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

Pull tile updates from Chris Metcalf:
 "These changes cover a range of new arch/tile features and
  optimizations.  They've been through LKML review and on linux-next for
  a month or so.  There's also one bug-fix that just missed 3.4, which
  I've marked for stable."

Fixed up trivial conflict in arch/tile/Kconfig (new added tile Kconfig
entries clashing with the generic timer/clockevents changes).

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
  tile: default to tilegx_defconfig for ARCH=tile
  tile: fix bug where fls(0) was not returning 0
  arch/tile: mark TILEGX as not EXPERIMENTAL
  tile/mm/fault.c: Port OOM changes to handle_page_fault
  arch/tile: add descriptive text if the kernel reports a bad trap
  arch/tile: allow querying cpu module information from the hypervisor
  arch/tile: fix hardwall for tilegx and generalize for idn and ipi
  arch/tile: support multiple huge page sizes dynamically
  mm: add new arch_make_huge_pte() method for tile support
  arch/tile: support kexec() for tilegx
  arch/tile: support <asm/cachectl.h> header for cacheflush() syscall
  arch/tile: Allow tilegx to build with either 16K or 64K page size
  arch/tile: optimize get_user/put_user and friends
  arch/tile: support building big-endian kernel
  arch/tile: allow building Linux with transparent huge pages enabled
  arch/tile: use interrupt critical sections less

arch/tile/Kconfig

@@ -48,6 +48,14 @@ config NEED_PER_CPU_PAGE_FIRST_CHUNK
 config SYS_SUPPORTS_HUGETLBFS
 	def_bool y
 
+# Support for additional huge page sizes besides HPAGE_SIZE.
+# The software support is currently only present in the TILE-Gx
+# hypervisor. TILEPro in any case does not support page sizes
+# larger than the default HPAGE_SIZE.
+config HUGETLB_SUPER_PAGES
+	depends on HUGETLB_PAGE && TILEGX
+	def_bool y
+
 # FIXME: tilegx can implement a more efficient rwsem.
 config RWSEM_GENERIC_SPINLOCK
 	def_bool y
@@ -107,16 +115,14 @@ config HVC_TILE
 	select HVC_DRIVER
 	def_bool y
 
-# Please note: TILE-Gx support is not yet finalized; this is
-# the preliminary support.  TILE-Gx drivers are only provided
-# with the alpha or beta test versions for Tilera customers.
 config TILEGX
-	depends on EXPERIMENTAL
 	bool "Building with TILE-Gx (64-bit) compiler and toolchain"
 
+config TILEPRO
+	def_bool !TILEGX
+
 config 64BIT
-	depends on TILEGX
-	def_bool y
+	def_bool TILEGX
 
 config ARCH_DEFCONFIG
 	string
@@ -137,6 +143,31 @@ config NR_CPUS
 	  smaller kernel memory footprint results from using a smaller
 	  value on chips with fewer tiles.
 
+if TILEGX
+
+choice
+	prompt "Kernel page size"
+	default PAGE_SIZE_64KB
+	help
+	  This lets you select the page size of the kernel.  For best
+	  performance on memory-intensive applications, a page size of 64KB
+	  is recommended.  For workloads involving many small files, many
+	  connections, etc., it may be better to select 16KB, which uses
+	  memory more efficiently at some cost in TLB performance.
+
+	  Note that this option is TILE-Gx specific; currently
+	  TILEPro page size is set by rebuilding the hypervisor.
+
+config PAGE_SIZE_16KB
+	bool "16KB"
+
+config PAGE_SIZE_64KB
+	bool "64KB"
+
+endchoice
+
+endif
+
 source "kernel/Kconfig.hz"
 
 config KEXEC

arch/tile/Makefile

@@ -34,7 +34,12 @@ LIBGCC_PATH     := \
   $(shell $(CC) $(KBUILD_CFLAGS) $(KCFLAGS) -print-libgcc-file-name)
 
 # Provide the path to use for "make defconfig".
-KBUILD_DEFCONFIG := $(ARCH)_defconfig
+# We default to the newer TILE-Gx architecture if only "tile" is given.
+ifeq ($(ARCH),tile)
+        KBUILD_DEFCONFIG := tilegx_defconfig
+else
+        KBUILD_DEFCONFIG := $(ARCH)_defconfig
+endif
 
