x86: Use syscore_ops instead of sysdev classes and sysdevs

Some subsystems in the x86 tree need to carry out suspend/resume and
shutdown operations with one CPU on-line and interrupts disabled and
they define sysdev classes and sysdevs or sysdev drivers for this
purpose. This leads to unnecessarily complicated code and excessive
memory usage, so switch them to using struct syscore_ops objects for
this purpose instead.

Generally, there are three categories of subsystems that use
sysdevs for implementing PM operations: (1) subsystems whose
suspend/resume callbacks ignore their arguments entirely (the
majority), (2) subsystems whose suspend/resume callbacks use their
struct sys_device argument, but don't really need to do that,
because they can be implemented differently in an arguably simpler
way (io_apic.c), and (3) subsystems whose suspend/resume callbacks
use their struct sys_device argument, but the value of that argument
is always the same and could be ignored (microcode_core.c). In all
of these cases the subsystems in question may be readily converted to
using struct syscore_ops objects for power management and shutdown.

Signed-off-by: Rafael J. Wysocki <[e-mail]>
Reviewed-by: Thomas Gleixner <[e-mail]>
Acked-by: Ingo Molnar <[e-mail]>

arch/x86/kernel/amd_iommu_init.c

···
 #include <linux/acpi.h>
 #include <linux/list.h>
 #include <linux/slab.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/interrupt.h>
 #include <linux/msi.h>
 #include <asm/pci-direct.h>
···
  * disable suspend until real resume implemented
  */
 
-static int amd_iommu_resume(struct sys_device *dev)
+static void amd_iommu_resume(void)
 {
 	struct amd_iommu *iommu;
 
···
 	 */
 	amd_iommu_flush_all_devices();
 	amd_iommu_flush_all_domains();
-
-	return 0;
 }
 
-static int amd_iommu_suspend(struct sys_device *dev, pm_message_t state)
+static int amd_iommu_suspend(void)
 {
 	/* disable IOMMUs to go out of the way for BIOS */
 	disable_iommus();
···
 	return 0;
 }
 
-static struct sysdev_class amd_iommu_sysdev_class = {
-	.name = "amd_iommu",
+static struct syscore_ops amd_iommu_syscore_ops = {
 	.suspend = amd_iommu_suspend,
 	.resume = amd_iommu_resume,
-};
-
-static struct sys_device device_amd_iommu = {
-	.id = 0,
-	.cls = &amd_iommu_sysdev_class,
 };
 
 /*
···
 		goto free;
 	}
 
-	ret = sysdev_class_register(&amd_iommu_sysdev_class);
-	if (ret)
-		goto free;
-
-	ret = sysdev_register(&device_amd_iommu);
-	if (ret)
-		goto free;
-
 	ret = amd_iommu_init_devices();
 	if (ret)
 		goto free;
···
 	amd_iommu_init_api();
 
 	amd_iommu_init_notifier();
+
+	register_syscore_ops(&amd_iommu_syscore_ops);
 
 	if (iommu_pass_through)
 		goto out;

arch/x86/kernel/apic/apic.c

···
 #include <linux/ftrace.h>
 #include <linux/ioport.h>
 #include <linux/module.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/delay.h>
 #include <linux/timex.h>
 #include <linux/dmar.h>
···
 	unsigned int apic_thmr;
 } apic_pm_state;
 
-static int lapic_suspend(struct sys_device *dev, pm_message_t state)
+static int lapic_suspend(void)
 {
 	unsigned long flags;
 	int maxlvt;
···
 	return 0;
 }
 
-static int lapic_resume(struct sys_device *dev)
+static void lapic_resume(void)
 {
 	unsigned int l, h;
 	unsigned long flags;
-	int maxlvt;
-	int ret = 0;
+	int maxlvt, ret;
 	struct IO_APIC_route_entry **ioapic_entries = NULL;
 
 	if (!apic_pm_state.active)
-		return 0;
+		return;
 
 	local_irq_save(flags);
 	if (intr_remapping_enabled) {
 		ioapic_entries = alloc_ioapic_entries();
 		if (!ioapic_entries) {
 			WARN(1, "Alloc ioapic_entries in lapic resume failed.");
-			ret = -ENOMEM;
 			goto restore;
 		}
 
