Merge tag 'arc-4.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc

Pull ARC architecture updates from Vineet Gupta:
 "ARC updates for 4.3:

   - perf support for ARCv2 based cores (sampling interrupt, SMP)
   - leftovers for ARCv2 support
   - futex fixes"

* tag 'arc-4.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  ARCv2: entry: Fix reserved handler
  ARCv2: perf: Finally introduce HS perf unit
  ARCv2: perf: SMP support
  ARCv2: perf: implement exclusion of event counting in user or kernel mode
  ARCv2: perf: Support sampling events using overflow interrupts
  ARCv2: perf: implement "event_set_period"
  ARC: perf: cap the number of counters to hardware max of 32
  ARC: Eliminate some ARCv2 specific code for ARCompact build
  ARC: add/fix some comments in code - no functional change
  ARC: change some branchs to jumps to resolve linkage errors
  ARC: ensure futex ops are atomic in !LLSC config
  ARC: Enable HAVE_FUTEX_CMPXCHG
  ARC: make futex_atomic_cmpxchg_inatomic() return bimodal
  ARC: futex cosmetics
  ARC: add barriers to futex code
  ARCv2: IOC: Allow boot time disable
  ARCv2: SLC: Allow boot time disable
  ARCv2: Support IO Coherency and permutations involving L1 and L2 caches
  ARC: Enable optimistic spinning for LLSC config
  MAINTAINERS: add git tree for the arc architecture

17 files changed, 524 insertions, 139 deletions

diff --git a/Documentation/devicetree/bindings/arc/archs-pct.txt b/Documentation/devicetree/bindings/arc/archs-pct.txt
new file mode 100644
index 000000000000..1ae98b87c640
--- /dev/null
+++ b/Documentation/devicetree/bindings/arc/archs-pct.txt
@@ -0,0 +1,17 @@
+* ARC HS Performance Counters
+
+The ARC HS can be configured with a pipeline performance monitor for counting
+CPU and cache events like cache misses and hits. Like conventional PCT there
+are 100+ hardware conditions dynamically mapped to upto 32 counters.
+It also supports overflow interrupts.
+
+Required properties:
+
+- compatible : should contain
+	"snps,archs-pct"
+
+Example:
+
+pmu {
+        compatible = "snps,archs-pct";
+};
diff --git a/MAINTAINERS b/MAINTAINERS
index 963233fc721b..4be50b79914e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9911,8 +9911,9 @@ SYNOPSYS ARC ARCHITECTURE
 M:	Vineet Gupta <vgupta@synopsys.com>
 S:	Supported
 F:	arch/arc/
-F:	Documentation/devicetree/bindings/arc/
+F:	Documentation/devicetree/bindings/arc/*
 F:	drivers/tty/serial/arc_uart.c
+T:	git git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc.git
 
 SYNOPSYS ARC SDP platform support
 M:	Alexey Brodkin <abrodkin@synopsys.com>
diff --git a/arch/arc/Kconfig b/arch/arc/Kconfig
index bd4670d1b89b..78c0621d5819 100644
--- a/arch/arc/Kconfig
+++ b/arch/arc/Kconfig
@@ -8,6 +8,7 @@
 
 config ARC
 	def_bool y
+	select ARCH_SUPPORTS_ATOMIC_RMW if ARC_HAS_LLSC
 	select BUILDTIME_EXTABLE_SORT
 	select COMMON_CLK
 	select CLONE_BACKWARDS
@@ -22,6 +23,7 @@ config ARC
 	select GENERIC_SMP_IDLE_THREAD
 	select HAVE_ARCH_KGDB
 	select HAVE_ARCH_TRACEHOOK
+	select HAVE_FUTEX_CMPXCHG
 	select HAVE_IOREMAP_PROT
 	select HAVE_KPROBES
 	select HAVE_KRETPROBES
diff --git a/arch/arc/boot/dts/axc003.dtsi b/arch/arc/boot/dts/axc003.dtsi
index 1cd5e82f5dc2..846481f37eef 100644
--- a/arch/arc/boot/dts/axc003.dtsi
+++ b/arch/arc/boot/dts/axc003.dtsi
@@ -72,12 +72,13 @@
 	};
 
 	/*
-	 * This INTC is actually connected to DW APB GPIO
-	 * which acts as a wire between MB INTC and CPU INTC.
-	 * GPIO INTC is configured in platform init code
-	 * and here we mimic direct connection from MB INTC to
-	 * CPU INTC, thus we set "interrupts = <7>" instead of
-	 * "interrupts = <12>"
+	 * The DW APB ICTL intc on MB is connected to CPU intc via a
+	 * DT "invisible" DW APB GPIO block, configured to simply pass thru
+	 * interrupts - setup accordinly in platform init (plat-axs10x/ax10x.c)
+	 *
+	 * So here we mimic a direct connection betwen them, ignoring the
+	 * ABPG GPIO. Thus set "interrupts = <24>" (DW APB GPIO to core)
+	 * instead of "interrupts = <12>" (DW APB ICTL to DW APB GPIO)
 	 *
 	 * This intc actually resides on MB, but we move it here to
 	 * avoid duplicating the MB dtsi file given that IRQ from
diff --git a/arch/arc/include/asm/arcregs.h b/arch/arc/include/asm/arcregs.h
index c8f57b8449dc..d8023bc8d1ad 100644
--- a/arch/arc/include/asm/arcregs.h
+++ b/arch/arc/include/asm/arcregs.h
@@ -35,6 +35,7 @@
 #define ARC_REG_RTT_BCR		0xF2
 #define ARC_REG_IRQ_BCR		0xF3
 #define ARC_REG_SMART_BCR	0xFF
+#define ARC_REG_CLUSTER_BCR	0xcf
 
 /* status32 Bits Positions */
 #define STATUS_AE_BIT		5	/* Exception active */
diff --git a/arch/arc/include/asm/cache.h b/arch/arc/include/asm/cache.h
index d67345d3e2d4..e23ea6e7633a 100644
--- a/arch/arc/include/asm/cache.h
+++ b/arch/arc/include/asm/cache.h
@@ -53,6 +53,8 @@ extern void arc_cache_init(void);
 extern char *arc_cache_mumbojumbo(int cpu_id, char *buf, int len);
 extern void read_decode_cache_bcr(void);
 
+extern int ioc_exists;
+
 #endif	/* !__ASSEMBLY__ */
 
 /* Instruction cache related Auxiliary registers */
@@ -94,4 +96,10 @@ extern void read_decode_cache_bcr(void);
 #define SLC_CTRL_BUSY		0x100
 #define SLC_CTRL_RGN_OP_INV	0x200
 
