LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [RFC] breakage in 4223cc34365e4 (h8300: uaccess.h update)
@ 2008-03-02 23:23 Al Viro
  2008-03-10 20:57 ` Yoshinori Sato
  0 siblings, 1 reply; 2+ messages in thread
From: Al Viro @ 2008-03-02 23:23 UTC (permalink / raw)
  To: Yoshinori Sato; +Cc: linux-kernel, Linus Torvalds

	After that commit in asm-h8300/uaccess.h we have 

#define get_user(x, ptr)                                        \
({                                                              \
    int __gu_err = 0;                                           \
    uint32_t __gu_val = 0;                              \
    ^^^^^^^^^^^^^^^^^
    switch (sizeof(*(ptr))) {                                   \
    case 1:                                                     \
    case 2:                                                     \
    case 4:                                                     \
        __gu_val = *(ptr);                                      \
        break;                                                  \
    case 8:                                                     \
        memcpy(&__gu_val, ptr, sizeof (*(ptr)));                \
                               ^^^^^^^^^^^^^^^^

which, of course, is FUBAR whenever we actually hit that case - memcpy of
8 bytes into uint32_t is obviously wrong.  Why don't we simply do

#define get_user(x, ptr)                                        \
({                                                              \
    int __gu_err = 0;                                           \
    typeof(*(ptr)) __gu_val = *ptr;                             \
    switch (sizeof(*(ptr))) {                                   \
    case 1:                                                     \
    case 2:                                                     \
    case 4:                                                     \
    case 8:                                                     \
        break;                                                  \
    default:                                                    \
        __gu_err = __get_user_bad();                            \
        break;                                                  \
    }                                                           \
    (x) = __gu_val;                                             \
    __gu_err;                                                   \
})

and be done with that, anyway?

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [RFC] breakage in 4223cc34365e4 (h8300: uaccess.h update)
  2008-03-02 23:23 [RFC] breakage in 4223cc34365e4 (h8300: uaccess.h update) Al Viro
@ 2008-03-10 20:57 ` Yoshinori Sato
  0 siblings, 0 replies; 2+ messages in thread
From: Yoshinori Sato @ 2008-03-10 20:57 UTC (permalink / raw)
  To: Al Viro; +Cc: lkml, Linus Torvalds, Andrew Morton

At Sun, 2 Mar 2008 23:23:14 +0000,
Al Viro wrote:
> 
> 	After that commit in asm-h8300/uaccess.h we have 
> 
> #define get_user(x, ptr)                                        \
> ({                                                              \
>     int __gu_err = 0;                                           \
>     uint32_t __gu_val = 0;                              \
>     ^^^^^^^^^^^^^^^^^
>     switch (sizeof(*(ptr))) {                                   \
>     case 1:                                                     \
>     case 2:                                                     \
>     case 4:                                                     \
>         __gu_val = *(ptr);                                      \
>         break;                                                  \
>     case 8:                                                     \
>         memcpy(&__gu_val, ptr, sizeof (*(ptr)));                \
>                                ^^^^^^^^^^^^^^^^
> 
> which, of course, is FUBAR whenever we actually hit that case - memcpy of
> 8 bytes into uint32_t is obviously wrong.  Why don't we simply do

Oops.

> 
> #define get_user(x, ptr)                                        \
> ({                                                              \
>     int __gu_err = 0;                                           \
>     typeof(*(ptr)) __gu_val = *ptr;                             \
>     switch (sizeof(*(ptr))) {                                   \
>     case 1:                                                     \
>     case 2:                                                     \
>     case 4:                                                     \
>     case 8:                                                     \
>         break;                                                  \
>     default:                                                    \
>         __gu_err = __get_user_bad();                            \
>         break;                                                  \
>     }                                                           \
>     (x) = __gu_val;                                             \
>     __gu_err;                                                   \
> })
> 
> and be done with that, anyway?

Good! no problem.
Thank you.

Signed-off-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>

 include/asm-h8300/uaccess.h |   11 ++++-------
 1 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/include/asm-h8300/uaccess.h b/include/asm-h8300/uaccess.h
index a22350e..356068c 100644
--- a/include/asm-h8300/uaccess.h
+++ b/include/asm-h8300/uaccess.h
@@ -91,22 +91,19 @@ extern int __put_user_bad(void);
 #define get_user(x, ptr)					\
 ({								\
     int __gu_err = 0;						\
-    uint32_t __gu_val = 0;				\
+    typeof(*(ptr)) __gu_val = *ptr;				\
     switch (sizeof(*(ptr))) {					\
     case 1:							\
     case 2:							\
     case 4:							\
-	__gu_val = *(ptr);	                		\
-	break;							\
-    case 8:							\
-	memcpy(&__gu_val, ptr, sizeof (*(ptr))); 		\
+    case 8: 							\
 	break;							\
     default:							\
-	__gu_val = 0;						\
 	__gu_err = __get_user_bad();				\
+	__gu_val = 0;						\
 	break;							\
     }								\
-    (x) = (typeof(*(ptr)))__gu_val;				\
+    (x) = __gu_val;						\
     __gu_err;							\
 })
 #define __get_user(x, ptr) get_user(x, ptr)
-- 
1.5.4.3

-- 
Yoshinori Sato
<ysato@users.sourceforge.jp>

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2008-03-10 20:57 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2008-03-02 23:23 [RFC] breakage in 4223cc34365e4 (h8300: uaccess.h update) Al Viro
2008-03-10 20:57 ` Yoshinori Sato

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).