···
 	}
 restore:
 	local_irq_restore(flags);
-
-	return ret;
 }
 
 /*
···
  * are needed on every CPU up until machine_halt/restart/poweroff.
  */
 
-static struct sysdev_class lapic_sysclass = {
-	.name		= "lapic",
+static struct syscore_ops lapic_syscore_ops = {
 	.resume		= lapic_resume,
 	.suspend	= lapic_suspend,
-};
-
-static struct sys_device device_lapic = {
-	.id	= 0,
-	.cls	= &lapic_sysclass,
 };
 
 static void __cpuinit apic_pm_activate(void)
···
 
 static int __init init_lapic_sysfs(void)
 {
-	int error;
-
-	if (!cpu_has_apic)
-		return 0;
 	/* XXX: remove suspend/resume procs if !apic_pm_state.active? */
+	if (cpu_has_apic)
+		register_syscore_ops(&lapic_syscore_ops);
 
-	error = sysdev_class_register(&lapic_sysclass);
-	if (!error)
-		error = sysdev_register(&device_lapic);
-	return error;
+	return 0;
 }
 
 /* local apic needs to resume before other devices access its registers. */

arch/x86/kernel/apic/io_apic.c

···
 #include <linux/compiler.h>
 #include <linux/acpi.h>
 #include <linux/module.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/msi.h>
 #include <linux/htirq.h>
 #include <linux/freezer.h>
···
 
 late_initcall(io_apic_bug_finalize);
 
-struct sysfs_ioapic_data {
-	struct sys_device dev;
-	struct IO_APIC_route_entry entry[0];
-};
-static struct sysfs_ioapic_data * mp_ioapic_data[MAX_IO_APICS];
+static struct IO_APIC_route_entry *ioapic_saved_data[MAX_IO_APICS];
 
-static int ioapic_suspend(struct sys_device *dev, pm_message_t state)
+static void suspend_ioapic(int ioapic_id)
 {
-	struct IO_APIC_route_entry *entry;
-	struct sysfs_ioapic_data *data;
+	struct IO_APIC_route_entry *saved_data = ioapic_saved_data[ioapic_id];
 	int i;
 
-	data = container_of(dev, struct sysfs_ioapic_data, dev);
-	entry = data->entry;
-	for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ )
-		*entry = ioapic_read_entry(dev->id, i);
+	if (!saved_data)
+		return;
+
+	for (i = 0; i < nr_ioapic_registers[ioapic_id]; i++)
+		saved_data[i] = ioapic_read_entry(ioapic_id, i);
+}
+
+static int ioapic_suspend(void)
+{
+	int ioapic_id;
+
+	for (ioapic_id = 0; ioapic_id < nr_ioapics; ioapic_id++)
+		suspend_ioapic(ioapic_id);
 
 	return 0;
 }
 
-static int ioapic_resume(struct sys_device *dev)
+static void resume_ioapic(int ioapic_id)
 {
-	struct IO_APIC_route_entry *entry;
-	struct sysfs_ioapic_data *data;
+	struct IO_APIC_route_entry *saved_data = ioapic_saved_data[ioapic_id];
 	unsigned long flags;
 	union IO_APIC_reg_00 reg_00;
 	int i;
 
