diff mbox

Do not build with -Winline

Message ID 553A08BE.4040700@redhat.com
State New
Headers show

Commit Message

Florian Weimer April 24, 2015, 9:11 a.m. UTC
Per the previous discussion.

Comments

Paul Eggert April 24, 2015, 3:01 p.m. UTC | #1
Florian Weimer wrote:
> Per the previous discussion.

This patch looks good to me.  In other GNU projects we've found that -Winline is 
more trouble than it's worth.  Like -Wunsafe-loop-optimizations, all that 
-Winline tells you is that generated code is not quite as efficient as it might 
be under ideal conditions, which is not a big deal, certainly not worth breaking 
a build over.
Florian Weimer April 24, 2015, 3:08 p.m. UTC | #2
On 04/24/2015 05:01 PM, Paul Eggert wrote:
> Florian Weimer wrote:
>> Per the previous discussion.
> 
> This patch looks good to me.  In other GNU projects we've found that
> -Winline is more trouble than it's worth.  Like
> -Wunsafe-loop-optimizations, all that -Winline tells you is that
> generated code is not quite as efficient as it might be under ideal
> conditions, which is not a big deal, certainly not worth breaking a
> build over.

Thanks, committed.

It's even worse, -Winline warns for cases where the code is, in fact,
optimal (not inlining a non-trivial function called multiple times from
unlikely code is often beneficial).
diff mbox

Patch

From 69372f19d39ac6cf3570cbf62ae4610a26c15e41 Mon Sep 17 00:00:00 2001
From: Florian Weimer <fweimer@redhat.com>
Date: Fri, 24 Apr 2015 11:10:24 +0200
Subject: [PATCH] Do not build with -Winline

-Winline causes architecture- and optimization-dependent build failures
due to -Werror.  -Winline warns about inlining decisions based on
branch hints, in effect preventing the use of inline functions in
header files (because they might be called on unlikely branches, leading
to a decision not to inline).

The option was apparently added to the glibc build at a time when GCC
did not support the always_inline attribute.  With current GCC versions,
inlining failure for functions declared always_inline will receive a
warning under -Wattributes, which is enabled by default, so -Winline
appears unnecessary.
---
 ChangeLog  | 4 ++++
 Makeconfig | 4 ++--
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 2c01eac..ace5fb3 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,7 @@ 
+2015-04-24  Florian Weimer  <fweimer@redhat.com>
+
+	* Makeconfig (+gccwarn): Remove -Winline.
+
 2015-04-22  Wilco Dijkstra  <wdijkstr@arm.com>
 
 	* math/k_casinh.c (__kernel_casinh): Use __copysign.
diff --git a/Makeconfig b/Makeconfig
index dc7f208..77752c0 100644
--- a/Makeconfig
+++ b/Makeconfig
@@ -739,9 +739,9 @@  endif
 
 # Extra flags to pass to GCC.
 ifeq ($(all-warnings),yes)
-+gccwarn := -Wall -Wwrite-strings -Winline -Wcast-qual -Wbad-function-cast -Wmissing-noreturn -Wmissing-prototypes -Wmissing-declarations -Wcomment -Wcomments -Wtrigraphs -Wsign-compare -Wfloat-equal -Wmultichar
++gccwarn := -Wall -Wwrite-strings -Wcast-qual -Wbad-function-cast -Wmissing-noreturn -Wmissing-prototypes -Wmissing-declarations -Wcomment -Wcomments -Wtrigraphs -Wsign-compare -Wfloat-equal -Wmultichar
 else
-+gccwarn := -Wall -Wwrite-strings -Winline
++gccwarn := -Wall -Wwrite-strings
 endif
 +gccwarn += -Wundef
 ifeq ($(enable-werror),yes)
-- 
2.1.0