From 0a89f3bc6e9b1e911e9efd8132377fe5c6838c66 Mon Sep 17 00:00:00 2001
From: Noah Misch <noah@leadboat.com>
Date: Wed, 29 Jul 2015 22:49:48 -0400
Subject: [PATCH] Blacklist xlc 32-bit inlining.
Per a suggestion from Tom Lane. Back-patch to 9.0 (all supported
versions). While only 9.4 and up have code known to elicit this
compiler bug, we were disabling inlining by accident until commit
43d89a23d59c487bc9258fad7a6187864cb8c0c0.
---
config/test_quiet_include.h | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/config/test_quiet_include.h b/config/test_quiet_include.h
index eb0515ede78..f059e1c2913 100644
--- a/config/test_quiet_include.h
+++ b/config/test_quiet_include.h
@@ -3,3 +3,12 @@
* of the PGAC_C_INLINE macro in config/c-compiler.m4.
*/
static inline int fun () { return 0; }
+
+/*
+ * "IBM XL C/C++ for AIX, V12.1" miscompiles, for 32-bit, some inline
+ * expansions of ginCompareItemPointers() "long long" arithmetic. To take
+ * advantage of inlining, build a 64-bit PostgreSQL.
+ */
+#if defined(__ILP32__) && defined(__IBMC__)
+#error "known inlining bug"
+#endif
--
GitLab