Index: head/sys/kern/init_main.c
===================================================================
--- head/sys/kern/init_main.c
+++ head/sys/kern/init_main.c
@@ -99,7 +99,7 @@
 static struct session session0;
 static struct pgrp pgrp0;
 struct	proc proc0;
-struct	thread thread0 __aligned(16);
+struct thread0_storage thread0_st __aligned(16);
 struct	vmspace vmspace0;
 struct	proc *initproc;
 
Index: head/sys/kern/kern_fork.c
===================================================================
--- head/sys/kern/kern_fork.c
+++ head/sys/kern/kern_fork.c
@@ -1011,7 +1011,7 @@
 	KASSERT(p->p_state == PRS_NORMAL, ("executing process is still new"));
 
 	CTR4(KTR_PROC, "fork_exit: new thread %p (td_sched %p, pid %d, %s)",
-		td, td->td_sched, p->p_pid, td->td_name);
+	    td, td_get_sched(td), p->p_pid, td->td_name);
 
 	sched_fork_exit(td);
 	/*
Index: head/sys/kern/kern_proc.c
===================================================================
--- head/sys/kern/kern_proc.c
+++ head/sys/kern/kern_proc.c
@@ -237,7 +237,6 @@
 
 	p = (struct proc *)mem;
 	SDT_PROBE3(proc, , init, entry, p, size, flags);
-	p->p_sched = (struct p_sched *)&p[1];
 	mtx_init(&p->p_mtx, "process lock", NULL, MTX_DEF | MTX_DUPOK | MTX_NEW);
 	mtx_init(&p->p_slock, "process slock", NULL, MTX_SPIN | MTX_NEW);
 	mtx_init(&p->p_statmtx, "pstatl", NULL, MTX_SPIN | MTX_NEW);
Index: head/sys/kern/kern_synch.c
===================================================================
--- head/sys/kern/kern_synch.c
+++ head/sys/kern/kern_synch.c
@@ -441,7 +441,7 @@
 	PCPU_INC(cnt.v_swtch);
 	PCPU_SET(switchticks, ticks);
 	CTR4(KTR_PROC, "mi_switch: old thread %ld (td_sched %p, pid %ld, %s)",
-	    td->td_tid, td->td_sched, td->td_proc->p_pid, td->td_name);
+	    td->td_tid, td_get_sched(td), td->td_proc->p_pid, td->td_name);
 #if (KTR_COMPILE & KTR_SCHED) != 0
 	if (TD_IS_IDLETHREAD(td))
 		KTR_STATE1(KTR_SCHED, "thread", sched_tdname(td), "idle",
@@ -457,7 +457,7 @@
 	    "prio:%d", td->td_priority);
 
 	CTR4(KTR_PROC, "mi_switch: new thread %ld (td_sched %p, pid %ld, %s)",
-	    td->td_tid, td->td_sched, td->td_proc->p_pid, td->td_name);
+	    td->td_tid, td_get_sched(td), td->td_proc->p_pid, td->td_name);
 
 	/* 
 	 * If the last thread was exiting, finish cleaning it up.
Index: head/sys/kern/kern_thread.c
===================================================================
--- head/sys/kern/kern_thread.c
+++ head/sys/kern/kern_thread.c
@@ -211,7 +211,6 @@
 	td->td_turnstile = turnstile_alloc();
 	td->td_rlqe = NULL;
 	EVENTHANDLER_INVOKE(thread_init, td);
-	td->td_sched = (struct td_sched *)&td[1];
 	umtx_thread_init(td);
 	td->td_kstack = 0;
 	td->td_sel = NULL;
Index: head/sys/kern/sched_4bsd.c
===================================================================
--- head/sys/kern/sched_4bsd.c
+++ head/sys/kern/sched_4bsd.c
@@ -117,7 +117,10 @@
 #define	THREAD_CAN_SCHED(td, cpu)	\
     CPU_ISSET((cpu), &(td)->td_cpuset->cs_mask)
 
