[PATCH] spinlock consolidation

This patch (written by me and also containing many suggestions of Arjan van
de Ven) does a major cleanup of the spinlock code.  It does the following
things:

 - consolidates and enhances the spinlock/rwlock debugging code

 - simplifies the asm/spinlock.h files

 - encapsulates the raw spinlock type and moves generic spinlock
   features (such as ->break_lock) into the generic code.

 - cleans up the spinlock code hierarchy to get rid of the spaghetti.

Most notably there's now only a single variant of the debugging code,
located in lib/spinlock_debug.c.  (previously we had one SMP debugging
variant per architecture, plus a separate generic one for UP builds)

Also, i've enhanced the rwlock debugging facility, it will now track
write-owners.  There is new spinlock-owner/CPU-tracking on SMP builds too.
All locks have lockup detection now, which will work for both soft and hard
spin/rwlock lockups.

The arch-level include files now only contain the minimally necessary
subset of the spinlock code - all the rest that can be generalized now
lives in the generic headers:

 include/asm-i386/spinlock_types.h       |   16
 include/asm-x86_64/spinlock_types.h     |   16

I have also split up the various spinlock variants into separate files,
making it easier to see which does what. The new layout is:

   SMP                         |  UP
   ----------------------------|-----------------------------------
   asm/spinlock_types_smp.h    |  linux/spinlock_types_up.h
   linux/spinlock_types.h      |  linux/spinlock_types.h
   asm/spinlock_smp.h          |  linux/spinlock_up.h
   linux/spinlock_api_smp.h    |  linux/spinlock_api_up.h
   linux/spinlock.h            |  linux/spinlock.h

/*
 * here's the role of the various spinlock/rwlock related include files:
 *
 * on SMP builds:
 *
 *  asm/spinlock_types.h: contains the raw_spinlock_t/raw_rwlock_t and the
 *                        initializers
 *
 *  linux/spinlock_types.h:
 *                        defines the generic type and initializers
 *
 *  asm/spinlock.h:       contains the __raw_spin_*()/etc. lowlevel
 *                        implementations, mostly inline assembly code
 *
 *   (also included on UP-debug builds:)
 *
 *  linux/spinlock_api_smp.h:
 *                        contains the prototypes for the _spin_*() APIs.
 *
 *  linux/spinlock.h:     builds the final spin_*() APIs.
 *
 * on UP builds:
 *
 *  linux/spinlock_type_up.h:
 *                        contains the generic, simplified UP spinlock type.
 *                        (which is an empty structure on non-debug builds)
 *
 *  linux/spinlock_types.h:
 *                        defines the generic type and initializers
 *
 *  linux/spinlock_up.h:
 *                        contains the __raw_spin_*()/etc. version of UP
 *                        builds. (which are NOPs on non-debug, non-preempt
 *                        builds)
 *
 *   (included on UP-non-debug builds:)
 *
 *  linux/spinlock_api_up.h:
 *                        builds the _spin_*() APIs.
 *
 *  linux/spinlock.h:     builds the final spin_*() APIs.
 */

All SMP and UP architectures are converted by this patch.

arm, i386, ia64, ppc, ppc64, s390/s390x, x64 was build-tested via
crosscompilers.  m32r, mips, sh, sparc, have not been tested yet, but should
be mostly fine.

From: Grant Grundler <grundler@parisc-linux.org>

  Booted and lightly tested on a500-44 (64-bit, SMP kernel, dual CPU).
  Builds 32-bit SMP kernel (not booted or tested).  I did not try to build
  non-SMP kernels.  That should be trivial to fix up later if necessary.

  I converted bit ops atomic_hash lock to raw_spinlock_t.  Doing so avoids
  some ugly nesting of linux/*.h and asm/*.h files.  Those particular locks
  are well tested and contained entirely inside arch specific code.  I do NOT
  expect any new issues to arise with them.

 If someone does ever need to use debug/metrics with them, then they will
  need to unravel this hairball between spinlocks, atomic ops, and bit ops
  that exist only because parisc has exactly one atomic instruction: LDCW
  (load and clear word).