+/* IO coherency related Auxiliary registers */
+#define ARC_REG_IO_COH_ENABLE	0x500
+#define ARC_REG_IO_COH_PARTIAL	0x501
+#define ARC_REG_IO_COH_AP0_BASE	0x508
+#define ARC_REG_IO_COH_AP0_SIZE	0x509
+
 #endif /* _ASM_CACHE_H */
diff --git a/arch/arc/include/asm/cmpxchg.h b/arch/arc/include/asm/cmpxchg.h
index 44fd531f4d7b..af7a2db139c9 100644
--- a/arch/arc/include/asm/cmpxchg.h
+++ b/arch/arc/include/asm/cmpxchg.h
@@ -110,18 +110,18 @@ static inline unsigned long __xchg(unsigned long val, volatile void *ptr,
 						 sizeof(*(ptr))))
 
 /*
- * On ARC700, EX insn is inherently atomic, so by default "vanilla" xchg() need
- * not require any locking. However there's a quirk.
- * ARC lacks native CMPXCHG, thus emulated (see above), using external locking -
- * incidently it "reuses" the same atomic_ops_lock used by atomic APIs.
- * Now, llist code uses cmpxchg() and xchg() on same data, so xchg() needs to
- * abide by same serializing rules, thus ends up using atomic_ops_lock as well.
+ * xchg() maps directly to ARC EX instruction which guarantees atomicity.
+ * However in !LLSC config, it also needs to be use @atomic_ops_lock spinlock
+ * due to a subtle reason:
+ *  - For !LLSC, cmpxchg() needs to use that lock (see above) and there is lot
+ *    of  kernel code which calls xchg()/cmpxchg() on same data (see llist.h)
+ *    Hence xchg() needs to follow same locking rules.
  *
- * This however is only relevant if SMP and/or ARC lacks LLSC
- *   if (UP or LLSC)
- *      xchg doesn't need serialization
- *   else <==> !(UP or LLSC) <==> (!UP and !LLSC) <==> (SMP and !LLSC)
- *      xchg needs serialization
+ * Technically the lock is also needed for UP (boils down to irq save/restore)
+ * but we can cheat a bit since cmpxchg() atomic_ops_lock() would cause irqs to
+ * be disabled thus can't possibly be interrpted/preempted/clobbered by xchg()
+ * Other way around, xchg is one instruction anyways, so can't be interrupted
+ * as such
  */
 
 #if !defined(CONFIG_ARC_HAS_LLSC) && defined(CONFIG_SMP)
diff --git a/arch/arc/include/asm/futex.h b/arch/arc/include/asm/futex.h
index 70cfe16b742d..11e1b1f3acda 100644
--- a/arch/arc/include/asm/futex.h
+++ b/arch/arc/include/asm/futex.h
@@ -20,6 +20,7 @@
 
 #define __futex_atomic_op(insn, ret, oldval, uaddr, oparg)\
 							\
+	smp_mb();					\
 	__asm__ __volatile__(				\
 	"1:	llock	%1, [%2]		\n"	\
 		insn				"\n"	\
@@ -30,7 +31,7 @@
 	"	.section .fixup,\"ax\"		\n"	\
 	"	.align  4			\n"	\
 	"4:	mov %0, %4			\n"	\
-	"	b   3b				\n"	\
+	"	j   3b				\n"	\
 	"	.previous			\n"	\
 	"	.section __ex_table,\"a\"	\n"	\
 	"	.align  4			\n"	\
@@ -40,12 +41,14 @@
 							\
 	: "=&r" (ret), "=&r" (oldval)			\
 	: "r" (uaddr), "r" (oparg), "ir" (-EFAULT)	\
-	: "cc", "memory")
+	: "cc", "memory");				\
+	smp_mb()					\
 
 #else	/* !CONFIG_ARC_HAS_LLSC */
 
 #define __futex_atomic_op(insn, ret, oldval, uaddr, oparg)\
 							\
+	smp_mb();					\
 	__asm__ __volatile__(				\
 	"1:	ld	%1, [%2]		\n"	\
 		insn				"\n"	\
@@ -55,7 +58,7 @@
 	"	.section .fixup,\"ax\"		\n"	\
 	"	.align  4			\n"	\
 	"4:	mov %0, %4			\n"	\
-	"	b   3b				\n"	\
+	"	j   3b				\n"	\
 	"	.previous			\n"	\
 	"	.section __ex_table,\"a\"	\n"	\
 	"	.align  4			\n"	\
@@ -65,7 +68,8 @@
 							\
 	: "=&r" (ret), "=&r" (oldval)			\
 	: "r" (uaddr), "r" (oparg), "ir" (-EFAULT)	\
-	: "cc", "memory")
+	: "cc", "memory");				\
+	smp_mb()					\
 
 #endif
 
@@ -83,6 +87,9 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))
 		return -EFAULT;
 
+#ifndef CONFIG_ARC_HAS_LLSC
+	preempt_disable();	/* to guarantee atomic r-m-w of futex op */
+#endif
 	pagefault_disable();
 
 	switch (op) {
@@ -90,6 +97,7 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 		__futex_atomic_op("mov %0, %3", ret, oldval, uaddr, oparg);
 		break;
 	case FUTEX_OP_ADD:
+		/* oldval = *uaddr; *uaddr += oparg ; ret = *uaddr */
 		__futex_atomic_op("add %0, %1, %3", ret, oldval, uaddr, oparg);
 		break;
 	case FUTEX_OP_OR:
@@ -106,6 +114,9 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	}
 
 	pagefault_enable();
+#ifndef CONFIG_ARC_HAS_LLSC
+	preempt_enable();
+#endif
 
 	if (!ret) {
 		switch (cmp) {
@@ -134,54 +145,57 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	return ret;
 }
 
-/* Compare-xchg with pagefaults disabled.
- *  Notes:
- *      -Best-Effort: Exchg happens only if compare succeeds.
- *          If compare fails, returns; leaving retry/looping to upper layers
- *      -successful cmp-xchg: return orig value in @addr (same as cmp val)
- *      -Compare fails: return orig value in @addr
- *      -user access r/w fails: return -EFAULT
+/*
+ * cmpxchg of futex (pagefaults disabled by caller)
+ * Return 0 for success, -EFAULT otherwise
  */
 static inline int
-futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval,
-					u32 newval)
+futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 expval,
+			      u32 newval)
 {
-	u32 val;
+	int ret = 0;
+	u32 existval;
 
-	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))
+	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
 		return -EFAULT;
 
-	pagefault_disable();
+#ifndef CONFIG_ARC_HAS_LLSC
+	preempt_disable();	/* to guarantee atomic r-m-w of futex op */
+#endif
+	smp_mb();
 
