public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Cyrill Gorcunov <gorcunov@gmail.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	"Maciej W. Rozycki" <macro@linux-mips.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH -tip] x86: apic LVTT - use APIC_DIVISOR on 64bit mode
Date: Sat, 12 Jul 2008 11:22:17 +0400	[thread overview]
Message-ID: <20080712072216.GA7010@asus> (raw)
In-Reply-To: <20080712060504.GB14714@elte.hu>

[Ingo Molnar - Sat, Jul 12, 2008 at 08:05:04AM +0200]
| 
| * Cyrill Gorcunov <gorcunov@gmail.com> wrote:
| 
| > We use APIC_TDR_DIV_16 while setting APIC_TDCR (divisor) it happened 
| > this moment is hidden from __setup_APIC_LVTT caller. So we better 
| > increment caller 'clocks' value to not change current behaviour and 
| > use APIC_DIVISOR (as already done in 32bit version).
| > 
| > The main benefit - unified procedure code for 32/64bit modes.
| 
| > Please review carefully. Any comments are welcome.
| > I'm _not_ 100% sure if this patch is safe.
| 
| hm, how about the other caller to __setup_APIC_LVTT(), 
| lapic_timer_setup():
| 
|         case CLOCK_EVT_MODE_ONESHOT:
|                 __setup_APIC_LVTT(calibration_result,
|                                   mode != CLOCK_EVT_MODE_PERIODIC, 1);
|                 break;
| 
| that will affect high-res timers. Wont calibration_result have the same 
| value after your patch as well, but only a 16th of it will be used in 
| __setup_APIC_LVTT(), causing 16 times shorter timeouts than intended?
| 
| I.e. are we fixing a bug, are we changing nothing, or are we introducing 
| a bug? :-)
| 
| 	Ingo
| 

well, the last I would say :). That happened I sent first version of
patch instead of modified second version. Thanks a lot Ingo for catching
it! Oh my, sorry... Here is an updated version.
---

We use APIC_TDR_DIV_16 while setting APIC_TDCR (divisor).
It happened this moment is hidden from __setup_APIC_LVTT
caller. So we better increment caller 'clocks' value to
not change current behaviour and use APIC_DIVISOR
(as already done in 32bit version).

The main benefit - unified procedure code for 32/64bit modes.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
CC: "Maciej W. Rozycki" <macro@linux-mips.org>
---

Index: linux-2.6.git/arch/x86/kernel/apic_64.c
===================================================================
--- linux-2.6.git.orig/arch/x86/kernel/apic_64.c	2008-07-12 11:11:33.000000000 +0400
+++ linux-2.6.git/arch/x86/kernel/apic_64.c	2008-07-12 11:14:22.000000000 +0400
@@ -165,6 +165,9 @@ int lapic_get_maxlvt(void)
 	return maxlvt;
 }
 
+/* Clock divisor is set to 16 */
+#define APIC_DIVISOR 16
+
 /*
  * This function sets up the local APIC timer, with a timeout of
  * 'clocks' APIC bus clock. During calibration we actually call
@@ -197,7 +200,7 @@ static void __setup_APIC_LVTT(unsigned i
 				| APIC_TDR_DIV_16);
 
 	if (!oneshot)
-		apic_write(APIC_TMICT, clocks);
+		apic_write(APIC_TMICT, clocks/APIC_DIVISOR);
 }
 
 /*
@@ -329,7 +332,7 @@ static void __init calibrate_APIC_clock(
 	 *
 	 * No interrupt enable !
 	 */
-	__setup_APIC_LVTT(250000000, 0, 0);
+	__setup_APIC_LVTT(4000000000, 0, 0);
 
 	apic_start = apic_read(APIC_TMCCT);
 #ifdef CONFIG_X86_PM_TIMER
@@ -367,7 +370,7 @@ static void __init calibrate_APIC_clock(
 	lapic_clockevent.min_delta_ns =
 		clockevent_delta2ns(0xF, &lapic_clockevent);
 
-	calibration_result = result / HZ;
+	calibration_result = result * APIC_DIVISOR / HZ;
 }
 
 /*

  reply	other threads:[~2008-07-12  7:22 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-09 19:02 [PATCH -tip] x86: apic LVTT - use APIC_DIVISOR on 64bit mode Cyrill Gorcunov
2008-07-12  6:05 ` Ingo Molnar
2008-07-12  7:22   ` Cyrill Gorcunov [this message]
2008-07-12  7:39     ` Ingo Molnar
2008-07-12  7:43       ` Cyrill Gorcunov
2008-07-12  7:37   ` 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=20080712072216.GA7010@asus \
    --to=gorcunov@gmail.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=macro@linux-mips.org \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    /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