mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [patch 5/6] lightweight robust futexes: i386
@ 2006-02-21  8:47 Ingo Molnar
  0 siblings, 0 replies; 5+ messages in thread
From: Ingo Molnar @ 2006-02-21  8:47 UTC (permalink / raw)
  To: linux-kernel
  Cc: Ulrich Drepper, Paul Jackson, Thomas Gleixner, Arjan van de Ven,
	Andrew Morton


i386: add the futex_atomic_cmpxchg_inuser() assembly implementation, and
wire up the new syscalls.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>

----

 arch/i386/kernel/syscall_table.S |    2 ++
 include/asm-i386/futex.h         |   23 ++++++++++++++++++++++-
 include/asm-i386/unistd.h        |    4 +++-
 3 files changed, 27 insertions(+), 2 deletions(-)

Index: linux/arch/i386/kernel/syscall_table.S
===================================================================
--- linux.orig/arch/i386/kernel/syscall_table.S
+++ linux/arch/i386/kernel/syscall_table.S
@@ -310,3 +310,5 @@ ENTRY(sys_call_table)
 	.long sys_pselect6
 	.long sys_ppoll
 	.long sys_unshare		/* 310 */
+	.long sys_set_robust_list
+	.long sys_get_robust_list
Index: linux/include/asm-i386/futex.h
===================================================================
--- linux.orig/include/asm-i386/futex.h
+++ linux/include/asm-i386/futex.h
@@ -107,7 +107,28 @@ futex_atomic_op_inuser (int encoded_op, 
 static inline int
 futex_atomic_cmpxchg_inuser(int __user *uaddr, int oldval, int newval)
 {
-	return -ENOSYS;
+	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))
+		return -EFAULT;
+
+	__asm__ __volatile__(
+		"1:	" LOCK_PREFIX "cmpxchgl %3, %1		\n"
+
+		"2:	.section .fixup, \"ax\"			\n"
+		"3:	mov     %2, %0				\n"
+		"	jmp     2b				\n"
+		"	.previous				\n"
+
+		"	.section __ex_table, \"a\"		\n"
+		"	.align  8				\n"
+		"	.long   1b,3b				\n"
+		"	.previous				\n"
+
+		: "=a" (oldval), "=m" (*uaddr)
+		: "i" (-EFAULT), "r" (newval), "0" (oldval)
+		: "memory"
+	);
+
+	return oldval;
 }
 
 #endif
Index: linux/include/asm-i386/unistd.h
===================================================================
--- linux.orig/include/asm-i386/unistd.h
+++ linux/include/asm-i386/unistd.h
@@ -316,8 +316,10 @@
 #define __NR_pselect6		308
 #define __NR_ppoll		309
 #define __NR_unshare		310
+#define __NR_set_robust_list	311
+#define __NR_get_robust_list	312
 
-#define NR_syscalls 311
+#define NR_syscalls 313
 
 /*
  * user-visible error numbers are in the range -1 - -128: see

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [patch 5/6] lightweight robust futexes: i386
@ 2006-02-22 21:45 Chuck Ebbert
  0 siblings, 0 replies; 5+ messages in thread
From: Chuck Ebbert @ 2006-02-22 21:45 UTC (permalink / raw)
  To: Ulrich Drepper
  Cc: linux-kernel, Ulrich Drepper, Arjan van de Ven, Andrew Morton,
	Thomas Gleixner, Ingo Molnar

In-Reply-To: <a36005b50602210838wa87764eof7b5d6e5a8a5ab3a@mail.gmail.com>

On Tue, 21 Feb 2006 at 08:38:04 -0800, Ulrich Drepper wrote:
> On 2/21/06, Chuck Ebbert <76306.1226@compuserve.com> wrote:
> > > +             : "=a" (oldval), "=m" (*uaddr)
> >                                  ^^^^
> >    Should be "+m" because it's both read and written.
> 
> No, this is why there is the "0" input parameter.

 But this is arg 1, not 0. With a memory clobber it's irrelevant
anyway though.

> > > +             : "memory"
> >                   ^^^^^^^^
> >    Is this necessary? Every possible memory location that could be
> > affected has been listed in the operands if the above change is made.
>
> This makes the asm a barrier for the compiler.

 A comment to that effect would be nice; IOW document the reason for it.

-- 
Chuck
"Equations are the Devil's sentences."  --Stephen Colbert


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [patch 5/6] lightweight robust futexes: i386
  2006-02-21 15:54 Chuck Ebbert
@ 2006-02-21 16:38 ` Ulrich Drepper
  0 siblings, 0 replies; 5+ messages in thread
From: Ulrich Drepper @ 2006-02-21 16:38 UTC (permalink / raw)
  To: Chuck Ebbert
  Cc: Ingo Molnar, Thomas Gleixner, Andrew Morton, Arjan van de Ven,
	Ulrich Drepper, linux-kernel

On 2/21/06, Chuck Ebbert <76306.1226@compuserve.com> wrote:
> > +             : "=a" (oldval), "=m" (*uaddr)
>                                  ^^^^
>    Should be "+m" because it's both read and written.

No, this is why there is the "0" input parameter.


\> > +             : "memory"
>                   ^^^^^^^^
>    Is this necessary? Every possible memory location that could be
> affected has been listed in the operands if the above change is made.

