From: Ingo Molnar <mingo@elte.hu>
To: Andrew Morton <akpm@osdl.org>
Cc: Valdis.Kletnieks@vt.edu, davidel@xmailserver.org,
jgarzik@pobox.com, greg@kroah.com, linux-kernel@vger.kernel.org,
netdev@oss.sgi.com
Subject: Re: MSEC_TO_JIFFIES is messed up...
Date: Wed, 12 May 2004 23:17:48 +0200 [thread overview]
Message-ID: <20040512211748.GB20800@elte.hu> (raw)
In-Reply-To: <20040512140729.476ace9e.akpm@osdl.org>
[-- Attachment #1: Type: text/plain, Size: 629 bytes --]
* Andrew Morton <akpm@osdl.org> wrote:
> Ingo Molnar <mingo@elte.hu> wrote:
> >
> > yet another patch - this time it's: complete, covers irda, accelerates
> > HZ=100, unifies the slightly differing namespaces and compiles/boots as
> > well.
> >
> > ...
> >
> > [hz-cleanup-2.6.6-A2 text/plain (2657 bytes)]
>
> This doesn't have the little round up which some implementations had,
> so someone who tries to sleep for 9 millisscondes on a 100HZ box may
> end up in a busywait. Looks risky.
ok. -A3 attached, it does the roundup in the msec->jiffies conversion.
(not the other way around though, and that's fine.)
Ingo
[-- Attachment #2: hz-cleanup-2.6.6-A3 --]
[-- Type: text/plain, Size: 2564 bytes --]
--- linux/include/linux/time.h.orig
+++ linux/include/linux/time.h
@@ -177,6 +177,24 @@ struct timezone {
(SH_DIV((MAX_JIFFY_OFFSET >> SEC_JIFFIE_SC) * TICK_NSEC, NSEC_PER_SEC, 1) - 1)
#endif
+
+/*
+ * Convert jiffies to milliseconds and back.
+ *
+ * Avoid unnecessary multiplications/divisions in the
+ * two most common HZ cases:
+ */
+#if HZ == 1000
+# define JIFFIES_TO_MSECS(x) (x)
+# define MSECS_TO_JIFFIES(x) (x)
+#elif HZ == 100
+# define JIFFIES_TO_MSECS(x) ((x) * 10)
+# define MSECS_TO_JIFFIES(x) (((x) + 9) / 10)
+#else
+# define JIFFIES_TO_MSECS(x) (((x) * 1000) / HZ)
+# define MSECS_TO_JIFFIES(x) (((x) * HZ + 999) / 1000)
+#endif
+
/*
* The TICK_NSEC - 1 rounds up the value to the next resolution. Note
* that a remainder subtract here would not do the right thing as the
--- linux/include/net/irda/irda.h.orig
+++ linux/include/net/irda/irda.h
@@ -83,8 +83,6 @@ if(!(expr)) do { \
#define MESSAGE(args...) printk(KERN_INFO args)
#define ERROR(args...) printk(KERN_ERR args)
-#define MSECS_TO_JIFFIES(ms) (((ms)*HZ+999)/1000)
-
/*
* Magic numbers used by Linux-IrDA. Random numbers which must be unique to
* give the best protection
--- linux/include/net/sctp/sctp.h.orig
+++ linux/include/net/sctp/sctp.h
@@ -116,11 +116,6 @@
#define SCTP_STATIC static
#endif
-#define MSECS_TO_JIFFIES(msec) \
- (((msec / 1000) * HZ) + ((msec % 1000) * HZ) / 1000)
-#define JIFFIES_TO_MSECS(jiff) \
- (((jiff / HZ) * 1000) + ((jiff % HZ) * 1000) / HZ)
-
/*
* Function declarations.
