Merge tag 'arc-4.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc

Pull ARC updates from Vineet Gupta:

 - AXS10x platform clk updates for I2S, PGU

 - add region based cache flush operation for ARCv2 cores

 - enforce PAE40 dependency on HIGHMEM

 - ptrace support for additional regs in ARCv2 cores

 - fix build failure in linux-next dut to a header include ordering
   change

* tag 'arc-4.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  Revert "ARCv2: Allow enabling PAE40 w/o HIGHMEM"
  ARC: mm: fix build failure in linux-next for UP builds
  ARCv2: ptrace: provide regset for accumulator/r30 regs
  elf: Add ARCv2 specific core note section
  ARCv2: mm: micro-optimize region flush generated code
  ARCv2: mm: Merge 2 updates to DC_CTRL for region flush
  ARCv2: mm: Implement cache region flush operations
  ARC: mm: Move full_page computation into cache version agnostic wrapper
  arc: axs10x: Fix ARC PGU default clock frequency
  arc: axs10x: Add DT bindings for I2S audio playback

1 files changed, 98 insertions, 13 deletions

diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index 928562967f3c..a867575a758b 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -21,6 +21,10 @@
 #include <asm/cachectl.h>
 #include <asm/setup.h>
 
+#ifdef CONFIG_ISA_ARCV2
+#define USE_RGN_FLSH	1
+#endif
+
 static int l2_line_sz;
 static int ioc_exists;
 int slc_enable = 1, ioc_enable = 1;
@@ -28,7 +32,7 @@ unsigned long perip_base = ARC_UNCACHED_ADDR_SPACE; /* legacy value for boot */
 unsigned long perip_end = 0xFFFFFFFF; /* legacy value */
 
 void (*_cache_line_loop_ic_fn)(phys_addr_t paddr, unsigned long vaddr,
-			       unsigned long sz, const int cacheop);
+			       unsigned long sz, const int op, const int full_page);
 
 void (*__dma_cache_wback_inv)(phys_addr_t start, unsigned long sz);
 void (*__dma_cache_inv)(phys_addr_t start, unsigned long sz);
@@ -233,11 +237,10 @@ slc_chk:
 
 static inline
 void __cache_line_loop_v2(phys_addr_t paddr, unsigned long vaddr,
-			  unsigned long sz, const int op)
+			  unsigned long sz, const int op, const int full_page)
 {
 	unsigned int aux_cmd;
 	int num_lines;
-	const int full_page = __builtin_constant_p(sz) && sz == PAGE_SIZE;
 
 	if (op == OP_INV_IC) {
 		aux_cmd = ARC_REG_IC_IVIL;
@@ -279,11 +282,10 @@ void __cache_line_loop_v2(phys_addr_t paddr, unsigned long vaddr,
  */
 static inline
 void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
-			  unsigned long sz, const int op)
+			  unsigned long sz, const int op, const int full_page)
 {
 	unsigned int aux_cmd, aux_tag;
 	int num_lines;
-	const int full_page = __builtin_constant_p(sz) && sz == PAGE_SIZE;
 
 	if (op == OP_INV_IC) {
 		aux_cmd = ARC_REG_IC_IVIL;
@@ -334,6 +336,8 @@ void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
 	}
 }
 
+#ifndef USE_RGN_FLSH
+
 /*
  * In HS38x (MMU v4), I-cache is VIPT (can alias), D-cache is PIPT
  * Here's how cache ops are implemented
@@ -349,17 +353,16 @@ void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
  */
 static inline
 void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
-			  unsigned long sz, const int cacheop)
+			  unsigned long sz, const int op, const int full_page)
 {
 	unsigned int aux_cmd;
 	int num_lines;
-	const int full_page_op = __builtin_constant_p(sz) && sz == PAGE_SIZE;
 
-	if (cacheop == OP_INV_IC) {
+	if (op == OP_INV_IC) {
 		aux_cmd = ARC_REG_IC_IVIL;
 	} else {
 		/* d$ cmd: INV (discard or wback-n-discard) OR FLUSH (wback) */
-		aux_cmd = cacheop & OP_INV ? ARC_REG_DC_IVDL : ARC_REG_DC_FLDL;
+		aux_cmd = op & OP_INV ? ARC_REG_DC_IVDL : ARC_REG_DC_FLDL;
 	}
 
 	/* Ensure we properly floor/ceil the non-line aligned/sized requests
@@ -368,7 +371,7 @@ void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
 	 *  -@paddr will be cache-line aligned already (being page aligned)
 	 *  -@sz will be integral multiple of line size (being page sized).
 	 */
-	if (!full_page_op) {
+	if (!full_page) {
 		sz += paddr & ~CACHE_LINE_MASK;
 		paddr &= CACHE_LINE_MASK;
 	}
@@ -381,7 +384,7 @@ void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
 	 *   - (and needs to be written before the lower 32 bits)
 	 */
 	if (is_pae40_enabled()) {
-		if (cacheop == OP_INV_IC)
+		if (op == OP_INV_IC)
 			/*
 			 * Non aliasing I-cache in HS38,
 			 * aliasing I-cache handled in __cache_line_loop_v3()
@@ -397,6 +400,55 @@ void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
 	}
 }
 
