Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1542901 > unrolled thread

[PATCH v3 0/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff

Back to article view | Back to linux.kernel

Contents

  [PATCH v3 0/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff Radim Krčmář <rkrcmar@redhat.com> - 2016-12-15 18:10 +0100
    [PATCH v3 4/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff Radim Krčmář <rkrcmar@redhat.com> - 2016-12-15 18:10 +0100
      Re: [PATCH v3 4/4] KVM: x86: allow hotplug of VCPU with APIC ID over  0xff David Hildenbrand <david@redhat.com> - 2016-12-16 17:50 +0100

#1542901 — [PATCH v3 0/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff

v3:
  * fixes off-by-one error in [2/4]
  * moves phys_map changes from [2/4] to [4/4]; [2/4] doesn't carry
    David's r-b to [v2 2/4] because of this change
  * expands a comment in [4/4]

v2: http://www.spinics.net/lists/kvm/msg142414.html
> Removes the capability that let userspace know of changes.

v1: http://www.spinics.net/lists/kvm/msg141944.html
> The problem is described in [4/4].
> 
> [1/4] is a prerequisite to allow a cleanup in [2/4].
> [2/4] hopefully makes [4/4] easier to understand.
> [3/4] fixes mistakes from dealing with the mixed mode.
> [4/4] allows the hotplug and and add a capability for it.


Radim Krčmář (4):
  KVM: x86: use delivery to self in hyperv synic
  KVM: x86: replace kvm_apic_id with kvm_{x,x2}apic_id
  KVM: x86: make interrupt delivery fast and slow path behave the same
  KVM: x86: allow hotplug of VCPU with APIC ID over 0xff

 arch/x86/kvm/hyperv.c |  4 +--
 arch/x86/kvm/lapic.c  | 67 +++++++++++++++++++++++++++++++++++++--------------
 arch/x86/kvm/lapic.h  | 11 ---------
 3 files changed, 51 insertions(+), 31 deletions(-)

-- 
2.11.0

[toc] | [next] | [standalone]

#1542902 — [PATCH v3 4/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff

LAPIC after reset is in xAPIC mode, which poses a problem for hotplug of
VCPUs with high APIC ID, because reset VCPU is waiting for INIT/SIPI,
but there is no way to uniquely address it using xAPIC.

From many possible options, we chose the one that also works on real
hardware: accepting interrupts addressed to LAPIC's x2APIC ID even in
xAPIC mode.

KVM intentionally differs from real hardware, because real hardware
(Knights Landing) does just "x2apic_id & 0xff" to decide whether to
accept the interrupt in xAPIC mode and it can deliver one interrupt to
more than one physical destination, e.g. 0x123 to 0x123 and 0x23.

Fixes: 682f732ecf73 ("KVM: x86: bump MAX_VCPUS to 288")
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
---
 v3:
 * move code that populated physical_map from patch 2 [David]
 * extend a comment kvm_apic_match_physical_addr()
 v2:
 * Do not add capability [David, Paolo]
 * Tag it as a fix [David]
---
 arch/x86/kvm/lapic.c | 30 +++++++++++++++++++++++++-----
 1 file changed, 25 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index 519b72bbf947..c6758bdc2c24 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -189,15 +189,26 @@ static void recalculate_apic_map(struct kvm *kvm)
 		struct kvm_lapic *apic = vcpu->arch.apic;
 		struct kvm_lapic **cluster;
 		u16 mask;
-		u32 ldr, aid;
+		u32 ldr;
+		u8 xapic_id;
+		u32 x2apic_id;
 
 		if (!kvm_apic_present(vcpu))
 			continue;
 
-		aid = apic_x2apic_mode(apic) ? kvm_x2apic_id(apic)
-		                             : kvm_xapic_id(apic);
-		if (aid <= new->max_apic_id)
-			new->phys_map[aid] = apic;
+		xapic_id = kvm_xapic_id(apic);
+		x2apic_id = kvm_x2apic_id(apic);
+
+		/* Hotplug hack: see kvm_apic_match_physical_addr(), ... */
+		if ((apic_x2apic_mode(apic) || x2apic_id > 0xff) &&
+				x2apic_id <= new->max_apic_id)
+			new->phys_map[x2apic_id] = apic;
+		/*
+		 * ... xAPIC ID of VCPUs with APIC ID > 0xff will wrap-around,
+		 * prevent them from masking VCPUs with APIC ID <= 0xff.
+		 */
+		if (!apic_x2apic_mode(apic) && !new->phys_map[xapic_id])
+			new->phys_map[xapic_id] = apic;
 
 		ldr = kvm_lapic_get_reg(apic, APIC_LDR);
 
@@ -606,6 +617,15 @@ static bool kvm_apic_match_physical_addr(struct kvm_lapic *apic, u32 mda)
 	if (apic_x2apic_mode(apic))
 		return mda == kvm_x2apic_id(apic);
 
+	/*
+	 * Hotplug hack: Make LAPIC in xAPIC mode also accept interrupts as if
+	 * it were in x2APIC mode.  Hotplugged VCPUs start in xAPIC mode and
+	 * this allows unique addressing of VCPUs with APIC ID over 0xff.
+	 * The 0xff condition is needed because writeable xAPIC ID.
+	 */
+	if (kvm_x2apic_id(apic) > 0xff && mda == kvm_x2apic_id(apic))
+		return true;
+
 	return mda == kvm_xapic_id(apic);
 }
 
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]

#1543583 — Re: [PATCH v3 4/4] KVM: x86: allow hotplug of VCPU with APIC ID over 0xff

Am 15.12.2016 um 18:06 schrieb Radim Krčmář:
> LAPIC after reset is in xAPIC mode, which poses a problem for hotplug of
> VCPUs with high APIC ID, because reset VCPU is waiting for INIT/SIPI,
> but there is no way to uniquely address it using xAPIC.
>
> From many possible options, we chose the one that also works on real
> hardware: accepting interrupts addressed to LAPIC's x2APIC ID even in
> xAPIC mode.
>
> KVM intentionally differs from real hardware, because real hardware
> (Knights Landing) does just "x2apic_id & 0xff" to decide whether to
> accept the interrupt in xAPIC mode and it can deliver one interrupt to
> more than one physical destination, e.g. 0x123 to 0x123 and 0x23.
>
> Fixes: 682f732ecf73 ("KVM: x86: bump MAX_VCPUS to 288")
> Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>

Reviewed-by: David Hildenbrand <david@redhat.com>

-- 

David

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web