On Wed, Nov 11, 2020 at 12:58:26AM +0000, Al Viro wrote:

On Wed, Nov 11, 2020 at 01:44:38AM +0100, Ansuel Smith wrote:

quoted

Fix a compilation error as PF_KTHREAD is defined in linux/sched.h and
this is missing.

Fixes: df325e05a682 ("arm64: Validate tagged addresses in access_ok()
called from kernel threads")
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 arch/arm64/include/asm/uaccess.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 991dd5f031e4..51a4f63f464a 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h

@@ -7,6 +7,8 @@
 #ifndef __ASM_UACCESS_H
 #define __ASM_UACCESS_H
 
+#include <linux/sched.h>
+
 #include <asm/alternative.h>
 #include <asm/kernel-pgtable.h>
 #include <asm/sysreg.h>

NAK.  The real bug is in arch/arm64/include/asm/asm-prototypes.h -
it has no business pulling asm/uaccess.h

Just include linux/uaccess.h instead.

BTW,
$ grep -n uaccess.h `find -name asm-prototypes.h`
./arch/alpha/include/asm/asm-prototypes.h:7:#include <linux/uaccess.h>
./arch/arm64/include/asm/asm-prototypes.h:18:#include <asm/uaccess.h>
./arch/ia64/include/asm/asm-prototypes.h:12:#include <linux/uaccess.h>
./arch/mips/include/asm/asm-prototypes.h:6:#include <linux/uaccess.h>
./arch/powerpc/include/asm/asm-prototypes.h:14:#include <linux/uaccess.h>
./arch/sparc/include/asm/asm-prototypes.h:9:#include <linux/uaccess.h>
./arch/x86/include/asm/asm-prototypes.h:3:#include <linux/uaccess.h>

Spot the irregularity...

While we are at it,
$ git grep -n -w '#.*include.*asm/uaccess.h'
arch/arm64/include/asm/asm-prototypes.h:18:#include <asm/uaccess.h>
arch/nds32/math-emu/fpuemu.c:5:#include <asm/uaccess.h>
arch/powerpc/kvm/book3s_xive_native.c:15:#include <asm/uaccess.h>
arch/powerpc/mm/book3s64/radix_pgtable.c:30:#include <asm/uaccess.h>
drivers/s390/net/ctcm_mpc.c:50:#include <linux/uaccess.h>       /* instead of <asm/uaccess.h> ok ? */
include/linux/uaccess.h:11:#include <asm/uaccess.h>

The last one is the only such include that should exist; drivers/s390 one
is obviously a false positive.  And IMO the right thing to do is to
replace the remaining arch/* instances with includes of linux/uaccess.h.

All of those are asking for trouble; any change moving e.g. a common
variant of some primitive into linux/uaccess.h might end up breaking
those.

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel