Merge tag 'arc-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc

Pull ARC updates from Vineet Gupta:

 - fixes for -Wmissing-prototype warnings

 - missing compiler barrier in relaxed atomics

 - some uaccess simplification, declutter

 - removal of massive glocal struct cpuinfo_arc from bootlog code

 - __switch_to consolidation (removal of inline asm variant)

 - use GP to cache task pointer (vs. r25)

 - misc rework of entry code

* tag 'arc-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc: (24 commits)
  ARC: boot log: fix warning
  arc: Explicitly include correct DT includes
  ARC: pt_regs: create seperate type for ecr
  ARCv2: entry: rearrange pt_regs slightly
  ARC: entry: replace 8 byte ADD.ne with 4 byte ADD2.ne
  ARC: entry: replace 8 byte OR with 4 byte BSET
  ARC: entry: Add more common chores to EXCEPTION_PROLOGUE
  ARC: entry: EV_MachineCheck dont re-read ECR
  ARC: entry: ARcompact EV_ProtV to use r10 directly
  ARC: entry: rework (non-functional)
  ARC: __switch_to: move ksp to thread_info from thread_struct
  ARC: __switch_to: asm with dwarf ops (vs. inline asm)
  ARC: kernel stack: INIT_THREAD need not setup @init_stack in @ksp
  ARC: entry: use gp to cache task pointer (vs. r25)
  ARC: boot log: eliminate struct cpuinfo_arc #4: boot log per ISA
  ARC: boot log: eliminate struct cpuinfo_arc #3: don't export
  ARC: boot log: eliminate struct cpuinfo_arc #2: cache
  ARC: boot log: eliminate struct cpuinfo_arc #1: mm
  ARCv2: memset: don't prefetch for len == 0 which happens a alot
  ARC: uaccess: elide unaliged handling if hardware supports
  ...

44 files changed, 724 insertions, 1033 deletions

diff --git a/arch/arc/Kconfig b/arch/arc/Kconfig
index 6f4995ad9873..3162db540ee9 100644
--- a/arch/arc/Kconfig
+++ b/arch/arc/Kconfig
@@ -27,6 +27,8 @@ config ARC
 	select GENERIC_SCHED_CLOCK
 	select GENERIC_SMP_IDLE_THREAD
 	select GENERIC_IOREMAP
+	select GENERIC_STRNCPY_FROM_USER if MMU
+	select GENERIC_STRNLEN_USER if MMU
 	select HAVE_ARCH_KGDB
 	select HAVE_ARCH_TRACEHOOK
 	select HAVE_ARCH_TRANSPARENT_HUGEPAGE if ARC_MMU_V4
@@ -491,11 +493,11 @@ config ARC_KVADDR_SIZE
 	  kernel-user gutter)
 
 config ARC_CURR_IN_REG
-	bool "Dedicate Register r25 for current_task pointer"
+	bool "cache current task pointer in gp"
 	default y
 	help
-	  This reserved Register R25 to point to Current Task in
-	  kernel mode. This saves memory access for each such access
+	  This reserves gp register to point to Current Task in
+	  kernel mode eliding memory access for each access
 
 
 config ARC_EMUL_UNALIGNED
diff --git a/arch/arc/Makefile b/arch/arc/Makefile
index 329400a1c355..2390dd042e36 100644
--- a/arch/arc/Makefile
+++ b/arch/arc/Makefile
@@ -28,14 +28,14 @@ cflags-y				+= $(tune-mcpu-def-y)
 endif
 endif
 
