diff --git a/sys/amd64/amd64/cpu_switch.S b/sys/amd64/amd64/cpu_switch.S
--- a/sys/amd64/amd64/cpu_switch.S
+++ b/sys/amd64/amd64/cpu_switch.S
@@ -31,6 +31,8 @@
  * SUCH DAMAGE.
  */
 
+#include "opt_global.h"
+
 #include <machine/asmacros.h>
 #include <machine/specialreg.h>
 
@@ -145,6 +147,10 @@
 	 * to load up the rest of the next context.
 	 */
 
+#ifdef PERTHREAD_SSP
+	movq	TD_MD_CANARY(%r12),%rax
+	movq	%rax,PCPU(CANARY)
+#endif
 	/* Skip loading LDT and user fsbase/gsbase for kthreads */
 	testl	$TDP_KTHREAD,TD_PFLAGS(%r12)
 	jnz	do_kthread
diff --git a/sys/amd64/amd64/genassym.c b/sys/amd64/amd64/genassym.c
--- a/sys/amd64/amd64/genassym.c
+++ b/sys/amd64/amd64/genassym.c
@@ -85,6 +85,7 @@
 ASSYM(TD_MD, offsetof(struct thread, td_md));
 ASSYM(TD_MD_PCB, offsetof(struct thread, td_md.md_pcb));
 ASSYM(TD_MD_STACK_BASE, offsetof(struct thread, td_md.md_stack_base));
+ASSYM(TD_MD_CANARY, offsetof(struct thread, td_md.md_canary));
 
 ASSYM(TDP_CALLCHAIN, TDP_CALLCHAIN);
 ASSYM(TDP_KTHREAD, TDP_KTHREAD);
@@ -257,6 +258,7 @@
 ASSYM(PC_MDS_TMP, offsetof(struct pcpu, pc_mds_tmp));
 ASSYM(PC_MDS_BUF, offsetof(struct pcpu, pc_mds_buf));
 ASSYM(PC_MDS_BUF64, offsetof(struct pcpu, pc_mds_buf64));
+ASSYM(PC_CANARY, offsetof(struct pcpu, pc_canary));
 
 ASSYM(LA_EOI, LAPIC_EOI * LAPIC_MEM_MUL);
 ASSYM(LA_ISR, LAPIC_ISR0 * LAPIC_MEM_MUL);
diff --git a/sys/amd64/amd64/locore.S b/sys/amd64/amd64/locore.S
--- a/sys/amd64/amd64/locore.S
+++ b/sys/amd64/amd64/locore.S
@@ -82,6 +82,10 @@
 	xorq	%rbp, %rbp
 
 	call	hammer_time		/* set up cpu for unix operation */
+#ifdef PERTHREAD_SSP
+	movq	boot_canary, %rbx
+	movq	%rbx, PCPU(CANARY)
+#endif
 	movq	%rax,%rsp		/* set up kstack for mi_startup() */
 	call	mi_startup		/* autoconfiguration, mountroot etc */
 0:	hlt
diff --git a/sys/amd64/amd64/machdep.c b/sys/amd64/amd64/machdep.c
--- a/sys/amd64/amd64/machdep.c
+++ b/sys/amd64/amd64/machdep.c
@@ -224,6 +224,10 @@
 
 bool efi_boot;
 
+#ifdef PERTHREAD_SSP
+uintptr_t boot_canary = 0x6543d292157d3053UL;
+#endif
+
 static void
 cpu_startup(void *dummy)
 {
@@ -1378,6 +1382,9 @@
 	bzero(thread0.td_kstack, kstack0_sz);
 	cpu_thread_new_kstack(&thread0);
 	physfree += kstack0_sz;
+#ifdef PERTHREAD_SSP
+	thread0.td_md.md_canary = boot_canary;
+#endif
 
 	/*
 	 * Initialize enough of thread0 for delayed invalidation to
diff --git a/sys/amd64/amd64/mp_machdep.c b/sys/amd64/amd64/mp_machdep.c
--- a/sys/amd64/amd64/mp_machdep.c
+++ b/sys/amd64/amd64/mp_machdep.c
@@ -105,6 +105,8 @@
 smp_targeted_tlb_shootdown_t smp_targeted_tlb_shootdown =
     &smp_targeted_tlb_shootdown_native;
 
+extern uintptr_t boot_canary;
+
 /*
  * Local data and functions.
  */
@@ -231,6 +233,7 @@
 	pc->pc_pcid_gen = 1;
 	pc->pc_kpmap_store.pm_pcid = PMAP_PCID_KERN;
 	pc->pc_kpmap_store.pm_gen = 1;
+	pc->pc_canary = boot_canary;
 
 	pc->pc_smp_tlb_gen = 1;
 
diff --git a/sys/amd64/amd64/vm_machdep.c b/sys/amd64/amd64/vm_machdep.c
--- a/sys/amd64/amd64/vm_machdep.c
+++ b/sys/amd64/amd64/vm_machdep.c
@@ -201,6 +201,9 @@
 		 */
 		td2->td_frame->tf_rflags &= ~PSL_T;
 	}
+#ifdef PERTHREAD_SSP
+	arc4random_buf(&td2->td_md.md_canary, sizeof(td2->td_md.md_canary));
+#endif
 }
 
