diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 74b925c8d534e76de6a3dcd41bba3ab739ad642c..b1c6cbe02db7a2c950a143279c95e80eec33ed2d 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -1813,7 +1813,7 @@ static void tracehook_report_syscall(struct pt_regs *regs,
 
 int syscall_trace_enter(struct pt_regs *regs)
 {
-	unsigned long flags = READ_ONCE(current_thread_info()->flags);
+	unsigned long flags = read_thread_flags();
 
 	if (flags & (_TIF_SYSCALL_EMU | _TIF_SYSCALL_TRACE)) {
 		tracehook_report_syscall(regs, PTRACE_SYSCALL_ENTER);
@@ -1836,7 +1836,7 @@ int syscall_trace_enter(struct pt_regs *regs)
 
 void syscall_trace_exit(struct pt_regs *regs)
 {
-	unsigned long flags = READ_ONCE(current_thread_info()->flags);
+	unsigned long flags = read_thread_flags();
 
 	audit_syscall_exit(regs);
 
diff --git a/arch/arm64/kernel/signal.c b/arch/arm64/kernel/signal.c
index 17cb54d1e420bc90a23ab90812bf34301de65786..3a99b096abec6f6b8ffd4e19e1210e78b0f42952 100644
--- a/arch/arm64/kernel/signal.c
+++ b/arch/arm64/kernel/signal.c
@@ -724,7 +724,7 @@ asmlinkage void do_notify_resume(struct pt_regs *regs,
 		}
 
 		local_daif_mask();
-		thread_flags = READ_ONCE(current_thread_info()->flags);
+		thread_flags = read_thread_flags();
 	} while (thread_flags & _TIF_WORK_MASK);
 }
 
diff --git a/arch/arm64/kernel/syscall.c b/arch/arm64/kernel/syscall.c
index 2a106e67a4cbf1c85ffc6146b6c3811da484b1c0..330cec1f7768b35b2303ab1388cfdb58f61a75fb 100644
--- a/arch/arm64/kernel/syscall.c
+++ b/arch/arm64/kernel/syscall.c
@@ -109,7 +109,7 @@ static void cortex_a76_erratum_1463225_svc_handler(void) { }
 static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
 			   const syscall_fn_t syscall_table[])
 {
-	unsigned long flags = current_thread_info()->flags;
+	unsigned long flags = read_thread_flags();
 
 	regs->orig_x0 = regs->regs[0];
 	regs->syscallno = scno;
@@ -176,11 +176,9 @@ static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
 	 * exit regardless, as the old entry assembly did.
 	 */
 	if (!has_syscall_work(flags) && !IS_ENABLED(CONFIG_DEBUG_RSEQ)) {
-		local_daif_mask();
-		flags = current_thread_info()->flags;
+		flags = read_thread_flags();
 		if (!has_syscall_work(flags) && !(flags & _TIF_SINGLESTEP))
 			return;
-		local_daif_restore(DAIF_PROCCTX);
 	}
 
 trace_exit:
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 76b37d1bcac16861ed16636a959c5350336fceda..7f2f299bff8d2365d1e23306e9b6d4db161e1941 100755
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -370,7 +370,7 @@ void arch_setup_new_exec(void)
 		clear_thread_flag(TIF_SSBD);
 		task_clear_spec_ssb_disable(current);
 		task_clear_spec_ssb_noexec(current);
-		speculation_ctrl_update(task_thread_info(current)->flags);
+		speculation_ctrl_update(read_thread_flags());
 	}
 }
 
@@ -622,7 +622,7 @@ static unsigned long speculation_ctrl_update_tif(struct task_struct *tsk)
 			clear_tsk_thread_flag(tsk, TIF_SPEC_IB);
 	}
 	/* Return the updated threadinfo flags*/
-	return task_thread_info(tsk)->flags;
+	return read_task_thread_flags(tsk);
 }
 
 void speculation_ctrl_update(unsigned long tif)
@@ -658,8 +658,8 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p)
 {
 	unsigned long tifp, tifn;
 