-static struct td_sched td_sched0;
+_Static_assert(sizeof(struct thread) + sizeof(struct td_sched) <=
+    sizeof(struct thread0_storage),
+    "increase struct thread0_storage.t0st_sched size");
+
 static struct mtx sched_lock;
 
 static int	realstathz = 127; /* stathz is sometimes 0 and run off of hz. */
@@ -491,8 +494,8 @@
 		}
 		FOREACH_THREAD_IN_PROC(p, td) {
 			awake = 0;
+			ts = td_get_sched(td);
 			thread_lock(td);
-			ts = td->td_sched;
 			/*
 			 * Increment sleep time (if sleeping).  We
 			 * ignore overflow, as above.
@@ -596,7 +599,7 @@
 	fixpt_t loadfac;
 	unsigned int newcpu;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	loadfac = loadfactor(averunnable.ldavg[0]);
 	if (ts->ts_slptime > 5 * loadfac)
 		ts->ts_estcpu = 0;
@@ -621,7 +624,8 @@
 
 	if (td->td_pri_class != PRI_TIMESHARE)
 		return;
-	newpriority = PUSER + td->td_sched->ts_estcpu / INVERSE_ESTCPU_WEIGHT +
+	newpriority = PUSER +
+	    td_get_sched(td)->ts_estcpu / INVERSE_ESTCPU_WEIGHT +
 	    NICE_WEIGHT * (td->td_proc->p_nice - PRIO_MIN);
 	newpriority = min(max(newpriority, PRI_MIN_TIMESHARE),
 	    PRI_MAX_TIMESHARE);
@@ -682,13 +686,12 @@
 void
 schedinit(void)
 {
+
 	/*
-	 * Set up the scheduler specific parts of proc0.
+	 * Set up the scheduler specific parts of thread0.
 	 */
-	proc0.p_sched = NULL; /* XXX */
-	thread0.td_sched = &td_sched0;
 	thread0.td_lock = &sched_lock;
-	td_sched0.ts_slice = sched_slice;
+	td_get_sched(&thread0)->ts_slice = sched_slice;
 	mtx_init(&sched_lock, "sched lock", NULL, MTX_SPIN | MTX_RECURSE);
 }
 
@@ -731,7 +734,7 @@
 	struct td_sched *ts;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 
 	ts->ts_cpticks++;
 	ts->ts_estcpu = ESTCPULIM(ts->ts_estcpu + 1);
@@ -775,8 +778,8 @@
 	KTR_STATE1(KTR_SCHED, "thread", sched_tdname(child), "exit",
 	    "prio:%d", child->td_priority);
 	thread_lock(td);
-	td->td_sched->ts_estcpu = ESTCPULIM(td->td_sched->ts_estcpu +
-	    child->td_sched->ts_estcpu);
+	td_get_sched(td)->ts_estcpu = ESTCPULIM(td_get_sched(td)->ts_estcpu +
+	    td_get_sched(child)->ts_estcpu);
 	thread_unlock(td);
 	thread_lock(child);
 	if ((child->td_flags & TDF_NOLOAD) == 0)
@@ -793,17 +796,18 @@
 void
 sched_fork_thread(struct thread *td, struct thread *childtd)
 {
-	struct td_sched *ts;
+	struct td_sched *ts, *tsc;
 
 	childtd->td_oncpu = NOCPU;
 	childtd->td_lastcpu = NOCPU;
 	childtd->td_lock = &sched_lock;
 	childtd->td_cpuset = cpuset_ref(td->td_cpuset);
 	childtd->td_priority = childtd->td_base_pri;
-	ts = childtd->td_sched;
+	ts = td_get_sched(childtd);
 	bzero(ts, sizeof(*ts));
-	ts->ts_estcpu = td->td_sched->ts_estcpu;
-	ts->ts_flags |= (td->td_sched->ts_flags & TSF_AFFINITY);
+	tsc = td_get_sched(td);
+	ts->ts_estcpu = tsc->ts_estcpu;
+	ts->ts_flags |= (tsc->ts_flags & TSF_AFFINITY);
 	ts->ts_slice = 1;
 }
 
@@ -952,7 +956,7 @@
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	td->td_slptick = ticks;
-	td->td_sched->ts_slptime = 0;
+	td_get_sched(td)->ts_slptime = 0;
 	if (pri != 0 && PRI_BASE(td->td_pri_class) == PRI_TIMESHARE)
 		sched_prio(td, pri);
 	if (TD_IS_SUSPENDED(td) || pri >= PSOCK)
@@ -968,7 +972,7 @@
 	int preempted;
 
