x86/process: Move the buffer clearing before MONITOR

Move the VERW clearing before the MONITOR so that VERW doesn't disarm it
and the machine never enters C1.

Original idea by Kim Phillips <kim.phillips@amd.com>.

Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>

Author: bp3tk0v

arch/x86/include/asm/mwait.h

@@ -43,8 +43,6 @@ static __always_inline void __monitorx(const void *eax, u32 ecx, u32 edx)
 
 static __always_inline void __mwait(u32 eax, u32 ecx)
 {
-	x86_idle_clear_cpu_buffers();
-
 	/*
 	 * Use the instruction mnemonic with implicit operands, as the LLVM
 	 * assembler fails to assemble the mnemonic with explicit operands:
@@ -98,7 +96,6 @@ static __always_inline void __mwaitx(u32 eax, u32 ebx, u32 ecx)
  */
 static __always_inline void __sti_mwait(u32 eax, u32 ecx)
 {
-	x86_idle_clear_cpu_buffers();
 
 	asm volatile("sti; mwait" :: "a" (eax), "c" (ecx));
 }
@@ -115,21 +112,29 @@ static __always_inline void __sti_mwait(u32 eax, u32 ecx)
  */
 static __always_inline void mwait_idle_with_hints(u32 eax, u32 ecx)
 {
+	if (need_resched())
+		return;
+
+	x86_idle_clear_cpu_buffers();
+
 	if (static_cpu_has_bug(X86_BUG_MONITOR) || !current_set_polling_and_test()) {
 		const void *addr = &current_thread_info()->flags;
 
 		alternative_input("", "clflush (%[addr])", X86_BUG_CLFLUSH_MONITOR, [addr] "a" (addr));
 		__monitor(addr, 0, 0);
 
-		if (!need_resched()) {
-			if (ecx & 1) {
-				__mwait(eax, ecx);
-			} else {
-				__sti_mwait(eax, ecx);
-				raw_local_irq_disable();
-			}
+		if (need_resched())
+			goto out;
+
+		if (ecx & 1) {
+			__mwait(eax, ecx);
+		} else {
+			__sti_mwait(eax, ecx);
+			raw_local_irq_disable();
 		}
 	}
+
+out:
 	current_clr_polling();
 }
 

arch/x86/kernel/process.c

@@ -907,16 +907,24 @@ static __init bool prefer_mwait_c1_over_halt(void)
  */
 static __cpuidle void mwait_idle(void)
 {
+	if (need_resched())
+		return;
+
+	x86_idle_clear_cpu_buffers();
+
 	if (!current_set_polling_and_test()) {
 		const void *addr = &current_thread_info()->flags;
 
 		alternative_input("", "clflush (%[addr])", X86_BUG_CLFLUSH_MONITOR, [addr] "a" (addr));
 		__monitor(addr, 0, 0);
-		if (!need_resched()) {
-			__sti_mwait(0, 0);
-			raw_local_irq_disable();
-		}
+		if (need_resched())
+			goto out;
+
+		__sti_mwait(0, 0);
+		raw_local_irq_disable();
 	}
+
+out:
 	__current_clr_polling();
 }