-	tifn = READ_ONCE(task_thread_info(next_p)->flags);
-	tifp = READ_ONCE(task_thread_info(prev_p)->flags);
+	tifn = read_task_thread_flags(next_p);
+	tifp = read_task_thread_flags(prev_p);
 
 	switch_to_bitmap(tifp);
 
diff --git a/arch/x86/kernel/process.h b/arch/x86/kernel/process.h
index 1d0797b2338a291a9ba6ec89f0ec7aead1b4fbd0..76b547b832320b0570ff8e0aa4a6ad1171db5348 100644
--- a/arch/x86/kernel/process.h
+++ b/arch/x86/kernel/process.h
@@ -13,8 +13,8 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p);
 static inline void switch_to_extra(struct task_struct *prev,
 				   struct task_struct *next)
 {
-	unsigned long next_tif = task_thread_info(next)->flags;
-	unsigned long prev_tif = task_thread_info(prev)->flags;
+	unsigned long next_tif = read_task_thread_flags(next);
+	unsigned long prev_tif = read_task_thread_flags(prev);
 
 	if (IS_ENABLED(CONFIG_SMP)) {
 		/*
diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c
index f4b162f273f52ca0ca6f404f6c30920886dab7fd..86a66efa26dc3394aa5269b4886f00669b2ab1ac 100644
--- a/arch/x86/mm/tlb.c
+++ b/arch/x86/mm/tlb.c
@@ -318,7 +318,7 @@ void switch_mm(struct mm_struct *prev, struct mm_struct *next,
 
 static unsigned long mm_mangle_tif_spec_ib(struct task_struct *next)
 {
-	unsigned long next_tif = task_thread_info(next)->flags;
+	unsigned long next_tif = read_task_thread_flags(next);
 	unsigned long ibpb = (next_tif >> TIF_SPEC_IB) & LAST_USER_MM_IBPB;
 
 	return (unsigned long)next->mm | ibpb;
diff --git a/drivers/irqchip/irq-gic-v3.c b/drivers/irqchip/irq-gic-v3.c
index 6cee83bf2e5e2dc4d42d3e7b05fcb3d80eff2a07..ef7ddfefcd00df93b1555cff1cad1a3430bb9fab 100644
--- a/drivers/irqchip/irq-gic-v3.c
+++ b/drivers/irqchip/irq-gic-v3.c
@@ -1299,7 +1299,7 @@ static void gic_ipi_send_mask(struct irq_data *d, const struct cpumask *mask)
 	 * Ensure that stores to Normal memory are visible to the
 	 * other CPUs before issuing the IPI.
 	 */
-	wmb();
+	dsb(ishst);
 
 	for_each_cpu(cpu, mask) {
 		u64 cluster_id = MPIDR_TO_SGI_CLUSTER_ID(cpu_logical_map(cpu));
diff --git a/include/linux/entry-kvm.h b/include/linux/entry-kvm.h
index 859fdfd7d46c97fd7b50dc3ee5fb09fe4c198a31..4fe38c2d27066ac136bd53efa7ee062e850ce536 100644
--- a/include/linux/entry-kvm.h
+++ b/include/linux/entry-kvm.h
@@ -60,7 +60,7 @@ int xfer_to_guest_mode_handle_work(struct kvm_vcpu *vcpu);
  */
 static inline bool __xfer_to_guest_mode_work_pending(void)
 {
-	unsigned long ti_work = READ_ONCE(current_thread_info()->flags);
+	unsigned long ti_work = read_thread_flags();
 
 	return !!(ti_work & XFER_TO_GUEST_MODE_WORK);
 }
diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
index 19f76d87f20fcd6aca9f00441ef5954c7c9ba1a5..af3ca6429d77e0e7957cb496b1eb1444a1c5fd85 100644
--- a/include/linux/thread_info.h
+++ b/include/linux/thread_info.h
@@ -36,6 +36,21 @@ static inline long set_restart_fn(struct restart_block *restart,
 
 #define THREADINFO_GFP		(GFP_KERNEL_ACCOUNT | __GFP_ZERO)
 
