mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Avi Kivity <avi@redhat.com>
To: kvm@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Subject: [PATCH 19/48] KVM: No disable_irq for MSI/MSI-X interrupt on device assignment
Date: Sun, 16 Aug 2009 12:29:39 +0300	[thread overview]
Message-ID: <1250415008-17175-20-git-send-email-avi@redhat.com> (raw)
In-Reply-To: <1250415008-17175-1-git-send-email-avi@redhat.com>

From: Sheng Yang <sheng@linux.intel.com>

Disable interrupt at interrupt handler and enable it when guest ack is for
the level triggered interrupt, to prevent reinjected interrupt. MSI/MSI-X don't
need it.

One possible problem is multiply same vector interrupt injected between irq
handler and scheduled work handler would be merged as one for MSI/MSI-X.
But AFAIK, the drivers handle it well.

The patch fixed the oplin card performance issue(MSI-X performance is half of
MSI/INTx).

Signed-off-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
---
 virt/kvm/kvm_main.c |   22 ++++++++--------------
 1 files changed, 8 insertions(+), 14 deletions(-)

diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index dee321e..2d9bc43 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -120,7 +120,7 @@ static void kvm_assigned_dev_interrupt_work_handler(struct work_struct *work)
 {
 	struct kvm_assigned_dev_kernel *assigned_dev;
 	struct kvm *kvm;
-	int irq, i;
+	int i;
 
 	assigned_dev = container_of(work, struct kvm_assigned_dev_kernel,
 				    interrupt_work);
@@ -143,20 +143,10 @@ static void kvm_assigned_dev_interrupt_work_handler(struct work_struct *work)
 			kvm_set_irq(assigned_dev->kvm,
 				    assigned_dev->irq_source_id,
 				    guest_entries[i].vector, 1);
-			irq = assigned_dev->host_msix_entries[i].vector;
-			if (irq != 0)
-				enable_irq(irq);
-			assigned_dev->host_irq_disabled = false;
 		}
-	} else {
+	} else
 		kvm_set_irq(assigned_dev->kvm, assigned_dev->irq_source_id,
 			    assigned_dev->guest_irq, 1);
-		if (assigned_dev->irq_requested_type &
-				KVM_DEV_IRQ_GUEST_MSI) {
-			enable_irq(assigned_dev->host_irq);
-			assigned_dev->host_irq_disabled = false;
-		}
-	}
 
 	spin_unlock_irq(&assigned_dev->assigned_dev_lock);
 	mutex_unlock(&assigned_dev->kvm->lock);
@@ -179,8 +169,10 @@ static irqreturn_t kvm_assigned_dev_intr(int irq, void *dev_id)
 
 	schedule_work(&assigned_dev->interrupt_work);
 
-	disable_irq_nosync(irq);
-	assigned_dev->host_irq_disabled = true;
+	if (assigned_dev->irq_requested_type & KVM_DEV_IRQ_GUEST_INTX) {
+		disable_irq_nosync(irq);
+		assigned_dev->host_irq_disabled = true;
+	}
 
 out:
 	spin_unlock_irqrestore(&assigned_dev->assigned_dev_lock, flags);
@@ -417,6 +409,7 @@ static int assigned_device_enable_guest_msi(struct kvm *kvm,
 {
 	dev->guest_irq = irq->guest_irq;
 	dev->ack_notifier.gsi = -1;
+	dev->host_irq_disabled = false;
 	return 0;
 }
 #endif
@@ -427,6 +420,7 @@ static int assigned_device_enable_guest_msix(struct kvm *kvm,
 {
 	dev->guest_irq = irq->guest_irq;
 	dev->ack_notifier.gsi = -1;
+	dev->host_irq_disabled = false;
 	return 0;
 }
 #endif
-- 
1.6.3.3


  parent reply	other threads:[~2009-08-16  9:37 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-08-16  9:29 [PATCH 00/48] KVM updates for 2.6.32 merge window (1/4) Avi Kivity
2009-08-16  9:29 ` [PATCH 01/48] KVM: VMX: Properly handle software interrupt re-injection in real mode Avi Kivity
2009-08-16  9:29 ` [PATCH 02/48] KVM: Replace MSR_IA32_TIME_STAMP_COUNTER with MSR_IA32_TSC of msr-index.h Avi Kivity
2009-08-16  9:29 ` [PATCH 03/48] KVM: Add MCE support Avi Kivity
2009-08-16  9:29 ` [PATCH 04/48] KVM: Use MSR names in place of address Avi Kivity
2009-08-16  9:29 ` [PATCH 05/48] KVM: fix cpuid E2BIG handling for extended request types Avi Kivity
2009-08-16  9:29 ` [PATCH 06/48] KVM: x86 emulator: Implement zero-extended immediate decoding Avi Kivity
2009-08-16  9:29 ` [PATCH 07/48] KVM: x86 emulator: fix jmp far decoding (opcode 0xea) Avi Kivity
2009-08-16  9:29 ` [PATCH 08/48] KVM: cleanup arch/x86/kvm/Makefile Avi Kivity
2009-08-16  9:29 ` [PATCH 09/48] KVM: Drop interrupt shadow when single stepping should be done only on VMX Avi Kivity
2009-08-16  9:29 ` [PATCH 10/48] KVM: Move common KVM Kconfig items to new file virt/kvm/Kconfig Avi Kivity
2009-08-16  9:29 ` [PATCH 11/48] KVM: irqfd Avi Kivity
2009-08-16  9:29 ` [PATCH 12/48] KVM: Allow PIT emulation without speaker port Avi Kivity
2009-08-16  9:29 ` [PATCH 13/48] KVM: ia64: Correct itc_offset calculations Avi Kivity
2009-08-16  9:29 ` [PATCH 14/48] KVM: s390: infrastructure to kick vcpus out of guest state Avi Kivity
2009-08-16  9:29 ` [PATCH 15/48] KVM: s390: fix signal handling Avi Kivity
2009-08-16  9:29 ` [PATCH 16/48] KVM: s390: streamline memslot handling Avi Kivity
2009-08-16  9:29 ` [PATCH 17/48] KVM: Downsize max support MSI-X entry to 256 Avi Kivity
2009-08-16  9:29 ` [PATCH 18/48] KVM: SVM: use explicit 64bit storage for sysenter values Avi Kivity
2009-08-16  9:29 ` Avi Kivity [this message]
2009-08-16  9:29 ` [PATCH 20/48] KVM: remove redundant declarations Avi Kivity
2009-08-16  9:29 ` [PATCH 21/48] KVM: SVM: Fold kvm_svm.h info svm.c Avi Kivity
2009-08-16  9:29 ` [PATCH 22/48] KVM: powerpc: fix some init/exit annotations Avi Kivity
2009-08-16  9:29 ` [PATCH 23/48] KVM: Clean up coalesced_mmio destruction Avi Kivity
2009-08-16  9:29 ` [PATCH 24/48] KVM: cleanup io_device code Avi Kivity
2009-08-16  9:29 ` [PATCH 25/48] KVM: do not register i8254 PIO regions until we are initialized Avi Kivity
2009-08-16  9:29 ` [PATCH 26/48] KVM: VMX: Avoid duplicate ept tlb flush when setting cr3 Avi Kivity
2009-08-16  9:29 ` [PATCH 27/48] KVM: VMX: Simplify pdptr and cr3 management Avi Kivity
2009-08-16  9:29 ` [PATCH 28/48] KVM: Cache pdptrs Avi Kivity
2009-08-16  9:29 ` [PATCH 29/48] KVM: VMX: Fix reporting of unhandled EPT violations Avi Kivity
2009-08-16  9:29 ` [PATCH 30/48] KVM: Calculate available entries in coalesced mmio ring Avi Kivity
2009-08-16  9:29 ` [PATCH 31/48] KVM: ppc: e500: Move to Book-3e MMU definitions Avi Kivity
2009-08-16  9:29 ` [PATCH 32/48] KVM: ppc: e500: Directly pass pvr to guest Avi Kivity
2009-08-16  9:29 ` [PATCH 33/48] KVM: ppc: e500: Add MMUCFG and PVR emulation Avi Kivity
2009-08-16  9:29 ` [PATCH 34/48] KVM: Cleanup LAPIC interface Avi Kivity
2009-08-16  9:29 ` [PATCH 35/48] KVM: Grab pic lock in kvm_pic_clear_isr_ack Avi Kivity
2009-08-16  9:29 ` [PATCH 36/48] KVM: move coalesced_mmio locking to its own device Avi Kivity
2009-08-16  9:29 ` [PATCH 37/48] KVM: introduce irq_lock, use it to protect ioapic Avi Kivity
2009-08-16  9:29 ` [PATCH 38/48] KVM: switch irq injection/acking data structures to irq_lock Avi Kivity
2009-08-16  9:29 ` [PATCH 39/48] KVM: VMX: Support Unrestricted Guest feature Avi Kivity
2009-08-16  9:30 ` [PATCH 40/48] KVM: Reorder ioctls in kvm.h Avi Kivity
2009-08-16  9:30 ` [PATCH 41/48] KVM: VMX: Move rmode structure to vmx-specific code Avi Kivity
2009-08-16  9:30 ` [PATCH 42/48] KVM: MMU: Fix is_dirty_pte() Avi Kivity
2009-08-16  9:30 ` [PATCH 43/48] KVM: MMU: Adjust pte accessors to explicitly indicate guest or shadow pte Avi Kivity
2009-08-16  9:30 ` [PATCH 44/48] KVM: MMU: s/shadow_pte/spte/ Avi Kivity
2009-08-16  9:30 ` [PATCH 45/48] KVM: Introduce kvm_vcpu_is_bsp() function Avi Kivity
2009-08-16  9:30 ` [PATCH 46/48] KVM: Use pointer to vcpu instead of vcpu_id in timer code Avi Kivity
2009-08-16  9:30 ` [PATCH 47/48] KVM: Break dependency between vcpu index in vcpus array and vcpu_id Avi Kivity
2009-08-16  9:30 ` [PATCH 48/48] KVM: Use macro to iterate over vcpus 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=1250415008-17175-20-git-send-email-avi@redhat.com \
    --to=avi@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

all inboxes | Powered by JetHome®