Merge git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile

Pull Tile arch updates from Chris Metcalf:
 "These changes bring in a bunch of new functionality that has been
  maintained internally at Tilera over the last year, plus other stray
  bits of work that I've taken into the tile tree from other folks.

  The changes include some PCI root complex work, interrupt-driven
  console support, support for performing fast-path unaligned data
  fixups by kernel-based JIT code generation, CONFIG_PREEMPT support,
  vDSO support for gettimeofday(), a serial driver for the tilegx
  on-chip UART, KGDB support, more optimized string routines, support
  for ftrace and kprobes, improved ASLR, and many bug fixes.

  We also remove support for the old TILE64 chip, which is no longer
  buildable"

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile: (85 commits)
  tile: refresh tile defconfig files
  tile: rework <asm/cmpxchg.h>
  tile PCI RC: make default consistent DMA mask 32-bit
  tile: add null check for kzalloc in tile/kernel/setup.c
  tile: make __write_once a synonym for __read_mostly
  tile: remove support for TILE64
  tile: use asm-generic/bitops/builtin-*.h
  tile: eliminate no-op "noatomichash" boot argument
  tile: use standard tile_bundle_bits type in traps.c
  tile: simplify code referencing hypervisor API addresses
  tile: change <asm/system.h> to <asm/switch_to.h> in comments
  tile: mark pcibios_init() as __init
  tile: check for correct compiler earlier in asm-offsets.c
  tile: use standard 'generic-y' model for <asm/hw_irq.h>
  tile: use asm-generic version of <asm/local64.h>
  tile PCI RC: add comment about "PCI hole" problem
  tile: remove DEBUG_EXTRA_FLAGS kernel config option
  tile: add virt_to_kpte() API and clean up and document behavior
  tile: support FRAME_POINTER
  tile: support reporting Tilera hypervisor statistics
  ...

1 files changed, 25 insertions, 21 deletions

diff --git a/arch/tile/kernel/head_64.S b/arch/tile/kernel/head_64.S
index 6093964fa5c7..bd0e12f283f3 100644
--- a/arch/tile/kernel/head_64.S
+++ b/arch/tile/kernel/head_64.S
@@ -25,6 +25,15 @@
 #include <arch/chip.h>
 #include <arch/spr_def.h>
 
+/* Extract two 32-bit bit values that were read into one register. */
+#ifdef __BIG_ENDIAN__
+#define GET_FIRST_INT(rd, rs) shrsi rd, rs, 32
+#define GET_SECOND_INT(rd, rs) addxi rd, rs, 0
+#else
+#define GET_FIRST_INT(rd, rs) addxi rd, rs, 0
+#define GET_SECOND_INT(rd, rs) shrsi rd, rs, 32
+#endif
+
 /*
  * This module contains the entry code for kernel images. It performs the
  * minimal setup needed to call the generic C routines.
@@ -46,11 +55,11 @@ ENTRY(_start)
 	  movei r2, TILE_CHIP_REV
 	  movei r3, KERNEL_PL
 	}
-	jal hv_init
+	jal _hv_init
 	/* Get a reasonable default ASID in r0 */
 	{
 	  move r0, zero
-	  jal hv_inquire_asid
+	  jal _hv_inquire_asid
 	}
 
 	/*
@@ -61,7 +70,7 @@ ENTRY(_start)
 	 * other CPUs should see a properly-constructed page table.
 	 */
 	{
-	  v4int_l r2, zero, r0    /* ASID for hv_install_context */
+	  GET_FIRST_INT(r2, r0)    /* ASID for hv_install_context */
 	  moveli r4, hw1_last(swapper_pgprot - PAGE_OFFSET)
 	}
 	{
@@ -77,7 +86,7 @@ ENTRY(_start)
 	{
 	  /* After initializing swapper_pgprot, HV_PTE_GLOBAL is set. */
 	  bfextu r7, r1, HV_PTE_INDEX_GLOBAL, HV_PTE_INDEX_GLOBAL
-	  inv r4
+	  finv r4
 	}
 	bnez r7, .Lno_write
 	{
@@ -121,29 +130,24 @@ ENTRY(_start)
 	}
 	{
 	  moveli r3, CTX_PAGE_FLAG
-	  j hv_install_context
+	  j _hv_install_context
 	}
 1:
 
 	/* Install the interrupt base. */
-	moveli r0, hw2_last(MEM_SV_START)
-	shl16insli r0, r0, hw1(MEM_SV_START)
-	shl16insli r0, r0, hw0(MEM_SV_START)
+	moveli r0, hw2_last(intrpt_start)
+	shl16insli r0, r0, hw1(intrpt_start)
+	shl16insli r0, r0, hw0(intrpt_start)
 	mtspr SPR_INTERRUPT_VECTOR_BASE_K, r0
 
-	/*
-	 * Get our processor number and save it away in SAVE_K_0.
-	 * Extract stuff from the topology structure: r4 = y, r6 = x,
-	 * r5 = width.  FIXME: consider whether we want to just make these
-	 * 64-bit values (and if so fix smp_topology write below, too).
-	 */
-	jal hv_inquire_topology
+	/* Get our processor number and save it away in SAVE_K_0. */
+	jal _hv_inquire_topology
 	{
-	  v4int_l r5, zero, r1    /* r5 = width */
-	  shrui r4, r0, 32        /* r4 = y */
+	  GET_FIRST_INT(r5, r1)   /* r5 = width */
+	  GET_SECOND_INT(r4, r0)  /* r4 = y */
 	}
 	{
-	  v4int_l r6, zero, r0    /* r6 = x */
+	  GET_FIRST_INT(r6, r0)   /* r6 = x */
 	  mul_lu_lu r4, r4, r5
 	}
 	{
@@ -154,7 +158,7 @@ ENTRY(_start)
 	/*
 	 * Load up our per-cpu offset.  When the first (master) tile
 	 * boots, this value is still zero, so we will load boot_pc
-	 * with start_kernel, and boot_sp with init_stack + THREAD_SIZE.
+	 * with start_kernel, and boot_sp with at the top of init_stack.
 	 * The master tile initializes the per-cpu offset array, so that
 	 * when subsequent (secondary) tiles boot, they will instead load
 	 * from their per-cpu versions of boot_sp and boot_pc.
@@ -198,9 +202,9 @@ ENTRY(_start)
 	}
 	ld r0, r0
 	ld sp, r1
-	or r4, sp, r4
+	shli r4, r4, CPU_SHIFT
+	bfins r4, sp, 0, CPU_SHIFT-1
 	mtspr SPR_SYSTEM_SAVE_K_0, r4  /* save ksp0 + cpu */
-	addi sp, sp, -STACK_TOP_DELTA
 	{
 	  move lr, zero   /* stop backtraces in the called function */
 	  jr r0