[patch 2.6.13-rc3] i386: clean up user_mode macros

[patch 2.6.13-rc3] i386: clean up user_mode macros

[Chuck Ebbert]

Recent patches from the Xen group changed the X86 user_mode macros.

This patch does the following:

        1. Makes the new user_mode() return 0 or 1 (same as x86_64)

        2. Removes conditional jump from user_mode_vm()
           (it's called every timer tick on each CPU on SMP)

I've been running this patch for a while now.  Please apply.

Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>

Index: 2.6.13-rc3a/include/asm-i386/ptrace.h
===================================================================
--- 2.6.13-rc3a.orig/include/asm-i386/ptrace.h	2005-07-13 16:20:26.000000000 -0400
+++ 2.6.13-rc3a/include/asm-i386/ptrace.h	2005-07-14 02:47:51.000000000 -0400
@@ -57,8 +57,8 @@
 #ifdef __KERNEL__
 struct task_struct;
 extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code);
-#define user_mode(regs)		(3 & (regs)->xcs)
-#define user_mode_vm(regs)	((VM_MASK & (regs)->eflags) || user_mode(regs))
+#define user_mode(regs)		(!!(3 & (regs)->xcs))
+#define user_mode_vm(regs)	(!!((VM_MASK & (regs)->eflags) | (3 & (regs)->xcs)))
 #define instruction_pointer(regs) ((regs)->eip)
 #if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
 extern unsigned long profile_pc(struct pt_regs *regs);
__
Chuck

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Linus Torvalds]

On Mon, 25 Jul 2005, Chuck Ebbert wrote:
> 
> Recent patches from the Xen group changed the X86 user_mode macros.
> 
> This patch does the following:
> 
>         1. Makes the new user_mode() return 0 or 1 (same as x86_64)

I _really_ prefer

	x != 0

over 

	!!x

since double negation is not only a bad habit in natural languages, it's a 
bad habit in computer languages too, for exactly the same reason. It's 
confusing.

Ask a hundred random C programmers what "!!x" means, versus what "x != 0"
means, and time their replies.

I will bet you $5 USD that even if they all give the right answer (and I
suspect you'll get a few wrogn answers in there too for the !! case),
they'll take a _lot_ longer answering the "!!x" version than they will the 
"x != 0" question.

And guess what? That means that the "!!x" version is worse. It means that
people don't "see" what it means - they have to think about it. And you
shouldn't have to think about something like that, you should write it in 
the obvious way in the first place.

		Linus

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Steven Rostedt]

On Mon, 2005-07-25 at 16:13 -0700, Linus Torvalds wrote:

> I _really_ prefer
> 
> 	x != 0
> 
> over 
> 
> 	!!x

Good to hear.  This means that you should have no problem accepting my
previous patch for signal.c that changed the x ^ y to x != y.  And I
would also assume that you prefer x *= 2 over x <<= 1 (also since the
first person to show this example used x <<= 2. Right Lee? :-)

-- Steve

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Miles Bader]

Linus Torvalds <torvalds@osdl.org> writes:
> Ask a hundred random C programmers what "!!x" means, versus what "x != 0"
> means, and time their replies.

I've always thought of "!!" as the "canonicalize boolean" operator...

Vaguely ugly, but I think it's actually _more clear_ than != 0 in
contexts where the value is already "boolean" in the C zero or not-zero
sense, but you want a "proper" boolean (0 or 1) for some reason.

-miles
-- 
Somebody has to do something, and it's just incredibly pathetic that it
has to be us.  -- Jerry Garcia

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Andrew Morton]

Miles Bader <miles@lsi.nec.co.jp> wrote:
>
> Linus Torvalds <torvalds@osdl.org> writes:
>  > Ask a hundred random C programmers what "!!x" means, versus what "x != 0"
>  > means, and time their replies.
> 
>  I've always thought of "!!" as the "canonicalize boolean" operator...

Me too.  Once you get used to it, it's just the "convert non-zero to 1"
operator.

But whatever.

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[randy_dunlap]

On Mon, 25 Jul 2005 18:34:13 -0700 Andrew Morton wrote:

> Miles Bader <miles@lsi.nec.co.jp> wrote:
> >
> > Linus Torvalds <torvalds@osdl.org> writes:
> >  > Ask a hundred random C programmers what "!!x" means, versus what "x != 0"
> >  > means, and time their replies.
> > 
> >  I've always thought of "!!" as the "canonicalize boolean" operator...
> 
> Me too.  Once you get used to it, it's just the "convert non-zero to 1"
> operator.
> 
> But whatever.

I call it "truth value" (true or false), but once I got used to it,
I thought it was OK too.

---
~Randy

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Lee Revell]

On Mon, 2005-07-25 at 20:20 -0400, Steven Rostedt wrote:
> And I
> would also assume that you prefer x *= 2 over x <<= 1 (also since the
> first person to show this example used x <<= 2. Right Lee? :-)

Let us never speak of that again.  These aren't the droids you're
looking for.

Lee

Re: [patch 2.6.13-rc3] i386: clean up user_mode macros

[Chuck Ebbert]

On Mon, 25 Jul 2005 at 16:13:13 -0700 (PDT), Linus Torvalds wrote:

> On Mon, 25 Jul 2005, Chuck Ebbert wrote:
> > 
> > Recent patches from the Xen group changed the X86 user_mode macros.
> > 
> > This patch does the following:
> > 
> >         1. Makes the new user_mode() return 0 or 1 (same as x86_64)
>
> I _really_ prefer
> 
>       x != 0
> 
> over 
> 
>       !!x


  Take 2:  compile tested only.


Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>
===================================================================
--- 2.6.13-rc3.orig/include/asm-i386/ptrace.h
+++ 2.6.13-rc3/include/asm-i386/ptrace.h
@@ -57,14 +57,21 @@
 #ifdef __KERNEL__
 struct task_struct;
 extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code);
-#define user_mode(regs)		(3 & (regs)->xcs)
-#define user_mode_vm(regs)	((VM_MASK & (regs)->eflags) || user_mode(regs))
+
+static inline int user_mode(struct pt_regs *regs)
+{
+	return (regs->xcs & 3) != 0;
+}
+static inline int user_mode_vm(struct pt_regs *regs)
+{
+	return ((regs->xcs & 3) | (regs->eflags & VM_MASK)) != 0;
+}
 #define instruction_pointer(regs) ((regs)->eip)
 #if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
 extern unsigned long profile_pc(struct pt_regs *regs);
 #else
 #define profile_pc(regs) instruction_pointer(regs)
 #endif
-#endif
+#endif /* __KERNEL__ */
 
 #endif
__
Chuck