Changeset 42046 in vbox

Timestamp:

Jul 9, 2012 7:29:55 AM (13 years ago)

Author:

vboxsync

Message:

VMM/HWVMXR0: comments.

File:

• trunk/src/VBox/VMM/VMMR0/HWVMXR0.cpp (modified) (3 diffs)

trunk/src/VBox/VMM/VMMR0/HWVMXR0.cpp

@@ -2395 +2395 @@
      * Force a TLB flush for the first world switch if the current CPU differs from the one we ran on last
      * This can happen both for start & resume due to long jumps back to ring-3.
-     * If the TLB flush count changed, another VM (VCPU rather) has hit the ASID limit while flushing the TLB,
-     * so we cannot reuse the current ASID anymore.
+     * If the TLB flush count changed, another VM (VCPU rather) has hit the ASID limit while flushing the TLB
+     * or the host Cpu is online after a suspend/resume, so we cannot reuse the current ASID anymore.
      */
     bool fNewASID = false;
@@ -2522 +2522 @@
      * Force a TLB flush for the first world switch if the current CPU differs from the one we ran on last
      * This can happen both for start & resume due to long jumps back to ring-3.
-     * If the TLB flush count shouldn't really change in this EPT-only case.
+     * A change in the TLB flush count implies the host Cpu is online after a suspend/resume.
      */
     if (   pVCpu->hwaccm.s.idLastCpu != pCpu->idCpu
@@ -2586 +2586 @@
      * Force a TLB flush for the first world switch if the current CPU differs from the one we ran on last
      * This can happen both for start & resume due to long jumps back to ring-3.
-     * If the TLB flush count changed, another VM (VCPU rather) has hit the ASID limit while flushing the TLB,
-     * so we cannot reuse the current ASID anymore.
+     * If the TLB flush count changed, another VM (VCPU rather) has hit the ASID limit while flushing the TLB
+     * or the host Cpu is online after a suspend/resume, so we cannot reuse the current ASID anymore.
      */
     if (   pVCpu->hwaccm.s.idLastCpu != pCpu->idCpu