mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Nicolas Pitre <nico@cam.org>
To: Ingo Molnar <mingo@elte.hu>
Cc: lkml <linux-kernel@vger.kernel.org>,
	Linus Torvalds <torvalds@osdl.org>, Andrew Morton <akpm@osdl.org>,
	Arjan van de Ven <arjan@infradead.org>,
	Jes Sorensen <jes@trained-monkey.org>,
	Al Viro <viro@ftp.linux.org.uk>, Oleg Nesterov <oleg@tv-sign.ru>,
	David Howells <dhowells@redhat.com>,
	Alan Cox <alan@lxorguk.ukuu.org.uk>,
	Christoph Hellwig <hch@infradead.org>, Andi Kleen <ak@suse.de>,
	Russell King <rmk+lkml@arm.linux.org.uk>
Subject: Re: [patch 06/21] mutex subsystem, add include/asm-arm/mutex.h
Date: Fri, 06 Jan 2006 22:04:34 -0500 (EST)	[thread overview]
Message-ID: <Pine.LNX.4.64.0601062149300.13870@localhost.localdomain> (raw)
In-Reply-To: <20060105153737.GG31013@elte.hu>

On Thu, 5 Jan 2006, Ingo Molnar wrote:

> From: Nicolas Pitre <nico@cam.org>
> 
> add the ARM version of mutex.h, which is optimized in assembly for
> ARMv6, and uses the xchg implementation on pre-ARMv6.
> 
> Signed-off-by: Ingo Molnar <mingo@elte.hu>

Please consider the ARM cleanups below.  I was reviewing the generated 
code and found out that, from my original version, you changed:

	if (unlikely(__res | __ex_flag))

for:

	if (unlikely(__res || __ex_flag))

which produces worse code on ARM.  I therefore made it more explicit:

	__res |= __ex_flag;
	if (unlikely(__res != 0))

that the single | (bitwise or) was not a mistake.

I also made everything static inline rather than macros for better 
readability (both produce the same code after all).

And finally added missing \t from multi-line assembly code.

Signed-off-by: Nicolas Pitre <nico@cam.org>

Index: linux-2.6/include/asm-arm/mutex.h
===================================================================
--- linux-2.6.orig/include/asm-arm/mutex.h
+++ linux-2.6/include/asm-arm/mutex.h
@@ -23,72 +23,71 @@
  * simply bail out immediately through the slow path where the lock will be
  * reattempted until it succeeds.
  */
-#define __mutex_fastpath_lock(count, fail_fn)				\
-do {									\
-	int __ex_flag, __res;						\
-									\
-	typecheck(atomic_t *, count);					\
-	typecheck_fn(fastcall void (*)(atomic_t *), fail_fn);		\
-									\
-	__asm__ (							\
-		"ldrex	%0, [%2]	\n"				\
-		"sub	%0, %0, #1	\n"				\
-		"strex	%1, %0, [%2]	\n"				\
-									\
-		: "=&r" (__res), "=&r" (__ex_flag)			\
-		: "r" (&(count)->counter)				\
-		: "cc","memory" );					\
-									\
-	if (unlikely(__res || __ex_flag))				\
-		fail_fn(count);						\
-} while (0)
-
-#define __mutex_fastpath_lock_retval(count, fail_fn)			\
-({									\
-	int __ex_flag, __res;						\
-									\
-	typecheck(atomic_t *, count);					\
-	typecheck_fn(fastcall int (*)(atomic_t *), fail_fn);		\
-									\
-	__asm__ (							\
-		"ldrex	%0, [%2]	\n"				\
-		"sub	%0, %0, #1	\n"				\
-		"strex	%1, %0, [%2]	\n"				\
-									\
-		: "=&r" (__res), "=&r" (__ex_flag)			\
-		: "r" (&(count)->counter)				\
-		: "cc","memory" );					\
-									\
-	__res |= __ex_flag;						\
-	if (unlikely(__res != 0))					\
-		__res = fail_fn(count);					\
-	__res;								\
-})
+static inline void
+__mutex_fastpath_lock(atomic_t *count, fastcall void (*fail_fn)(atomic_t *))
+{
+	int __ex_flag, __res;
+
+	__asm__ (
+
+		"ldrex	%0, [%2]	\n\t"
+		"sub	%0, %0, #1	\n\t"
+		"strex	%1, %0, [%2]	"
+
+		: "=&r" (__res), "=&r" (__ex_flag)
+		: "r" (&(count)->counter)
+		: "cc","memory" );
+
+	__res |= __ex_flag;
+	if (unlikely(__res != 0))
+		fail_fn(count);
+}
+
+static inline int
+__mutex_fastpath_lock_retval(atomic_t *count, fastcall int (*fail_fn)(atomic_t *))
+{
+	int __ex_flag, __res;
+
+	__asm__ (
+
+		"ldrex	%0, [%2]	\n\t"
+		"sub	%0, %0, #1	\n\t"
+		"strex	%1, %0, [%2]	"
+
+		: "=&r" (__res), "=&r" (__ex_flag)
+		: "r" (&(count)->counter)
+		: "cc","memory" );
+
+	__res |= __ex_flag;
+	if (unlikely(__res != 0))
+		__res = fail_fn(count);
+	return __res;
+}
 
