patch for virtual machine oriented scheduling(5)

Commit Message

alex April 22, 2009, 2:58 p.m. UTC

changes to standard KVM files:
--------------------------------------------------------------------------------------------

               else {
@@ -3297,13 +3303,9 @@ static int __vcpu_run(struct kvm_vcpu *vcpu,
struct kvm_run *kvm_run)
                       kvm_run->exit_reason = KVM_EXIT_INTR;
                       ++vcpu->stat.signal_exits;
               }
-               if (need_resched()) {
-                       up_read(&vcpu->kvm->slots_lock);
-                       kvm_resched(vcpu);
-                       down_read(&vcpu->kvm->slots_lock);
-               }
       }

+       set_bit(_VPF_blocked, &vcpu->pause_flags);
       up_read(&vcpu->kvm->slots_lock);
       post_kvm_run_save(vcpu, kvm_run);

@@ -4382,16 +4384,17 @@ static void kvm_free_vcpus(struct kvm *kvm)
       /*
        * Unpin any mmu pages first.
        */
-       for (i = 0; i < KVM_MAX_VCPUS; ++i)
-               if (kvm->vcpus[i])
-                       kvm_unload_vcpu_mmu(kvm->vcpus[i]);
+       if ((!is_idle_vm(kvm)) && (!is_host_vm(kvm))) {
+           for (i = 0; i < KVM_MAX_VCPUS; ++i)
+                   if (kvm->vcpus[i])
+                           kvm_unload_vcpu_mmu(kvm->vcpus[i]);
+       }
       for (i = 0; i < KVM_MAX_VCPUS; ++i) {
               if (kvm->vcpus[i]) {
                       kvm_arch_vcpu_free(kvm->vcpus[i]);
                       kvm->vcpus[i] = NULL;
               }
       }
-
 }

 void kvm_arch_sync_events(struct kvm *kvm)
@@ -4401,16 +4404,20 @@ void kvm_arch_sync_events(struct kvm *kvm)

 void kvm_arch_destroy_vm(struct kvm *kvm)
 {
-       kvm_iommu_unmap_guest(kvm);
-       kvm_free_pit(kvm);
-       kfree(kvm->arch.vpic);
-       kfree(kvm->arch.vioapic);
-       kvm_free_vcpus(kvm);
-       kvm_free_physmem(kvm);
-       if (kvm->arch.apic_access_page)
-               put_page(kvm->arch.apic_access_page);
-       if (kvm->arch.ept_identity_pagetable)
-               put_page(kvm->arch.ept_identity_pagetable);
+       if ((!is_idle_vm(kvm)) && (!is_host_vm(kvm))) {
+           kvm_iommu_unmap_guest(kvm);
+           kvm_free_pit(kvm);
+           kfree(kvm->arch.vpic);
+           kfree(kvm->arch.vioapic);
+           kvm_free_vcpus(kvm);
+           kvm_free_physmem(kvm);
+           if (kvm->arch.apic_access_page)
+                   put_page(kvm->arch.apic_access_page);
+           if (kvm->arch.ept_identity_pagetable)
+                   put_page(kvm->arch.ept_identity_pagetable);
+       }else
+            kvm_free_vcpus(kvm);
+
       kfree(kvm);
 }
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index c6997c0..ec69310 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -32,6 +32,7 @@ 
 #include <asm/desc.h>
 #include <asm/vmx.h>
 #include <asm/virtext.h>
+#include <linux/sched-if.h>

 #define __ex(x) __kvm_handle_fault_on_reboot(x)

@@ -360,7 +361,6 @@  static void __vcpu_clear(void *arg)
       if (per_cpu(current_vmcs, cpu) == vmx->vmcs)
               per_cpu(current_vmcs, cpu) = NULL;
       rdtscll(vmx->vcpu.arch.host_tsc);
-       list_del(&vmx->local_vcpus_link);
       vmx->vcpu.cpu = -1;
       vmx->launched = 0;
 }
@@ -369,6 +369,7 @@  static void vcpu_clear(struct vcpu_vmx *vmx)
 {
       if (vmx->vcpu.cpu == -1)
               return;
+       list_del(&vmx->local_vcpus_link);
       smp_call_function_single(vmx->vcpu.cpu, __vcpu_clear, vmx, 1);
 }