From: "Luck, Tony" <tony.luck@intel.com>

   ia64 fix

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjanv@infradead.org>
Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Cc: Matthew Wilcox <willy@debian.org>
Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Signed-off-by: Mikael Pettersson <mikpe@csd.uu.se>
Signed-off-by: Benoit Boissinot <benoit.boissinot@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

7 files changed, 59 insertions, 165 deletions

diff --git a/include/asm-parisc/atomic.h b/include/asm-parisc/atomic.h
index e24f7579adb0..048a2c7fd0c0 100644
--- a/include/asm-parisc/atomic.h
+++ b/include/asm-parisc/atomic.h
@@ -24,19 +24,19 @@
 #  define ATOMIC_HASH_SIZE 4
 #  define ATOMIC_HASH(a) (&(__atomic_hash[ (((unsigned long) a)/L1_CACHE_BYTES) & (ATOMIC_HASH_SIZE-1) ]))
 
-extern spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] __lock_aligned;
+extern raw_spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] __lock_aligned;
 
-/* Can't use _raw_spin_lock_irq because of #include problems, so
+/* Can't use raw_spin_lock_irq because of #include problems, so
  * this is the substitute */
 #define _atomic_spin_lock_irqsave(l,f) do {	\
-	spinlock_t *s = ATOMIC_HASH(l);		\
+	raw_spinlock_t *s = ATOMIC_HASH(l);		\
 	local_irq_save(f);			\
-	_raw_spin_lock(s);			\
+	__raw_spin_lock(s);			\
 } while(0)
 
 #define _atomic_spin_unlock_irqrestore(l,f) do {	\
-	spinlock_t *s = ATOMIC_HASH(l);			\
-	_raw_spin_unlock(s);				\
+	raw_spinlock_t *s = ATOMIC_HASH(l);			\
+	__raw_spin_unlock(s);				\
 	local_irq_restore(f);				\
 } while(0)
 
diff --git a/include/asm-parisc/bitops.h b/include/asm-parisc/bitops.h
index 928e5ef850bd..af7db694b22d 100644
--- a/include/asm-parisc/bitops.h
+++ b/include/asm-parisc/bitops.h
@@ -2,7 +2,7 @@
 #define _PARISC_BITOPS_H
 
 #include <linux/compiler.h>
-#include <asm/system.h>
+#include <asm/spinlock.h>
 #include <asm/byteorder.h>
 #include <asm/atomic.h>
 
diff --git a/include/asm-parisc/cacheflush.h b/include/asm-parisc/cacheflush.h
index 06732719d927..aa592d8c0e39 100644
--- a/include/asm-parisc/cacheflush.h
+++ b/include/asm-parisc/cacheflush.h
@@ -3,6 +3,7 @@
 
 #include <linux/config.h>
 #include <linux/mm.h>
+#include <asm/cache.h>	/* for flush_user_dcache_range_asm() proto */
 
 /* The usual comment is "Caches aren't brain-dead on the <architecture>".
  * Unfortunately, that doesn't apply to PA-RISC. */
diff --git a/include/asm-parisc/processor.h b/include/asm-parisc/processor.h
index 0b61f51d8467..a9dfadd05658 100644
--- a/include/asm-parisc/processor.h
+++ b/include/asm-parisc/processor.h
@@ -11,6 +11,7 @@
 #ifndef __ASSEMBLY__
 #include <linux/config.h>
 #include <linux/threads.h>
+#include <linux/spinlock_types.h>
 
 #include <asm/hardware.h>
 #include <asm/page.h>
diff --git a/include/asm-parisc/spinlock.h b/include/asm-parisc/spinlock.h
index 679ea1c651ef..43eaa6e742e0 100644
--- a/include/asm-parisc/spinlock.h
+++ b/include/asm-parisc/spinlock.h
@@ -2,30 +2,25 @@
 #define __ASM_SPINLOCK_H
 
 #include <asm/system.h>
+#include <asm/processor.h>
+#include <asm/spinlock_types.h>
 
 /* Note that PA-RISC has to use `1' to mean unlocked and `0' to mean locked
  * since it only has load-and-zero. Moreover, at least on some PA processors,
  * the semaphore address has to be 16-byte aligned.
  */
 
-#ifndef CONFIG_DEBUG_SPINLOCK
-
-#define __SPIN_LOCK_UNLOCKED	{ { 1, 1, 1, 1 } }
-#undef SPIN_LOCK_UNLOCKED
-#define SPIN_LOCK_UNLOCKED (spinlock_t) __SPIN_LOCK_UNLOCKED
-
-#define spin_lock_init(x)	do { *(x) = SPIN_LOCK_UNLOCKED; } while(0)
-
-static inline int spin_is_locked(spinlock_t *x)
+static inline int __raw_spin_is_locked(raw_spinlock_t *x)
 {
 	volatile unsigned int *a = __ldcw_align(x);
 	return *a == 0;
 }
 
