Re: [patch 7/7] Make "inline" no longer mandatory for gcc 4.x

From: Arjan van de Ven
Date: Fri Jan 06 2006 - 14:34:50 EST


On Fri, 2006-01-06 at 12:31 -0500, Jeff Garzik wrote:
> Arjan van de Ven wrote:
> > Subject: when CONFIG_CC_OPTIMIZE_FOR_SIZE, allow gcc4 to control inlining
> > From: Ingo Molnar <mingo@xxxxxxx>
> >
> > if optimizing for size (CONFIG_CC_OPTIMIZE_FOR_SIZE), allow gcc4 compilers
> > to decide what to inline and what not - instead of the kernel forcing gcc
> > to inline all the time. This requires several places that require to be
> > inlined to be marked as such, previous patches in this series do that.
> > This is probably the most flame-worthy patch of the series.
> >
> > Signed-off-by: Ingo Molnar <mingo@xxxxxxx>
> > Signed-off-by: Arjan van de Ven <arjan@xxxxxxxxxxxxx>
>
> NAK, for what it's worth... This should be first integrated with its
> own "off switch", and then later added to optimze-for-size.

at the same time I'd not want such a config option to linger around
forever; if we do one config option per patch it gets unmanagable.
How about this?





if optimizing for size (CONFIG_CC_OPTIMIZE_FOR_SIZE), allow gcc4 compilers
to decide what to inline and what not - instead of the kernel forcing gcc
to inline all the time. This requires several places that require to be
inlined to be marked as such, previous patches in this series do that.


Signed-off-by: Arjan van de Ven <arjan@xxxxxxxxxxxxx>

----

include/linux/compiler-gcc4.h | 13 +++++++++----
1 files changed, 9 insertions(+), 4 deletions(-)

Index: linux-2.6.15/include/linux/compiler-gcc4.h
===================================================================
--- linux-2.6.15.orig/include/linux/compiler-gcc4.h
+++ linux-2.6.15/include/linux/compiler-gcc4.h
@@ -3,9 +3,11 @@
/* These definitions are for GCC v4.x. */
#include <linux/compiler-gcc.h>

-#define inline inline __attribute__((always_inline))
-#define __inline__ __inline__ __attribute__((always_inline))
-#define __inline __inline __attribute__((always_inline))
+#ifdef CONFIG_FORCED_INLINING
+# define inline inline __attribute__((always_inline))
+# define __inline__ __inline__ __attribute__((always_inline))
+# define __inline __inline __attribute__((always_inline))
+#endif
#define __always_inline inline __attribute__((always_inline))
#define __deprecated __attribute__((deprecated))
#define __attribute_used__ __attribute__((__used__))
Index: linux-2.6.15/lib/Kconfig.debug
===================================================================
--- linux-2.6.15.orig/lib/Kconfig.debug
+++ linux-2.6.15/lib/Kconfig.debug
@@ -186,6 +186,20 @@ config FRAME_POINTER
some architectures or if you use external debuggers.
If you don't debug the kernel, you can say N.

+config FORCED_INLINING
+ bool "Force gcc to inline functions marked 'inline'"
+ depends on DEBUG_KERNEL
+ default y
+ help
+ This option determines if the kernel forces gcc to inline the functions
+ developers have marked 'inline'. Doing so takes away freedom from gcc to
+ do what it thinks is best, which is desirable for the gcc 3.x series of
+ compilers. The gcc 4.x series have a rewritten inlining algorithm and
+ disabling this option will generate a smaller kernel there. Hopefully
+ this algorithm is so good that allowing gcc4 to make the decision can
+ become the default in the future, until then this option is there to
+ test gcc for this.
+
config RCU_TORTURE_TEST
tristate "torture tests for RCU"
depends on DEBUG_KERNEL
Index: linux-2.6.15/Documentation/feature-removal-schedule.txt
===================================================================
--- linux-2.6.15.orig/Documentation/feature-removal-schedule.txt
+++ linux-2.6.15/Documentation/feature-removal-schedule.txt
@@ -143,6 +143,15 @@ Who: Christoph Hellwig <hch@xxxxxx>

---------------------------

+What: CONFIG_FORCED_INLINING
+When: June 2006
+Why: Config option is there to see if gcc is good enough. (in january
+ 2006). If it is, the behavior should just be the default. If it's not,
+ the option should just go away entirely.
+Who: Arjan van de Ven <arjan@xxxxxxxxxxxxx>
+
+---------------------------
+
What: START_ARRAY ioctl for md
When: July 2006
Files: drivers/md/md.c


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/