]> pilppa.com Git - linux-2.6-omap-h63xx.git/commitdiff
Revert "[PATCH] x86: Drop cc-options call for all options supported in gcc 3.2+"
authorLinus Torvalds <torvalds@woody.linux-foundation.org>
Fri, 18 May 2007 03:18:11 +0000 (20:18 -0700)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Fri, 18 May 2007 03:18:11 +0000 (20:18 -0700)
This reverts commit c8fdd247255a3a027cd9f66dcf93e6847d1d2f85.

It turns out the kernel was correct, and the gcc complaint was a gcc
bug.  The preferred stack boundary is expressed not in bytes, but in the
the log2() of the preferred boundary, so "-mpreferred-stack-boundary=2"
is in fact exactly what we want, but a gcc that is compiled for x86-64
will consider it an error (because the 64-bit calling sequence says that
the stack should be 16-byte aligned) even if we are then using "-m32" to
generate 32-bit code.

Noted-by: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz>
Cc: Jan Hubicka <jh@suse.cz>
Acked-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Makefile
arch/i386/Makefile

index e6990e2cdafcf870e3ce7f844b4f08750d9735c5..948fa09442f18b7dba70b2c1f91c0b62c4e8e58b 100644 (file)
--- a/Makefile
+++ b/Makefile
@@ -491,7 +491,7 @@ endif
 include $(srctree)/arch/$(ARCH)/Makefile
 
 ifdef CONFIG_FRAME_POINTER
-CFLAGS         += -fno-omit-frame-pointer -fno-optimize-sibling-calls
+CFLAGS         += -fno-omit-frame-pointer $(call cc-option,-fno-optimize-sibling-calls,)
 else
 CFLAGS         += -fomit-frame-pointer
 endif
index 6dc5e5d90fec10d694b030a14ce837b67d2684d2..bd28f9f9b4b7bfa020a2bb364561abb14db1a03e 100644 (file)
@@ -34,7 +34,7 @@ CHECKFLAGS    += -D__i386__
 CFLAGS += -pipe -msoft-float -mregparm=3 -freg-struct-return
 
 # prevent gcc from keeping the stack 16 byte aligned
-CFLAGS += -mpreferred-stack-boundary=4
+CFLAGS += $(call cc-option,-mpreferred-stack-boundary=2)
 
 # CPU-specific tuning. Anything which can be shared with UML should go here.
 include $(srctree)/arch/i386/Makefile.cpu