[PATCH] kbuild: Add option to turn incompatible pointer check into error

From: Daniel Wagner
Date: Tue Mar 08 2016 - 03:29:57 EST

From: Daniel Wagner <daniel.wagner@xxxxxxxxxxxx>

With the introduction of the simple wait API we have two very
similar APIs in the kernel. For example wake_up() and swake_up()
is only one character away. Although the compiler will warn
happily the wrong usage it keeps on going an even links the kernel.
Thomas and Peter would rather like to see early missuses reported
as error early on.

In a first attempt we tried to wrap all swait and wait calls
into a macro which has an compile time type assertion. The result
was pretty ugly and wasn't able to catch all wrong usages.
woken_wake_function(), autoremove_wake_function() and wake_bit_function()
are assigned as function pointers. Wrapping them with a macro around is
not possible. Prefixing them with '_' was also not a real option
because there some users in the kernel which do use them as well.
All in all this attempt looked to intrusive and too ugly.

An alternative is to turn the pointer type check into an error which
catches wrong type uses. Obviously not only the swait/wait ones. That
isn't a bad thing either.

Signed-off-by: Daniel Wagner <daniel.wagner@xxxxxxxxxxxx>
Acked-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>


This patch was part of the swait patchset [1]. Ingo reverted it later

"So adding -Werror=incompatible-pointer-types wasn't a bad idea, but it
should really not be done in the scheduler tree: it exposes us to a
number of pre-existing warnings (most of them harmless), now upgraded
to build failures...

This should be done via the kbuild tree."

I tried to build a bunch of different architectures and configurations
to find any hold offs. For those architecture I had a working
toolchain it looks good. All but one are fixed or on the way to

Here is the list of archs and config I tried out. First run was with the
config listed, followed by a allyesconfig and allmodconfig build.

# name ARCH CROSS_COMPILE config
declare -a config=(
"alpha alpha alpha-linux-gnu- defconfig"
"arm32 arm arm-linux-gnu- versatile_defconfig"
"arm64 arm64 aarch64-linux-gnu- defconfig"
"cris10 cris cris-linux-gnu- etrax-100lx_defconfig"
"frv frv frv-linux-gnu- defconfig"
"ia64 ia64 ia64-linux-gnu- generic_defconfig"
"m68k m68k m68k-linux-gnu- amiga_defconfig"
"mips32 mips mips64-linux-gnu- ar7_defconfig"
"mips64 mips mips64-linux-gnu- bigsur_defconfig"
"ppc64 powerpc powerpc64-linux-gnu- pseries_defconfig"
"s390 s390 s390x-linux-gnu- defconfig"
"sparc32 sparc sparc64-linux-gnu- sparc32_defconfig"
"sparc64 sparc sparc64-linux-gnu- sparc64_defconfig"
"um_x86_64 um gcc defconfig"
"x86_64 x86_64 gcc defconfig"

As said above for the missing arch I didn't get hold on a working
toolchain with gcc >= 5.2.

On alpha I found [2]. So far no feedback on the patch. Propoably missing
the right CCs.

x86 is breaking for allyesconfig in gma500. The fix is linus-next [3].


[1] http://thread.gmane.org/gmane.linux.kernel/2156371
[2] http://thread.gmane.org/gmane.linux.ports.alpha/3494
[3] https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/diff/drivers/gpu/drm/gma500/mdfld_dsi_output.c?id=075f82f5db32848f5cbcdf5d7d6668b7c087b49f

Makefile | 3 +++
1 file changed, 3 insertions(+)

diff --git a/Makefile b/Makefile
index 2d519d2..a0ff507 100644
--- a/Makefile
+++ b/Makefile
@@ -773,6 +773,9 @@ KBUILD_CFLAGS += $(call cc-option,-Werror=strict-prototypes)
# Prohibit date/time macros, which would make the build non-deterministic
KBUILD_CFLAGS += $(call cc-option,-Werror=date-time)

+# enforce correct pointer usage
+KBUILD_CFLAGS += $(call cc-option,-Werror=incompatible-pointer-types)
# use the deterministic mode of AR if available
KBUILD_ARFLAGS := $(call ar-option,D)