-
 ifdef CONFIG_ARC_CURR_IN_REG
 # For a global register definition, make sure it gets passed to every file
 # We had a customer reported bug where some code built in kernel was NOT using
-# any kernel headers, and missing the r25 global register
+# any kernel headers, and missing the global register
 # Can't do unconditionally because of recursive include issues
 # due to <linux/thread_info.h>
 LINUXINCLUDE	+=  -include $(srctree)/arch/arc/include/asm/current.h
+cflags-y	+= -ffixed-gp
 endif
 
 cflags-y				+= -fsection-anchors
@@ -67,7 +67,7 @@ cflags-$(CONFIG_ARC_DW2_UNWIND)		+= -fasynchronous-unwind-tables $(cfi)
 # small data is default for elf32 tool-chain. If not usable, disable it
 # This also allows repurposing GP as scratch reg to gcc reg allocator
 disable_small_data := y
-cflags-$(disable_small_data)		+= -mno-sdata -fcall-used-gp
+cflags-$(disable_small_data)		+= -mno-sdata
 
 cflags-$(CONFIG_CPU_BIG_ENDIAN)		+= -mbig-endian
 ldflags-$(CONFIG_CPU_BIG_ENDIAN)	+= -EB
diff --git a/arch/arc/include/asm/arcregs.h b/arch/arc/include/asm/arcregs.h
index 2162023195c5..4b13f60fe7ca 100644
--- a/arch/arc/include/asm/arcregs.h
+++ b/arch/arc/include/asm/arcregs.h
@@ -23,7 +23,7 @@
 #define ARC_REG_ICCM_BUILD	0x78	/* ICCM size (common) */
 #define ARC_REG_XY_MEM_BCR	0x79
 #define ARC_REG_MAC_BCR		0x7a
-#define ARC_REG_MUL_BCR		0x7b
+#define ARC_REG_MPY_BCR		0x7b
 #define ARC_REG_SWAP_BCR	0x7c
 #define ARC_REG_NORM_BCR	0x7d
 #define ARC_REG_MIXMAX_BCR	0x7e
@@ -177,7 +177,7 @@ struct bcr_isa_arcv2 {
 #endif
 };
 
-struct bcr_uarch_build_arcv2 {
+struct bcr_uarch_build {
 #ifdef CONFIG_CPU_BIG_ENDIAN
 	unsigned int pad:8, prod:8, maj:8, min:8;
 #else
@@ -185,6 +185,59 @@ struct bcr_uarch_build_arcv2 {
 #endif
 };
 
+struct bcr_mmu_3 {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int ver:8, ways:4, sets:4, res:3, sasid:1, pg_sz:4,
+		     u_itlb:4, u_dtlb:4;
+#else
+	unsigned int u_dtlb:4, u_itlb:4, pg_sz:4, sasid:1, res:3, sets:4,
+		     ways:4, ver:8;
+#endif
+};
+
+struct bcr_mmu_4 {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int ver:8, sasid:1, sz1:4, sz0:4, res:2, pae:1,
+		     n_ways:2, n_entry:2, n_super:2, u_itlb:3, u_dtlb:3;
+#else
+	/*           DTLB      ITLB      JES        JE         JA      */
+	unsigned int u_dtlb:3, u_itlb:3, n_super:2, n_entry:2, n_ways:2,
+		     pae:1, res:2, sz0:4, sz1:4, sasid:1, ver:8;
+#endif
+};
+
+struct bcr_cache {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int pad:12, line_len:4, sz:4, config:4, ver:8;
+#else
+	unsigned int ver:8, config:4, sz:4, line_len:4, pad:12;
+#endif
+};
+
+struct bcr_slc_cfg {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int pad:24, way:2, lsz:2, sz:4;
+#else
+	unsigned int sz:4, lsz:2, way:2, pad:24;
+#endif
+};
+
+struct bcr_clust_cfg {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int pad:7, c:1, num_entries:8, num_cores:8, ver:8;
+#else
+	unsigned int ver:8, num_cores:8, num_entries:8, c:1, pad:7;
+#endif
+};
+
+struct bcr_volatile {
+#ifdef CONFIG_CPU_BIG_ENDIAN
+	unsigned int start:4, limit:4, pad:22, order:1, disable:1;
+#else
+	unsigned int disable:1, order:1, pad:22, limit:4, start:4;
+#endif
+};
+
 struct bcr_mpy {
 #ifdef CONFIG_CPU_BIG_ENDIAN
 	unsigned int pad:8, x1616:8, dsp:4, cycles:2, type:2, ver:8;
@@ -302,48 +355,6 @@ struct bcr_generic {
 #endif
 };
 