 # Used as a file extension when useful, e.g. head_$(BITS).o
 # Not needed for (e.g.) "$(CC) -m32" since the compiler automatically

arch/tile/include/arch/spr_def_32.h

@@ -65,6 +65,31 @@
 #define SPR_EX_CONTEXT_2_1__ICS_RMASK 0x1
 #define SPR_EX_CONTEXT_2_1__ICS_MASK  0x4
 #define SPR_FAIL 0x4e09
+#define SPR_IDN_AVAIL_EN 0x3e05
+#define SPR_IDN_CA_DATA 0x0b00
+#define SPR_IDN_DATA_AVAIL 0x0b03
+#define SPR_IDN_DEADLOCK_TIMEOUT 0x3406
+#define SPR_IDN_DEMUX_CA_COUNT 0x0a05
+#define SPR_IDN_DEMUX_COUNT_0 0x0a06
+#define SPR_IDN_DEMUX_COUNT_1 0x0a07
+#define SPR_IDN_DEMUX_CTL 0x0a08
+#define SPR_IDN_DEMUX_QUEUE_SEL 0x0a0a
+#define SPR_IDN_DEMUX_STATUS 0x0a0b
+#define SPR_IDN_DEMUX_WRITE_FIFO 0x0a0c
+#define SPR_IDN_DIRECTION_PROTECT 0x2e05
+#define SPR_IDN_PENDING 0x0a0e
+#define SPR_IDN_REFILL_EN 0x0e05
+#define SPR_IDN_SP_FIFO_DATA 0x0a0f
+#define SPR_IDN_SP_FIFO_SEL 0x0a10
+#define SPR_IDN_SP_FREEZE 0x0a11
+#define SPR_IDN_SP_FREEZE__SP_FRZ_MASK  0x1
+#define SPR_IDN_SP_FREEZE__DEMUX_FRZ_MASK  0x2
+#define SPR_IDN_SP_FREEZE__NON_DEST_EXT_MASK  0x4
+#define SPR_IDN_SP_STATE 0x0a12
+#define SPR_IDN_TAG_0 0x0a13
+#define SPR_IDN_TAG_1 0x0a14
+#define SPR_IDN_TAG_VALID 0x0a15
+#define SPR_IDN_TILE_COORD 0x0a16
 #define SPR_INTCTRL_0_STATUS 0x4a07
 #define SPR_INTCTRL_1_STATUS 0x4807
 #define SPR_INTCTRL_2_STATUS 0x4607
@@ -87,12 +112,36 @@
 #define SPR_INTERRUPT_MASK_SET_1_1 0x480e
 #define SPR_INTERRUPT_MASK_SET_2_0 0x460c
 #define SPR_INTERRUPT_MASK_SET_2_1 0x460d
+#define SPR_MPL_AUX_PERF_COUNT_SET_0 0x6000
+#define SPR_MPL_AUX_PERF_COUNT_SET_1 0x6001
+#define SPR_MPL_AUX_PERF_COUNT_SET_2 0x6002
 #define SPR_MPL_DMA_CPL_SET_0 0x5800
 #define SPR_MPL_DMA_CPL_SET_1 0x5801
 #define SPR_MPL_DMA_CPL_SET_2 0x5802
 #define SPR_MPL_DMA_NOTIFY_SET_0 0x3800
 #define SPR_MPL_DMA_NOTIFY_SET_1 0x3801
 #define SPR_MPL_DMA_NOTIFY_SET_2 0x3802