 	__asm__ __volatile__(
 #ifdef CONFIG_ARC_HAS_LLSC
-	"1:	llock	%0, [%3]		\n"
-	"	brne	%0, %1, 3f		\n"
-	"2:	scond	%2, [%3]		\n"
+	"1:	llock	%1, [%4]		\n"
+	"	brne	%1, %2, 3f		\n"
+	"2:	scond	%3, [%4]		\n"
 	"	bnz	1b			\n"
 #else
-	"1:	ld	%0, [%3]		\n"
-	"	brne	%0, %1, 3f		\n"
-	"2:	st	%2, [%3]		\n"
+	"1:	ld	%1, [%4]		\n"
+	"	brne	%1, %2, 3f		\n"
+	"2:	st	%3, [%4]		\n"
 #endif
 	"3:	\n"
 	"	.section .fixup,\"ax\"	\n"
-	"4:	mov %0, %4	\n"
-	"	b   3b	\n"
+	"4:	mov %0, %5	\n"
+	"	j   3b	\n"
 	"	.previous	\n"
 	"	.section __ex_table,\"a\"	\n"
 	"	.align  4	\n"
 	"	.word   1b, 4b	\n"
 	"	.word   2b, 4b	\n"
 	"	.previous\n"
-	: "=&r"(val)
-	: "r"(oldval), "r"(newval), "r"(uaddr), "ir"(-EFAULT)
+	: "+&r"(ret), "=&r"(existval)
+	: "r"(expval), "r"(newval), "r"(uaddr), "ir"(-EFAULT)
 	: "cc", "memory");
 
-	pagefault_enable();
+	smp_mb();
 
-	*uval = val;
-	return val;
+#ifndef CONFIG_ARC_HAS_LLSC
+	preempt_enable();
+#endif
+	*uval = existval;
+	return ret;
 }
 
 #endif
diff --git a/arch/arc/include/asm/perf_event.h b/arch/arc/include/asm/perf_event.h
index 2b8880e953a2..5f071762fb1c 100644
--- a/arch/arc/include/asm/perf_event.h
+++ b/arch/arc/include/asm/perf_event.h
@@ -1,6 +1,7 @@
 /*
  * Linux performance counter support for ARC
  *
+ * Copyright (C) 2014-2015 Synopsys, Inc. (www.synopsys.com)
  * Copyright (C) 2011-2013 Synopsys, Inc. (www.synopsys.com)
  *
  * This program is free software; you can redistribute it and/or modify
@@ -12,8 +13,8 @@
 #ifndef __ASM_PERF_EVENT_H
 #define __ASM_PERF_EVENT_H
 
-/* real maximum varies per CPU, this is the maximum supported by the driver */
-#define ARC_PMU_MAX_HWEVENTS	64
+/* Max number of counters that PCT block may ever have */
+#define ARC_PERF_MAX_COUNTERS	32
 
 #define ARC_REG_CC_BUILD	0xF6
 #define ARC_REG_CC_INDEX	0x240
@@ -28,15 +29,22 @@
 #define ARC_REG_PCT_CONFIG	0x254
 #define ARC_REG_PCT_CONTROL	0x255
 #define ARC_REG_PCT_INDEX	0x256
+#define ARC_REG_PCT_INT_CNTL	0x25C
+#define ARC_REG_PCT_INT_CNTH	0x25D
+#define ARC_REG_PCT_INT_CTRL	0x25E
+#define ARC_REG_PCT_INT_ACT	0x25F
+
+#define ARC_REG_PCT_CONFIG_USER	(1 << 18)	/* count in user mode */
+#define ARC_REG_PCT_CONFIG_KERN	(1 << 19)	/* count in kernel mode */
 
 #define ARC_REG_PCT_CONTROL_CC	(1 << 16)	/* clear counts */
 #define ARC_REG_PCT_CONTROL_SN	(1 << 17)	/* snapshot */
 
 struct arc_reg_pct_build {
 #ifdef CONFIG_CPU_BIG_ENDIAN
-	unsigned int m:8, c:8, r:6, s:2, v:8;
+	unsigned int m:8, c:8, r:5, i:1, s:2, v:8;
 #else
-	unsigned int v:8, s:2, r:6, c:8, m:8;
+	unsigned int v:8, s:2, i:1, r:5, c:8, m:8;
 #endif
 };
 
@@ -95,10 +103,13 @@ static const char * const arc_pmu_ev_hw_map[] = {
 
 	/* counts condition */
 	[PERF_COUNT_HW_INSTRUCTIONS] = "iall",
-	[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = "ijmp",
+	[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = "ijmp", /* Excludes ZOL jumps */
 	[PERF_COUNT_ARC_BPOK]         = "bpok",	  /* NP-NT, PT-T, PNT-NT */
+#ifdef CONFIG_ISA_ARCV2
+	[PERF_COUNT_HW_BRANCH_MISSES] = "bpmp",
+#else
 	[PERF_COUNT_HW_BRANCH_MISSES] = "bpfail", /* NP-T, PT-NT, PNT-T */
-
+#endif
 	[PERF_COUNT_ARC_LDC] = "imemrdc",	/* Instr: mem read cached */
 	[PERF_COUNT_ARC_STC] = "imemwrc",	/* Instr: mem write cached */
 
diff --git a/arch/arc/kernel/entry-arcv2.S b/arch/arc/kernel/entry-arcv2.S
index bd7105d3172f..8fa76567e402 100644
--- a/arch/arc/kernel/entry-arcv2.S
+++ b/arch/arc/kernel/entry-arcv2.S
@@ -57,13 +57,8 @@ VECTOR	handle_interrupt	; (23) End of fixed IRQs
 
 	.section .text, "ax",@progbits
 
-res_service:		; processor restart
-	flag    0x1     ; not implemented
-	nop
-	nop
-
-reserved:		; processor restart
-	rtie            ; jump to processor initializations
+reserved:
+	flag 1		; Unexpected event, halt
 
 ;##################### Interrupt Handling ##############################
 
diff --git a/arch/arc/kernel/entry.S b/arch/arc/kernel/entry.S
index f7a82fd4d601..589abf5172d6 100644
--- a/arch/arc/kernel/entry.S
+++ b/arch/arc/kernel/entry.S
@@ -42,7 +42,7 @@ ENTRY(ret_from_fork)
 	; when the forked child comes here from the __switch_to function
 	; r0 has the last task pointer.
 	; put last task in scheduler queue
-	bl   @schedule_tail
+	jl   @schedule_tail
 
 	ld   r9, [sp, PT_status32]
 	brne r9, 0, 1f
@@ -320,7 +320,7 @@ resume_user_mode_begin:
 	; --- (Slow Path #1) task preemption ---
 	bbit0  r9, TIF_NEED_RESCHED, .Lchk_pend_signals
 	mov    blink, resume_user_mode_begin  ; tail-call to U mode ret chks
-	b      @schedule 	; BTST+Bnz causes relo error in link
+	j      @schedule 	; BTST+Bnz causes relo error in link
 