This makes the asm a barrier for the compiler.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [patch 5/6] lightweight robust futexes: i386
@ 2006-02-21 15:54 Chuck Ebbert
  2006-02-21 16:38 ` Ulrich Drepper
  0 siblings, 1 reply; 5+ messages in thread
From: Chuck Ebbert @ 2006-02-21 15:54 UTC (permalink / raw)
  To: Ingo Molnar
  Cc: Thomas Gleixner, Andrew Morton, Arjan van de Ven, Ulrich Drepper,
	linux-kernel

In-Reply-To: <20060216094234.GF29716@elte.hu>

On Thu, 16 Feb 2006 at 10:42:34 +0100, Ingo Molnar wrote:

> --- linux-robust-list.q.orig/include/asm-i386/futex.h
> +++ linux-robust-list.q/include/asm-i386/futex.h
> @@ -107,7 +107,25 @@ futex_atomic_op_inuser (int encoded_op, 
>  static inline int
>  futex_atomic_cmpxchg_inuser(int __user *uaddr, int oldval, int newval)
>  {
> -     return -ENOSYS;
> +     __asm__ __volatile__(
> +             "1:     " LOCK_PREFIX "cmpxchgl %3, %1          \n"
> +
> +             "2:     .section .fixup, \"ax\"                 \n"
> +             "3:     mov     %2, %0                          \n"
> +             "       jmp     2b                              \n"
> +             "       .previous                               \n"
> +
> +             "       .section __ex_table, \"a\"              \n"
> +             "       .align  8                               \n"
> +             "       .long   1b,3b                           \n"
> +             "       .previous                               \n"
> +
> +             : "=a" (oldval), "=m" (*uaddr)
                                 ^^^^
   Should be "+m" because it's both read and written.

> +             : "i" (-EFAULT), "r" (newval), "0" (oldval)
> +             : "memory"
                  ^^^^^^^^
   Is this necessary? Every possible memory location that could be
affected has been listed in the operands if the above change is made.

> +     );
> +
> +     return oldval;
>  }
>  
>  #endif
-- 
Chuck
"Equations are the Devil's sentences."  --Stephen Colbert

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [patch 5/6] lightweight robust futexes: i386
@ 2006-02-16  9:42 Ingo Molnar
  0 siblings, 0 replies; 5+ messages in thread
From: Ingo Molnar @ 2006-02-16  9:42 UTC (permalink / raw)
  To: linux-kernel
  Cc: Ulrich Drepper, Thomas Gleixner, Arjan van de Ven, Andrew Morton


i386: add the futex_atomic_cmpxchg_inuser() assembly implementation, and
wire up the new syscalls.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>

----

 arch/i386/kernel/syscall_table.S |    2 ++
 include/asm-i386/futex.h         |   20 +++++++++++++++++++-
 include/asm-i386/unistd.h        |    4 +++-
 3 files changed, 24 insertions(+), 2 deletions(-)

Index: linux-robust-list.q/arch/i386/kernel/syscall_table.S
===================================================================
--- linux-robust-list.q.orig/arch/i386/kernel/syscall_table.S
+++ linux-robust-list.q/arch/i386/kernel/syscall_table.S
@@ -310,3 +310,5 @@ ENTRY(sys_call_table)
 	.long sys_pselect6
 	.long sys_ppoll
 	.long sys_unshare		/* 310 */
+	.long sys_set_robust_list
+	.long sys_get_robust_list
Index: linux-robust-list.q/include/asm-i386/futex.h
===================================================================
--- linux-robust-list.q.orig/include/asm-i386/futex.h
+++ linux-robust-list.q/include/asm-i386/futex.h
@@ -107,7 +107,25 @@ futex_atomic_op_inuser (int encoded_op, 
 static inline int
 futex_atomic_cmpxchg_inuser(int __user *uaddr, int oldval, int newval)
 {
-	return -ENOSYS;
+	__asm__ __volatile__(
+		"1:	" LOCK_PREFIX "cmpxchgl %3, %1		\n"
+
+		"2:	.section .fixup, \"ax\"			\n"
+		"3:	mov     %2, %0				\n"
+		"	jmp     2b				\n"
+		"	.previous				\n"
+
+		"	.section __ex_table, \"a\"		\n"
+		"	.align  8				\n"
+		"	.long   1b,3b				\n"
+		"	.previous				\n"
+
+		: "=a" (oldval), "=m" (*uaddr)
+		: "i" (-EFAULT), "r" (newval), "0" (oldval)
+		: "memory"
+	);
+
+	return oldval;
 }
 
 #endif
Index: linux-robust-list.q/include/asm-i386/unistd.h
===================================================================
--- linux-robust-list.q.orig/include/asm-i386/unistd.h
+++ linux-robust-list.q/include/asm-i386/unistd.h
@@ -316,8 +316,10 @@
 #define __NR_pselect6		308
 #define __NR_ppoll		309
 #define __NR_unshare		310
+#define __NR_set_robust_list	311
+#define __NR_get_robust_list	312
 
-#define NR_syscalls 311
+#define NR_syscalls 313
 
 /*
  * user-visible error numbers are in the range -1 - -128: see

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2006-02-22 21:51 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-02-21  8:47 [patch 5/6] lightweight robust futexes: i386 Ingo Molnar
  -- strict thread matches above, loose matches on Subject: below --
2006-02-22 21:45 Chuck Ebbert
2006-02-21 15:54 Chuck Ebbert
2006-02-21 16:38 ` Ulrich Drepper
2006-02-16  9:42 Ingo Molnar

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome