All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dimitri Sivanich <sivanich@sgi.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: linux-kernel@vger.kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH 1/4 v2] X86: UV RTC fix early expiry handling
Date: Wed, 14 Oct 2009 09:16:30 -0500	[thread overview]
Message-ID: <20091014141630.GB11048@sgi.com> (raw)
In-Reply-To: <20091014141438.GA11048@sgi.com>

Tune/fix early timer expiry handling and return correct early timeout value
for set_next_event.

Signed-off-by: Dimitri Sivanich <sivanich@sgi.com>

---

 arch/x86/kernel/uv_time.c |   29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)

Index: linux/arch/x86/kernel/uv_time.c
===================================================================
--- linux.orig/arch/x86/kernel/uv_time.c	2009-10-13 08:50:44.000000000 -0500
+++ linux/arch/x86/kernel/uv_time.c	2009-10-13 10:41:44.000000000 -0500
@@ -123,7 +123,10 @@ static int uv_setup_intr(int cpu, u64 ex
 	/* Initialize comparator value */
 	uv_write_global_mmr64(pnode, UVH_INT_CMPB, expires);
 
-	return (expires < uv_read_rtc(NULL) && !uv_intr_pending(pnode));
+	if (uv_read_rtc(NULL) <= expires)
+		return 0;
+
+	return !uv_intr_pending(pnode);
 }
 
 /*
@@ -223,6 +226,7 @@ static int uv_rtc_set_timer(int cpu, u64
 
 	next_cpu = head->next_cpu;
 	*t = expires;
+
 	/* Will this one be next to go off? */
 	if (next_cpu < 0 || bcpu == next_cpu ||
 			expires < head->cpu[next_cpu].expires) {
@@ -231,7 +235,7 @@ static int uv_rtc_set_timer(int cpu, u64
 			*t = ULLONG_MAX;
 			uv_rtc_find_next_timer(head, pnode);
 			spin_unlock_irqrestore(&head->lock, flags);
-			return 1;
+			return -ETIME;
 		}
 	}
 
@@ -244,7 +248,7 @@ static int uv_rtc_set_timer(int cpu, u64
  *
  * Returns 1 if this timer was pending.
  */
-static int uv_rtc_unset_timer(int cpu)
+static int uv_rtc_unset_timer(int cpu, int force)
 {
 	int pnode = uv_cpu_to_pnode(cpu);
 	int bid = uv_cpu_to_blade_id(cpu);
@@ -256,14 +260,15 @@ static int uv_rtc_unset_timer(int cpu)
 
 	spin_lock_irqsave(&head->lock, flags);
 
-	if (head->next_cpu == bcpu && uv_read_rtc(NULL) >= *t)
+	if ((head->next_cpu == bcpu && uv_read_rtc(NULL) >= *t) || force)
 		rc = 1;
 
-	*t = ULLONG_MAX;
-
-	/* Was the hardware setup for this timer? */
-	if (head->next_cpu == bcpu)
-		uv_rtc_find_next_timer(head, pnode);
+	if (rc) {
+		*t = ULLONG_MAX;
+		/* Was the hardware setup for this timer? */
+		if (head->next_cpu == bcpu)
+			uv_rtc_find_next_timer(head, pnode);
+	}
 
 	spin_unlock_irqrestore(&head->lock, flags);
 
@@ -310,20 +315,20 @@ static void uv_rtc_timer_setup(enum cloc
 		break;
 	case CLOCK_EVT_MODE_UNUSED:
 	case CLOCK_EVT_MODE_SHUTDOWN:
-		uv_rtc_unset_timer(ced_cpu);
+		uv_rtc_unset_timer(ced_cpu, 1);
 		break;
 	}
 }
 
 static void uv_rtc_interrupt(void)
 {
-	struct clock_event_device *ced = &__get_cpu_var(cpu_ced);
 	int cpu = smp_processor_id();
+	struct clock_event_device *ced = &per_cpu(cpu_ced, cpu);
 
 	if (!ced || !ced->event_handler)
 		return;
 
-	if (uv_rtc_unset_timer(cpu) != 1)
+	if (uv_rtc_unset_timer(cpu, 0) != 1)
 		return;
 
 	ced->event_handler(ced);

  reply	other threads:[~2009-10-14 14:18 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-10-09 13:47 [PATCH] x86 UV: Fixes for UV rtc timers Dimitri Sivanich
2009-10-12 16:23 ` Ingo Molnar
2009-10-12 17:52   ` Dimitri Sivanich
2009-10-12 17:59     ` Ingo Molnar
2009-10-14 14:14       ` [PATCH 0/4 v2] X86: UV RTC fixes/cleanup Dimitri Sivanich
2009-10-14 14:16         ` Dimitri Sivanich [this message]
2009-10-14 14:18           ` [PATCH 2/4 v2] X86: UV RTC add clocksource only boot option Dimitri Sivanich
2009-10-14 14:21             ` [PATCH 3/4 v2] X86: UV RTC cleanup error handling Dimitri Sivanich
2009-10-14 14:22               ` [PATCH 4/4 v2] X86: UV RTC rename generic_interrupt Dimitri Sivanich
2009-10-14 16:33                 ` [tip:x86/uv] x86: UV RTC: Rename generic_interrupt to x86_platform_ipi tip-bot for Dimitri Sivanich
2009-10-14 16:33               ` [tip:x86/uv] x86: UV RTC: Clean up error handling tip-bot for Dimitri Sivanich
2009-10-14 16:33             ` [tip:x86/uv] x86: UV RTC: Add clocksource only boot option tip-bot for Dimitri Sivanich
2009-10-14 16:32           ` [tip:x86/uv] x86: UV RTC: Fix early expiry handling tip-bot for Dimitri Sivanich

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=20091014141630.GB11048@sgi.com \
    --to=sivanich@sgi.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.