-/*
- *******************************************************************
- * Generic structures to hold build configuration used at runtime
- */
-
-struct cpuinfo_arc_mmu {
-	unsigned int ver:4, pg_sz_k:8, s_pg_sz_m:8, pad:10, sasid:1, pae:1;
-	unsigned int sets:12, ways:4, u_dtlb:8, u_itlb:8;
-};
-
-struct cpuinfo_arc_cache {
-	unsigned int sz_k:14, line_len:8, assoc:4, alias:1, vipt:1, pad:4;
-};
-
-struct cpuinfo_arc_bpu {
-	unsigned int ver, full, num_cache, num_pred, ret_stk;
-};
-
-struct cpuinfo_arc_ccm {
-	unsigned int base_addr, sz;
-};
-
-struct cpuinfo_arc {
-	struct cpuinfo_arc_cache icache, dcache, slc;
-	struct cpuinfo_arc_mmu mmu;
-	struct cpuinfo_arc_bpu bpu;
-	struct bcr_identity core;
-	struct bcr_isa_arcv2 isa;
-	const char *release, *name;
-	unsigned int vec_base;
-	struct cpuinfo_arc_ccm iccm, dccm;
-	struct {
-		unsigned int swap:1, norm:1, minmax:1, barrel:1, crc:1, swape:1, pad1:2,
-			     fpu_sp:1, fpu_dp:1, dual:1, dual_enb:1, pad2:4,
-			     ap_num:4, ap_full:1, smart:1, rtt:1, pad3:1,
-			     timer0:1, timer1:1, rtc:1, gfrc:1, pad4:4;
-	} extn;
-	struct bcr_mpy extn_mpy;
-};
-
-extern struct cpuinfo_arc cpuinfo_arc700[];
-
 static inline int is_isa_arcv2(void)
 {
 	return IS_ENABLED(CONFIG_ISA_ARCV2);
diff --git a/arch/arc/include/asm/atomic-llsc.h b/arch/arc/include/asm/atomic-llsc.h
index 1b0ffaeee16d..5258cb81a16b 100644
--- a/arch/arc/include/asm/atomic-llsc.h
+++ b/arch/arc/include/asm/atomic-llsc.h
@@ -18,7 +18,7 @@ static inline void arch_atomic_##op(int i, atomic_t *v)			\
 	: [val]	"=&r"	(val) /* Early clobber to prevent reg reuse */	\
 	: [ctr]	"r"	(&v->counter), /* Not "m": llock only supports reg direct addr mode */	\
 	  [i]	"ir"	(i)						\
-	: "cc");							\
+	: "cc", "memory");						\
 }									\
 
 #define ATOMIC_OP_RETURN(op, asm_op)				\
@@ -34,7 +34,7 @@ static inline int arch_atomic_##op##_return_relaxed(int i, atomic_t *v)	\
 	: [val]	"=&r"	(val)						\
 	: [ctr]	"r"	(&v->counter),					\
 	  [i]	"ir"	(i)						\
-	: "cc");							\
+	: "cc", "memory");						\
 									\
 	return val;							\
 }
@@ -56,7 +56,7 @@ static inline int arch_atomic_fetch_##op##_relaxed(int i, atomic_t *v)	\
 	  [orig] "=&r" (orig)						\
 	: [ctr]	"r"	(&v->counter),					\
 	  [i]	"ir"	(i)						\
-	: "cc");							\
+	: "cc", "memory");						\
 									\
 	return orig;							\
 }
diff --git a/arch/arc/include/asm/atomic64-arcv2.h b/arch/arc/include/asm/atomic64-arcv2.h
index 6b6db981967a..9b5791b85471 100644
--- a/arch/arc/include/asm/atomic64-arcv2.h
+++ b/arch/arc/include/asm/atomic64-arcv2.h
@@ -60,7 +60,7 @@ static inline void arch_atomic64_##op(s64 a, atomic64_t *v)		\
 	"	bnz     1b		\n"				\
 	: "=&r"(val)							\
 	: "r"(&v->counter), "ir"(a)					\
-	: "cc");							\
+	: "cc", "memory");						\
 }									\
 
