public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Avi Kivity <avi@redhat.com>
To: linux-kernel@vger.kernel.org
Cc: kvm@vger.kernel.org, Glauber Costa <gcosta@redhat.com>
Subject: [PATCH 34/40] x86: KVM guest: use paravirt function to calculate cpu khz
Date: Tue, 23 Sep 2008 16:46:48 +0300	[thread overview]
Message-ID: <1222177614-26669-35-git-send-email-avi@redhat.com> (raw)
In-Reply-To: <1222177614-26669-1-git-send-email-avi@redhat.com>

From: Glauber Costa <gcosta@redhat.com>

We're currently facing timing problems in guests that do
calibration under heavy load, and then the load vanishes.
This means we'll have a much lower lpj than we actually should,
and delays end up taking less time than they should, which is a
nasty bug.

Solution is to pass on the lpj value from host to guest, and have it
preset.

Signed-off-by: Glauber Costa <gcosta@redhat.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
---
 arch/x86/kernel/kvmclock.c |   30 ++++++++++++++++++++++++++++++
 1 files changed, 30 insertions(+), 0 deletions(-)

diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c
index d02def0..774ac49 100644
--- a/arch/x86/kernel/kvmclock.c
+++ b/arch/x86/kernel/kvmclock.c
@@ -78,6 +78,34 @@ static cycle_t kvm_clock_read(void)
 	return ret;
 }
 
+/*
+ * If we don't do that, there is the possibility that the guest
+ * will calibrate under heavy load - thus, getting a lower lpj -
+ * and execute the delays themselves without load. This is wrong,
+ * because no delay loop can finish beforehand.
+ * Any heuristics is subject to fail, because ultimately, a large
+ * poll of guests can be running and trouble each other. So we preset
+ * lpj here
+ */
+static unsigned long kvm_get_tsc_khz(void)
+{
+	return preset_lpj;
+}
+
+static void kvm_get_preset_lpj(void)
+{
+	struct pvclock_vcpu_time_info *src;
+	unsigned long khz;
+	u64 lpj;
+
+	src = &per_cpu(hv_clock, 0);
+	khz = pvclock_tsc_khz(src);
+
+	lpj = ((u64)khz * 1000);
+	do_div(lpj, HZ);
+	preset_lpj = lpj;
+}
+
 static struct clocksource kvm_clock = {
 	.name = "kvm-clock",
 	.read = kvm_clock_read,
@@ -153,6 +181,7 @@ void __init kvmclock_init(void)
 		pv_time_ops.get_wallclock = kvm_get_wallclock;
 		pv_time_ops.set_wallclock = kvm_set_wallclock;
 		pv_time_ops.sched_clock = kvm_clock_read;
+		pv_time_ops.get_tsc_khz = kvm_get_tsc_khz;
 #ifdef CONFIG_X86_LOCAL_APIC
 		pv_apic_ops.setup_secondary_clock = kvm_setup_secondary_clock;
 #endif
@@ -163,6 +192,7 @@ void __init kvmclock_init(void)
 #ifdef CONFIG_KEXEC
 		machine_ops.crash_shutdown  = kvm_crash_shutdown;
 #endif
+		kvm_get_preset_lpj();
 		clocksource_register(&kvm_clock);
 	}
 }
-- 
1.6.0.1


