fix arm atomic store and generate simpler/less-bloated/faster code

atomic store was lacking a barrier, which was fine for legacy arm with
no real smp and kernel-emulated cas, but unsuitable for more modern
systems. the kernel provides another "kuser" function, at 0xffff0fa0,
which could be used for the barrier, but using that would drop support
for kernels 2.6.12 through 2.6.14 unless an extra conditional were
added to check for barrier availability. just using the barrier in the
kernel cas is easier, and, based on my reading of the assembly code in
the kernel, does not appear to be significantly slower.

at the same time, other atomic operations are adapted to call the
kernel cas function directly rather than using a_cas; due to small
differences in their interface contracts, this makes the generated
code much simpler.

diff --git a/arch/arm/atomic.h b/arch/arm/atomic.h
index 94efe5a0675c9f69f96dc390b4e8541c11f548d0..734d2871787a85bc5ebc79f0dcceda0f42edeae3 100644 (file)
--- a/arch/arm/atomic.h
+++ b/arch/arm/atomic.h
@@ -22,11 +22,13 @@ static inline int a_ctz_64(uint64_t x)
        return a_ctz_l(y);
 }
 
+#define __k_cas ((int (*)(int, int, volatile int *))0xffff0fc0)
+
 static inline int a_cas(volatile int *p, int t, int s)
 {
        int old;
        for (;;) {
-               if (!((int (*)(int, int, volatile int *))0xffff0fc0)(t, s, p))
+               if (!__k_cas(t, s, p))
                        return t;
                if ((old=*p) != t)
                        return old;
@@ -47,7 +49,7 @@ static inline int a_swap(volatile int *x, int v)
 {
        int old;
        do old = *x;
-       while (a_cas(x, old, v) != old);
+       while (__k_cas(old, v, x));
        return old;
 }
 
@@ -55,7 +57,7 @@ static inline int a_fetch_add(volatile int *x, int v)
 {
        int old;
        do old = *x;
-       while (a_cas(x, old, old+v) != old);
+       while (__k_cas(old, old+v, x));
        return old;
 }
 
@@ -71,7 +73,7 @@ static inline void a_dec(volatile int *x)
 
 static inline void a_store(volatile int *p, int x)
 {
-       *p=x;
+       while (__k_cas(*p, x, p));
 }
 
 static inline void a_spin()
@@ -87,14 +89,14 @@ static inline void a_and(volatile int *p, int v)
 {
        int old;
        do old = *p;
-       while (a_cas(p, old, old&v) != old);
+       while (__k_cas(old, old&v, p));
 }
 
 static inline void a_or(volatile int *p, int v)
 {
        int old;
        do old = *p;
-       while (a_cas(p, old, old|v) != old);
+       while (__k_cas(old, old|v, p));
 }
 
 static inline void a_or_l(volatile void *p, long v)