diff options
author | Michael S. Tsirkin <mst@redhat.com> | 2012-06-24 12:24:49 -0400 |
---|---|---|
committer | Avi Kivity <avi@redhat.com> | 2012-06-25 05:40:34 -0400 |
commit | c1af87dc96cd0f8f17694d0cd9be01b80b2c7a6a (patch) | |
tree | df243a9a8ac7c5d4e825a37f8afb1bdc9eb09924 /Documentation/virtual | |
parent | d0a69d6321ca759bb8d47803d06ba8571ab42d07 (diff) |
KVM: eoi msi documentation
Document the new EOI MSR. Couldn't decide whether this change belongs
conceptually on guest or host side, so a separate patch.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Diffstat (limited to 'Documentation/virtual')
-rw-r--r-- | Documentation/virtual/kvm/msr.txt | 33 |
1 files changed, 33 insertions, 0 deletions
diff --git a/Documentation/virtual/kvm/msr.txt b/Documentation/virtual/kvm/msr.txt index 96b41bd97523..730471048583 100644 --- a/Documentation/virtual/kvm/msr.txt +++ b/Documentation/virtual/kvm/msr.txt | |||
@@ -223,3 +223,36 @@ MSR_KVM_STEAL_TIME: 0x4b564d03 | |||
223 | steal: the amount of time in which this vCPU did not run, in | 223 | steal: the amount of time in which this vCPU did not run, in |
224 | nanoseconds. Time during which the vcpu is idle, will not be | 224 | nanoseconds. Time during which the vcpu is idle, will not be |
225 | reported as steal time. | 225 | reported as steal time. |
226 | |||
227 | MSR_KVM_EOI_EN: 0x4b564d04 | ||
228 | data: Bit 0 is 1 when PV end of interrupt is enabled on the vcpu; 0 | ||
229 | when disabled. Bit 1 is reserved and must be zero. When PV end of | ||
230 | interrupt is enabled (bit 0 set), bits 63-2 hold a 4-byte aligned | ||
231 | physical address of a 4 byte memory area which must be in guest RAM and | ||
232 | must be zeroed. | ||
233 | |||
234 | The first, least significant bit of 4 byte memory location will be | ||
235 | written to by the hypervisor, typically at the time of interrupt | ||
236 | injection. Value of 1 means that guest can skip writing EOI to the apic | ||
237 | (using MSR or MMIO write); instead, it is sufficient to signal | ||
238 | EOI by clearing the bit in guest memory - this location will | ||
239 | later be polled by the hypervisor. | ||
240 | Value of 0 means that the EOI write is required. | ||
241 | |||
242 | It is always safe for the guest to ignore the optimization and perform | ||
243 | the APIC EOI write anyway. | ||
244 | |||
245 | Hypervisor is guaranteed to only modify this least | ||
246 | significant bit while in the current VCPU context, this means that | ||
247 | guest does not need to use either lock prefix or memory ordering | ||
248 | primitives to synchronise with the hypervisor. | ||
249 | |||
250 | However, hypervisor can set and clear this memory bit at any time: | ||
251 | therefore to make sure hypervisor does not interrupt the | ||
252 | guest and clear the least significant bit in the memory area | ||
253 | in the window between guest testing it to detect | ||
254 | whether it can skip EOI apic write and between guest | ||
255 | clearing it to signal EOI to the hypervisor, | ||
256 | guest must both read the least significant bit in the memory area and | ||
257 | clear it using a single CPU instruction, such as test and clear, or | ||
258 | compare and exchange. | ||