  parent reply	other threads:[~2008-09-23 13:47 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-09-23 13:46 [PATCH 00/40] KVM Updates for 2.6.28 merge window (part 1 of 3) Avi Kivity
2008-09-23 13:46 ` [PATCH 01/40] KVM: VMX: Rename misnamed msr bits Avi Kivity
2008-09-23 13:46 ` [PATCH 02/40] KVM: x86: accessors for guest registers Avi Kivity
2008-09-23 13:46 ` [PATCH 03/40] KVM: Move KVM TRACE DEFINITIONS to common header Avi Kivity
2008-09-23 13:46 ` [PATCH 04/40] KVM: Introduce kvm_set_irq to inject interrupts in guests Avi Kivity
2008-09-23 13:46 ` [PATCH 05/40] KVM: MMU: Separate the code for unlinking a shadow page from its parents Avi Kivity
2008-09-23 13:46 ` [PATCH 06/40] KVM: MMU: Simplify kvm_mmu_zap_page() Avi Kivity
2008-09-23 13:46 ` [PATCH 07/40] KVM: Move NMI IRET fault processing to new vmx_complete_interrupts() Avi Kivity
2008-09-23 13:46 ` [PATCH 08/40] KVM: VMX: Move nmi injection failure processing to vm exit path Avi Kivity
2008-09-23 13:46 ` [PATCH 09/40] KVM: Clear exception queue before emulating an instruction Avi Kivity
2008-09-23 13:46 ` [PATCH 10/40] KVM: VMX: Fix pending exception processing Avi Kivity
2008-09-23 13:46 ` [PATCH 11/40] KVM: Add a pending interrupt queue Avi Kivity
2008-09-23 13:46 ` [PATCH 12/40] KVM: VMX: Move interrupt post-processing to vmx_complete_interrupts() Avi Kivity
2008-09-23 13:46 ` [PATCH 13/40] KVM: VMX: Remove redundant check in handle_rmode_exception Avi Kivity
2008-09-23 13:46 ` [PATCH 14/40] KVM: Consolidate PIC isr clearing into a function Avi Kivity
2008-09-23 13:46 ` [PATCH 15/40] KVM: Consolidate XX_VECTOR defines Avi Kivity
2008-09-23 13:46 ` [PATCH 16/40] KVM: VMX: Reinject real mode exception Avi Kivity
2008-09-23 13:46 ` [PATCH 17/40] KVM: VMX: Unify register save/restore across 32 and 64 bit hosts Avi Kivity
2008-09-23 13:46 ` [PATCH 18/40] KVM: SVM: " Avi Kivity
2008-09-23 13:46 ` [PATCH 19/40] KVM: kvmtrace: Remove use of bit fields in kvm trace structure Avi Kivity
2008-09-23 13:46 ` [PATCH 20/40] KVM: kvmtrace: replace get_cycles with ktime_get v3 Avi Kivity
2008-09-23 13:46 ` [PATCH 21/40] KVM: ppc: enable KVM_TRACE building for powerpc Avi Kivity
2008-09-23 13:46 ` [PATCH 22/40] KVM: ppc: adds trace points for ppc tlb activity Avi Kivity
2008-09-23 13:46 ` [PATCH 23/40] KVM: ppc: trace powerpc instruction emulation Avi Kivity
2008-09-23 13:46 ` [PATCH 24/40] KVM: VMX: Avoid vmwrite(HOST_RSP) when possible Avi Kivity
2008-09-23 13:46 ` [PATCH 25/40] KVM: Ignore DEBUGCTL MSRs with no effect Avi Kivity
2008-09-23 13:46 ` [PATCH 26/40] KVM: ppc: guest breakpoint support Avi Kivity
2008-09-23 13:46 ` [PATCH 27/40] KVM: ppc: Stop saving host TLB state Avi Kivity
2008-09-23 13:46 ` [PATCH 28/40] KVM: ppc: Write only modified shadow entries into the TLB on exit Avi Kivity
2008-09-23 13:46 ` [PATCH 29/40] KVM: powerpc: Map guest userspace with TID=0 mappings Avi Kivity
2008-09-23 13:46 ` [PATCH 30/40] KVM: Add irq ack notifier list Avi Kivity
2008-09-23 13:46 ` [PATCH 31/40] KVM: irq ack notification Avi Kivity
2008-09-23 13:46 ` [PATCH 32/40] KVM: PIT: fix injection logic and count Avi Kivity
2008-09-23 13:46 ` [PATCH 33/40] x86: paravirt: factor out cpu_khz to common code Avi Kivity
2008-09-23 13:46 ` Avi Kivity [this message]
2008-09-23 13:46 ` [PATCH 35/40] KVM: direct mmio pfn check Avi Kivity
2008-09-23 13:46 ` [PATCH 36/40] KVM: pci device assignment Avi Kivity
2008-09-23 13:46 ` [PATCH 37/40] KVM: Reduce kvm stack usage in kvm_arch_vm_ioctl() Avi Kivity
2008-09-23 13:46 ` [PATCH 38/40] KVM: Reduce stack usage in kvm_vcpu_ioctl() Avi Kivity
2008-09-23 13:46 ` [PATCH 39/40] KVM: Reduce stack usage in kvm_arch_vcpu_ioctl() Avi Kivity
2008-09-23 13:46 ` [PATCH 40/40] KVM: Reduce stack usage in kvm_pv_mmu_op() Avi Kivity

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=1222177614-26669-35-git-send-email-avi@redhat.com \
    --to=avi@redhat.com \
    --cc=gcosta@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    /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