 /*
  * Same trick is used for the unlock fast path. However the original value,
  * rather than the result, is used to test for success in order to have
  * better generated assembly.
  */
-#define __mutex_fastpath_unlock(count, fail_fn)				\
-do {									\
-	int __ex_flag, __res, __orig;					\
-									\
-	typecheck(atomic_t *, count);					\
-	typecheck_fn(fastcall void (*)(atomic_t *), fail_fn);		\
-									\
-	__asm__ (							\
-		"ldrex	%0, [%3]	\n"				\
-		"add	%1, %0, #1	\n"				\
-		"strex	%2, %1, [%3]	\n"				\
-									\
-		: "=&r" (__orig), "=&r" (__res), "=&r" (__ex_flag)	\
-		: "r" (&(count)->counter)				\
-		: "cc","memory" );					\
-									\
-	if (unlikely(__orig || __ex_flag))				\
-		fail_fn(count);						\
-} while (0)
+static inline void
+__mutex_fastpath_unlock(atomic_t *count, fastcall void (*fail_fn)(atomic_t *))
+{
+	int __ex_flag, __res, __orig;
+
+	__asm__ (
+
+		"ldrex	%0, [%3]	\n\t"
+		"add	%1, %0, #1	\n\t"
+		"strex	%2, %1, [%3]	"
+
+		: "=&r" (__orig), "=&r" (__res), "=&r" (__ex_flag)
+		: "r" (&(count)->counter)
+		: "cc","memory" );
+
+	__orig |= __ex_flag;
+	if (unlikely(__orig != 0))
+		fail_fn(count);
+}
 
 /*
  * If the unlock was done on a contended lock, or if the unlock simply fails
@@ -110,12 +109,12 @@ __mutex_fastpath_trylock(atomic_t *count
 
 	__asm__ (
 
-		"1: ldrex	%0, [%3]	\n"
-		"subs		%1, %0, #1	\n"
-		"strexeq	%2, %1, [%3]	\n"
-		"movlt		%0, #0		\n"
-		"cmpeq		%2, #0		\n"
-		"bgt		1b		\n"
+		"1: ldrex	%0, [%3]	\n\t"
+		"subs		%1, %0, #1	\n\t"
+		"strexeq	%2, %1, [%3]	\n\t"
+		"movlt		%0, #0		\n\t"
+		"cmpeq		%2, #0		\n\t"
+		"bgt		1b		"
 
 		: "=&r" (__orig), "=&r" (__res), "=&r" (__ex_flag)
 		: "r" (&count->counter)

  reply	other threads:[~2006-01-07  3:05 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-01-05 15:37 Ingo Molnar
2006-01-07  3:04 ` Nicolas Pitre [this message]
  -- strict thread matches above, loose matches on Subject: below --
2006-01-04 14:42 Ingo Molnar

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Pine.LNX.4.64.0601062149300.13870@localhost.localdomain \
    --to=nico@cam.org \
    --cc=ak@suse.de \
    --cc=akpm@osdl.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=arjan@infradead.org \
    --cc=dhowells@redhat.com \
    --cc=hch@infradead.org \
    --cc=jes@trained-monkey.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=oleg@tv-sign.ru \
    --cc=rmk+lkml@arm.linux.org.uk \
    --cc=torvalds@osdl.org \
    --cc=viro@ftp.linux.org.uk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®