 #define ATOMIC64_OP_RETURN(op, op1, op2)		        	\
@@ -77,7 +77,7 @@ static inline s64 arch_atomic64_##op##_return_relaxed(s64 a, atomic64_t *v)	\
 	"	bnz     1b		\n"				\
 	: [val] "=&r"(val)						\
 	: "r"(&v->counter), "ir"(a)					\
-	: "cc");	/* memory clobber comes from smp_mb() */	\
+	: "cc", "memory");						\
 									\
 	return val;							\
 }
@@ -99,7 +99,7 @@ static inline s64 arch_atomic64_fetch_##op##_relaxed(s64 a, atomic64_t *v)	\
 	"	bnz     1b		\n"				\
 	: "=&r"(orig), "=&r"(val)					\
 	: "r"(&v->counter), "ir"(a)					\
-	: "cc");	/* memory clobber comes from smp_mb() */	\
+	: "cc", "memory");						\
 									\
 	return orig;							\
 }
diff --git a/arch/arc/include/asm/current.h b/arch/arc/include/asm/current.h
index 9b9bdd3e6538..06be89f6f2f0 100644
--- a/arch/arc/include/asm/current.h
+++ b/arch/arc/include/asm/current.h
@@ -13,7 +13,7 @@
 
 #ifdef CONFIG_ARC_CURR_IN_REG
 
-register struct task_struct *curr_arc asm("r25");
+register struct task_struct *curr_arc asm("gp");
 #define current (curr_arc)
 
 #else
diff --git a/arch/arc/include/asm/dwarf.h b/arch/arc/include/asm/dwarf.h
index 5f4de05bd4ee..a0d5ebe1bc3f 100644
--- a/arch/arc/include/asm/dwarf.h
+++ b/arch/arc/include/asm/dwarf.h
@@ -10,23 +10,31 @@
 
 #ifdef ARC_DW2_UNWIND_AS_CFI
 
-#define CFI_STARTPROC	.cfi_startproc
-#define CFI_ENDPROC	.cfi_endproc
-#define CFI_DEF_CFA	.cfi_def_cfa
-#define CFI_REGISTER	.cfi_register
-#define CFI_REL_OFFSET	.cfi_rel_offset
-#define CFI_UNDEFINED	.cfi_undefined
+#define CFI_STARTPROC		.cfi_startproc
+#define CFI_ENDPROC		.cfi_endproc
+#define CFI_DEF_CFA		.cfi_def_cfa
+#define CFI_DEF_CFA_OFFSET	.cfi_def_cfa_offset
+#define CFI_DEF_CFA_REGISTER	.cfi_def_cfa_register
+#define CFI_OFFSET		.cfi_offset
+#define CFI_REL_OFFSET		.cfi_rel_offset
+#define CFI_REGISTER		.cfi_register
+#define CFI_RESTORE		.cfi_restore
+#define CFI_UNDEFINED		.cfi_undefined
 
 #else
 
 #define CFI_IGNORE	#
 
-#define CFI_STARTPROC	CFI_IGNORE
-#define CFI_ENDPROC	CFI_IGNORE
-#define CFI_DEF_CFA	CFI_IGNORE
-#define CFI_REGISTER	CFI_IGNORE
-#define CFI_REL_OFFSET	CFI_IGNORE
-#define CFI_UNDEFINED	CFI_IGNORE
+#define CFI_STARTPROC		CFI_IGNORE
+#define CFI_ENDPROC		CFI_IGNORE
+#define CFI_DEF_CFA		CFI_IGNORE
+#define CFI_DEF_CFA_OFFSET	CFI_IGNORE
+#define CFI_DEF_CFA_REGISTER	CFI_IGNORE
+#define CFI_OFFSET		CFI_IGNORE
+#define CFI_REL_OFFSET		CFI_IGNORE
+#define CFI_REGISTER		CFI_IGNORE
+#define CFI_RESTORE		CFI_IGNORE
+#define CFI_UNDEFINED		CFI_IGNORE
 