+#define SPR_MPL_IDN_ACCESS_SET_0 0x0a00
+#define SPR_MPL_IDN_ACCESS_SET_1 0x0a01
+#define SPR_MPL_IDN_ACCESS_SET_2 0x0a02
+#define SPR_MPL_IDN_AVAIL_SET_0 0x3e00
+#define SPR_MPL_IDN_AVAIL_SET_1 0x3e01
+#define SPR_MPL_IDN_AVAIL_SET_2 0x3e02
+#define SPR_MPL_IDN_CA_SET_0 0x3a00
+#define SPR_MPL_IDN_CA_SET_1 0x3a01
+#define SPR_MPL_IDN_CA_SET_2 0x3a02
+#define SPR_MPL_IDN_COMPLETE_SET_0 0x1200
+#define SPR_MPL_IDN_COMPLETE_SET_1 0x1201
+#define SPR_MPL_IDN_COMPLETE_SET_2 0x1202
+#define SPR_MPL_IDN_FIREWALL_SET_0 0x2e00
+#define SPR_MPL_IDN_FIREWALL_SET_1 0x2e01
+#define SPR_MPL_IDN_FIREWALL_SET_2 0x2e02
+#define SPR_MPL_IDN_REFILL_SET_0 0x0e00
+#define SPR_MPL_IDN_REFILL_SET_1 0x0e01
+#define SPR_MPL_IDN_REFILL_SET_2 0x0e02
+#define SPR_MPL_IDN_TIMER_SET_0 0x3400
+#define SPR_MPL_IDN_TIMER_SET_1 0x3401
+#define SPR_MPL_IDN_TIMER_SET_2 0x3402
 #define SPR_MPL_INTCTRL_0_SET_0 0x4a00
 #define SPR_MPL_INTCTRL_0_SET_1 0x4a01
 #define SPR_MPL_INTCTRL_0_SET_2 0x4a02
@@ -102,6 +151,9 @@
 #define SPR_MPL_INTCTRL_2_SET_0 0x4600
 #define SPR_MPL_INTCTRL_2_SET_1 0x4601
 #define SPR_MPL_INTCTRL_2_SET_2 0x4602
+#define SPR_MPL_PERF_COUNT_SET_0 0x4200
+#define SPR_MPL_PERF_COUNT_SET_1 0x4201
+#define SPR_MPL_PERF_COUNT_SET_2 0x4202
 #define SPR_MPL_SN_ACCESS_SET_0 0x0800
 #define SPR_MPL_SN_ACCESS_SET_1 0x0801
 #define SPR_MPL_SN_ACCESS_SET_2 0x0802
@@ -181,6 +233,7 @@
 #define SPR_UDN_DEMUX_STATUS 0x0c0d
 #define SPR_UDN_DEMUX_WRITE_FIFO 0x0c0e
 #define SPR_UDN_DIRECTION_PROTECT 0x3005
+#define SPR_UDN_PENDING 0x0c10
 #define SPR_UDN_REFILL_EN 0x1005
 #define SPR_UDN_SP_FIFO_DATA 0x0c11
 #define SPR_UDN_SP_FIFO_SEL 0x0c12
@@ -195,6 +248,9 @@
 #define SPR_UDN_TAG_3 0x0c18
 #define SPR_UDN_TAG_VALID 0x0c19
 #define SPR_UDN_TILE_COORD 0x0c1a
+#define SPR_WATCH_CTL 0x4209
+#define SPR_WATCH_MASK 0x420a
+#define SPR_WATCH_VAL 0x420b
 
