linux-stable.git/include/linux/kvm_host.h, branch linux-2.6.39.y

KVM: Fix race between nmi injection and enabling nmi window

2011-03-17T16:08:30+00:00

The interrupt injection logic looks something like

  if an nmi is pending, and nmi injection allowed
    inject nmi
  if an nmi is pending
    request exit on nmi window

the problem is that "nmi is pending" can be set asynchronously by
the PIT; if it happens to fire between the two if statements, we
will request an nmi window even though nmi injection is allowed.  On
SVM, this has disasterous results, since it causes eflags.TF to be
set in random guest code.

The fix is simple; make nmi_pending synchronous using the standard
vcpu->requests mechanism; this ensures the code above is completely
synchronous wrt nmi_pending.

Signed-off-by: Avi Kivity

KVM: use yield_to instead of sleep in kvm_vcpu_on_spin

2011-03-17T16:08:29+00:00

Instead of sleeping in kvm_vcpu_on_spin, which can cause gigantic
slowdowns of certain workloads, we instead use yield_to to get
another VCPU in the same KVM guest to run sooner.

This seems to give a 10-15% speedup in certain workloads.

Signed-off-by: Rik van Riel 
Signed-off-by: Marcelo Tosatti 
Signed-off-by: Avi Kivity

KVM: keep track of which task is running a KVM vcpu

2011-03-17T16:08:29+00:00

Keep track of which task is running a KVM vcpu.  This helps us
figure out later what task to wake up if we want to boost a
vcpu that got preempted.

Unfortunately there are no guarantees that the same task
always keeps the same vcpu, so we can only track the task
across a single "run" of the vcpu.

Signed-off-by: Rik van Riel 
Signed-off-by: Avi Kivity

KVM: make make_all_cpus_request() lockless

2011-03-17T16:08:26+00:00

Now, we have 'vcpu->mode' to judge whether need to send ipi to other
cpus, this way is very exact, so checking request bit is needless,
then we can drop the spinlock let it's collateral

Signed-off-by: Xiao Guangrong 
Signed-off-by: Avi Kivity

KVM: Add "exiting guest mode" state

2011-03-17T16:08:26+00:00

Currently we keep track of only two states: guest mode and host
mode.  This patch adds an "exiting guest mode" state that tells
us that an IPI will happen soon, so unless we need to wait for the
IPI, we can avoid it completely.

Also
1: No need atomically to read/write ->mode in vcpu's thread

2: reorganize struct kvm_vcpu to make ->mode and ->requests
   in the same cache line explicitly

Signed-off-by: Xiao Guangrong 
Signed-off-by: Avi Kivity

KVM: Fix build error on s390 due to missing tlbs_dirty

2011-01-12T09:30:50+00:00

Make it available for all archs.

Signed-off-by: Avi Kivity

KVM: MMU: Make the way of accessing lpage_info more generic

2011-01-12T09:30:47+00:00

Large page information has two elements but one of them, write_count, alone
is accessed by a helper function.

This patch replaces this helper function with more generic one which returns
newly named kvm_lpage_info structure and use it to access the other element
rmap_pde.

Signed-off-by: Takuya Yoshikawa 
Signed-off-by: Avi Kivity

KVM: MMU: delay flush all tlbs on sync_page path

2011-01-12T09:29:51+00:00

Quote from Avi:
| I don't think we need to flush immediately; set a "tlb dirty" bit somewhere
| that is cleareded when we flush the tlb.  kvm_mmu_notifier_invalidate_page()
| can consult the bit and force a flush if set.

Signed-off-by: Xiao Guangrong 
Signed-off-by: Marcelo Tosatti

KVM: PPC: Fix compile warning

2011-01-12T09:29:42+00:00

KVM compilation fails with the following warning:

include/linux/kvm_host.h: In function 'kvm_irq_routing_update':
include/linux/kvm_host.h:679:2: error: 'struct kvm' has no member named 'irq_routing'

That function is only used and reasonable to have on systems that implement
an in-kernel interrupt chip. PPC doesn't.

Fix by #ifdef'ing it out when no irqchip is available.

Signed-off-by: Alexander Graf 
Signed-off-by: Avi Kivity

KVM: fast-path msi injection with irqfd

2011-01-12T09:29:38+00:00

Store irq routing table pointer in the irqfd object,
and use that to inject MSI directly without bouncing out to
a kernel thread.

While we touch this structure, rearrange irqfd fields to make fastpath
better packed for better cache utilization.

This also adds some comments about locking rules and rcu usage in code.

Some notes on the design:
- Use pointer into the rt instead of copying an entry,
  to make it possible to use rcu, thus side-stepping
  locking complexities.  We also save some memory this way.
- Old workqueue code is still used for level irqs.
  I don't think we DTRT with level anyway, however,
  it seems easier to keep the code around as
  it has been thought through and debugged, and fix level later than
  rip out and re-instate it later.

Signed-off-by: Michael S. Tsirkin 
Acked-by: Marcelo Tosatti 
Acked-by: Gregory Haskins 
Signed-off-by: Avi Kivity