 .Lchk_pend_signals:
 	IRQ_ENABLE	r10
@@ -381,7 +381,7 @@ resume_kernel_mode:
 	bbit0  r9, TIF_NEED_RESCHED, .Lrestore_regs
 
 	; Invoke PREEMPTION
-	bl      preempt_schedule_irq
+	jl      preempt_schedule_irq
 
 	; preempt_schedule_irq() always returns with IRQ disabled
 #endif
diff --git a/arch/arc/kernel/perf_event.c b/arch/arc/kernel/perf_event.c
index 1287388c258a..0c08bb1ce15a 100644
--- a/arch/arc/kernel/perf_event.c
+++ b/arch/arc/kernel/perf_event.c
@@ -1,7 +1,7 @@
 /*
  * Linux performance counter support for ARC700 series
  *
- * Copyright (C) 2013 Synopsys, Inc. (www.synopsys.com)
+ * Copyright (C) 2013-2015 Synopsys, Inc. (www.synopsys.com)
  *
  * This code is inspired by the perf support of various other architectures.
  *
@@ -11,6 +11,7 @@
  *
  */
 #include <linux/errno.h>
+#include <linux/interrupt.h>
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/perf_event.h>
@@ -20,12 +21,25 @@
 
 struct arc_pmu {
 	struct pmu	pmu;
-	int		counter_size;	/* in bits */
+	unsigned int	irq;
 	int		n_counters;
-	unsigned long	used_mask[BITS_TO_LONGS(ARC_PMU_MAX_HWEVENTS)];
+	u64		max_period;
 	int		ev_hw_idx[PERF_COUNT_ARC_HW_MAX];
 };
 
+struct arc_pmu_cpu {
+	/*
+	 * A 1 bit for an index indicates that the counter is being used for
+	 * an event. A 0 means that the counter can be used.
+	 */
+	unsigned long	used_mask[BITS_TO_LONGS(ARC_PERF_MAX_COUNTERS)];
+
+	/*
+	 * The events that are active on the PMU for the given index.
+	 */
+	struct perf_event *act_counter[ARC_PERF_MAX_COUNTERS];
+};
+
 struct arc_callchain_trace {
 	int depth;
 	void *perf_stuff;
@@ -65,6 +79,7 @@ perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs)
 }
 
 static struct arc_pmu *arc_pmu;
+static DEFINE_PER_CPU(struct arc_pmu_cpu, arc_pmu_cpu);
 
 /* read counter #idx; note that counter# != event# on ARC! */
 static uint64_t arc_pmu_read_counter(int idx)
@@ -88,18 +103,15 @@ static uint64_t arc_pmu_read_counter(int idx)
 static void arc_perf_event_update(struct perf_event *event,
 				  struct hw_perf_event *hwc, int idx)
 {
-	uint64_t prev_raw_count, new_raw_count;
-	int64_t delta;
-
-	do {
-		prev_raw_count = local64_read(&hwc->prev_count);
-		new_raw_count = arc_pmu_read_counter(idx);
-	} while (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
-				 new_raw_count) != prev_raw_count);
-
-	delta = (new_raw_count - prev_raw_count) &
-		((1ULL << arc_pmu->counter_size) - 1ULL);
+	uint64_t prev_raw_count = local64_read(&hwc->prev_count);
+	uint64_t new_raw_count = arc_pmu_read_counter(idx);
+	int64_t delta = new_raw_count - prev_raw_count;
 
+	/*
+	 * We don't afaraid of hwc->prev_count changing beneath our feet
+	 * because there's no way for us to re-enter this function anytime.
+	 */
+	local64_set(&hwc->prev_count, new_raw_count);
 	local64_add(delta, &event->count);
 	local64_sub(delta, &hwc->period_left);
 }
@@ -142,22 +154,41 @@ static int arc_pmu_event_init(struct perf_event *event)
 	struct hw_perf_event *hwc = &event->hw;
 	int ret;
 
+	if (!is_sampling_event(event)) {
+		hwc->sample_period  = arc_pmu->max_period;
+		hwc->last_period = hwc->sample_period;
+		local64_set(&hwc->period_left, hwc->sample_period);
+	}
+
+	hwc->config = 0;
+
+	if (is_isa_arcv2()) {
+		/* "exclude user" means "count only kernel" */
+		if (event->attr.exclude_user)
+			hwc->config |= ARC_REG_PCT_CONFIG_KERN;
+
+		/* "exclude kernel" means "count only user" */
+		if (event->attr.exclude_kernel)
+			hwc->config |= ARC_REG_PCT_CONFIG_USER;
+	}
+
 	switch (event->attr.type) {
 	case PERF_TYPE_HARDWARE:
 		if (event->attr.config >= PERF_COUNT_HW_MAX)
 			return -ENOENT;
 		if (arc_pmu->ev_hw_idx[event->attr.config] < 0)
 			return -ENOENT;
-		hwc->config = arc_pmu->ev_hw_idx[event->attr.config];
+		hwc->config |= arc_pmu->ev_hw_idx[event->attr.config];
 		pr_debug("init event %d with h/w %d \'%s\'\n",
 			 (int) event->attr.config, (int) hwc->config,
 			 arc_pmu_ev_hw_map[event->attr.config]);
 		return 0;
+
 	case PERF_TYPE_HW_CACHE:
 		ret = arc_pmu_cache_event(event->attr.config);
 		if (ret < 0)
 			return ret;
-		hwc->config = arc_pmu->ev_hw_idx[ret];
+		hwc->config |= arc_pmu->ev_hw_idx[ret];
 		return 0;
 	default:
 		return -ENOENT;
@@ -180,6 +211,47 @@ static void arc_pmu_disable(struct pmu *pmu)
 	write_aux_reg(ARC_REG_PCT_CONTROL, (tmp & 0xffff0000) | 0x0);
 }
 
+static int arc_pmu_event_set_period(struct perf_event *event)
+{
+	struct hw_perf_event *hwc = &event->hw;
+	s64 left = local64_read(&hwc->period_left);
+	s64 period = hwc->sample_period;
+	int idx = hwc->idx;
+	int overflow = 0;
+	u64 value;
+
+	if (unlikely(left <= -period)) {
+		/* left underflowed by more than period. */
+		left = period;
+		local64_set(&hwc->period_left, left);
+		hwc->last_period = period;
+		overflow = 1;
+	} else	if (unlikely(left <= 0)) {
+		/* left underflowed by less than period. */
+		left += period;
+		local64_set(&hwc->period_left, left);
+		hwc->last_period = period;
+		overflow = 1;
+	}
+
+	if (left > arc_pmu->max_period)
+		left = arc_pmu->max_period;
+
+	value = arc_pmu->max_period - left;
+	local64_set(&hwc->prev_count, value);
+
+	/* Select counter */
+	write_aux_reg(ARC_REG_PCT_INDEX, idx);
+
+	/* Write value */
+	write_aux_reg(ARC_REG_PCT_COUNTL, (u32)value);
+	write_aux_reg(ARC_REG_PCT_COUNTH, (value >> 32));
+
+	perf_event_update_userpage(event);
+
+	return overflow;
+}
+
 /*
  * Assigns hardware counter to hardware condition.
  * Note that there is no separate start/stop mechanism;
@@ -194,13 +266,20 @@ static void arc_pmu_start(struct perf_event *event, int flags)
 		return;
 
 	if (flags & PERF_EF_RELOAD)
-		WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE));
+		WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));
+
+	hwc->state = 0;
 