 	tmtx = NULL;
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	p = td->td_proc;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
@@ -1095,7 +1099,7 @@
 	struct td_sched *ts;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	td->td_flags &= ~TDF_CANSWAP;
 	if (ts->ts_slptime > 1) {
 		updatepri(td);
@@ -1266,7 +1270,7 @@
 	int forwarded = 0;
 	int single_cpu = 0;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	KASSERT((td->td_inhibitors == 0),
 	    ("sched_add: trying to run inhibited thread"));
@@ -1361,7 +1365,7 @@
 {
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	KASSERT((td->td_inhibitors == 0),
 	    ("sched_add: trying to run inhibited thread"));
@@ -1414,7 +1418,7 @@
 {
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	KASSERT(td->td_flags & TDF_INMEM,
 	    ("sched_rem: thread swapped out"));
 	KASSERT(TD_ON_RUNQ(td),
@@ -1527,7 +1531,7 @@
 	THREAD_LOCK_ASSERT(td, MA_OWNED|MA_NOTRECURSED);
 	KASSERT(td == curthread, ("sched_bind: can only bind curthread"));
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 
 	td->td_flags |= TDF_BOUND;
 #ifdef SMP
@@ -1586,7 +1590,7 @@
 	struct td_sched *ts;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	return (ts->ts_pctcpu);
 }
 
@@ -1603,7 +1607,7 @@
 	int realstathz;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	delta = 0;
 	realstathz = stathz ? stathz : hz;
 	if (ts->ts_cpticks != 0) {
@@ -1628,7 +1632,7 @@
 sched_estcpu(struct thread *td)
 {
 	
-	return (td->td_sched->ts_estcpu);
+	return (td_get_sched(td)->ts_estcpu);
 }
 
 /*
@@ -1707,7 +1711,7 @@
 #ifdef KTR
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if (ts->ts_name[0] == '\0')
 		snprintf(ts->ts_name, sizeof(ts->ts_name),
 		    "%s tid %d", td->td_name, td->td_tid);
@@ -1723,7 +1727,7 @@
 {
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	ts->ts_name[0] = '\0';
 }
 #endif
@@ -1741,7 +1745,7 @@
 	 * Set the TSF_AFFINITY flag if there is at least one CPU this
 	 * thread can't run on.
 	 */
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	ts->ts_flags &= ~TSF_AFFINITY;
 	CPU_FOREACH(cpu) {
 		if (!THREAD_CAN_SCHED(td, cpu)) {
Index: head/sys/kern/sched_ule.c
===================================================================
--- head/sys/kern/sched_ule.c
+++ head/sys/kern/sched_ule.c
@@ -106,12 +106,14 @@
 #define	TSF_BOUND	0x0001		/* Thread can not migrate. */
 #define	TSF_XFERABLE	0x0002		/* Thread was added as transferable. */
 
-static struct td_sched td_sched0;
-
 #define	THREAD_CAN_MIGRATE(td)	((td)->td_pinned == 0)
 #define	THREAD_CAN_SCHED(td, cpu)	\
     CPU_ISSET((cpu), &(td)->td_cpuset->cs_mask)
 
+_Static_assert(sizeof(struct thread) + sizeof(struct td_sched) <=
+    sizeof(struct thread0_storage),
+    "increase struct thread0_storage.t0st_sched size");
+
 /*
  * Priority ranges used for interactive and non-interactive timeshare
  * threads.  The timeshare priorities are split up into four ranges.
@@ -460,7 +462,7 @@
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 
 	pri = td->td_priority;
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	TD_SET_RUNQ(td);
 	if (THREAD_CAN_MIGRATE(td)) {
 		tdq->tdq_transferable++;
@@ -506,7 +508,7 @@
 {
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	TDQ_LOCK_ASSERT(tdq, MA_OWNED);
 	KASSERT(ts->ts_runq != NULL,
 	    ("tdq_runq_remove: thread %p null ts_runq", td));
@@ -962,7 +964,7 @@
 	td = tdq_steal(tdq, cpu);
 	if (td == NULL)
 		return (0);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	/*
 	 * Although the run queue is locked the thread may be blocked.  Lock
 	 * it to clear this and acquire the run-queue lock.
@@ -1046,7 +1048,7 @@
 
 	if (tdq->tdq_ipipending)
 		return;
-	cpu = td->td_sched->ts_cpu;
+	cpu = td_get_sched(td)->ts_cpu;
 	pri = td->td_priority;
 	ctd = pcpu_find(cpu)->pc_curthread;
 	if (!sched_shouldpreempt(pri, ctd->td_priority, 1))
@@ -1174,7 +1176,7 @@
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	tdq = TDQ_CPU(cpu);
-	td->td_sched->ts_cpu = cpu;
+	td_get_sched(td)->ts_cpu = cpu;
 	/*
 	 * If the lock matches just return the queue.
 	 */
@@ -1221,7 +1223,7 @@
 	int cpu, pri, self;
 
