3 files changed, 68 insertions, 46 deletions

diff --git a/arch/x86/kernel/cpu/microcode/amd.c b/arch/x86/kernel/cpu/microcode/amd.c
index a998e1a7d46f..48179928ff38 100644
--- a/arch/x86/kernel/cpu/microcode/amd.c
+++ b/arch/x86/kernel/cpu/microcode/amd.c
@@ -339,7 +339,7 @@ int __init save_microcode_in_initrd_amd(unsigned int cpuid_1_eax)
 		return -EINVAL;
 
 	ret = load_microcode_amd(true, x86_family(cpuid_1_eax), desc.data, desc.size);
-	if (ret != UCODE_OK)
+	if (ret > UCODE_UPDATED)
 		return -EINVAL;
 
 	return 0;
@@ -683,27 +683,35 @@ static enum ucode_state __load_microcode_amd(u8 family, const u8 *data,
 static enum ucode_state
 load_microcode_amd(bool save, u8 family, const u8 *data, size_t size)
 {
+	struct ucode_patch *p;
 	enum ucode_state ret;
 
 	/* free old equiv table */
 	free_equiv_cpu_table();
 
 	ret = __load_microcode_amd(family, data, size);
-
-	if (ret != UCODE_OK)
+	if (ret != UCODE_OK) {
 		cleanup();
+		return ret;
+	}
 
-#ifdef CONFIG_X86_32
-	/* save BSP's matching patch for early load */
-	if (save) {
-		struct ucode_patch *p = find_patch(0);
-		if (p) {
-			memset(amd_ucode_patch, 0, PATCH_MAX_SIZE);
-			memcpy(amd_ucode_patch, p->data, min_t(u32, ksize(p->data),
-							       PATCH_MAX_SIZE));
-		}
+	p = find_patch(0);
+	if (!p) {
+		return ret;
+	} else {
+		if (boot_cpu_data.microcode == p->patch_id)
+			return ret;
+
+		ret = UCODE_NEW;
 	}
-#endif
+
+	/* save BSP's matching patch for early load */
+	if (!save)
+		return ret;
+
+	memset(amd_ucode_patch, 0, PATCH_MAX_SIZE);
+	memcpy(amd_ucode_patch, p->data, min_t(u32, ksize(p->data), PATCH_MAX_SIZE));
+
 	return ret;
 }
 
diff --git a/arch/x86/kernel/cpu/microcode/core.c b/arch/x86/kernel/cpu/microcode/core.c
index 70ecbc8099c9..10c4fc2c91f8 100644
--- a/arch/x86/kernel/cpu/microcode/core.c
+++ b/arch/x86/kernel/cpu/microcode/core.c
@@ -517,7 +517,29 @@ static int check_online_cpus(void)
 	return -EINVAL;
 }
 
-static atomic_t late_cpus;
+static atomic_t late_cpus_in;
+static atomic_t late_cpus_out;
+
+static int __wait_for_cpus(atomic_t *t, long long timeout)
+{
+	int all_cpus = num_online_cpus();
+
+	atomic_inc(t);
+
+	while (atomic_read(t) < all_cpus) {
+		if (timeout < SPINUNIT) {
+			pr_err("Timeout while waiting for CPUs rendezvous, remaining: %d\n",
+				all_cpus - atomic_read(t));
+			return 1;
+		}
+
+		ndelay(SPINUNIT);
+		timeout -= SPINUNIT;
+
+		touch_nmi_watchdog();
+	}
+	return 0;
+}
 
 /*
  * Returns:
@@ -527,30 +549,16 @@ static atomic_t late_cpus;
  */
 static int __reload_late(void *info)
 {
-	unsigned int timeout = NSEC_PER_SEC;
-	int all_cpus = num_online_cpus();
 	int cpu = smp_processor_id();
 	enum ucode_state err;
 	int ret = 0;
 
