Messages in this thread Patch in this message | | | From | "Paul E. McKenney" <> | Subject | [PATCH tip/core/rcu 9/9] doc: Update control-dependencies section of memory-barriers.txt | Date | Wed, 12 Apr 2017 09:40:22 -0700 |
| |
From: pierre Kuo <vichy.kuo@gmail.com>
In the following example, if MAX is defined to be 1, then the compiler knows (Q % MAX) is equal to zero. The compiler can therefore throw away the "then" branch (and the "if"), retaining only the "else" branch.
q = READ_ONCE(a); if (q % MAX) { WRITE_ONCE(b, 1); do_something(); } else { WRITE_ONCE(b, 2); do_something_else(); }
It is therefore necessary to modify the example like this:
q = READ_ONCE(a); - WRITE_ONCE(b, 1); + WRITE_ONCE(b, 2); do_something_else();
Signed-off-by: pierre Kuo <vichy.kuo@gmail.com> Acked-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> --- Documentation/memory-barriers.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Documentation/memory-barriers.txt b/Documentation/memory-barriers.txt index d2b0a8d81258..08329cb857ed 100644 --- a/Documentation/memory-barriers.txt +++ b/Documentation/memory-barriers.txt @@ -768,7 +768,7 @@ equal to zero, in which case the compiler is within its rights to transform the above code into the following: q = READ_ONCE(a); - WRITE_ONCE(b, 1); + WRITE_ONCE(b, 2); do_something_else(); Given this transformation, the CPU is not required to respect the ordering -- 2.5.2
| |