Merge tag 'xtensa-20161005' of git://github.com/jcmvbkbc/linux-xtensa

Pull Xtensa updates from Max Filippov:
 "Updates for the xtensa architecture.  It is a combined set of patches
  for 4.8 that never got to the mainline and new patches for 4.9.

   - add new kernel memory layouts for MMUv3 cores: with 256MB and 512MB
     KSEG size, starting at physical address other than 0

   - make kernel load address configurable

   - clean up kernel memory layout macros

   - drop sysmem early allocator and switch to memblock

   - enable kmemleak and memory reservation from the device tree

   - wire up new syscalls: userfaultfd, membarrier, mlock2,
     copy_file_range, preadv2 and pwritev2

   - add new platform: Cadence Configurable System Platform (CSP) and
     new core variant for it: xt_lnx

   - rearrange CCOUNT calibration code, make most of it generic

   - improve machine reset code (XTFPGA now reboots reliably with MMUv3
     cores)

   - provide default memmap command line option for configurations
     without device tree support

   - ISS fixes: simdisk is now capable of using highmem pages, panic
     correctly terminates simulator"

* tag 'xtensa-20161005' of git://github.com/jcmvbkbc/linux-xtensa: (24 commits)
  xtensa: disable MMU initialization option on MMUv2 cores
  xtensa: add default memmap and mmio32native options to defconfigs
  xtensa: add default memmap option to common_defconfig
  xtensa: add default memmap option to iss_defconfig
  xtensa: ISS: allow simdisk to use high memory buffers
  xtensa: ISS: define simc_exit and use it instead of inline asm
  xtensa: xtfpga: group platform_* functions together
  xtensa: rearrange CCOUNT calibration
  xtensa: xtfpga: use clock provider, don't update DT
  xtensa: Tweak xuartps UART driver Rx watermark for Cadence CSP config.
  xtensa: initialize MMU before jumping to reset vector
  xtensa: fix icountlevel setting in cpu_reset
  xtensa: extract common CPU reset code into separate function
  xtensa: Added Cadence CSP kernel configuration for Xtensa
  xtensa: fix default kernel load address
  xtensa: wire up new syscalls
  xtensa: support reserved-memory DT node
  xtensa: drop sysmem and switch to memblock
  xtensa: minimize use of PLATFORM_DEFAULT_MEM_{ADDR,SIZE}
  xtensa: cleanup MMU setup and kernel layout macros
  ...

arch/xtensa/kernel/setup.c

@@ -7,6 +7,7 @@
  *
  * Copyright (C) 1995  Linus Torvalds
  * Copyright (C) 2001 - 2005  Tensilica Inc.
+ * Copyright (C) 2014 - 2016  Cadence Design Systems Inc.
  *
  * Chris Zankel	<chris@zankel.net>
  * Joe Taylor	<joe@tensilica.com, joetylr@yahoo.com>
@@ -22,7 +23,6 @@
 #include <linux/bootmem.h>
 #include <linux/kernel.h>
 #include <linux/percpu.h>
-#include <linux/clk-provider.h>
 #include <linux/cpu.h>
 #include <linux/of.h>
 #include <linux/of_fdt.h>
@@ -114,7 +114,7 @@ static int __init parse_tag_mem(const bp_tag_t *tag)
 	if (mi->type != MEMORY_TYPE_CONVENTIONAL)
 		return -1;
 
-	return add_sysmem_bank(mi->start, mi->end);
+	return memblock_add(mi->start, mi->end - mi->start);
 }
 
 __tagtable(BP_TAG_MEMORY, parse_tag_mem);
@@ -188,7 +188,6 @@ static int __init parse_bootparam(const bp_tag_t* tag)
 }
 
 #ifdef CONFIG_OF
-bool __initdata dt_memory_scan = false;
 
 #if !XCHAL_HAVE_PTP_MMU || XCHAL_HAVE_SPANNING_WAY
 unsigned long xtensa_kio_paddr = XCHAL_KIO_DEFAULT_PADDR;
@@ -228,11 +227,8 @@ static int __init xtensa_dt_io_area(unsigned long node, const char *uname,
 
 void __init early_init_dt_add_memory_arch(u64 base, u64 size)
 {
-	if (!dt_memory_scan)
-		return;
-
 	size &= PAGE_MASK;
-	add_sysmem_bank(base, base + size);
+	memblock_add(base, size);
 }
 
 void * __init early_init_dt_alloc_memory_arch(u64 size, u64 align)
@@ -242,9 +238,6 @@ void * __init early_init_dt_alloc_memory_arch(u64 size, u64 align)
 
 void __init early_init_devtree(void *params)
 {
-	if (sysmem.nr_banks == 0)
-		dt_memory_scan = true;
-
 	early_init_dt_scan(params);
 	of_scan_flat_dt(xtensa_dt_io_area, NULL);
 
@@ -252,14 +245,6 @@ void __init early_init_devtree(void *params)
 		strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
 }
 
-static int __init xtensa_device_probe(void)
-{
-	of_clk_init(NULL);
-	return 0;
-}
-
-device_initcall(xtensa_device_probe);
-
 #endif /* CONFIG_OF */
 
