linux-toradex.git/virt/kvm/ioapic.h, branch v3.6

KVM: fix race with level interrupts

2012-07-20T19:12:00+00:00

When more than 1 source id is in use for the same GSI, we have the
following race related to handling irq_states race:

CPU 0 clears bit 0. CPU 0 read irq_state as 0. CPU 1 sets level to 1.
CPU 1 calls kvm_ioapic_set_irq(1). CPU 0 calls kvm_ioapic_set_irq(0).
Now ioapic thinks the level is 0 but irq_state is not 0.

Fix by performing all irq_states bitmap handling under pic/ioapic lock.
This also removes the need for atomics with irq_states handling.

Reported-by: Gleb Natapov 
Signed-off-by: Michael S. Tsirkin 
Signed-off-by: Marcelo Tosatti

KVM: dont clear TMR on EOI

2012-04-16T23:36:38+00:00

Intel spec says that TMR needs to be set/cleared
when IRR is set, but kvm also clears it on  EOI.

I did some tests on a real (AMD based) system,
and I see same TMR values both before
and after EOI, so I think it's a minor bug in kvm.

This patch fixes TMR to be set/cleared on IRR set
only as per spec.

And now that we don't clear TMR, we can save
an atomic read of TMR on EOI that's not propagated
to ioapic, by checking whether ioapic needs
a specific vector first and calculating
the mode afterwards.

Signed-off-by: Michael S. Tsirkin 
Signed-off-by: Marcelo Tosatti

KVM: convert ioapic lock to spinlock

2010-05-13T04:23:55+00:00

kvm_set_irq is used from non sleepable contexes, so convert ioapic from
mutex to spinlock.

KVM-Stable-Tag.
Tested-by: Ralf Bonenkamp 
Signed-off-by: Marcelo Tosatti

KVM: cleanup the failure path of KVM_CREATE_IRQCHIP ioctrl

2010-03-01T15:36:10+00:00

If we fail to init ioapic device or the fail to setup the default irq
routing, the device register by kvm_create_pic() and kvm_ioapic_init()
remain unregister. This patch fixed to do this.

Signed-off-by: Wei Yongjun 
Signed-off-by: Avi Kivity

KVM: avoid taking ioapic mutex for non-ioapic EOIs

2010-03-01T15:35:46+00:00

When the guest acknowledges an interrupt, it sends an EOI message to the local
apic, which broadcasts it to the ioapic.  To handle the EOI, we need to take
the ioapic mutex.

On large guests, this causes a lot of contention on this mutex.  Since large
guests usually don't route interrupts via the ioapic (they use msi instead),
this is completely unnecessary.

Avoid taking the mutex by introducing a handled_vectors bitmap.  Before taking
the mutex, check if the ioapic was actually responsible for the acked vector.
If not, we can return early.

Signed-off-by: Avi Kivity 
Signed-off-by: Marcelo Tosatti

KVM: Move IO APIC to its own lock

2009-12-03T07:32:08+00:00

The allows removal of irq_lock from the injection path.

Signed-off-by: Gleb Natapov 
Signed-off-by: Avi Kivity

KVM: Move irq sharing information to irqchip level

2009-12-03T07:32:06+00:00

This removes assumptions that max GSIs is smaller than number of pins.
Sharing is tracked on pin level not GSI level.

[avi: no PIC on ia64]

Signed-off-by: Gleb Natapov 
Signed-off-by: Avi Kivity

KVM: APIC: get rid of deliver_bitmask

2009-06-10T08:48:27+00:00

Deliver interrupt during destination matching loop.

Signed-off-by: Gleb Natapov 
Acked-by: Xiantao Zhang 
Signed-off-by: Marcelo Tosatti

KVM: change the way how lowest priority vcpu is calculated

2009-06-10T08:48:27+00:00

The new way does not require additional loop over vcpus to calculate
the one with lowest priority as one is chosen during delivery bitmap
construction.

Signed-off-by: Gleb Natapov 
Signed-off-by: Marcelo Tosatti

KVM: consolidate ioapic/ipi interrupt delivery logic

2009-06-10T08:48:27+00:00

Use kvm_apic_match_dest() in kvm_get_intr_delivery_bitmask() instead
of duplicating the same code. Use kvm_get_intr_delivery_bitmask() in
apic_send_ipi() to figure out ipi destination instead of reimplementing
the logic.

Signed-off-by: Gleb Natapov 
Signed-off-by: Marcelo Tosatti