -349,19 +292,24 @@ static int
 __ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 {
 	unsigned int op[2];
-	unsigned long ip = rec->ip;
+	void *ip = (void *)rec->ip;
 
 	/* read where this goes */
-	if (probe_kernel_read(op, (void *)ip, MCOUNT_INSN_SIZE * 2))
+	if (probe_kernel_read(op, ip, sizeof(op)))
 		return -EFAULT;
 
 	/*
-	 * It should be pointing to two nops or
-	 *  b +8; ld r2,40(r1)
+	 * We expect to see:
+	 *
+	 * b +8
+	 * ld r2,XX(r1)
+	 *
+	 * The load offset is different depending on the ABI. For simplicity
+	 * just mask it out when doing the compare.
 	 */
-	if (((op[0] != 0x48000008) || (op[1] != 0xe8410028)) &&
-	    ((op[0] != PPC_INST_NOP) || (op[1] != PPC_INST_NOP))) {
-		printk(KERN_ERR "Expected NOPs but have %x %x\n", op[0], op[1]);
+	if ((op[0] != 0x48000008) || ((op[1] & 0xffff00000) != 0xe8410000)) {
+		printk(KERN_ERR "Unexpected call sequence: %x %x\n",
+			op[0], op[1]);
 		return -EINVAL;
 	}
 
@@ -371,23 +319,16 @@ __ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
 		return -EINVAL;
 	}
 
-	/* create the branch to the trampoline */
-	op[0] = create_branch((unsigned int *)ip,
-			      rec->arch.mod->arch.tramp, BRANCH_SET_LINK);
-	if (!op[0]) {
-		printk(KERN_ERR "REL24 out of range!\n");
+	/* Ensure branch is within 24 bits */
+	if (create_branch(ip, rec->arch.mod->arch.tramp, BRANCH_SET_LINK)) {
+		printk(KERN_ERR "Branch out of range");
 		return -EINVAL;
 	}
 
-	/* ld r2,40(r1) */
-	op[1] = 0xe8410028;
-
-	pr_devel("write to %lx\n", rec->ip);
-
-	if (probe_kernel_write((void *)ip, op, MCOUNT_INSN_SIZE * 2))
-		return -EPERM;
-
-	flush_icache_range(ip, ip + 8);
+	if (patch_branch(ip, rec->arch.mod->arch.tramp, BRANCH_SET_LINK)) {
+		printk(KERN_ERR "REL24 out of range!\n");
+		return -EINVAL;
+	}
 
 	return 0;
 }