 #endif	/* !ARC_DW2_UNWIND_AS_CFI */
 
diff --git a/arch/arc/include/asm/entry-arcv2.h b/arch/arc/include/asm/entry-arcv2.h
index 0ff4c0610561..4d13320e0c1b 100644
--- a/arch/arc/include/asm/entry-arcv2.h
+++ b/arch/arc/include/asm/entry-arcv2.h
@@ -18,7 +18,6 @@
  *              |      orig_r0      |
  *              |      event/ECR    |
  *              |      bta          |
- *              |      user_r25     |
  *              |      gp           |
  *              |      fp           |
  *              |      sp           |
@@ -49,14 +48,18 @@
 /*------------------------------------------------------------------------*/
 .macro INTERRUPT_PROLOGUE
 
-	; (A) Before jumping to Interrupt Vector, hardware micro-ops did following:
+	; Before jumping to Interrupt Vector, hardware micro-ops did following:
 	;   1. SP auto-switched to kernel mode stack
 	;   2. STATUS32.Z flag set if in U mode at time of interrupt (U:1,K:0)
 	;   3. Auto save: (mandatory) Push PC and STAT32 on stack
 	;                 hardware does even if CONFIG_ARC_IRQ_NO_AUTOSAVE
-	;   4. Auto save: (optional) r0-r11, blink, LPE,LPS,LPC, JLI,LDI,EI
+	;  4a. Auto save: (optional) r0-r11, blink, LPE,LPS,LPC, JLI,LDI,EI
 	;
-	; (B) Manually saved some regs: r12,r25,r30, sp,fp,gp, ACCL pair
+	; Now
+	;  4b. If Auto-save (optional) not enabled in hw, manually save them
+	;   5. Manually save: r12,r30, sp,fp,gp, ACCL pair
+	;
+	; At the end, SP points to pt_regs
 
 #ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
 	; carve pt_regs on stack (case #3), PC/STAT32 already on stack
@@ -72,15 +75,16 @@
 .endm
 
 /*------------------------------------------------------------------------*/
-.macro EXCEPTION_PROLOGUE
+.macro EXCEPTION_PROLOGUE_KEEP_AE
 
-	; (A) Before jumping to Exception Vector, hardware micro-ops did following:
+	; Before jumping to Exception Vector, hardware micro-ops did following:
 	;   1. SP auto-switched to kernel mode stack
 	;   2. STATUS32.Z flag set if in U mode at time of exception (U:1,K:0)
 	;
-	; (B) Manually save the complete reg file below
+	; Now manually save rest of reg file
+	; At the end, SP points to pt_regs
 
-	sub	sp, sp, SZ_PT_REGS	; carve pt_regs
+	sub	sp, sp, SZ_PT_REGS	; carve space for pt_regs
 
 	; _HARD saves r10 clobbered by _SOFT as scratch hence comes first
 
@@ -100,6 +104,16 @@
 	; OUTPUT: r10 has ECR expected by EV_Trap
 .endm
 
+.macro EXCEPTION_PROLOGUE
+
+	EXCEPTION_PROLOGUE_KEEP_AE	; return ECR in r10
+
+	lr  r0, [efa]
+	mov r1, sp
+
+	FAKE_RET_FROM_EXCPN		; clobbers r9
+.endm
+
 /*------------------------------------------------------------------------
  * This macro saves the registers manually which would normally be autosaved
  * by hardware on taken interrupts. It is used by
@@ -135,10 +149,10 @@
  */
 .macro __SAVE_REGFILE_SOFT
 