-	event->hw.state = 0;
+	arc_pmu_event_set_period(event);
+
+	/* Enable interrupt for this counter */
+	if (is_sampling_event(event))
+		write_aux_reg(ARC_REG_PCT_INT_CTRL,
+			      read_aux_reg(ARC_REG_PCT_INT_CTRL) | (1 << idx));
 
 	/* enable ARC pmu here */
-	write_aux_reg(ARC_REG_PCT_INDEX, idx);
-	write_aux_reg(ARC_REG_PCT_CONFIG, hwc->config);
+	write_aux_reg(ARC_REG_PCT_INDEX, idx);		/* counter # */
+	write_aux_reg(ARC_REG_PCT_CONFIG, hwc->config);	/* condition */
 }
 
 static void arc_pmu_stop(struct perf_event *event, int flags)
@@ -208,6 +287,17 @@ static void arc_pmu_stop(struct perf_event *event, int flags)
 	struct hw_perf_event *hwc = &event->hw;
 	int idx = hwc->idx;
 
+	/* Disable interrupt for this counter */
+	if (is_sampling_event(event)) {
+		/*
+		 * Reset interrupt flag by writing of 1. This is required
+		 * to make sure pending interrupt was not left.
+		 */
+		write_aux_reg(ARC_REG_PCT_INT_ACT, 1 << idx);
+		write_aux_reg(ARC_REG_PCT_INT_CTRL,
+			      read_aux_reg(ARC_REG_PCT_INT_CTRL) & ~(1 << idx));
+	}
+
 	if (!(event->hw.state & PERF_HES_STOPPED)) {
 		/* stop ARC pmu here */
 		write_aux_reg(ARC_REG_PCT_INDEX, idx);
@@ -227,8 +317,12 @@ static void arc_pmu_stop(struct perf_event *event, int flags)
 
 static void arc_pmu_del(struct perf_event *event, int flags)
 {
+	struct arc_pmu_cpu *pmu_cpu = this_cpu_ptr(&arc_pmu_cpu);
+
 	arc_pmu_stop(event, PERF_EF_UPDATE);
-	__clear_bit(event->hw.idx, arc_pmu->used_mask);
+	__clear_bit(event->hw.idx, pmu_cpu->used_mask);
+
+	pmu_cpu->act_counter[event->hw.idx] = 0;
 
 	perf_event_update_userpage(event);
 }
@@ -236,20 +330,31 @@ static void arc_pmu_del(struct perf_event *event, int flags)
 /* allocate hardware counter and optionally start counting */
 static int arc_pmu_add(struct perf_event *event, int flags)
 {
+	struct arc_pmu_cpu *pmu_cpu = this_cpu_ptr(&arc_pmu_cpu);
 	struct hw_perf_event *hwc = &event->hw;
 	int idx = hwc->idx;
 
-	if (__test_and_set_bit(idx, arc_pmu->used_mask)) {
-		idx = find_first_zero_bit(arc_pmu->used_mask,
+	if (__test_and_set_bit(idx, pmu_cpu->used_mask)) {
+		idx = find_first_zero_bit(pmu_cpu->used_mask,
 					  arc_pmu->n_counters);
 		if (idx == arc_pmu->n_counters)
 			return -EAGAIN;
 
-		__set_bit(idx, arc_pmu->used_mask);
+		__set_bit(idx, pmu_cpu->used_mask);
 		hwc->idx = idx;
 	}
 
 	write_aux_reg(ARC_REG_PCT_INDEX, idx);
+
+	pmu_cpu->act_counter[idx] = event;
+
+	if (is_sampling_event(event)) {
+		/* Mimic full counter overflow as other arches do */
+		write_aux_reg(ARC_REG_PCT_INT_CNTL, (u32)arc_pmu->max_period);
+		write_aux_reg(ARC_REG_PCT_INT_CNTH,
+			      (arc_pmu->max_period >> 32));
+	}
+
 	write_aux_reg(ARC_REG_PCT_CONFIG, 0);
 	write_aux_reg(ARC_REG_PCT_COUNTL, 0);
 	write_aux_reg(ARC_REG_PCT_COUNTH, 0);
@@ -264,11 +369,82 @@ static int arc_pmu_add(struct perf_event *event, int flags)
 	return 0;
 }
 
+#ifdef CONFIG_ISA_ARCV2
+static irqreturn_t arc_pmu_intr(int irq, void *dev)
+{
+	struct perf_sample_data data;
+	struct arc_pmu_cpu *pmu_cpu = this_cpu_ptr(&arc_pmu_cpu);
+	struct pt_regs *regs;
+	int active_ints;
+	int idx;
+
+	arc_pmu_disable(&arc_pmu->pmu);
+
+	active_ints = read_aux_reg(ARC_REG_PCT_INT_ACT);
+
+	regs = get_irq_regs();
+
+	for (idx = 0; idx < arc_pmu->n_counters; idx++) {
+		struct perf_event *event = pmu_cpu->act_counter[idx];
+		struct hw_perf_event *hwc;
+
+		if (!(active_ints & (1 << idx)))
+			continue;
+
+		/* Reset interrupt flag by writing of 1 */
+		write_aux_reg(ARC_REG_PCT_INT_ACT, 1 << idx);
+
+		/*
+		 * On reset of "interrupt active" bit corresponding
+		 * "interrupt enable" bit gets automatically reset as well.
+		 * Now we need to re-enable interrupt for the counter.
+		 */
+		write_aux_reg(ARC_REG_PCT_INT_CTRL,
+			read_aux_reg(ARC_REG_PCT_INT_CTRL) | (1 << idx));
+
+		hwc = &event->hw;
+
+		WARN_ON_ONCE(hwc->idx != idx);
+
+		arc_perf_event_update(event, &event->hw, event->hw.idx);
+		perf_sample_data_init(&data, 0, hwc->last_period);
+		if (!arc_pmu_event_set_period(event))
+			continue;
+
+		if (perf_event_overflow(event, &data, regs))
+			arc_pmu_stop(event, 0);
+	}
+
+	arc_pmu_enable(&arc_pmu->pmu);
+
+	return IRQ_HANDLED;
+}
+#else
+
+static irqreturn_t arc_pmu_intr(int irq, void *dev)
+{
+	return IRQ_NONE;
+}
+
+#endif /* CONFIG_ISA_ARCV2 */
+
+void arc_cpu_pmu_irq_init(void)
+{
+	struct arc_pmu_cpu *pmu_cpu = this_cpu_ptr(&arc_pmu_cpu);
+
+	arc_request_percpu_irq(arc_pmu->irq, smp_processor_id(), arc_pmu_intr,
+			       "ARC perf counters", pmu_cpu);
+
+	/* Clear all pending interrupt flags */
+	write_aux_reg(ARC_REG_PCT_INT_ACT, 0xffffffff);
+}
+
 static int arc_pmu_device_probe(struct platform_device *pdev)
 {
 	struct arc_reg_pct_build pct_bcr;
 	struct arc_reg_cc_build cc_bcr;
-	int i, j;
+	int i, j, has_interrupts;
+	int counter_size;	/* in bits */
 