 #endif /* !defined(__ARCH_SPR_DEF_H__) */
 

arch/tile/include/arch/spr_def_64.h

@@ -52,6 +52,13 @@
 #define SPR_EX_CONTEXT_2_1__ICS_RMASK 0x1
 #define SPR_EX_CONTEXT_2_1__ICS_MASK  0x4
 #define SPR_FAIL 0x2707
+#define SPR_IDN_AVAIL_EN 0x1a05
+#define SPR_IDN_DATA_AVAIL 0x0a80
+#define SPR_IDN_DEADLOCK_TIMEOUT 0x1806
+#define SPR_IDN_DEMUX_COUNT_0 0x0a05
+#define SPR_IDN_DEMUX_COUNT_1 0x0a06
+#define SPR_IDN_DIRECTION_PROTECT 0x1405
+#define SPR_IDN_PENDING 0x0a08
 #define SPR_ILL_TRANS_REASON__I_STREAM_VA_RMASK 0x1
 #define SPR_INTCTRL_0_STATUS 0x2505
 #define SPR_INTCTRL_1_STATUS 0x2405
@@ -88,9 +95,27 @@
 #define SPR_IPI_MASK_SET_0 0x1f0a
 #define SPR_IPI_MASK_SET_1 0x1e0a
 #define SPR_IPI_MASK_SET_2 0x1d0a
+#define SPR_MPL_AUX_PERF_COUNT_SET_0 0x2100
+#define SPR_MPL_AUX_PERF_COUNT_SET_1 0x2101
+#define SPR_MPL_AUX_PERF_COUNT_SET_2 0x2102
 #define SPR_MPL_AUX_TILE_TIMER_SET_0 0x1700
 #define SPR_MPL_AUX_TILE_TIMER_SET_1 0x1701
 #define SPR_MPL_AUX_TILE_TIMER_SET_2 0x1702
+#define SPR_MPL_IDN_ACCESS_SET_0 0x0a00
+#define SPR_MPL_IDN_ACCESS_SET_1 0x0a01
+#define SPR_MPL_IDN_ACCESS_SET_2 0x0a02
+#define SPR_MPL_IDN_AVAIL_SET_0 0x1a00
+#define SPR_MPL_IDN_AVAIL_SET_1 0x1a01
+#define SPR_MPL_IDN_AVAIL_SET_2 0x1a02
+#define SPR_MPL_IDN_COMPLETE_SET_0 0x0500
+#define SPR_MPL_IDN_COMPLETE_SET_1 0x0501
+#define SPR_MPL_IDN_COMPLETE_SET_2 0x0502
+#define SPR_MPL_IDN_FIREWALL_SET_0 0x1400
+#define SPR_MPL_IDN_FIREWALL_SET_1 0x1401
+#define SPR_MPL_IDN_FIREWALL_SET_2 0x1402
+#define SPR_MPL_IDN_TIMER_SET_0 0x1800
+#define SPR_MPL_IDN_TIMER_SET_1 0x1801
+#define SPR_MPL_IDN_TIMER_SET_2 0x1802
 #define SPR_MPL_INTCTRL_0_SET_0 0x2500
 #define SPR_MPL_INTCTRL_0_SET_1 0x2501
 #define SPR_MPL_INTCTRL_0_SET_2 0x2502
@@ -100,6 +125,21 @@
 #define SPR_MPL_INTCTRL_2_SET_0 0x2300
 #define SPR_MPL_INTCTRL_2_SET_1 0x2301
 #define SPR_MPL_INTCTRL_2_SET_2 0x2302
+#define SPR_MPL_IPI_0 0x1f04
+#define SPR_MPL_IPI_0_SET_0 0x1f00
+#define SPR_MPL_IPI_0_SET_1 0x1f01
+#define SPR_MPL_IPI_0_SET_2 0x1f02
+#define SPR_MPL_IPI_1 0x1e04
+#define SPR_MPL_IPI_1_SET_0 0x1e00
+#define SPR_MPL_IPI_1_SET_1 0x1e01
+#define SPR_MPL_IPI_1_SET_2 0x1e02
+#define SPR_MPL_IPI_2 0x1d04
+#define SPR_MPL_IPI_2_SET_0 0x1d00
+#define SPR_MPL_IPI_2_SET_1 0x1d01
+#define SPR_MPL_IPI_2_SET_2 0x1d02
+#define SPR_MPL_PERF_COUNT_SET_0 0x2000
+#define SPR_MPL_PERF_COUNT_SET_1 0x2001
+#define SPR_MPL_PERF_COUNT_SET_2 0x2002
 #define SPR_MPL_UDN_ACCESS_SET_0 0x0b00
 #define SPR_MPL_UDN_ACCESS_SET_1 0x0b01
 #define SPR_MPL_UDN_ACCESS_SET_2 0x0b02
@@ -167,6 +207,9 @@
 #define SPR_UDN_DEMUX_COUNT_2 0x0b07
 #define SPR_UDN_DEMUX_COUNT_3 0x0b08
 #define SPR_UDN_DIRECTION_PROTECT 0x1505
+#define SPR_UDN_PENDING 0x0b0a
+#define SPR_WATCH_MASK 0x200a
+#define SPR_WATCH_VAL 0x200b
 
 #endif /* !defined(__ARCH_SPR_DEF_H__) */
 

arch/tile/include/asm/Kbuild

@@ -2,6 +2,7 @@ include include/asm-generic/Kbuild.asm
 
 header-y += ../arch/
 
+header-y += cachectl.h
 header-y += ucontext.h
 header-y += hardwall.h
 
@@ -21,7 +22,6 @@ generic-y += ipcbuf.h
 generic-y += irq_regs.h
 generic-y += kdebug.h
 generic-y += local.h
-generic-y += module.h
 generic-y += msgbuf.h
 generic-y += mutex.h
 generic-y += param.h

arch/tile/include/asm/atomic_32.h

@@ -303,7 +303,14 @@ void __init_atomic_per_cpu(void);
 void __atomic_fault_unlock(int *lock_ptr);
 #endif
 
+/* Return a pointer to the lock for the given address. */
+int *__atomic_hashed_lock(volatile void *v);
+
 /* Private helper routines in lib/atomic_asm_32.S */
+struct __get_user {
+	unsigned long val;
+	int err;
+};
 extern struct __get_user __atomic_cmpxchg(volatile int *p,
 					  int *lock, int o, int n);
 extern struct __get_user __atomic_xchg(volatile int *p, int *lock, int n);
@@ -319,6 +326,9 @@ extern u64 __atomic64_xchg_add(volatile u64 *p, int *lock, u64 n);
 extern u64 __atomic64_xchg_add_unless(volatile u64 *p,
 				      int *lock, u64 o, u64 n);
 