-	ST2	gp, fp, PT_r26		; gp (r26), fp (r27)
-
-	st	r12, [sp, PT_sp + 4]
-	st	r30, [sp, PT_sp + 8]
+	st	fp,  [sp, PT_fp]	; r27
+	st	r30, [sp, PT_r30]
+	st	r12, [sp, PT_r12]
+	st	r26, [sp, PT_r26]	; gp
 
 	; Saving pt_regs->sp correctly requires some extra work due to the way
 	; Auto stack switch works
@@ -153,30 +167,30 @@
 
 	; ISA requires ADD.nz to have same dest and src reg operands
 	mov.nz	r10, sp
-	add.nz	r10, r10, SZ_PT_REGS	; K mode SP
+	add2.nz	r10, r10, SZ_PT_REGS/4	; K mode SP
 
 	st	r10, [sp, PT_sp]	; SP (pt_regs->sp)
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	st	r25, [sp, PT_user_r25]
-	GET_CURR_TASK_ON_CPU	r25
-#endif
-
 #ifdef CONFIG_ARC_HAS_ACCL_REGS
 	ST2	r58, r59, PT_r58
 #endif
 
 	/* clobbers r10, r11 registers pair */
 	DSP_SAVE_REGFILE_IRQ
+
+#ifdef CONFIG_ARC_CURR_IN_REG
+	GET_CURR_TASK_ON_CPU	gp
+#endif
+
 .endm
 
 /*------------------------------------------------------------------------*/
 .macro __RESTORE_REGFILE_SOFT
 
-	LD2	gp, fp, PT_r26		; gp (r26), fp (r27)
-
-	ld	r12, [sp, PT_r12]
+	ld	fp,  [sp, PT_fp]
 	ld	r30, [sp, PT_r30]
+	ld	r12, [sp, PT_r12]
+	ld	r26, [sp, PT_r26]
 
 	; Restore SP (into AUX_USER_SP) only if returning to U mode
 	;  - for K mode, it will be implicitly restored as stack is unwound
@@ -188,10 +202,6 @@
 	sr	r10, [AUX_USER_SP]
 1:
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	ld	r25, [sp, PT_user_r25]
-#endif
-
 	/* clobbers r10, r11 registers pair */
 	DSP_RESTORE_REGFILE_IRQ
 
@@ -249,7 +259,7 @@
 
 	btst	r0, STATUS_U_BIT	; Z flag set if K, used in restoring SP
 
-	ld	r10, [sp, PT_event + 4]
+	ld	r10, [sp, PT_bta]
 	sr	r10, [erbta]
 
 	LD2	r10, r11, PT_ret
@@ -264,8 +274,8 @@
 
 .macro FAKE_RET_FROM_EXCPN
 	lr      r9, [status32]
-	bic     r9, r9, STATUS_AE_MASK
-	or      r9, r9, STATUS_IE_MASK
+	bclr    r9, r9, STATUS_AE_BIT
+	bset    r9, r9, STATUS_IE_BIT
 	kflag   r9
 .endm
 
diff --git a/arch/arc/include/asm/entry-compact.h b/arch/arc/include/asm/entry-compact.h
index 67ff06e15cea..a0e760eb35a8 100644
--- a/arch/arc/include/asm/entry-compact.h
+++ b/arch/arc/include/asm/entry-compact.h
@@ -140,7 +140,7 @@
  *
  * After this it is safe to call the "C" handlers
  *-------------------------------------------------------------*/
-.macro EXCEPTION_PROLOGUE
+.macro EXCEPTION_PROLOGUE_KEEP_AE
 
 	/* Need at least 1 reg to code the early exception prologue */
 	PROLOG_FREEUP_REG r9, @ex_saved_reg1
@@ -151,14 +151,6 @@
 	/* ARC700 doesn't provide auto-stack switching */
 	SWITCH_TO_KERNEL_STK
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	/* Treat r25 as scratch reg (save on stack) and load with "current" */
-	PUSH    r25
-	GET_CURR_TASK_ON_CPU   r25
-#else
-	sub     sp, sp, 4
-#endif
-
 	st.a	r0, [sp, -8]    /* orig_r0 needed for syscall (skip ECR slot) */
 	sub	sp, sp, 4	/* skip pt_regs->sp, already saved above */
 