+/*
+ * This may be used in noinstr code, and needs to be __always_inline to prevent
+ * inadvertent instrumentation.
+ */
+static __always_inline unsigned long read_ti_thread_flags(struct thread_info *ti)
+{
+	return READ_ONCE(ti->flags);
+}
+
+#define read_thread_flags() \
+	read_ti_thread_flags(current_thread_info())
+
+#define read_task_thread_flags(t) \
+	read_ti_thread_flags(task_thread_info(t))
+
 #define tif_need_resched() test_thread_flag(TIF_NEED_RESCHED)
 
 #ifndef CONFIG_HAVE_ARCH_WITHIN_STACK_FRAMES
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index a028b28daed5b587334d5880b1680adfb2ee2625..7e4fc453da7da5fea335b2b894c605309800a960 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -186,7 +186,7 @@ static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
 		/* Check if any of the above work has queued a deferred wakeup */
 		rcu_nocb_flush_deferred_wakeup();
 
-		ti_work = READ_ONCE(current_thread_info()->flags);
+		ti_work = read_thread_flags();
 	}
 
 	/* Return the latest work state for arch_exit_to_user_mode() */
@@ -195,7 +195,7 @@ static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
 
 static void exit_to_user_mode_prepare(struct pt_regs *regs)
 {
-	unsigned long ti_work = READ_ONCE(current_thread_info()->flags);
+	unsigned long ti_work = read_thread_flags();
 
 	lockdep_assert_irqs_disabled();
 
diff --git a/kernel/entry/kvm.c b/kernel/entry/kvm.c
index 049fd06b4c3de348d3e47a9691880c3272a81c8f..6995b7ee63b9784c2105cf0174c166a503ab5a61 100644
--- a/kernel/entry/kvm.c
+++ b/kernel/entry/kvm.c
@@ -28,7 +28,7 @@ static int xfer_to_guest_mode_work(struct kvm_vcpu *vcpu, unsigned long ti_work)
 		if (ret)
 			return ret;
 
-		ti_work = READ_ONCE(current_thread_info()->flags);
+		ti_work = read_thread_flags();
 	} while (ti_work & XFER_TO_GUEST_MODE_WORK || need_resched());
 	return 0;
 }
@@ -45,7 +45,7 @@ int xfer_to_guest_mode_handle_work(struct kvm_vcpu *vcpu)
 	 * disabled in the inner loop before going into guest mode. No need
 	 * to disable interrupts here.
 	 */
-	ti_work = READ_ONCE(current_thread_info()->flags);
+	ti_work = read_thread_flags();
 	if (!(ti_work & XFER_TO_GUEST_MODE_WORK))
 		return 0;
 
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 20efc7cccf199e417e338d59a1b4b55aff1e3652..81b58ce710ba8de04792ee48f9a37630619f8656 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7524,7 +7524,7 @@ void sched_show_task(struct task_struct *p)
 	rcu_read_unlock();
 	pr_cont(" stack:%5lu pid:%5d ppid:%6d flags:0x%08lx\n",
 		free, task_pid_nr(p), ppid,
-		(unsigned long)task_thread_info(p)->flags);
+		read_task_thread_flags(p));
 
 	print_worker_info(KERN_INFO, p);
 	show_stack(p, NULL, KERN_INFO);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index d3c4b945c0196204c80ff36f51189171d5592768..fbabad1e193aa7a7deb8fb31814c708f24d138b3 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6190,9 +6190,11 @@ static inline bool test_idle_cores(int cpu, bool def)
 {
 	struct sched_domain_shared *sds;
 
-	sds = rcu_dereference(per_cpu(sd_llc_shared, cpu));
-	if (sds)
-		return READ_ONCE(sds->has_idle_cores);
+	if (static_branch_likely(&sched_smt_present)) {
+		sds = rcu_dereference(per_cpu(sd_llc_shared, cpu));
+		if (sds)
+			return READ_ONCE(sds->has_idle_cores);
+	}
 
 	return def;
 }