-	atomic_dec(&late_cpus);
-
 	/*
 	 * Wait for all CPUs to arrive. A load will not be attempted unless all
 	 * CPUs show up.
 	 * */
-	while (atomic_read(&late_cpus)) {
-		if (timeout < SPINUNIT) {
-			pr_err("Timeout while waiting for CPUs rendezvous, remaining: %d\n",
-				atomic_read(&late_cpus));
-			return -1;
-		}
-
-		ndelay(SPINUNIT);
-		timeout -= SPINUNIT;
-
-		touch_nmi_watchdog();
-	}
+	if (__wait_for_cpus(&late_cpus_in, NSEC_PER_SEC))
+		return -1;
 
 	spin_lock(&update_lock);
 	apply_microcode_local(&err);
@@ -558,15 +566,22 @@ static int __reload_late(void *info)
 
 	if (err > UCODE_NFOUND) {
 		pr_warn("Error reloading microcode on CPU %d\n", cpu);
-		ret = -1;
-	} else if (err == UCODE_UPDATED) {
+		return -1;
+	/* siblings return UCODE_OK because their engine got updated already */
+	} else if (err == UCODE_UPDATED || err == UCODE_OK) {
 		ret = 1;
+	} else {
+		return ret;
 	}
 
-	atomic_inc(&late_cpus);
-
-	while (atomic_read(&late_cpus) != all_cpus)
-		cpu_relax();
+	/*
+	 * Increase the wait timeout to a safe value here since we're
+	 * serializing the microcode update and that could take a while on a
+	 * large number of CPUs. And that is fine as the *actual* timeout will
+	 * be determined by the last CPU finished updating and thus cut short.
+	 */
+	if (__wait_for_cpus(&late_cpus_out, NSEC_PER_SEC * num_online_cpus()))
+		panic("Timeout during microcode update!\n");
 
 	return ret;
 }
@@ -579,12 +594,11 @@ static int microcode_reload_late(void)
 {
 	int ret;
 
-	atomic_set(&late_cpus, num_online_cpus());
+	atomic_set(&late_cpus_in,  0);
+	atomic_set(&late_cpus_out, 0);
 
 	ret = stop_machine_cpuslocked(__reload_late, NULL, cpu_online_mask);
-	if (ret < 0)
-		return ret;
-	else if (ret > 0)
+	if (ret > 0)
 		microcode_check();
 
 	return ret;
@@ -607,7 +621,7 @@ static ssize_t reload_store(struct device *dev,
 		return size;
 
 	tmp_ret = microcode_ops->request_microcode_fw(bsp, &microcode_pdev->dev, true);
-	if (tmp_ret != UCODE_OK)
+	if (tmp_ret != UCODE_NEW)
 		return size;
 
 	get_online_cpus();
@@ -691,10 +705,8 @@ static enum ucode_state microcode_init_cpu(int cpu, bool refresh_fw)
 	if (system_state != SYSTEM_RUNNING)
 		return UCODE_NFOUND;
 
-	ustate = microcode_ops->request_microcode_fw(cpu, &microcode_pdev->dev,
-						     refresh_fw);
-
-	if (ustate == UCODE_OK) {
+	ustate = microcode_ops->request_microcode_fw(cpu, &microcode_pdev->dev, refresh_fw);
+	if (ustate == UCODE_NEW) {
 		pr_debug("CPU%d updated upon init\n", cpu);
 		apply_microcode_on_target(cpu);
 	}
diff --git a/arch/x86/kernel/cpu/microcode/intel.c b/arch/x86/kernel/cpu/microcode/intel.c
index 2aded9db1d42..32b8e5724f96 100644
--- a/arch/x86/kernel/cpu/microcode/intel.c
+++ b/arch/x86/kernel/cpu/microcode/intel.c
@@ -862,6 +862,7 @@ static enum ucode_state generic_load_microcode(int cpu, void *data, size_t size,
 	unsigned int leftover = size;
 	unsigned int curr_mc_size = 0, new_mc_size = 0;
 	unsigned int csig, cpf;
+	enum ucode_state ret = UCODE_OK;
 
 	while (leftover) {
 		struct microcode_header_intel mc_header;
@@ -903,6 +904,7 @@ static enum ucode_state generic_load_microcode(int cpu, void *data, size_t size,
 			new_mc  = mc;
 			new_mc_size = mc_size;
 			mc = NULL;	/* trigger new vmalloc */
+			ret = UCODE_NEW;
 		}
 
 		ucode_ptr += mc_size;
@@ -932,7 +934,7 @@ static enum ucode_state generic_load_microcode(int cpu, void *data, size_t size,
 	pr_debug("CPU%d found a matching microcode update with version 0x%x (current=0x%x)\n",
 		 cpu, new_rev, uci->cpu_sig.rev);
 
-	return UCODE_OK;
+	return ret;
 }
 
 static int get_ucode_fw(void *to, const void *from, size_t n)