+#else
+
+/*
+ * optimized flush operation which takes a region as opposed to iterating per line
+ */
+static inline
+void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
+			  unsigned long sz, const int op, const int full_page)
+{
+	unsigned int s, e;
+
+	/* Only for Non aliasing I-cache in HS38 */
+	if (op == OP_INV_IC) {
+		s = ARC_REG_IC_IVIR;
+		e = ARC_REG_IC_ENDR;
+	} else {
+		s = ARC_REG_DC_STARTR;
+		e = ARC_REG_DC_ENDR;
+	}
+
+	if (!full_page) {
+		/* for any leading gap between @paddr and start of cache line */
+		sz += paddr & ~CACHE_LINE_MASK;
+		paddr &= CACHE_LINE_MASK;
+
+		/*
+		 *  account for any trailing gap to end of cache line
+		 *  this is equivalent to DIV_ROUND_UP() in line ops above
+		 */
+		sz += L1_CACHE_BYTES - 1;
+	}
+
+	if (is_pae40_enabled()) {
+		/* TBD: check if crossing 4TB boundary */
+		if (op == OP_INV_IC)
+			write_aux_reg(ARC_REG_IC_PTAG_HI, (u64)paddr >> 32);
+		else
+			write_aux_reg(ARC_REG_DC_PTAG_HI, (u64)paddr >> 32);
+	}
+
+	/* ENDR needs to be set ahead of START */
+	write_aux_reg(e, paddr + sz);	/* ENDR is exclusive */
+	write_aux_reg(s, paddr);
+
+	/* caller waits on DC_CTRL.FS */
+}
+
+#endif
+
 #if (CONFIG_ARC_MMU_VER < 3)
 #define __cache_line_loop	__cache_line_loop_v2
 #elif (CONFIG_ARC_MMU_VER == 3)
@@ -411,6 +463,11 @@ void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
  * Machine specific helpers for Entire D-Cache or Per Line ops
  */
 
+#ifndef USE_RGN_FLSH
+/*
+ * this version avoids extra read/write of DC_CTRL for flush or invalid ops
+ * in the non region flush regime (such as for ARCompact)
+ */
 static inline void __before_dc_op(const int op)
 {
 	if (op == OP_FLUSH_N_INV) {
@@ -424,6 +481,32 @@ static inline void __before_dc_op(const int op)
 	}
 }
 
+#else
+
+static inline void __before_dc_op(const int op)
+{
+	const unsigned int ctl = ARC_REG_DC_CTRL;
+	unsigned int val = read_aux_reg(ctl);
+
+	if (op == OP_FLUSH_N_INV) {
+		val |= DC_CTRL_INV_MODE_FLUSH;
+	}
+
+	if (op != OP_INV_IC) {
+		/*
+		 * Flush / Invalidate is provided by DC_CTRL.RNG_OP 0 or 1
+		 * combined Flush-n-invalidate uses DC_CTRL.IM = 1 set above
+		 */
+		val &= ~DC_CTRL_RGN_OP_MSK;
+		if (op & OP_INV)
+			val |= DC_CTRL_RGN_OP_INV;
+	}
+	write_aux_reg(ctl, val);
+}
+
+#endif
+
+
 static inline void __after_dc_op(const int op)
 {
 	if (op & OP_FLUSH) {
@@ -486,13 +569,14 @@ static void __dc_enable(void)
 static inline void __dc_line_op(phys_addr_t paddr, unsigned long vaddr,
 				unsigned long sz, const int op)
 {
+	const int full_page = __builtin_constant_p(sz) && sz == PAGE_SIZE;
 	unsigned long flags;
 
 	local_irq_save(flags);
 
 	__before_dc_op(op);
 
-	__cache_line_loop(paddr, vaddr, sz, op);
+	__cache_line_loop(paddr, vaddr, sz, op, full_page);
 
 	__after_dc_op(op);
 
@@ -521,10 +605,11 @@ static inline void
 __ic_line_inv_vaddr_local(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz)
 {
+	const int full_page = __builtin_constant_p(sz) && sz == PAGE_SIZE;
 	unsigned long flags;
 
 	local_irq_save(flags);
-	(*_cache_line_loop_ic_fn)(paddr, vaddr, sz, OP_INV_IC);
+	(*_cache_line_loop_ic_fn)(paddr, vaddr, sz, OP_INV_IC, full_page);
 	local_irq_restore(flags);
 }