 	self = PCPU_GET(cpuid);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if (smp_started == 0)
 		return (self);
 	/*
@@ -1472,7 +1474,7 @@
 	struct td_sched *ts;
 	int div;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	/*
 	 * The score is only needed if this is likely to be an interactive
 	 * task.  Don't go through the expense of computing it if there's
@@ -1537,16 +1539,16 @@
 		    pri, score));
 	} else {
 		pri = SCHED_PRI_MIN;
-		if (td->td_sched->ts_ticks)
-			pri += min(SCHED_PRI_TICKS(td->td_sched),
+		if (td_get_sched(td)->ts_ticks)
+			pri += min(SCHED_PRI_TICKS(td_get_sched(td)),
 			    SCHED_PRI_RANGE - 1);
 		pri += SCHED_PRI_NICE(td->td_proc->p_nice);
 		KASSERT(pri >= PRI_MIN_BATCH && pri <= PRI_MAX_BATCH,
 		    ("sched_priority: invalid priority %d: nice %d, " 
 		    "ticks %d ftick %d ltick %d tick pri %d",
-		    pri, td->td_proc->p_nice, td->td_sched->ts_ticks,
-		    td->td_sched->ts_ftick, td->td_sched->ts_ltick,
-		    SCHED_PRI_TICKS(td->td_sched)));
+		    pri, td->td_proc->p_nice, td_get_sched(td)->ts_ticks,
+		    td_get_sched(td)->ts_ftick, td_get_sched(td)->ts_ltick,
+		    SCHED_PRI_TICKS(td_get_sched(td))));
 	}
 	sched_user_prio(td, pri);
 
@@ -1564,7 +1566,7 @@
 	struct td_sched *ts;
 	u_int sum;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	sum = ts->ts_runtime + ts->ts_slptime;
 	if (sum < SCHED_SLP_RUN_MAX)
 		return;
@@ -1606,14 +1608,16 @@
 static void
 sched_interact_fork(struct thread *td)
 {
+	struct td_sched *ts;
 	int ratio;
 	int sum;
 
-	sum = td->td_sched->ts_runtime + td->td_sched->ts_slptime;
+	ts = td_get_sched(td);
+	sum = ts->ts_runtime + ts->ts_slptime;
 	if (sum > SCHED_SLP_RUN_FORK) {
 		ratio = sum / SCHED_SLP_RUN_FORK;
-		td->td_sched->ts_runtime /= ratio;
-		td->td_sched->ts_slptime /= ratio;
+		ts->ts_runtime /= ratio;
+		ts->ts_slptime /= ratio;
 	}
 }
 
@@ -1623,15 +1627,15 @@
 void
 schedinit(void)
 {
+	struct td_sched *ts0;
 
 	/*
-	 * Set up the scheduler specific parts of proc0.
+	 * Set up the scheduler specific parts of thread0.
 	 */
-	proc0.p_sched = NULL; /* XXX */
-	thread0.td_sched = &td_sched0;
-	td_sched0.ts_ltick = ticks;
-	td_sched0.ts_ftick = ticks;
-	td_sched0.ts_slice = 0;
+	ts0 = td_get_sched(&thread0);
+	ts0->ts_ltick = ticks;
+	ts0->ts_ftick = ticks;
+	ts0->ts_slice = 0;
 }
 
 /*
@@ -1694,7 +1698,7 @@
 		SDT_PROBE4(sched, , , lend__pri, td, td->td_proc, prio, 
 		    curthread);
 	} 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	if (td->td_priority == prio)
 		return;
@@ -1829,7 +1833,7 @@
 {
 	struct tdq *tdn;
 