-	data = container_of(dev, struct sysfs_ioapic_data, dev);
-	entry = data->entry;
+	if (!saved_data)
+		return;
 
 	raw_spin_lock_irqsave(&ioapic_lock, flags);
-	reg_00.raw = io_apic_read(dev->id, 0);
-	if (reg_00.bits.ID != mp_ioapics[dev->id].apicid) {
-		reg_00.bits.ID = mp_ioapics[dev->id].apicid;
-		io_apic_write(dev->id, 0, reg_00.raw);
+	reg_00.raw = io_apic_read(ioapic_id, 0);
+	if (reg_00.bits.ID != mp_ioapics[ioapic_id].apicid) {
+		reg_00.bits.ID = mp_ioapics[ioapic_id].apicid;
+		io_apic_write(ioapic_id, 0, reg_00.raw);
 	}
 	raw_spin_unlock_irqrestore(&ioapic_lock, flags);
-	for (i = 0; i < nr_ioapic_registers[dev->id]; i++)
-		ioapic_write_entry(dev->id, i, entry[i]);
-
-	return 0;
+	for (i = 0; i < nr_ioapic_registers[ioapic_id]; i++)
+		ioapic_write_entry(ioapic_id, i, saved_data[i]);
 }
 
-static struct sysdev_class ioapic_sysdev_class = {
-	.name = "ioapic",
+static void ioapic_resume(void)
+{
+	int ioapic_id;
+
+	for (ioapic_id = nr_ioapics - 1; ioapic_id >= 0; ioapic_id--)
+		resume_ioapic(ioapic_id);
+}
+
+static struct syscore_ops ioapic_syscore_ops = {
 	.suspend = ioapic_suspend,
 	.resume = ioapic_resume,
 };
 
-static int __init ioapic_init_sysfs(void)
+static int __init ioapic_init_ops(void)
 {
-	struct sys_device * dev;
-	int i, size, error;
+	int i;
 
-	error = sysdev_class_register(&ioapic_sysdev_class);
-	if (error)
-		return error;
+	for (i = 0; i < nr_ioapics; i++) {
+		unsigned int size;
 
-	for (i = 0; i < nr_ioapics; i++ ) {
-		size = sizeof(struct sys_device) + nr_ioapic_registers[i]
+		size = nr_ioapic_registers[i]
 			* sizeof(struct IO_APIC_route_entry);
-		mp_ioapic_data[i] = kzalloc(size, GFP_KERNEL);
-		if (!mp_ioapic_data[i]) {
-			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
-			continue;
-		}
-		dev = &mp_ioapic_data[i]->dev;
-		dev->id = i;
-		dev->cls = &ioapic_sysdev_class;
-		error = sysdev_register(dev);
-		if (error) {
-			kfree(mp_ioapic_data[i]);
-			mp_ioapic_data[i] = NULL;
-			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
-			continue;
-		}
+		ioapic_saved_data[i] = kzalloc(size, GFP_KERNEL);
+		if (!ioapic_saved_data[i])
+			pr_err("IOAPIC %d: suspend/resume impossible!\n", i);
 	}
+
+	register_syscore_ops(&ioapic_syscore_ops);
 
 	return 0;
 }
 
-device_initcall(ioapic_init_sysfs);
+device_initcall(ioapic_init_ops);
 
 /*
  * Dynamic irq allocate and deallocation

arch/x86/kernel/cpu/mcheck/mce.c

···
 #include <linux/percpu.h>
 #include <linux/string.h>
 #include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/delay.h>
 #include <linux/ctype.h>
 #include <linux/sched.h>
···
 	return 0;
 }
 
-static int mce_suspend(struct sys_device *dev, pm_message_t state)
+static int mce_suspend(void)
 {
 	return mce_disable_error_reporting();
 }
 
-static int mce_shutdown(struct sys_device *dev)
+static void mce_shutdown(void)
 {
-	return mce_disable_error_reporting();
+	mce_disable_error_reporting();
 }
 
 /*
···
  * Only one CPU is active at this time, the others get re-added later using
  * CPU hotplug:
  */
-static int mce_resume(struct sys_device *dev)
+static void mce_resume(void)
 {
 	__mcheck_cpu_init_generic();
 	__mcheck_cpu_init_vendor(__this_cpu_ptr(&cpu_info));
-
-	return 0;
 }
+
+static struct syscore_ops mce_syscore_ops = {
+	.suspend	= mce_suspend,
+	.shutdown	= mce_shutdown,
+	.resume		= mce_resume,
+};
 
 static void mce_cpu_restart(void *data)
 {
···
 }
 
 static struct sysdev_class mce_sysclass = {
-	.suspend	= mce_suspend,
-	.shutdown	= mce_shutdown,
-	.resume		= mce_resume,
 	.name		= "machinecheck",
 };
 
