[2.6 patch] chsc headers userspace cleanup

[2.6 patch] chsc headers userspace cleanup

[Adrian Bunk]

Kernel headers shouldn't expose functions to userspace.

Signed-off-by: Adrian Bunk <bunk@kernel.org>

---

BTW: For compile-testing my patches it would be nice if s390 had
     more defconfigs (e.g. also a non-64bit one).

 include/asm-s390/Kbuild  |    2 +-
 include/asm-s390/chpid.h |    6 +++---
 include/asm-s390/schid.h |    3 +++
 3 files changed, 7 insertions(+), 4 deletions(-)

0ee038c850fe402d0a8a548b438ae045e8e06f50 
diff --git a/include/asm-s390/Kbuild b/include/asm-s390/Kbuild
index 09f3125..bb5e9ed 100644
--- a/include/asm-s390/Kbuild
+++ b/include/asm-s390/Kbuild
@@ -8,9 +8,9 @@ header-y += ucontext.h
 header-y += vtoc.h
 header-y += zcrypt.h
 header-y += kvm.h
-header-y += schid.h
 header-y += chsc.h
 
 unifdef-y += cmb.h
 unifdef-y += debug.h
 unifdef-y += chpid.h
+unifdef-y += schid.h
diff --git a/include/asm-s390/chpid.h b/include/asm-s390/chpid.h
index 606844d..dfe3c7f 100644
--- a/include/asm-s390/chpid.h
+++ b/include/asm-s390/chpid.h
@@ -20,6 +20,9 @@ struct chp_id {
 	u8 id;
 } __attribute__((packed));
 
+#ifdef __KERNEL__
+#include <asm/cio.h>
+
 static inline void chp_id_init(struct chp_id *chpid)
 {
 	memset(chpid, 0, sizeof(struct chp_id));
@@ -40,9 +43,6 @@ static inline void chp_id_next(struct chp_id *chpid)
 	}
 }
 
-#ifdef __KERNEL__
-#include <asm/cio.h>
-
 static inline int chp_id_is_valid(struct chp_id *chpid)
 {
 	return (chpid->cssid <= __MAX_CSSID);
diff --git a/include/asm-s390/schid.h b/include/asm-s390/schid.h
index 5017ffa..7707eb9 100644
--- a/include/asm-s390/schid.h
+++ b/include/asm-s390/schid.h
@@ -10,6 +10,7 @@ struct subchannel_id {
 	__u32 sch_no : 16;
 } __attribute__ ((packed, aligned(4)));
 
+#ifdef  __KERNEL__
 
 /* Helper function for sane state of pre-allocated subchannel_id. */
 static inline void
@@ -25,4 +26,6 @@ schid_equal(struct subchannel_id *schid1, struct subchannel_id *schid2)
 	return !memcmp(schid1, schid2, sizeof(struct subchannel_id));
 }
 
+#endif /* __KERNEL__ */
+
 #endif /* ASM_SCHID_H */

Re: [2.6 patch] chsc headers userspace cleanup

[Cornelia Huck]

On Tue, 15 Jul 2008 15:58:18 +0300,
Adrian Bunk <bunk@kernel.org> wrote:

> Kernel headers shouldn't expose functions to userspace.

Could you please elaborate? Especially as I see e.g.
include/linux/virtio_ring.h exporting functions outside #ifdef
__KERNEL__ as well...

Re: [2.6 patch] chsc headers userspace cleanup

[Adrian Bunk]

On Tue, Jul 15, 2008 at 07:17:48PM +0200, Cornelia Huck wrote:
> On Tue, 15 Jul 2008 15:58:18 +0300,
> Adrian Bunk <bunk@kernel.org> wrote:
> 
> > Kernel headers shouldn't expose functions to userspace.
> 
> Could you please elaborate? Especially as I see e.g.
> include/linux/virtio_ring.h exporting functions outside #ifdef
> __KERNEL__ as well...

Our headers are in a bad shape...

The userspace headers should contain everything that is part of the ABI 
between the kernel and userspace.

Nothing more.

The kernel is heavily changing with each release while the userspace ABI 
is cast in stone, and leaking more stuff to userspace than required only 
increases the chances of some userspace programmer using it and some 
kernel programmer then changing it.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: [2.6 patch] chsc headers userspace cleanup

[Heiko Carstens]

On Tue, Jul 15, 2008 at 03:58:18PM +0300, Adrian Bunk wrote:
> Kernel headers shouldn't expose functions to userspace.
> 
> Signed-off-by: Adrian Bunk <bunk@kernel.org>

Makes sense. Applied -- Thanks.

> BTW: For compile-testing my patches it would be nice if s390 had
>      more defconfigs (e.g. also a non-64bit one).

Yes, will consider that ;)