-#define spin_unlock_wait(x)	do { barrier(); } while(spin_is_locked(x))
-#define _raw_spin_lock_flags(lock, flags) _raw_spin_lock(lock)
+#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
+#define __raw_spin_unlock_wait(x) \
+		do { cpu_relax(); } while (__raw_spin_is_locked(x))
 
-static inline void _raw_spin_lock(spinlock_t *x)
+static inline void __raw_spin_lock(raw_spinlock_t *x)
 {
 	volatile unsigned int *a;
 
@@ -36,7 +31,7 @@ static inline void _raw_spin_lock(spinlock_t *x)
 	mb();
 }
 
-static inline void _raw_spin_unlock(spinlock_t *x)
+static inline void __raw_spin_unlock(raw_spinlock_t *x)
 {
 	volatile unsigned int *a;
 	mb();
@@ -45,7 +40,7 @@ static inline void _raw_spin_unlock(spinlock_t *x)
 	mb();
 }
 
-static inline int _raw_spin_trylock(spinlock_t *x)
+static inline int __raw_spin_trylock(raw_spinlock_t *x)
 {
 	volatile unsigned int *a;
 	int ret;
@@ -57,131 +52,38 @@ static inline int _raw_spin_trylock(spinlock_t *x)
 
 	return ret;
 }
-	
-#define spin_lock_own(LOCK, LOCATION)	((void)0)
-
-#else /* !(CONFIG_DEBUG_SPINLOCK) */
-
-#define SPINLOCK_MAGIC	0x1D244B3C
-
-#define __SPIN_LOCK_UNLOCKED	{ { 1, 1, 1, 1 }, SPINLOCK_MAGIC, 10, __FILE__ , NULL, 0, -1, NULL, NULL }
-#undef SPIN_LOCK_UNLOCKED
-#define SPIN_LOCK_UNLOCKED (spinlock_t) __SPIN_LOCK_UNLOCKED
-
-#define spin_lock_init(x)	do { *(x) = SPIN_LOCK_UNLOCKED; } while(0)
-
-#define CHECK_LOCK(x)							\
-	do {								\
-	 	if (unlikely((x)->magic != SPINLOCK_MAGIC)) {			\
-			printk(KERN_ERR "%s:%d: spin_is_locked"		\
-			" on uninitialized spinlock %p.\n",		\
-				__FILE__, __LINE__, (x)); 		\
-		} 							\
-	} while(0)
-
-#define spin_is_locked(x)						\
-	({								\
-	 	CHECK_LOCK(x);						\
-		volatile unsigned int *a = __ldcw_align(x);		\
-		if (unlikely((*a == 0) && (x)->babble)) {				\
-			(x)->babble--;					\
-			printk("KERN_WARNING				\
-				%s:%d: spin_is_locked(%s/%p) already"	\
-				" locked by %s:%d in %s at %p(%d)\n",	\
-				__FILE__,__LINE__, (x)->module,	(x),	\
-				(x)->bfile, (x)->bline, (x)->task->comm,\
-				(x)->previous, (x)->oncpu);		\
-		}							\
-		*a == 0;						\
-	})
-
-#define spin_unlock_wait(x)						\
-	do {								\
-	 	CHECK_LOCK(x);						\
-		volatile unsigned int *a = __ldcw_align(x);		\
-		if (unlikely((*a == 0) && (x)->babble)) {				\
-			(x)->babble--;					\
-			printk("KERN_WARNING				\
-				%s:%d: spin_unlock_wait(%s/%p)"		\
-				" owned by %s:%d in %s at %p(%d)\n",	\
-				__FILE__,__LINE__, (x)->module, (x),	\
-				(x)->bfile, (x)->bline, (x)->task->comm,\
-				(x)->previous, (x)->oncpu);		\
-		}							\
-		barrier();						\
-	} while (*((volatile unsigned char *)(__ldcw_align(x))) == 0)
-
-extern void _dbg_spin_lock(spinlock_t *lock, const char *base_file, int line_no);
-extern void _dbg_spin_unlock(spinlock_t *lock, const char *, int);
-extern int _dbg_spin_trylock(spinlock_t * lock, const char *, int);
-
-#define _raw_spin_lock_flags(lock, flags) _raw_spin_lock(lock)
-
-#define _raw_spin_unlock(lock)	_dbg_spin_unlock(lock, __FILE__, __LINE__)
-#define _raw_spin_lock(lock) _dbg_spin_lock(lock, __FILE__, __LINE__)
-#define _raw_spin_trylock(lock) _dbg_spin_trylock(lock, __FILE__, __LINE__)
-
-/* just in case we need it */
-#define spin_lock_own(LOCK, LOCATION)					\
-do {									\
-	volatile unsigned int *a = __ldcw_align(LOCK);			\
-	if (!((*a == 0) && ((LOCK)->oncpu == smp_processor_id())))	\
-		printk("KERN_WARNING					\
-			%s: called on %d from %p but lock %s on %d\n",	\
-			LOCATION, smp_processor_id(),			\
-			__builtin_return_address(0),			\
-			(*a == 0) ? "taken" : "freed", (LOCK)->on_cpu);	\
-} while (0)
-
-#endif /* !(CONFIG_DEBUG_SPINLOCK) */
 
