diff --git a/kernel/arch/i386/arch_clock.c b/kernel/arch/i386/arch_clock.c
index b07dacf91..34faee56a 100644
--- a/kernel/arch/i386/arch_clock.c
+++ b/kernel/arch/i386/arch_clock.c
@@ -241,6 +241,18 @@ PUBLIC void context_stop(struct proc * p)
 	
 	tsc_delta = sub64(tsc, *__tsc_ctr_switch);
 
+	if(kbill_ipc) {
+		kbill_ipc->p_kipc_cycles =
+			add64(kbill_ipc->p_kipc_cycles, tsc_delta);
+		kbill_ipc = NULL;
+	}
+
+	if(kbill_kcall) {
+		kbill_kcall->p_kcall_cycles =
+			add64(kbill_kcall->p_kcall_cycles, tsc_delta);
+		kbill_kcall = NULL;
+	}
+
 	/*
 	 * deduct the just consumed cpu cycles from the cpu time left for this
 	 * process during its current quantum. Skip IDLE and other pseudo kernel
diff --git a/kernel/glo.h b/kernel/glo.h
index 906fd5ffa..2446ca795 100644
--- a/kernel/glo.h
+++ b/kernel/glo.h
@@ -29,6 +29,8 @@ EXTERN struct loadinfo kloadinfo;	/* status of load average */
 EXTERN struct proc *vmrequest;  /* first process on vmrequest queue */
 EXTERN unsigned lost_ticks;	/* clock ticks counted outside clock task */
 EXTERN char *ipc_call_names[IPCNO_HIGHEST+1]; /* human-readable call names */
+EXTERN struct proc *kbill_kcall; /* process that made kernel call */
+EXTERN struct proc *kbill_ipc; /* process that invoked ipc */
 
 /* Interrupt related variables. */
 EXTERN irq_hook_t irq_hooks[NR_IRQ_HOOKS];	/* hooks for general use */
diff --git a/kernel/proc.c b/kernel/proc.c
index ec85c9343..cb972bec1 100644
--- a/kernel/proc.c
+++ b/kernel/proc.c
@@ -539,6 +539,9 @@ PUBLIC int do_ipc(reg_t r1, reg_t r2, reg_t r3)
 
   assert(!RTS_ISSET(caller_ptr, RTS_SLOT_FREE));
 
+  /* bill kernel time to this process. */
+  kbill_ipc = caller_ptr;
+
   /* If this process is subject to system call tracing, handle that first. */
   if (caller_ptr->p_misc_flags & (MF_SC_TRACE | MF_SC_DEFER)) {
 	/* Are we tracing this process, and is it the first sys_call entry? */
diff --git a/kernel/proc.h b/kernel/proc.h
index dfc332698..d8f50d088 100644
--- a/kernel/proc.h
+++ b/kernel/proc.h
@@ -58,6 +58,8 @@ struct proc {
   clock_t p_prof_left;		/* number of ticks left on profile timer */
 
   u64_t p_cycles;		/* how many cycles did the process use */
+  u64_t p_kcall_cycles;		/* kernel cycles caused by this proc (kcall) */
+  u64_t p_kipc_cycles;		/* cycles caused by this proc (ipc) */
 
   struct proc *p_nextready;	/* pointer to next ready process */
   struct proc *p_caller_q;	/* head of list of procs wishing to send */
diff --git a/kernel/system.c b/kernel/system.c
index eeae82807..029738b7e 100644
--- a/kernel/system.c
+++ b/kernel/system.c
@@ -156,6 +156,10 @@ PUBLIC void kernel_call(message *m_user, struct proc * caller)
 	  result = EBADREQUEST;
   }
 
+  
+  /* remember who invoked the kcall so we can bill it its time */
+  kbill_kcall = caller;
+
   kernel_call_finish(caller, &msg, result);
 }
 
diff --git a/kernel/system/do_fork.c b/kernel/system/do_fork.c
index 631252a6b..eac87ea00 100644
--- a/kernel/system/do_fork.c
+++ b/kernel/system/do_fork.c
@@ -90,6 +90,8 @@ PUBLIC int do_fork(struct proc * caller, message * m_ptr)
 
   make_zero64(rpc->p_cpu_time_left);
   make_zero64(rpc->p_cycles);
+  make_zero64(rpc->p_kcall_cycles);
+  make_zero64(rpc->p_kipc_cycles);
 
   /* If the parent is a privileged process, take away the privileges from the 
    * child process and inhibit it from running by setting the NO_PRIV flag.