Re: const from include/asm-i386/byteorder.h

[Nikita Danilov <Nikita@Namesys.COM>]

William Lee Irwin III writes:
 > On Mon, Jun 02, 2003 at 03:51:39PM +0400, Nikita Danilov wrote:
 > > Gcc info page:
 > > `const'
 > >      Many functions do not examine any values except their arguments,
 > >      and have no effects except the return value.  Basically this is
 > >      just slightly more strict class than the `pure' attribute above,
 > >      since function is not allowed to read global memory.
 > > So, it seems byte swapping functions should be __attribute__((const))
 > > Here is a patch:
 > 
 > I'm very skeptical, in particular, about these hunks:
 > 
 > 
 > On Mon, Jun 02, 2003 at 03:51:39PM +0400, Nikita Danilov wrote:
 > >  #include <linux/thread_info.h>
 > > -static inline struct task_struct *get_current(void) __attribute__ (( __const__ ));
 > > +static inline struct task_struct *get_current(void) __attribute_const;
 > >  static inline struct task_struct *get_current(void)
 > >  {
 > > ===== include/asm-arm/thread_info.h 1.6 vs edited =====
 > > --- 1.6/include/asm-arm/thread_info.h	Sat Dec 28 19:26:45 2002
 > > +++ edited/include/asm-arm/thread_info.h	Mon Jun  2 14:44:24 2003
 > > @@ -74,7 +74,7 @@
 > >  /*
 > >   * how to get the thread information struct from C
 > >   */
 > > -static inline struct thread_info *current_thread_info(void) __attribute__ (( __const__ ));
 > > +static inline struct thread_info *current_thread_info(void) __attribute_const;
 > >  
 > >  static inline struct thread_info *current_thread_info(void)
 > >  {
 > 
 > Someone needs to doublecheck whether this actually works. Last I heard,

I don't quite understand. __attribute_const is defined as

#define __attribute_const __attribute__ ((__const__))

so, after preprocessing prototypes of get_current() and
current_thread_info() will be the same as before patch, modulo spacing.

 > it did not, but that could have changed since. It vaguely appears some
 > assumption about it working was made recently since __const__ was there.
 > 

I am currently running ./fsstress -p 111 on patched kernel on 2*XEON
2.20GHz with hyper threading.

 > 
 > -- wli

Nikita.