 /*
@@ -277,12 +262,6 @@ void __init init_arch(bp_tag_t *bp_start)
 	early_init_devtree(dtb_start);
 #endif
 
-	if (sysmem.nr_banks == 0) {
-		add_sysmem_bank(PLATFORM_DEFAULT_MEM_START,
-				PLATFORM_DEFAULT_MEM_START +
-				PLATFORM_DEFAULT_MEM_SIZE);
-	}
-
 #ifdef CONFIG_CMDLINE_BOOL
 	if (!command_line[0])
 		strlcpy(command_line, default_command_line, COMMAND_LINE_SIZE);
@@ -452,6 +431,10 @@ static int __init check_s32c1i(void)
 early_initcall(check_s32c1i);
 #endif /* CONFIG_S32C1I_SELFTEST */
 
+static inline int mem_reserve(unsigned long start, unsigned long end)
+{
+	return memblock_reserve(start, end - start);
+}
 
 void __init setup_arch(char **cmdline_p)
 {
@@ -463,54 +446,54 @@ void __init setup_arch(char **cmdline_p)
 #ifdef CONFIG_BLK_DEV_INITRD
 	if (initrd_start < initrd_end) {
 		initrd_is_mapped = mem_reserve(__pa(initrd_start),
-					       __pa(initrd_end), 0) == 0;
+					       __pa(initrd_end)) == 0;
 		initrd_below_start_ok = 1;
 	} else {
 		initrd_start = 0;
 	}
 #endif
 
-	mem_reserve(__pa(&_stext),__pa(&_end), 1);
+	mem_reserve(__pa(&_stext), __pa(&_end));
 
 	mem_reserve(__pa(&_WindowVectors_text_start),
-		    __pa(&_WindowVectors_text_end), 0);
+		    __pa(&_WindowVectors_text_end));
 
 	mem_reserve(__pa(&_DebugInterruptVector_literal_start),
-		    __pa(&_DebugInterruptVector_text_end), 0);
+		    __pa(&_DebugInterruptVector_text_end));
 
 	mem_reserve(__pa(&_KernelExceptionVector_literal_start),
-		    __pa(&_KernelExceptionVector_text_end), 0);
+		    __pa(&_KernelExceptionVector_text_end));
 
 	mem_reserve(__pa(&_UserExceptionVector_literal_start),
-		    __pa(&_UserExceptionVector_text_end), 0);
+		    __pa(&_UserExceptionVector_text_end));
 
 	mem_reserve(__pa(&_DoubleExceptionVector_literal_start),
-		    __pa(&_DoubleExceptionVector_text_end), 0);
+		    __pa(&_DoubleExceptionVector_text_end));
 
 #if XCHAL_EXCM_LEVEL >= 2
 	mem_reserve(__pa(&_Level2InterruptVector_text_start),
-		    __pa(&_Level2InterruptVector_text_end), 0);
+		    __pa(&_Level2InterruptVector_text_end));
 #endif
 #if XCHAL_EXCM_LEVEL >= 3
 	mem_reserve(__pa(&_Level3InterruptVector_text_start),
-		    __pa(&_Level3InterruptVector_text_end), 0);
+		    __pa(&_Level3InterruptVector_text_end));
 #endif
 #if XCHAL_EXCM_LEVEL >= 4
 	mem_reserve(__pa(&_Level4InterruptVector_text_start),
-		    __pa(&_Level4InterruptVector_text_end), 0);
+		    __pa(&_Level4InterruptVector_text_end));
 #endif
 #if XCHAL_EXCM_LEVEL >= 5
 	mem_reserve(__pa(&_Level5InterruptVector_text_start),
-		    __pa(&_Level5InterruptVector_text_end), 0);
+		    __pa(&_Level5InterruptVector_text_end));
 #endif
 #if XCHAL_EXCM_LEVEL >= 6
 	mem_reserve(__pa(&_Level6InterruptVector_text_start),
-		    __pa(&_Level6InterruptVector_text_end), 0);
+		    __pa(&_Level6InterruptVector_text_end));
 #endif
 
 #ifdef CONFIG_SMP
 	mem_reserve(__pa(&_SecondaryResetVector_text_start),
-		    __pa(&_SecondaryResetVector_text_end), 0);
+		    __pa(&_SecondaryResetVector_text_end));
 #endif
 	parse_early_param();
 	bootmem_init();
@@ -555,6 +538,137 @@ static int __init topology_init(void)
 }
 subsys_initcall(topology_init);
 
