Use hrtimers for the decrementer

Following S390's good example we should use hrtimers for the decrementer too!
This patch converts the timer from the old mechanism to hrtimers.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

2 files changed, 29 insertions, 9 deletions

diff --git a/arch/powerpc/kvm/emulate.c b/arch/powerpc/kvm/emulate.c
index 1ec5e07b81eb..4a9ac6640fad 100644
--- a/arch/powerpc/kvm/emulate.c
+++ b/arch/powerpc/kvm/emulate.c
@@ -18,7 +18,7 @@
  */
 
 #include <linux/jiffies.h>
-#include <linux/timer.h>
+#include <linux/hrtimer.h>
 #include <linux/types.h>
 #include <linux/string.h>
 #include <linux/kvm_host.h>
@@ -79,12 +79,13 @@ static int kvmppc_dec_enabled(struct kvm_vcpu *vcpu)
 
 void kvmppc_emulate_dec(struct kvm_vcpu *vcpu)
 {
-        unsigned long nr_jiffies;
+        unsigned long dec_nsec;
 
+        pr_debug("mtDEC: %x\n", vcpu->arch.dec);
 #ifdef CONFIG_PPC64
         /* POWER4+ triggers a dec interrupt if the value is < 0 */
         if (vcpu->arch.dec & 0x80000000) {
-                del_timer(&vcpu->arch.dec_timer);
+                hrtimer_try_to_cancel(&vcpu->arch.dec_timer);
                 kvmppc_core_queue_dec(vcpu);
                 return;
         }
@@ -94,12 +95,15 @@ void kvmppc_emulate_dec(struct kvm_vcpu *vcpu)
                  * that's how we convert the guest DEC value to the number of
                  * host ticks. */
 
+                hrtimer_try_to_cancel(&vcpu->arch.dec_timer);
+                dec_nsec = vcpu->arch.dec;
+                dec_nsec *= 1000;
+                dec_nsec /= tb_ticks_per_usec;
+                hrtimer_start(&vcpu->arch.dec_timer, ktime_set(0, dec_nsec),
+                              HRTIMER_MODE_REL);
                 vcpu->arch.dec_jiffies = get_tb();
-                nr_jiffies = vcpu->arch.dec / tb_ticks_per_jiffy;
-                mod_timer(&vcpu->arch.dec_timer,
-                          get_jiffies_64() + nr_jiffies);
         } else {
-                del_timer(&vcpu->arch.dec_timer);
+                hrtimer_try_to_cancel(&vcpu->arch.dec_timer);
         }
 }
 
diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c
index a06ecc3401fd..692c3709011e 100644
--- a/arch/powerpc/kvm/powerpc.c
+++ b/arch/powerpc/kvm/powerpc.c
@@ -23,6 +23,7 @@
 #include <linux/kvm_host.h>
 #include <linux/module.h>
 #include <linux/vmalloc.h>
+#include <linux/hrtimer.h>
 #include <linux/fs.h>
 #include <asm/cputable.h>
 #include <asm/uaccess.h>
@@ -208,10 +209,25 @@ static void kvmppc_decrementer_func(unsigned long data)
         }
 }
 
+/*
+ * low level hrtimer wake routine. Because this runs in hardirq context
+ * we schedule a tasklet to do the real work.
+ */
+enum hrtimer_restart kvmppc_decrementer_wakeup(struct hrtimer *timer)
+{
+        struct kvm_vcpu *vcpu;
+
+        vcpu = container_of(timer, struct kvm_vcpu, arch.dec_timer);
+        tasklet_schedule(&vcpu->arch.tasklet);
+
+        return HRTIMER_NORESTART;
+}
+
 int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
 {
-        setup_timer(&vcpu->arch.dec_timer, kvmppc_decrementer_func,
-                    (unsigned long)vcpu);
+        hrtimer_init(&vcpu->arch.dec_timer, CLOCK_REALTIME, HRTIMER_MODE_ABS);
+        tasklet_init(&vcpu->arch.tasklet, kvmppc_decrementer_func, (ulong)vcpu);
+        vcpu->arch.dec_timer.function = kvmppc_decrementer_wakeup;
 
         return 0;
 }