 	union cc_name {
 		struct {
@@ -284,7 +460,7 @@ static int arc_pmu_device_probe(struct platform_device *pdev)
 		pr_err("This core does not have performance counters!\n");
 		return -ENODEV;
 	}
-	BUG_ON(pct_bcr.c > ARC_PMU_MAX_HWEVENTS);
+	BUG_ON(pct_bcr.c > ARC_PERF_MAX_COUNTERS);
 
 	READ_BCR(ARC_REG_CC_BUILD, cc_bcr);
 	BUG_ON(!cc_bcr.v); /* Counters exist but No countable conditions ? */
@@ -293,11 +469,16 @@ static int arc_pmu_device_probe(struct platform_device *pdev)
 	if (!arc_pmu)
 		return -ENOMEM;
 
+	has_interrupts = is_isa_arcv2() ? pct_bcr.i : 0;
+
 	arc_pmu->n_counters = pct_bcr.c;
-	arc_pmu->counter_size = 32 + (pct_bcr.s << 4);
+	counter_size = 32 + (pct_bcr.s << 4);
 
-	pr_info("ARC perf\t: %d counters (%d bits), %d countable conditions\n",
-		arc_pmu->n_counters, arc_pmu->counter_size, cc_bcr.c);
+	arc_pmu->max_period = (1ULL << counter_size) / 2 - 1ULL;
+
+	pr_info("ARC perf\t: %d counters (%d bits), %d conditions%s\n",
+		arc_pmu->n_counters, counter_size, cc_bcr.c,
+		has_interrupts ? ", [overflow IRQ support]":"");
 
 	cc_name.str[8] = 0;
 	for (i = 0; i < PERF_COUNT_ARC_HW_MAX; i++)
@@ -332,8 +513,37 @@ static int arc_pmu_device_probe(struct platform_device *pdev)
 		.read		= arc_pmu_read,
 	};
 
-	/* ARC 700 PMU does not support sampling events */
-	arc_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
+	if (has_interrupts) {
+		int irq = platform_get_irq(pdev, 0);
+		unsigned long flags;
+
+		if (irq < 0) {
+			pr_err("Cannot get IRQ number for the platform\n");
+			return -ENODEV;
+		}
+
+		arc_pmu->irq = irq;
+
+		/*
+		 * arc_cpu_pmu_irq_init() needs to be called on all cores for
+		 * their respective local PMU.
+		 * However we use opencoded on_each_cpu() to ensure it is called
+		 * on core0 first, so that arc_request_percpu_irq() sets up
+		 * AUTOEN etc. Otherwise enable_percpu_irq() fails to enable
+		 * perf IRQ on non master cores.
+		 * see arc_request_percpu_irq()
+		 */
+		preempt_disable();
+		local_irq_save(flags);
+		arc_cpu_pmu_irq_init();
+		local_irq_restore(flags);
+		smp_call_function((smp_call_func_t)arc_cpu_pmu_irq_init, 0, 1);
+		preempt_enable();
+
+		/* Clean all pending interrupt flags */
+		write_aux_reg(ARC_REG_PCT_INT_ACT, 0xffffffff);
+	} else
+		arc_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
 
 	return perf_pmu_register(&arc_pmu->pmu, pdev->name, PERF_TYPE_RAW);
 }
@@ -341,6 +551,7 @@ static int arc_pmu_device_probe(struct platform_device *pdev)
 #ifdef CONFIG_OF
 static const struct of_device_id arc_pmu_match[] = {
 	{ .compatible = "snps,arc700-pct" },
+	{ .compatible = "snps,archs-pct" },
 	{},
 };
 MODULE_DEVICE_TABLE(of, arc_pmu_match);
@@ -348,7 +559,7 @@ MODULE_DEVICE_TABLE(of, arc_pmu_match);
 
 static struct platform_driver arc_pmu_driver = {
 	.driver	= {
-		.name		= "arc700-pct",
+		.name		= "arc-pct",
 		.of_match_table = of_match_ptr(arc_pmu_match),
 	},
 	.probe		= arc_pmu_device_probe,
diff --git a/arch/arc/kernel/process.c b/arch/arc/kernel/process.c
index 44092456776f..91d5a0f1f3f7 100644
--- a/arch/arc/kernel/process.c
+++ b/arch/arc/kernel/process.c
@@ -65,7 +65,7 @@ asmlinkage void ret_from_fork(void);
  * ------------------
  * |     r25        |   <==== top of Stack (thread.ksp)
  * ~                ~
- * |    --to--      |   (CALLEE Regs of user mode)
+ * |    --to--      |   (CALLEE Regs of kernel mode)
  * |     r13        |
  * ------------------
  * |     fp         |
diff --git a/arch/arc/kernel/unaligned.c b/arch/arc/kernel/unaligned.c
index 74db59b6f392..abd961f3e763 100644
--- a/arch/arc/kernel/unaligned.c
+++ b/arch/arc/kernel/unaligned.c
@@ -34,7 +34,7 @@
 	"	.section .fixup,\"ax\"\n"		\
 	"	.align	4\n"				\
 	"3:	mov	%0, 1\n"			\
-	"	b	2b\n"				\
+	"	j	2b\n"				\
 	"	.previous\n"				\
 	"	.section __ex_table,\"a\"\n"		\
 	"	.align	4\n"				\
@@ -82,7 +82,7 @@
 		"	.section .fixup,\"ax\"\n"	\
 		"	.align	4\n"			\
 		"4:	mov	%0, 1\n"		\
-		"	b	3b\n"			\
+		"	j	3b\n"			\
 		"	.previous\n"			\
 		"	.section __ex_table,\"a\"\n"	\
 		"	.align	4\n"			\
@@ -113,7 +113,7 @@
 		"	.section .fixup,\"ax\"\n"	\
 		"	.align	4\n"			\
 		"6:	mov	%0, 1\n"		\
-		"	b	5b\n"			\
+		"	j	5b\n"			\
 		"	.previous\n"			\
 		"	.section __ex_table,\"a\"\n"	\
 		"	.align	4\n"			\
diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index 1cd6695b6ab5..0d1a6e96839f 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -22,15 +22,22 @@
 #include <asm/setup.h>
 
 static int l2_line_sz;
+int ioc_exists;
+volatile int slc_enable = 1, ioc_enable = 1;
 
 void (*_cache_line_loop_ic_fn)(unsigned long paddr, unsigned long vaddr,
 			       unsigned long sz, const int cacheop);
 