+void cpu_reset(void)
+{
+#if XCHAL_HAVE_PTP_MMU
+	local_irq_disable();
+	/*
+	 * We have full MMU: all autoload ways, ways 7, 8 and 9 of DTLB must
+	 * be flushed.
+	 * Way 4 is not currently used by linux.
+	 * Ways 5 and 6 shall not be touched on MMUv2 as they are hardwired.
+	 * Way 5 shall be flushed and way 6 shall be set to identity mapping
+	 * on MMUv3.
+	 */
+	local_flush_tlb_all();
+	invalidate_page_directory();
+#if XCHAL_HAVE_SPANNING_WAY
+	/* MMU v3 */
+	{
+		unsigned long vaddr = (unsigned long)cpu_reset;
+		unsigned long paddr = __pa(vaddr);
+		unsigned long tmpaddr = vaddr + SZ_512M;
+		unsigned long tmp0, tmp1, tmp2, tmp3;
+
+		/*
+		 * Find a place for the temporary mapping. It must not be
+		 * in the same 512MB region with vaddr or paddr, otherwise
+		 * there may be multihit exception either on entry to the
+		 * temporary mapping, or on entry to the identity mapping.
+		 * (512MB is the biggest page size supported by TLB.)
+		 */
+		while (((tmpaddr ^ paddr) & -SZ_512M) == 0)
+			tmpaddr += SZ_512M;
+
+		/* Invalidate mapping in the selected temporary area */
+		if (itlb_probe(tmpaddr) & 0x8)
+			invalidate_itlb_entry(itlb_probe(tmpaddr));
+		if (itlb_probe(tmpaddr + PAGE_SIZE) & 0x8)
+			invalidate_itlb_entry(itlb_probe(tmpaddr + PAGE_SIZE));
+
+		/*
+		 * Map two consecutive pages starting at the physical address
+		 * of this function to the temporary mapping area.
+		 */
+		write_itlb_entry(__pte((paddr & PAGE_MASK) |
+				       _PAGE_HW_VALID |
+				       _PAGE_HW_EXEC |
+				       _PAGE_CA_BYPASS),
+				 tmpaddr & PAGE_MASK);
+		write_itlb_entry(__pte(((paddr & PAGE_MASK) + PAGE_SIZE) |
+				       _PAGE_HW_VALID |
+				       _PAGE_HW_EXEC |
+				       _PAGE_CA_BYPASS),
+				 (tmpaddr & PAGE_MASK) + PAGE_SIZE);
+
+		/* Reinitialize TLB */
+		__asm__ __volatile__ ("movi	%0, 1f\n\t"
+				      "movi	%3, 2f\n\t"
+				      "add	%0, %0, %4\n\t"
+				      "add	%3, %3, %5\n\t"
+				      "jx	%0\n"
+				      /*
+				       * No literal, data or stack access
+				       * below this point
+				       */
+				      "1:\n\t"
+				      /* Initialize *tlbcfg */
+				      "movi	%0, 0\n\t"
+				      "wsr	%0, itlbcfg\n\t"
+				      "wsr	%0, dtlbcfg\n\t"
+				      /* Invalidate TLB way 5 */
+				      "movi	%0, 4\n\t"
+				      "movi	%1, 5\n"
+				      "1:\n\t"
+				      "iitlb	%1\n\t"
+				      "idtlb	%1\n\t"
+				      "add	%1, %1, %6\n\t"
+				      "addi	%0, %0, -1\n\t"
+				      "bnez	%0, 1b\n\t"
+				      /* Initialize TLB way 6 */
+				      "movi	%0, 7\n\t"
+				      "addi	%1, %9, 3\n\t"
+				      "addi	%2, %9, 6\n"
+				      "1:\n\t"
+				      "witlb	%1, %2\n\t"
+				      "wdtlb	%1, %2\n\t"
+				      "add	%1, %1, %7\n\t"
+				      "add	%2, %2, %7\n\t"
+				      "addi	%0, %0, -1\n\t"
+				      "bnez	%0, 1b\n\t"
+				      /* Jump to identity mapping */
+				      "jx	%3\n"
+				      "2:\n\t"
+				      /* Complete way 6 initialization */
+				      "witlb	%1, %2\n\t"
+				      "wdtlb	%1, %2\n\t"
+				      /* Invalidate temporary mapping */
+				      "sub	%0, %9, %7\n\t"
+				      "iitlb	%0\n\t"
+				      "add	%0, %0, %8\n\t"
+				      "iitlb	%0"
+				      : "=&a"(tmp0), "=&a"(tmp1), "=&a"(tmp2),
+					"=&a"(tmp3)
+				      : "a"(tmpaddr - vaddr),
+					"a"(paddr - vaddr),
+					"a"(SZ_128M), "a"(SZ_512M),
+					"a"(PAGE_SIZE),
+					"a"((tmpaddr + SZ_512M) & PAGE_MASK)
+				      : "memory");
+	}
+#endif
+#endif
+	__asm__ __volatile__ ("movi	a2, 0\n\t"
+			      "wsr	a2, icountlevel\n\t"
+			      "movi	a2, 0\n\t"
+			      "wsr	a2, icount\n\t"
+#if XCHAL_NUM_IBREAK > 0
+			      "wsr	a2, ibreakenable\n\t"
+#endif
+#if XCHAL_HAVE_LOOPS
+			      "wsr	a2, lcount\n\t"
+#endif
+			      "movi	a2, 0x1f\n\t"
+			      "wsr	a2, ps\n\t"
+			      "isync\n\t"
+			      "jx	%0\n\t"
+			      :
+			      : "a" (XCHAL_RESET_VECTOR_VADDR)
+			      : "a2");
+	for (;;)
+		;
+}
+
 void machine_restart(char * cmd)
 {
 	platform_restart();