tile: rework <asm/cmpxchg.h>

The macrology in cmpxchg.h was designed to allow arbitrary pointer
and integer values to be passed through the routines.  To support
cmpxchg() on 64-bit values on the 32-bit tilepro architecture, we
used the idiom "(typeof(val))(typeof(val-val))".  This way, in the
"size 8" branch of the switch, when the underlying cmpxchg routine
returns a 64-bit quantity, we cast it first to a typeof(val-val)
quantity (i.e. size_t if "val" is a pointer) with no warnings about
casting between pointers and integers of different sizes, then cast
onwards to typeof(val), again with no warnings.  If val is not a
pointer type, the additional cast is a no-op.  We can't replace the
typeof(val-val) cast with (for example) unsigned long, since then if
"val" is really a 64-bit type, we cast away the high bits.

HOWEVER, this fails with current gcc (through 4.7 at least) if "val"
is a pointer to an incomplete type.  Unfortunately gcc isn't smart
enough to realize that "val - val" will always be a size_t type
even if it's an incomplete type pointer.

Accordingly, I've reworked the way we handle the casting.  We have
given up the ability to use cmpxchg() on 64-bit values on tilepro,
which is OK in the kernel since we should use cmpxchg64() explicitly
on such values anyway.  As a result, I can just use simple "unsigned
long" casts internally.

As I reworked it, I realized it would be cleaner to move the
architecture-specific conditionals for cmpxchg and xchg out of the
atomic.h headers and into cmpxchg, and then use the cmpxchg() and
xchg() primitives directly in atomic.h and elsewhere.  This allowed
the cmpxchg.h header to stand on its own without relying on the
implicit include of it that is performed by <asm/atomic.h>.
It also allowed collapsing the atomic_xchg/atomic_cmpxchg routines
from atomic_{32,64}.h into atomic.h.

I improved the tests that guard the allowed size of the arguments
to the routines to use a __compiletime_error() test.  (By avoiding
the use of BUILD_BUG, I could include cmpxchg.h into bitops.h as
well and use the macros there, which is otherwise impossible due
to include order dependency issues.)

The tilepro _atomic_xxx internal methods were previously set up to
take atomic_t and atomic64_t arguments, which isn't as convenient
with the new model, so I modified them to take int or u64 arguments,
which is consistent with how they used the arguments internally
anyway, so provided some nice simplification there too.

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

1 files changed, 73 insertions, 24 deletions

diff --git a/arch/tile/include/asm/cmpxchg.h b/arch/tile/include/asm/cmpxchg.h
index 1da5bfb..4001d5e 100644
--- a/arch/tile/include/asm/cmpxchg.h
+++ b/arch/tile/include/asm/cmpxchg.h
@@ -20,59 +20,108 @@
 
 #ifndef __ASSEMBLY__
 
-/* Nonexistent functions intended to cause link errors. */
-extern unsigned long __xchg_called_with_bad_pointer(void);
-extern unsigned long __cmpxchg_called_with_bad_pointer(void);
+#include <asm/barrier.h>
 