-	tdn = TDQ_CPU(td->td_sched->ts_cpu);
+	tdn = TDQ_CPU(td_get_sched(td)->ts_cpu);
 #ifdef SMP
 	tdq_load_rem(tdq, td);
 	/*
@@ -1888,7 +1892,7 @@
 
 	cpuid = PCPU_GET(cpuid);
 	tdq = TDQ_CPU(cpuid);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	mtx = td->td_lock;
 	sched_pctcpu_update(ts, 1);
 	ts->ts_rltick = ticks;
@@ -1948,7 +1952,7 @@
 		SDT_PROBE2(sched, , , off__cpu, newtd, newtd->td_proc);
 		lock_profile_release_lock(&TDQ_LOCKPTR(tdq)->lock_object);
 		TDQ_LOCKPTR(tdq)->mtx_lock = (uintptr_t)newtd;
-		sched_pctcpu_update(newtd->td_sched, 0);
+		sched_pctcpu_update(td_get_sched(newtd), 0);
 
 #ifdef KDTRACE_HOOKS
 		/*
@@ -2038,7 +2042,7 @@
 	int slptick;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	td->td_flags &= ~TDF_CANSWAP;
 	/*
 	 * If we slept for more than a tick update our interactivity and
@@ -2066,14 +2070,14 @@
 sched_fork(struct thread *td, struct thread *child)
 {
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	sched_pctcpu_update(td->td_sched, 1);
+	sched_pctcpu_update(td_get_sched(td), 1);
 	sched_fork_thread(td, child);
 	/*
 	 * Penalize the parent and child for forking.
 	 */
 	sched_interact_fork(child);
 	sched_priority(child);
-	td->td_sched->ts_runtime += tickincr;
+	td_get_sched(td)->ts_runtime += tickincr;
 	sched_interact_update(td);
 	sched_priority(td);
 }
@@ -2093,8 +2097,8 @@
 	/*
 	 * Initialize child.
 	 */
-	ts = td->td_sched;
-	ts2 = child->td_sched;
+	ts = td_get_sched(td);
+	ts2 = td_get_sched(child);
 	child->td_oncpu = NOCPU;
 	child->td_lastcpu = NOCPU;
 	child->td_lock = TDQ_LOCKPTR(tdq);
@@ -2169,7 +2173,7 @@
 	 * launch expensive things to mark their children as expensive.
 	 */
 	thread_lock(td);
-	td->td_sched->ts_runtime += child->td_sched->ts_runtime;
+	td_get_sched(td)->ts_runtime += td_get_sched(child)->ts_runtime;
 	sched_interact_update(td);
 	sched_priority(td);
 	thread_unlock(td);
@@ -2264,7 +2268,7 @@
 		if (TAILQ_EMPTY(&tdq->tdq_timeshare.rq_queues[tdq->tdq_ridx]))
 			tdq->tdq_ridx = tdq->tdq_idx;
 	}
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	sched_pctcpu_update(ts, 1);
 	if (td->td_pri_class & PRI_FIFO_BIT)
 		return;
@@ -2273,7 +2277,7 @@
 		 * We used a tick; charge it to the thread so
 		 * that we can compute our interactivity.
 		 */
-		td->td_sched->ts_runtime += tickincr;
+		td_get_sched(td)->ts_runtime += tickincr;
 		sched_interact_update(td);
 		sched_priority(td);
 	}
@@ -2455,7 +2459,7 @@
 	KTR_STATE1(KTR_SCHED, "thread", sched_tdname(td), "runq rem",
 	    "prio:%d", td->td_priority);
 	SDT_PROBE3(sched, , , dequeue, td, td->td_proc, NULL);
-	tdq = TDQ_CPU(td->td_sched->ts_cpu);
+	tdq = TDQ_CPU(td_get_sched(td)->ts_cpu);
 	TDQ_LOCK_ASSERT(tdq, MA_OWNED);
 	MPASS(td->td_lock == TDQ_LOCKPTR(tdq));
 	KASSERT(TD_ON_RUNQ(td),
@@ -2477,9 +2481,7 @@
 	struct td_sched *ts;
 
 	pctcpu = 0;
-	ts = td->td_sched;
-	if (ts == NULL)
-		return (0);
+	ts = td_get_sched(td);
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	sched_pctcpu_update(ts, TD_IS_RUNNING(td));
@@ -2505,7 +2507,7 @@
 	struct td_sched *ts;
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if (THREAD_CAN_SCHED(td, ts->ts_cpu))
 		return;
 	if (TD_ON_RUNQ(td)) {
@@ -2536,7 +2538,7 @@
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED|MA_NOTRECURSED);
 	KASSERT(td == curthread, ("sched_bind: can only bind curthread"));
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if (ts->ts_flags & TSF_BOUND)
 		sched_unbind(td);
 	KASSERT(THREAD_CAN_MIGRATE(td), ("%p must be migratable", td));
@@ -2559,7 +2561,7 @@
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
 	KASSERT(td == curthread, ("sched_unbind: can only bind curthread"));
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if ((ts->ts_flags & TSF_BOUND) == 0)
 		return;
 	ts->ts_flags &= ~TSF_BOUND;
@@ -2570,7 +2572,7 @@
 sched_is_bound(struct thread *td)
 {
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
-	return (td->td_sched->ts_flags & TSF_BOUND);
+	return (td_get_sched(td)->ts_flags & TSF_BOUND);
 }
 