@@ -178,7 +170,23 @@
 	PUSHAX	erbta
 
 	lr	r10, [ecr]
-	st      r10, [sp, PT_event]    /* EV_Trap expects r10 to have ECR */
+	st      r10, [sp, PT_event]
+
+#ifdef CONFIG_ARC_CURR_IN_REG
+	/* gp already saved on stack: now load with "current" */
+	GET_CURR_TASK_ON_CPU   gp
+#endif
+	; OUTPUT: r10 has ECR expected by EV_Trap
+.endm
+
+.macro EXCEPTION_PROLOGUE
+
+	EXCEPTION_PROLOGUE_KEEP_AE	; return ECR in r10
+
+	lr  r0, [efa]
+	mov r1, sp
+
+	FAKE_RET_FROM_EXCPN		; clobbers r9
 .endm
 
 /*--------------------------------------------------------------
@@ -208,11 +216,8 @@
 	POP	gp
 	RESTORE_R12_TO_R0
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	ld	r25, [sp, 12]
-#endif
 	ld  sp, [sp] /* restore original sp */
-	/* orig_r0, ECR, user_r25 skipped automatically */
+	/* orig_r0, ECR skipped automatically */
 .endm
 
 /* Dummy ECR values for Interrupts */
@@ -229,13 +234,6 @@
 
 	SWITCH_TO_KERNEL_STK
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	/* Treat r25 as scratch reg (save on stack) and load with "current" */
-	PUSH    r25
-	GET_CURR_TASK_ON_CPU   r25
-#else
-	sub     sp, sp, 4
-#endif
 
 	PUSH	0x003\LVL\()abcd    /* Dummy ECR */
 	sub	sp, sp, 8	    /* skip orig_r0 (not needed)
@@ -255,6 +253,10 @@
 	PUSHAX	lp_start
 	PUSHAX	bta_l\LVL\()
 
+#ifdef CONFIG_ARC_CURR_IN_REG
+	/* gp already saved on stack: now load with "current" */
+	GET_CURR_TASK_ON_CPU   gp
+#endif
 .endm
 
 /*--------------------------------------------------------------
@@ -282,11 +284,7 @@
 	POP	gp
 	RESTORE_R12_TO_R0
 
-#ifdef CONFIG_ARC_CURR_IN_REG
-	ld	r25, [sp, 12]
-#endif
-	ld  sp, [sp] /* restore original sp */
-	/* orig_r0, ECR, user_r25 skipped automatically */
+	ld  sp, [sp] /* restore original sp; orig_r0, ECR skipped implicitly */
 .endm
 
 /* Get thread_info of "current" tsk */
diff --git a/arch/arc/include/asm/entry.h b/arch/arc/include/asm/entry.h
index fcdd59d77f42..49c2e090cb5c 100644
--- a/arch/arc/include/asm/entry.h
+++ b/arch/arc/include/asm/entry.h
@@ -13,6 +13,8 @@
 #include <asm/processor.h>	/* For VMALLOC_START */
 #include <asm/mmu.h>
 
+#ifdef __ASSEMBLY__
+
 #ifdef CONFIG_ISA_ARCOMPACT
 #include <asm/entry-compact.h>	/* ISA specific bits */
 #else
@@ -89,7 +91,7 @@
  * Helpers to save/restore callee-saved regs:
  * used by several macros below
  *-------------------------------------------------------------*/
-.macro SAVE_R13_TO_R24
+.macro SAVE_R13_TO_R25
 	PUSH	r13
 	PUSH	r14
 	PUSH	r15
@@ -102,9 +104,11 @@
 	PUSH	r22
 	PUSH	r23
 	PUSH	r24
+	PUSH	r25
 .endm
 
-.macro RESTORE_R24_TO_R13
+.macro RESTORE_R25_TO_R13
+	POP	r25
 	POP	r24
 	POP	r23
 	POP	r22
@@ -119,81 +123,31 @@
 	POP	r13
 .endm
 