+/* Return failure from the atomic wrappers. */
+struct __get_user __atomic_bad_address(int __user *addr);
+
 #endif /* !__ASSEMBLY__ */
 
 #endif /* _ASM_TILE_ATOMIC_32_H */

arch/tile/include/asm/bitops.h

@@ -77,6 +77,11 @@ static inline int ffs(int x)
 	return __builtin_ffs(x);
 }
 
+static inline int fls64(__u64 w)
+{
+	return (sizeof(__u64) * 8) - __builtin_clzll(w);
+}
+
 /**
  * fls - find last set bit in word
  * @x: the word to search
@@ -90,12 +95,7 @@ static inline int ffs(int x)
  */
 static inline int fls(int x)
 {
-	return (sizeof(int) * 8) - __builtin_clz(x);
-}
-
-static inline int fls64(__u64 w)
-{
-	return (sizeof(__u64) * 8) - __builtin_clzll(w);
+	return fls64((unsigned int) x);
 }
 
 static inline unsigned int __arch_hweight32(unsigned int w)

arch/tile/include/asm/byteorder.h

@@ -1 +1,21 @@
+/*
+ * Copyright 2011 Tilera Corporation. All Rights Reserved.
+ *
+ *   This program is free software; you can redistribute it and/or
+ *   modify it under the terms of the GNU General Public License
+ *   as published by the Free Software Foundation, version 2.
+ *
+ *   This program is distributed in the hope that it will be useful, but
+ *   WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ *   NON INFRINGEMENT.  See the GNU General Public License for
+ *   more details.
+ */
+
+#if defined (__BIG_ENDIAN__)
+#include <linux/byteorder/big_endian.h>
+#elif defined (__LITTLE_ENDIAN__)
 #include <linux/byteorder/little_endian.h>
+#else
+#error "__BIG_ENDIAN__ or __LITTLE_ENDIAN__ must be defined."
+#endif

arch/tile/include/asm/cachectl.h

@@ -0,0 +1,42 @@
+/*
+ * Copyright 2011 Tilera Corporation. All Rights Reserved.
+ *
+ *   This program is free software; you can redistribute it and/or
+ *   modify it under the terms of the GNU General Public License
+ *   as published by the Free Software Foundation, version 2.
+ *
+ *   This program is distributed in the hope that it will be useful, but
+ *   WITHOUT ANY WARRANTY; without even the implied warranty of
+ *   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ *   NON INFRINGEMENT.  See the GNU General Public License for
+ *   more details.
+ */
+
+#ifndef _ASM_TILE_CACHECTL_H
+#define _ASM_TILE_CACHECTL_H
+
+/*
+ * Options for cacheflush system call.
+ *
+ * The ICACHE flush is performed on all cores currently running the
+ * current process's address space.  The intent is for user
+ * applications to be able to modify code, invoke the system call,
+ * then allow arbitrary other threads in the same address space to see
+ * the newly-modified code.  Passing a length of CHIP_L1I_CACHE_SIZE()
+ * or more invalidates the entire icache on all cores in the address
+ * spaces.  (Note: currently this option invalidates the entire icache
+ * regardless of the requested address and length, but we may choose
+ * to honor the arguments at some point.)
+ *
+ * Flush and invalidation of memory can normally be performed with the
+ * __insn_flush(), __insn_inv(), and __insn_finv() instructions from
+ * userspace.  The DCACHE option to the system call allows userspace
+ * to flush the entire L1+L2 data cache from the core.  In this case,
+ * the address and length arguments are not used.  The DCACHE flush is
+ * restricted to the current core, not all cores in the address space.
+ */
+#define	ICACHE	(1<<0)		/* invalidate L1 instruction cache */
+#define	DCACHE	(1<<1)		/* flush and invalidate data cache */
+#define	BCACHE	(ICACHE|DCACHE)	/* flush both caches               */
+
+#endif	/* _ASM_TILE_CACHECTL_H */

arch/tile/include/asm/compat.h

@@ -242,9 +242,6 @@ long compat_sys_fallocate(int fd, int mode,
 long compat_sys_sched_rr_get_interval(compat_pid_t pid,
 				      struct compat_timespec __user *interval);
 
-/* Tilera Linux syscalls that don't have "compat" versions. */
-#define compat_sys_flush_cache sys_flush_cache
-
 /* These are the intvec_64.S trampolines. */
 long _compat_sys_execve(const char __user *path,
 			const compat_uptr_t __user *argv,