Re: [PR c/68966] Restore atomic builtins usage in libstdc++-v3

[Jonathan Wakely]

On 05/04/16 19:08 +0100, Jonathan Wakely wrote:

On 05/04/16 12:01 +0100, Jonathan Wakely wrote:

Well I guess it's mine, and this is a fairly serious regression (is it
tracked in Bugzilla anywhere?) so the patch is OK for trunk.

This is now https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70554

I've committed this smaller change, which still tests atomics for
bool, but simply removes the uses of __atomic_fetch_add for bool.

During stage 1 we can revisit whether we want to do something more
invasive and stop testing for atomics that we don't actually need.

Tested ppc64le-linux, x86_64-linux, comitted to trunk.



commit 7591475c21a34bbcce6d225ea7640d35022fb8ff
Author: Jonathan Wakely <jwak...@redhat.com>
Date:   Tue Apr 5 19:16:05 2016 +0100

    Restore atomic builtins usage in libstdc++-v3
    
    	PR libstdc++/70554
    	* acinclude.m4 (GLIBCXX_ENABLE_ATOMIC_BUILTINS): Don't test
    	__atomic_fetch_add for bool.
    	* configure: Regenerate.

diff --git a/libstdc++-v3/acinclude.m4 b/libstdc++-v3/acinclude.m4
index 95df24a..b0f88cb 100644
--- a/libstdc++-v3/acinclude.m4
+++ b/libstdc++-v3/acinclude.m4
@@ -3290,7 +3290,7 @@ AC_DEFUN([GLIBCXX_ENABLE_ATOMIC_BUILTINS], [
        atomic_type c1;
        atomic_type c2;
        atomic_type c3(0);
-       __atomic_fetch_add(&c1, c2, __ATOMIC_RELAXED);
+       // N.B. __atomic_fetch_add is not supported for bool.
        __atomic_compare_exchange_n(&c1, &c2, c3, true, __ATOMIC_ACQ_REL,
 				   __ATOMIC_RELAXED);
        __atomic_test_and_set(&c1, __ATOMIC_RELAXED);
@@ -3375,7 +3375,7 @@ int main()
   atomic_type c1;
   atomic_type c2;
   atomic_type c3(0);
-  __atomic_fetch_add(&c1, c2, __ATOMIC_RELAXED);
+  // N.B. __atomic_fetch_add is not supported for bool.
   __atomic_compare_exchange_n(&c1, &c2, c3, true, __ATOMIC_ACQ_REL,
 			      __ATOMIC_RELAXED);
   __atomic_test_and_set(&c1, __ATOMIC_RELAXED);