+void (*__dma_cache_wback_inv)(unsigned long start, unsigned long sz);
+void (*__dma_cache_inv)(unsigned long start, unsigned long sz);
+void (*__dma_cache_wback)(unsigned long start, unsigned long sz);
+
 char *arc_cache_mumbojumbo(int c, char *buf, int len)
 {
 	int n = 0;
 	struct cpuinfo_arc_cache *p;
 
+#define IS_USED_RUN(v)		((v) ? "" : "(disabled) ")
 #define PR_CACHE(p, cfg, str)						\
 	if (!(p)->ver)							\
 		n += scnprintf(buf + n, len - n, str"\t\t: N/A\n");	\
@@ -45,10 +52,18 @@ char *arc_cache_mumbojumbo(int c, char *buf, int len)
 	PR_CACHE(&cpuinfo_arc700[c].icache, CONFIG_ARC_HAS_ICACHE, "I-Cache");
 	PR_CACHE(&cpuinfo_arc700[c].dcache, CONFIG_ARC_HAS_DCACHE, "D-Cache");
 
+	if (!is_isa_arcv2())
+                return buf;
+
 	p = &cpuinfo_arc700[c].slc;
 	if (p->ver)
 		n += scnprintf(buf + n, len - n,
-			"SLC\t\t: %uK, %uB Line\n", p->sz_k, p->line_len);
+			       "SLC\t\t: %uK, %uB Line%s\n",
+			       p->sz_k, p->line_len, IS_USED_RUN(slc_enable));
+
+	if (ioc_exists)
+		n += scnprintf(buf + n, len - n, "IOC\t\t:%s\n",
+				IS_USED_RUN(ioc_enable));
 
 	return buf;
 }
@@ -58,18 +73,9 @@ char *arc_cache_mumbojumbo(int c, char *buf, int len)
  * the cpuinfo structure for later use.
  * No Validation done here, simply read/convert the BCRs
  */
-void read_decode_cache_bcr(void)
+static void read_decode_cache_bcr_arcv2(int cpu)
 {
-	struct cpuinfo_arc_cache *p_ic, *p_dc, *p_slc;
-	unsigned int cpu = smp_processor_id();
-	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
-	} ibcr, dbcr;
-
+	struct cpuinfo_arc_cache *p_slc = &cpuinfo_arc700[cpu].slc;
 	struct bcr_generic sbcr;
 
 	struct bcr_slc_cfg {
@@ -80,6 +86,39 @@ void read_decode_cache_bcr(void)
 #endif
 	} slc_cfg;
 
+	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
+	} cbcr;
+
+	READ_BCR(ARC_REG_SLC_BCR, sbcr);
+	if (sbcr.ver) {
+		READ_BCR(ARC_REG_SLC_CFG, slc_cfg);
+		p_slc->ver = sbcr.ver;
+		p_slc->sz_k = 128 << slc_cfg.sz;
+		l2_line_sz = p_slc->line_len = (slc_cfg.lsz == 0) ? 128 : 64;
+	}
+
+	READ_BCR(ARC_REG_CLUSTER_BCR, cbcr);
+	if (cbcr.c && ioc_enable)
+		ioc_exists = 1;
+}
+
+void read_decode_cache_bcr(void)
+{
+	struct cpuinfo_arc_cache *p_ic, *p_dc;
+	unsigned int cpu = smp_processor_id();
+	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
+	} ibcr, dbcr;
+
 	p_ic = &cpuinfo_arc700[cpu].icache;
 	READ_BCR(ARC_REG_IC_BCR, ibcr);
 
@@ -122,17 +161,8 @@ dc_chk:
 	p_dc->ver = dbcr.ver;
 
 slc_chk:
-	if (!is_isa_arcv2())
-		return;
-
-	p_slc = &cpuinfo_arc700[cpu].slc;
-	READ_BCR(ARC_REG_SLC_BCR, sbcr);
-	if (sbcr.ver) {
-		READ_BCR(ARC_REG_SLC_CFG, slc_cfg);
-		p_slc->ver = sbcr.ver;
-		p_slc->sz_k = 128 << slc_cfg.sz;
-		l2_line_sz = p_slc->line_len = (slc_cfg.lsz == 0) ? 128 : 64;
-	}
+	if (is_isa_arcv2())
+                read_decode_cache_bcr_arcv2(cpu);
 }
 
 /*
@@ -516,11 +546,6 @@ noinline void slc_op(unsigned long paddr, unsigned long sz, const int op)
 #endif
 }
 
-static inline int need_slc_flush(void)
-{
-	return is_isa_arcv2() && l2_line_sz;
-}
-
 /***********************************************************
  * Exported APIs
  */
@@ -569,30 +594,74 @@ void flush_dcache_page(struct page *page)
 }
 EXPORT_SYMBOL(flush_dcache_page);
 
-void dma_cache_wback_inv(unsigned long start, unsigned long sz)
+/*
+ * DMA ops for systems with L1 cache only
+ * Make memory coherent with L1 cache by flushing/invalidating L1 lines
+ */
+static void __dma_cache_wback_inv_l1(unsigned long start, unsigned long sz)
 {
 	__dc_line_op_k(start, sz, OP_FLUSH_N_INV);
+}
 
-	if (need_slc_flush())
-		slc_op(start, sz, OP_FLUSH_N_INV);
+static void __dma_cache_inv_l1(unsigned long start, unsigned long sz)
+{
+	__dc_line_op_k(start, sz, OP_INV);
 }
-EXPORT_SYMBOL(dma_cache_wback_inv);
 