*/
--- linux/include/asm-i386/param.h.orig
+++ linux/include/asm-i386/param.h
@@ -5,8 +5,6 @@
# define HZ 1000 /* Internal kernel timer frequency */
# define USER_HZ 100 /* .. some user interfaces are in "ticks" */
# define CLOCKS_PER_SEC (USER_HZ) /* like times() */
-# define JIFFIES_TO_MSEC(x) (x)
-# define MSEC_TO_JIFFIES(x) (x)
#endif
#ifndef HZ
--- linux/kernel/sched.c.orig
+++ linux/kernel/sched.c
@@ -75,13 +75,6 @@
#define NS_TO_JIFFIES(TIME) ((TIME) / (1000000000 / HZ))
#define JIFFIES_TO_NS(TIME) ((TIME) * (1000000000 / HZ))
-#ifndef JIFFIES_TO_MSEC
-# define JIFFIES_TO_MSEC(x) ((x) * 1000 / HZ)
-#endif
-#ifndef MSEC_TO_JIFFIES
-# define MSEC_TO_JIFFIES(x) ((x) * HZ / 1000)
-#endif
-
/*
* These are the 'tuning knobs' of the scheduler:
*
@@ -1880,7 +1873,7 @@ static void rebalance_tick(int this_cpu,
interval *= sd->busy_factor;
/* scale ms to jiffies */
- interval = MSEC_TO_JIFFIES(interval);
+ interval = MSECS_TO_JIFFIES(interval);
if (unlikely(!interval))
interval = 1;
next prev parent reply other threads:[~2004-05-12 21:17 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20040512020700.6f6aa61f.akpm@osdl.org>
[not found] ` <20040512181903.GG13421@kroah.com>
2004-05-12 18:42 ` MSEC_TO_JIFFIES is messed up Jeff Garzik
2004-05-12 19:33 ` Ingo Molnar
2004-05-12 19:47 ` Valdis.Kletnieks
2004-05-12 19:56 ` Davide Libenzi
2004-05-12 20:07 ` Valdis.Kletnieks
2004-05-12 20:28 ` Ingo Molnar
2004-05-12 20:35 ` Ingo Molnar
2004-05-12 20:50 ` Ingo Molnar
2004-05-12 21:03 ` Valdis.Kletnieks
2004-05-12 21:33 ` Davide Libenzi
2004-05-12 21:07 ` Andrew Morton
2004-05-12 21:17 ` Ingo Molnar [this message]
2004-05-12 22:18 ` William Lee Irwin III
2004-05-13 17:38 ` Paul Wagland
2004-05-13 19:11 ` Andrew Morton
2004-05-13 19:19 ` Jeff Garzik
2004-05-13 22:40 ` Andrew Morton
2004-05-13 22:40 ` Andrew Morton
2004-05-13 22:41 ` Andrew Morton
2004-05-13 23:02 ` Jeff Garzik
2004-05-13 19:50 ` Paul Wagland
2004-05-12 23:33 ` Peter Williams
2004-05-12 19:49 ` Davide Libenzi
2004-05-12 20:03 ` Ingo Molnar
2004-05-12 20:18 ` Valdis.Kletnieks
2004-05-12 20:20 ` Andrew Morton
2004-05-12 20:24 ` Jeff Garzik
2004-05-12 20:35 ` Andrew Morton
2004-05-12 20:44 ` Jeff Garzik
2004-05-12 21:03 ` Sridhar Samudrala
2004-05-12 20:32 ` Greg KH
2004-05-12 20:38 ` William Lee Irwin III
2004-05-12 20:47 ` Andrew Morton
2004-05-12 20:58 ` Jeff Garzik
2004-05-12 20:59 ` William Lee Irwin III
2004-05-12 20:55 ` Ingo Molnar
2004-05-12 21:01 ` Davide Libenzi
2004-05-12 21:12 ` Ingo Molnar
2004-05-12 21:40 ` Davide Libenzi
2004-05-12 21:49 ` Zan Lynx
2004-05-12 22:05 ` Roland Dreier
2004-05-12 21:56 ` Zan Lynx
2004-05-12 21:39 ` J. Bruce Fields
2004-05-12 21:55 ` Andreas Schwab
2004-05-12 22:07 ` J. Bruce Fields
2004-05-16 3:48 ` Chris Wedgwood
2004-05-16 12:10 ` Paul Wagland
2004-05-12 20:17 ` Jeff Garzik
2004-05-12 20:54 ` Bill Rugolsky Jr.
2004-05-12 22:44 ` Bill Rugolsky Jr.
2004-05-12 20:40 Jan Olderdissen
2004-05-12 20:46 ` Jeff Garzik
2004-05-12 20:49 ` Andreas Schwab
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=20040512211748.GB20800@elte.hu \
--to=mingo@elte.hu \
--cc=Valdis.Kletnieks@vt.edu \
--cc=akpm@osdl.org \
--cc=davidel@xmailserver.org \
--cc=greg@kroah.com \
--cc=jgarzik@pobox.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@oss.sgi.com \
/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;
as well as URLs for NNTP newsgroup(s).