-/*--------------------------------------------------------------
- * Collect User Mode callee regs as struct callee_regs - needed by
- * fork/do_signal/unaligned-access-emulation.
- * (By default only scratch regs are saved on entry to kernel)
- *
- * Special handling for r25 if used for caching Task Pointer.
- * It would have been saved in task->thread.user_r25 already, but to keep
- * the interface same it is copied into regular r25 placeholder in
- * struct callee_regs.
- *-------------------------------------------------------------*/
+/*
+ * save user mode callee regs as struct callee_regs
+ *  - needed by fork/do_signal/unaligned-access-emulation.
+ */
 .macro SAVE_CALLEE_SAVED_USER
+	SAVE_R13_TO_R25
+.endm
 
-	mov	r12, sp		; save SP as ref to pt_regs
-	SAVE_R13_TO_R24
-
-#ifdef CONFIG_ARC_CURR_IN_REG
-	; Retrieve orig r25 and save it with rest of callee_regs
-	ld	r12, [r12, PT_user_r25]
-	PUSH	r12
-#else
-	PUSH	r25
-#endif
-
+/*
+ * restore user mode callee regs as struct callee_regs
+ *  - could have been changed by ptrace tracer or unaligned-access fixup
+ */
+.macro RESTORE_CALLEE_SAVED_USER
+	RESTORE_R25_TO_R13
 .endm
 
-/*--------------------------------------------------------------
- * Save kernel Mode callee regs at the time of Contect Switch.
- *
- * Special handling for r25 if used for caching Task Pointer.
- * Kernel simply skips saving it since it will be loaded with
- * incoming task pointer anyways
- *-------------------------------------------------------------*/
+/*
+ * save/restore kernel mode callee regs at the time of context switch
+ */
 .macro SAVE_CALLEE_SAVED_KERNEL
-
-	SAVE_R13_TO_R24
-
-#ifdef CONFIG_ARC_CURR_IN_REG
-	sub     sp, sp, 4
-#else
-	PUSH	r25
-#endif
+	SAVE_R13_TO_R25
 .endm
 
-/*--------------------------------------------------------------
- * Opposite of SAVE_CALLEE_SAVED_KERNEL
- *-------------------------------------------------------------*/
 .macro RESTORE_CALLEE_SAVED_KERNEL
-
-#ifdef CONFIG_ARC_CURR_IN_REG
-	add     sp, sp, 4  /* skip usual r25 placeholder */
-#else
-	POP	r25
-#endif
-	RESTORE_R24_TO_R13
-.endm
-
-/*--------------------------------------------------------------
- * Opposite of SAVE_CALLEE_SAVED_USER
- *
- * ptrace tracer or unaligned-access fixup might have changed a user mode
- * callee reg which is saved back to usual r25 storage location
- *-------------------------------------------------------------*/
-.macro RESTORE_CALLEE_SAVED_USER
-
-#ifdef CONFIG_ARC_CURR_IN_REG
-	POP	r12
-#else
-	POP	r25
-#endif
-	RESTORE_R24_TO_R13
-
-	; SP is back to start of pt_regs
-#ifdef CONFIG_ARC_CURR_IN_REG
-	st	r12, [sp, PT_user_r25]
-#endif
+	RESTORE_R25_TO_R13
 .endm
 
 /*--------------------------------------------------------------
@@ -229,10 +183,10 @@
 
 #ifdef CONFIG_SMP
 
-/*-------------------------------------------------
+/*
  * Retrieve the current running task on this CPU
- * 1. Determine curr CPU id.
- * 2. Use it to index into _current_task[ ]
+ *  - loads it from backing _current_task[] (and can't use the
+ *    caching reg for current task
  */
 .macro  GET_CURR_TASK_ON_CPU   reg
 	GET_CPU_ID  \reg
@@ -254,7 +208,7 @@
 	add2 \tmp, @_current_task, \tmp
 	st   \tsk, [\tmp]
 #