 /*
diff --git a/sys/amd64/conf/GENERIC b/sys/amd64/conf/GENERIC
--- a/sys/amd64/conf/GENERIC
+++ b/sys/amd64/conf/GENERIC
@@ -91,6 +91,7 @@
 options 	RACCT			# Resource accounting framework
 options 	RACCT_DEFAULT_TO_DISABLED # Set kern.racct.enable=0 by default
 options 	RCTL			# Resource limits
+options 	PERTHREAD_SSP		# Per-thread SSP canary
 
 # Debugging support.  Always need this:
 options 	KDB			# Enable kernel debugger support.
diff --git a/sys/amd64/include/pcpu.h b/sys/amd64/include/pcpu.h
--- a/sys/amd64/include/pcpu.h
+++ b/sys/amd64/include/pcpu.h
@@ -101,7 +101,8 @@
 	u_int	pc_small_core;						\
 	u_int	pc_pcid_invlpg_workaround;				\
 	struct pmap_pcid pc_kpmap_store;				\
-	char	__pad[2900]		/* pad to UMA_PCPU_ALLOC_SIZE */
+	uintptr_t pc_canary;						\
+	char	__pad[2892]		/* pad to UMA_PCPU_ALLOC_SIZE */
 
 #define	PC_DBREG_CMD_NONE	0
 #define	PC_DBREG_CMD_LOAD	1
diff --git a/sys/amd64/include/proc.h b/sys/amd64/include/proc.h
--- a/sys/amd64/include/proc.h
+++ b/sys/amd64/include/proc.h
@@ -77,6 +77,7 @@
 	struct pcb md_pcb;
 	void *md_stack_base;
 	void *md_usr_fpu_save;
+	register_t md_canary;
 };
 
 struct mdproc {
diff --git a/sys/conf/Makefile.amd64 b/sys/conf/Makefile.amd64
--- a/sys/conf/Makefile.amd64
+++ b/sys/conf/Makefile.amd64
@@ -35,6 +35,23 @@
 CFLAGS+=	-fno-omit-frame-pointer -mno-omit-leaf-frame-pointer
 .endif
 
+.if ${MK_SSP} != "no"
+PERTHREAD_SSP_ENABLED!=	grep PERTHREAD_SSP opt_global.h || true ; echo
+. if !empty(PERTHREAD_SSP_ENABLED)
+.  if (${COMPILER_TYPE} == "clang" && ${OSRELDATE} >= 1600019) || \
+  ${COMPILER_TYPE} == "gcc"
+PCPU_CANARY_OFFSET!= grep PC_CANARY assym.inc | cut -f3 | xargs printf "%d"
+AMD64_SSP_CFLAGS+= -mstack-protector-guard=tls
+AMD64_SSP_CFLAGS+= -mstack-protector-guard-reg=gs
+AMD64_SSP_CFLAGS+= -mstack-protector-guard-offset=${PCPU_CANARY_OFFSET}
+.  else
+.   warning "Compiler is too old to support PERTHREAD_SSP"
+.  endif
+CFLAGS+= ${AMD64_SSP_CFLAGS}
+ARCH_FLAGS+= ${AMD64_SSP_CFLAGS}
+. endif
+.endif
+
 %BEFORE_DEPEND
 
 %OBJS
diff --git a/sys/conf/options.amd64 b/sys/conf/options.amd64
--- a/sys/conf/options.amd64
+++ b/sys/conf/options.amd64
@@ -8,6 +8,7 @@
 MAXMEM
 MPTABLE_FORCE_HTT
 NKPT			opt_pmap.h
+PERTHREAD_SSP		opt_global.h
 PV_STATS		opt_pmap.h
 
 # Options for emulators.  These should only be used at config time, so
diff --git a/sys/kern/kern_thread.c b/sys/kern/kern_thread.c
--- a/sys/kern/kern_thread.c
+++ b/sys/kern/kern_thread.c
@@ -92,7 +92,7 @@
     "struct thread KBI td_pflags");
 _Static_assert(offsetof(struct thread, td_frame) == 0x4e8,
     "struct thread KBI td_frame");
-_Static_assert(offsetof(struct thread, td_emuldata) == 0x6f0,
+_Static_assert(offsetof(struct thread, td_emuldata) == 0x700,
     "struct thread KBI td_emuldata");
 _Static_assert(offsetof(struct proc, p_flag) == 0xb8,
     "struct proc KBI p_flag");