 /*
  * Read-write spinlocks, allowing multiple readers
  * but only one writer.
  */
-typedef struct {
-	spinlock_t lock;
-	volatile int counter;
-#ifdef CONFIG_PREEMPT
-	unsigned int break_lock;
-#endif
-} rwlock_t;
-
-#define RW_LOCK_UNLOCKED (rwlock_t) { __SPIN_LOCK_UNLOCKED, 0 }
-
-#define rwlock_init(lp)	do { *(lp) = RW_LOCK_UNLOCKED; } while (0)
 
-#define _raw_read_trylock(lock) generic_raw_read_trylock(lock)
+#define __raw_read_trylock(lock) generic__raw_read_trylock(lock)
 
 /* read_lock, read_unlock are pretty straightforward.  Of course it somehow
  * sucks we end up saving/restoring flags twice for read_lock_irqsave aso. */
 
-#ifdef CONFIG_DEBUG_RWLOCK
-extern void _dbg_read_lock(rwlock_t * rw, const char *bfile, int bline);
-#define _raw_read_lock(rw) _dbg_read_lock(rw, __FILE__, __LINE__)
-#else
-static  __inline__ void _raw_read_lock(rwlock_t *rw)
+static  __inline__ void __raw_read_lock(raw_rwlock_t *rw)
 {
 	unsigned long flags;
 	local_irq_save(flags);
-	_raw_spin_lock(&rw->lock); 
+	__raw_spin_lock(&rw->lock);
 
 	rw->counter++;
 
-	_raw_spin_unlock(&rw->lock);
+	__raw_spin_unlock(&rw->lock);
 	local_irq_restore(flags);
 }
-#endif	/* CONFIG_DEBUG_RWLOCK */
 
-static  __inline__ void _raw_read_unlock(rwlock_t *rw)
+static  __inline__ void __raw_read_unlock(raw_rwlock_t *rw)
 {
 	unsigned long flags;
 	local_irq_save(flags);
-	_raw_spin_lock(&rw->lock); 
+	__raw_spin_lock(&rw->lock);
 
 	rw->counter--;
 
-	_raw_spin_unlock(&rw->lock);
+	__raw_spin_unlock(&rw->lock);
 	local_irq_restore(flags);
 }
 
@@ -194,20 +96,17 @@ static  __inline__ void _raw_read_unlock(rwlock_t *rw)
  * writers) in interrupt handlers someone fucked up and we'd dead-lock
  * sooner or later anyway.   prumpf */
 
-#ifdef CONFIG_DEBUG_RWLOCK
-extern void _dbg_write_lock(rwlock_t * rw, const char *bfile, int bline);
-#define _raw_write_lock(rw) _dbg_write_lock(rw, __FILE__, __LINE__)
-#else
-static  __inline__ void _raw_write_lock(rwlock_t *rw)
+static  __inline__ void __raw_write_lock(raw_rwlock_t *rw)
 {
 retry:
-	_raw_spin_lock(&rw->lock);
+	__raw_spin_lock(&rw->lock);
 
 	if(rw->counter != 0) {
 		/* this basically never happens */
-		_raw_spin_unlock(&rw->lock);
+		__raw_spin_unlock(&rw->lock);
 