@@ -648,6 +649,7 @@  static void vmx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
       struct vcpu_vmx *vmx = to_vmx(vcpu);
       u64 phys_addr = __pa(vmx->vmcs);
       u64 tsc_this, delta, new_offset;
+       BUG_ON(is_host_vcpu(vcpu) || (is_idle_vcpu(vcpu)));

       if (vcpu->cpu != cpu) {
               vcpu_clear(vmx);
@@ -1102,8 +1104,10 @@  static void vmclear_local_vcpus(void)
       struct vcpu_vmx *vmx, *n;

       list_for_each_entry_safe(vmx, n, &per_cpu(vcpus_on_cpu, cpu),
-                                local_vcpus_link)
+                                local_vcpus_link) {
+               list_del(&vmx->local_vcpus_link);
               __vcpu_clear(vmx);
+       }
 }


@@ -3587,13 +3591,15 @@  static void vmx_free_vcpu(struct kvm_vcpu *vcpu)
 {
       struct vcpu_vmx *vmx = to_vmx(vcpu);

-       spin_lock(&vmx_vpid_lock);
-       if (vmx->vpid != 0)
-               __clear_bit(vmx->vpid, vmx_vpid_bitmap);
-       spin_unlock(&vmx_vpid_lock);
-       vmx_free_vmcs(vcpu);
-       kfree(vmx->host_msrs);
-       kfree(vmx->guest_msrs);
+       if ((!is_host_vcpu(vcpu)) && (!is_idle_vcpu(vcpu))) {
+           spin_lock(&vmx_vpid_lock);
+           if (vmx->vpid != 0)
+                   __clear_bit(vmx->vpid, vmx_vpid_bitmap);
+           spin_unlock(&vmx_vpid_lock);
+           vmx_free_vmcs(vcpu);
+           kfree(vmx->host_msrs);
+           kfree(vmx->guest_msrs);
+       }
       kvm_vcpu_uninit(vcpu);
       kmem_cache_free(kvm_vcpu_cache, vmx);
 }
@@ -3613,6 +3619,9 @@  static struct kvm_vcpu *vmx_create_vcpu(struct
kvm *kvm, unsigned int id)
       if (err)
               goto free_vcpu;

+       if (is_idle_vm(kvm) || is_host_vm(kvm))
+           return &vmx->vcpu;
+
       vmx->guest_msrs = kmalloc(PAGE_SIZE, GFP_KERNEL);
       if (!vmx->guest_msrs) {
               err = -ENOMEM;
@@ -3739,11 +3748,24 @@  static struct kvm_x86_ops vmx_x86_ops = {
       .get_tdp_level = get_ept_level,
       .get_mt_mask_shift = vmx_get_mt_mask_shift,
 };
-
+#include <linux/moduleparam.h>
+extern long (*sched_setaffinity_p)(pid_t pid, cpumask_t* in_mask);
+static int setaffinity = -1;
+module_param(setaffinity, int, 0);
 static int __init vmx_init(void)
 {
       int r;

+       sched_setaffinity_p = (void*)setaffinity;
+       if( setaffinity == -1 ){
+           printk("Please insert this module with parameters\n");
+           printk("to notify me the address of sched_setaffinity\n");
+           printk("Example:\n  insmod kvm-intel.ko
setaffinity=0xffffffff9002ecd9\n");
+           return -EINVAL;
+       }else{
+           printk("the address of function sched_setaffinity is %p \n",
sched_setaffinity_p);
+       }
+
       vmx_io_bitmap_a = (unsigned long *)__get_free_page(GFP_KERNEL);
       if (!vmx_io_bitmap_a)
               return -ENOMEM;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 0bb4131..3b19058 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -43,6 +43,8 @@ 
 #include <asm/desc.h>
 #include <asm/mtrr.h>

+#include <linux/sched-if.h>
+
 #define MAX_IO_MSRS 256
 #define CR0_RESERVED_BITS                                              \
       (~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
@@ -3256,8 +3258,12 @@  static int __vcpu_run(struct kvm_vcpu *vcpu,
struct kvm_run *kvm_run)
       down_read(&vcpu->kvm->slots_lock);
       vapic_enter(vcpu);

-       r = 1;
+       r = 1;
+       if(test_and_clear_bit(_VPF_blocked, &vcpu->pause_flags))
+           vcpu_wake(vcpu);
+
       while (r > 0) {
+               wait_event(vcpu->wq, (vcpu->status == VCPU_RUNNING));
               if (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE)
                       r = vcpu_enter_guest(vcpu, kvm_run);