 /*
@@ -2761,7 +2763,7 @@
 #ifdef KTR
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	if (ts->ts_name[0] == '\0')
 		snprintf(ts->ts_name, sizeof(ts->ts_name),
 		    "%s tid %d", td->td_name, td->td_tid);
@@ -2777,7 +2779,7 @@
 {
 	struct td_sched *ts;
 
-	ts = td->td_sched;
+	ts = td_get_sched(td);
 	ts->ts_name[0] = '\0';
 }
 #endif
Index: head/sys/mips/mips/locore.S
===================================================================
--- head/sys/mips/mips/locore.S
+++ head/sys/mips/mips/locore.S
@@ -173,7 +173,7 @@
 	jal	_C_LABEL(platform_start)
 	nop
 
-	PTR_LA	sp, _C_LABEL(thread0)
+	PTR_LA	sp, _C_LABEL(thread0_st)
 	PTR_L	a0, TD_PCB(sp)
 	REG_LI	t0, ~7
 	and	a0, a0, t0
Index: head/sys/sys/proc.h
===================================================================
--- head/sys/sys/proc.h
+++ head/sys/sys/proc.h
@@ -325,7 +325,6 @@
 	int		td_kstack_pages; /* (a) Size of the kstack. */
 	volatile u_int	td_critnest;	/* (k*) Critical section nest level. */
 	struct mdthread td_md;		/* (k) Any machine-dependent fields. */
-	struct td_sched	*td_sched;	/* (*) Scheduler-specific data. */
 	struct kaudit_record	*td_ar;	/* (k) Active audit record, if any. */
 	struct lpohead	td_lprof[2];	/* (a) lock profiling objects. */
 	struct kdtrace_thread	*td_dtrace; /* (*) DTrace-specific data. */
@@ -341,6 +340,11 @@
 	int		td_oncpu;	/* (t) Which cpu we are on. */
 };
 
+struct thread0_storage {
+	struct thread t0st_thread;
+	uint64_t t0st_sched[10];
+};
+
 struct mtx *thread_lock_block(struct thread *);
 void thread_lock_unblock(struct thread *, struct mtx *);
 void thread_lock_set(struct thread *, struct mtx *);
@@ -616,7 +620,6 @@
 	struct proc	*p_leader;	/* (b) */
 	void		*p_emuldata;	/* (c) Emulator state data. */
 	struct label	*p_label;	/* (*) Proc (not subject) MAC label. */
-	struct p_sched	*p_sched;	/* (*) Scheduler-specific data. */
 	STAILQ_HEAD(, ktr_request)	p_ktr;	/* (o) KTR event queue. */
 	LIST_HEAD(, mqueue_notifier)	p_mqnotifier; /* (c) mqueue notifiers.*/
 	struct kdtrace_proc	*p_dtrace; /* (*) DTrace-specific data. */
@@ -890,7 +893,8 @@
 extern struct sx proctree_lock;
 extern struct mtx ppeers_lock;
 extern struct proc proc0;		/* Process slot for swapper. */
-extern struct thread thread0;		/* Primary thread in proc0. */
+extern struct thread0_storage thread0_st;	/* Primary thread in proc0. */
+#define	thread0 (thread0_st.t0st_thread)
 extern struct vmspace vmspace0;		/* VM space for proc0. */
 extern int hogticks;			/* Limit on kernel cpu hogs. */
 extern int lastpid;
@@ -1065,6 +1069,13 @@
 	curthread->td_pflags &= save;
 }
 
+static __inline __pure2 struct td_sched *
+td_get_sched(struct thread *td)
+{
+
+	return ((struct td_sched *)&td[1]);
+}
+
 #endif	/* _KERNEL */
 
 #endif	/* !_SYS_PROC_H_ */