-		while(rw->counter != 0);
+		while (rw->counter != 0)
+			cpu_relax();
 
 		goto retry;
 	}
@@ -215,26 +114,21 @@ retry:
 	/* got it.  now leave without unlocking */
 	rw->counter = -1; /* remember we are locked */
 }
-#endif /* CONFIG_DEBUG_RWLOCK */
 
 /* write_unlock is absolutely trivial - we don't have to wait for anything */
 
-static  __inline__ void _raw_write_unlock(rwlock_t *rw)
+static  __inline__ void __raw_write_unlock(raw_rwlock_t *rw)
 {
 	rw->counter = 0;
-	_raw_spin_unlock(&rw->lock);
+	__raw_spin_unlock(&rw->lock);
 }
 
-#ifdef CONFIG_DEBUG_RWLOCK
-extern int _dbg_write_trylock(rwlock_t * rw, const char *bfile, int bline);
-#define _raw_write_trylock(rw) _dbg_write_trylock(rw, __FILE__, __LINE__)
-#else
-static  __inline__ int _raw_write_trylock(rwlock_t *rw)
+static  __inline__ int __raw_write_trylock(raw_rwlock_t *rw)
 {
-	_raw_spin_lock(&rw->lock);
+	__raw_spin_lock(&rw->lock);
 	if (rw->counter != 0) {
 		/* this basically never happens */
-		_raw_spin_unlock(&rw->lock);
+		__raw_spin_unlock(&rw->lock);
 
 		return 0;
 	}
@@ -243,14 +137,13 @@ static  __inline__ int _raw_write_trylock(rwlock_t *rw)
 	rw->counter = -1; /* remember we are locked */
 	return 1;
 }
-#endif /* CONFIG_DEBUG_RWLOCK */
 
-static __inline__ int is_read_locked(rwlock_t *rw)
+static __inline__ int __raw_is_read_locked(raw_rwlock_t *rw)
 {
 	return rw->counter > 0;
 }
 
-static __inline__ int is_write_locked(rwlock_t *rw)
+static __inline__ int __raw_is_write_locked(raw_rwlock_t *rw)
 {
 	return rw->counter < 0;
 }
diff --git a/include/asm-parisc/spinlock_types.h b/include/asm-parisc/spinlock_types.h
new file mode 100644
index 000000000000..785bba822fbf
--- /dev/null
+++ b/include/asm-parisc/spinlock_types.h
@@ -0,0 +1,21 @@
+#ifndef __ASM_SPINLOCK_TYPES_H
+#define __ASM_SPINLOCK_TYPES_H
+
+#ifndef __LINUX_SPINLOCK_TYPES_H
+# error "please don't include this file directly"
+#endif
+
+typedef struct {
+	volatile unsigned int lock[4];
+} raw_spinlock_t;
+
+#define __RAW_SPIN_LOCK_UNLOCKED	{ { 1, 1, 1, 1 } }
+
+typedef struct {
+	raw_spinlock_t lock;
+	volatile int counter;
+} raw_rwlock_t;
+
+#define __RAW_RW_LOCK_UNLOCKED		{ __RAW_SPIN_LOCK_UNLOCKED, 0 }
+
+#endif
diff --git a/include/asm-parisc/system.h b/include/asm-parisc/system.h
index 81c543339036..26ff844a21c1 100644
--- a/include/asm-parisc/system.h
+++ b/include/asm-parisc/system.h
@@ -160,29 +160,7 @@ static inline void set_eiem(unsigned long val)
 })
 
 #ifdef CONFIG_SMP
-/*
- * Your basic SMP spinlocks, allowing only a single CPU anywhere
- */
-
-typedef struct {
-	volatile unsigned int lock[4];
-#ifdef CONFIG_DEBUG_SPINLOCK
-	unsigned long magic;
-	volatile unsigned int babble;
-	const char *module;
-	char *bfile;
-	int bline;
-	int oncpu;
-	void *previous;
-	struct task_struct * task;
-#endif
-#ifdef CONFIG_PREEMPT
-	unsigned int break_lock;
-#endif
-} spinlock_t;
-
-#define __lock_aligned __attribute__((__section__(".data.lock_aligned")))
-
+# define __lock_aligned __attribute__((__section__(".data.lock_aligned")))
 #endif
 
 #define KERNEL_START (0x10100000 - 0x1000)