ia64: atomic.h: fix atomic_exchange_and_add 64bit handling

Way back in 2005 the atomic_exchange_and_add function was cleaned up to
avoid the explicit size checking and instead let gcc handle things itself.
Unfortunately that change ended up leaving beyond a cast to int, even when
the incoming value was a long.  This has flown under the radar for a long
time due to the function not being heavily used in the tree (especially as
a full 64bit field), but a recent change to semaphores made some nptl tests
fail reliably.  This is due to the code packing two 32bit values into one
64bit variable (where the high 32bits contained the number of waiters), and
then the whole variable being atomically updated between threads.  On ia64,
that meant we never atomically updated the count, so sometimes the sem_post
would not wake up the waiters.

1 files changed, 1 insertions, 3 deletions

diff --git a/sysdeps/ia64/bits/atomic.h b/sysdeps/ia64/bits/atomic.h
index 0e9dfc7..4c2b540 100644
--- a/sysdeps/ia64/bits/atomic.h
+++ b/sysdeps/ia64/bits/atomic.h
@@ -82,9 +82,7 @@ typedef uintmax_t uatomic_max_t;
   (__sync_synchronize (), __sync_lock_test_and_set (mem, value))
 
 #define atomic_exchange_and_add(mem, value) \
-  ({ __typeof (*mem) __result;						      \
-     __result = __sync_fetch_and_add ((mem), (int) (value));		      \
-     __result; })
+  __sync_fetch_and_add ((mem), (value))
 
 #define atomic_decrement_if_positive(mem) \
   ({ __typeof (*mem) __oldval, __val;					      \