-void dma_cache_inv(unsigned long start, unsigned long sz)
+static void __dma_cache_wback_l1(unsigned long start, unsigned long sz)
+{
+	__dc_line_op_k(start, sz, OP_FLUSH);
+}
+
+/*
+ * DMA ops for systems with both L1 and L2 caches, but without IOC
+ * Both L1 and L2 lines need to be explicity flushed/invalidated
+ */
+static void __dma_cache_wback_inv_slc(unsigned long start, unsigned long sz)
+{
+	__dc_line_op_k(start, sz, OP_FLUSH_N_INV);
+	slc_op(start, sz, OP_FLUSH_N_INV);
+}
+
+static void __dma_cache_inv_slc(unsigned long start, unsigned long sz)
 {
 	__dc_line_op_k(start, sz, OP_INV);
+	slc_op(start, sz, OP_INV);
+}
+
+static void __dma_cache_wback_slc(unsigned long start, unsigned long sz)
+{
+	__dc_line_op_k(start, sz, OP_FLUSH);
+	slc_op(start, sz, OP_FLUSH);
+}
+
+/*
+ * DMA ops for systems with IOC
+ * IOC hardware snoops all DMA traffic keeping the caches consistent with
+ * memory - eliding need for any explicit cache maintenance of DMA buffers
+ */
+static void __dma_cache_wback_inv_ioc(unsigned long start, unsigned long sz) {}
+static void __dma_cache_inv_ioc(unsigned long start, unsigned long sz) {}
+static void __dma_cache_wback_ioc(unsigned long start, unsigned long sz) {}
+
+/*
+ * Exported DMA API
+ */
+void dma_cache_wback_inv(unsigned long start, unsigned long sz)
+{
+	__dma_cache_wback_inv(start, sz);
+}
+EXPORT_SYMBOL(dma_cache_wback_inv);
 
-	if (need_slc_flush())
-		slc_op(start, sz, OP_INV);
+void dma_cache_inv(unsigned long start, unsigned long sz)
+{
+	__dma_cache_inv(start, sz);
 }
 EXPORT_SYMBOL(dma_cache_inv);
 
 void dma_cache_wback(unsigned long start, unsigned long sz)
 {
-	__dc_line_op_k(start, sz, OP_FLUSH);
-
-	if (need_slc_flush())
-		slc_op(start, sz, OP_FLUSH);
+	__dma_cache_wback(start, sz);
 }
 EXPORT_SYMBOL(dma_cache_wback);
 
@@ -848,4 +917,41 @@ void arc_cache_init(void)
 				panic("Disable CONFIG_ARC_CACHE_VIPT_ALIASING\n");
 		}
 	}
+
+	if (is_isa_arcv2() && l2_line_sz && !slc_enable) {
+
+		/* IM set : flush before invalidate */
+		write_aux_reg(ARC_REG_SLC_CTRL,
+			read_aux_reg(ARC_REG_SLC_CTRL) | SLC_CTRL_IM);
+
+		write_aux_reg(ARC_REG_SLC_INVALIDATE, 1);
+
+		/* Important to wait for flush to complete */
+		while (read_aux_reg(ARC_REG_SLC_CTRL) & SLC_CTRL_BUSY);
+		write_aux_reg(ARC_REG_SLC_CTRL,
+			read_aux_reg(ARC_REG_SLC_CTRL) | SLC_CTRL_DISABLE);
+	}
+
+	if (is_isa_arcv2() && ioc_exists) {
+		/* IO coherency base - 0x8z */
+		write_aux_reg(ARC_REG_IO_COH_AP0_BASE, 0x80000);
+		/* IO coherency aperture size - 512Mb: 0x8z-0xAz */
+		write_aux_reg(ARC_REG_IO_COH_AP0_SIZE, 0x11);
+		/* Enable partial writes */
+		write_aux_reg(ARC_REG_IO_COH_PARTIAL, 1);
+		/* Enable IO coherency */
+		write_aux_reg(ARC_REG_IO_COH_ENABLE, 1);
+
+		__dma_cache_wback_inv = __dma_cache_wback_inv_ioc;
+		__dma_cache_inv = __dma_cache_inv_ioc;
+		__dma_cache_wback = __dma_cache_wback_ioc;
+	} else if (is_isa_arcv2() && l2_line_sz && slc_enable) {
+		__dma_cache_wback_inv = __dma_cache_wback_inv_slc;
+		__dma_cache_inv = __dma_cache_inv_slc;
+		__dma_cache_wback = __dma_cache_wback_slc;
+	} else {
+		__dma_cache_wback_inv = __dma_cache_wback_inv_l1;
+		__dma_cache_inv = __dma_cache_inv_l1;
+		__dma_cache_wback = __dma_cache_wback_l1;
+	}
 }
diff --git a/arch/arc/mm/dma.c b/arch/arc/mm/dma.c
index 57706a9c6948..29a46bb198cc 100644
--- a/arch/arc/mm/dma.c
+++ b/arch/arc/mm/dma.c
@@ -19,6 +19,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/dma-debug.h>
 #include <linux/export.h>
+#include <asm/cache.h>
 #include <asm/cacheflush.h>
 
 /*
@@ -53,6 +54,20 @@ void *dma_alloc_coherent(struct device *dev, size_t size,
 {
 	void *paddr, *kvaddr;
 
+	/*
+	 * IOC relies on all data (even coherent DMA data) being in cache
+	 * Thus allocate normal cached memory
+	 *
+	 * The gains with IOC are two pronged:
+	 *   -For streaming data, elides needs for cache maintenance, saving
+	 *    cycles in flush code, and bus bandwidth as all the lines of a
+	 *    buffer need to be flushed out to memory
+	 *   -For coherent data, Read/Write to buffers terminate early in cache
+	 *   (vs. always going to memory - thus are faster)
+	 */
+	if (is_isa_arcv2() && ioc_exists)
+		return dma_alloc_noncoherent(dev, size, dma_handle, gfp);
+
 	/* This is linear addr (0x8000_0000 based) */
 	paddr = alloc_pages_exact(size, gfp);
 	if (!paddr)
@@ -85,6 +100,9 @@ EXPORT_SYMBOL(dma_alloc_coherent);
 void dma_free_coherent(struct device *dev, size_t size, void *kvaddr,
 		       dma_addr_t dma_handle)
 {
+	if (is_isa_arcv2() && ioc_exists)
+		return dma_free_noncoherent(dev, size, kvaddr, dma_handle);
+
 	iounmap((void __force __iomem *)kvaddr);
 
 	free_pages_exact((void *)dma_handle, size);
diff --git a/arch/arc/plat-axs10x/axs10x.c b/arch/arc/plat-axs10x/axs10x.c
index e7769c3ab5f2..ad9825d4026a 100644
--- a/arch/arc/plat-axs10x/axs10x.c
+++ b/arch/arc/plat-axs10x/axs10x.c
@@ -46,7 +46,7 @@ static void __init axs10x_enable_gpio_intc_wire(void)
 	 * -------------------   -------------------
 	 * | snps,dw-apb-gpio |  | snps,dw-apb-gpio |
 	 * -------------------   -------------------
-	 *        |                         |
+	 *        | #12                     |
 	 *        |                 [ Debug UART on cpu card ]
 	 *        |
 	 * ------------------------