-#define xchg(ptr, x)							\
+/* Nonexistent functions intended to cause compile errors. */
+extern void __xchg_called_with_bad_pointer(void)
+	__compiletime_error("Bad argument size for xchg");
+extern void __cmpxchg_called_with_bad_pointer(void)
+	__compiletime_error("Bad argument size for cmpxchg");
+
+#ifndef __tilegx__
+
+/* Note the _atomic_xxx() routines include a final mb(). */
+int _atomic_xchg(int *ptr, int n);
+int _atomic_xchg_add(int *v, int i);
+int _atomic_xchg_add_unless(int *v, int a, int u);
+int _atomic_cmpxchg(int *ptr, int o, int n);
+u64 _atomic64_xchg(u64 *v, u64 n);
+u64 _atomic64_xchg_add(u64 *v, u64 i);
+u64 _atomic64_xchg_add_unless(u64 *v, u64 a, u64 u);
+u64 _atomic64_cmpxchg(u64 *v, u64 o, u64 n);
+
+#define xchg(ptr, n)							\
+	({								\
+		if (sizeof(*(ptr)) != 4)				\
+			__xchg_called_with_bad_pointer();		\
+		smp_mb();						\
+		(typeof(*(ptr)))_atomic_xchg((int *)(ptr), (int)(n));	\
+	})
+
+#define cmpxchg(ptr, o, n)						\
+	({								\
+		if (sizeof(*(ptr)) != 4)				\
+			__cmpxchg_called_with_bad_pointer();		\
+		smp_mb();						\
+		(typeof(*(ptr)))_atomic_cmpxchg((int *)ptr, (int)o, (int)n); \
+	})
+
+#define xchg64(ptr, n)							\
+	({								\
+		if (sizeof(*(ptr)) != 8)				\
+			__xchg_called_with_bad_pointer();		\
+		smp_mb();						\
+		(typeof(*(ptr)))_atomic64_xchg((u64 *)(ptr), (u64)(n));	\
+	})
+
+#define cmpxchg64(ptr, o, n)						\
+	({								\
+		if (sizeof(*(ptr)) != 8)				\
+			__cmpxchg_called_with_bad_pointer();		\
+		smp_mb();						\
+		(typeof(*(ptr)))_atomic64_cmpxchg((u64 *)ptr, (u64)o, (u64)n); \
+	})
+
+#else
+
+#define xchg(ptr, n)							\
 	({								\
 		typeof(*(ptr)) __x;					\
+		smp_mb();						\
 		switch (sizeof(*(ptr))) {				\
 		case 4:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic_xchg( \
-				(atomic_t *)(ptr),			\
-				(u32)(typeof((x)-(x)))(x));		\
+			__x = (typeof(__x))(unsigned long)		\
+				__insn_exch4((ptr), (u32)(unsigned long)(n)); \
 			break;						\
 		case 8:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic64_xchg( \
-				(atomic64_t *)(ptr),			\
-				(u64)(typeof((x)-(x)))(x));		\
+			__x = (typeof(__x))			\
+				__insn_exch((ptr), (unsigned long)(n));	\
 			break;						\
 		default:						\
 			__xchg_called_with_bad_pointer();		\
+			break;						\
 		}							\
+		smp_mb();						\
 		__x;							\
 	})
 
 #define cmpxchg(ptr, o, n)						\
 	({								\
 		typeof(*(ptr)) __x;					\
+		__insn_mtspr(SPR_CMPEXCH_VALUE, (unsigned long)(o));	\
+		smp_mb();						\
 		switch (sizeof(*(ptr))) {				\
 		case 4:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic_cmpxchg( \
-				(atomic_t *)(ptr),			\
-				(u32)(typeof((o)-(o)))(o),		\
-				(u32)(typeof((n)-(n)))(n));		\
+			__x = (typeof(__x))(unsigned long)		\
+				__insn_cmpexch4((ptr), (u32)(unsigned long)(n)); \
 			break;						\
 		case 8:							\
-			__x = (typeof(__x))(typeof(__x-__x))atomic64_cmpxchg( \
-				(atomic64_t *)(ptr),			\
-				(u64)(typeof((o)-(o)))(o),		\
-				(u64)(typeof((n)-(n)))(n));		\
+			__x = (typeof(__x))__insn_cmpexch((ptr), (u64)(n)); \
 			break;						\
 		default:						\
 			__cmpxchg_called_with_bad_pointer();		\
+			break;						\
 		}							\
+		smp_mb();						\
 		__x;							\
 	})
 
-#define tas(ptr) (xchg((ptr), 1))
+#define xchg64 xchg
+#define cmpxchg64 cmpxchg
 
-#define cmpxchg64(ptr, o, n)						\
-({									\
-	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
-	cmpxchg((ptr), (o), (n));					\
-})
+#endif
+
+#define tas(ptr) xchg((ptr), 1)
 
 #endif /* __ASSEMBLY__ */