···
 			return err;
 	}
 
+	register_syscore_ops(&mce_syscore_ops);
 	register_hotcpu_notifier(&mce_cpu_notifier);
 	misc_register(&mce_log_device);
 

arch/x86/kernel/cpu/mtrr/main.c

···
 #include <linux/cpu.h>
 #include <linux/pci.h>
 #include <linux/smp.h>
+#include <linux/syscore_ops.h>
 
 #include <asm/processor.h>
 #include <asm/e820.h>
···
 
 static struct mtrr_value mtrr_value[MTRR_MAX_VAR_RANGES];
 
-static int mtrr_save(struct sys_device *sysdev, pm_message_t state)
+static int mtrr_save(void)
 {
 	int i;
 
···
 	return 0;
 }
 
-static int mtrr_restore(struct sys_device *sysdev)
+static void mtrr_restore(void)
 {
 	int i;
 
···
 				    mtrr_value[i].ltype);
 		}
 	}
-	return 0;
 }
 
 
 
-static struct sysdev_driver mtrr_sysdev_driver = {
+static struct syscore_ops mtrr_syscore_ops = {
 	.suspend	= mtrr_save,
 	.resume		= mtrr_restore,
 };
···
 	 * TBD: is there any system with such CPU which supports
 	 * suspend/resume? If no, we should remove the code.
 	 */
-	sysdev_driver_register(&cpu_sysdev_class, &mtrr_sysdev_driver);
+	register_syscore_ops(&mtrr_syscore_ops);
 
 	return 0;
 }

arch/x86/kernel/i8237.c

···
  */
 
 #include <linux/init.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 
 #include <asm/dma.h>
 
···
  * in asm/dma.h.
  */
 
-static int i8237A_resume(struct sys_device *dev)
+static void i8237A_resume(void)
 {
 	unsigned long flags;
 	int i;
···
 	enable_dma(4);
 
 	release_dma_lock(flags);
-
-	return 0;
 }
 
-static int i8237A_suspend(struct sys_device *dev, pm_message_t state)
-{
-	return 0;
-}
-
-static struct sysdev_class i8237_sysdev_class = {
-	.name		= "i8237",
-	.suspend	= i8237A_suspend,
+static struct syscore_ops i8237_syscore_ops = {
 	.resume		= i8237A_resume,
 };
 
-static struct sys_device device_i8237A = {
-	.id		= 0,
-	.cls		= &i8237_sysdev_class,
-};
-
-static int __init i8237A_init_sysfs(void)
+static int __init i8237A_init_ops(void)
 {
-	int error = sysdev_class_register(&i8237_sysdev_class);
-	if (!error)
-		error = sysdev_register(&device_i8237A);
-	return error;
+	register_syscore_ops(&i8237_syscore_ops);
+	return 0;
 }
-device_initcall(i8237A_init_sysfs);
+device_initcall(i8237A_init_ops);

arch/x86/kernel/i8259.c

···
 #include <linux/random.h>
 #include <linux/init.h>
 #include <linux/kernel_stat.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/bitops.h>
 #include <linux/acpi.h>
 #include <linux/io.h>
···
 	trigger[1] = inb(0x4d1) & 0xDE;
 }
 
-static int i8259A_resume(struct sys_device *dev)
+static void i8259A_resume(void)
 {
 	init_8259A(i8259A_auto_eoi);
 	restore_ELCR(irq_trigger);
-	return 0;
 }
 
-static int i8259A_suspend(struct sys_device *dev, pm_message_t state)
+static int i8259A_suspend(void)
 {
 	save_ELCR(irq_trigger);
 	return 0;
 }
 
-static int i8259A_shutdown(struct sys_device *dev)
+static void i8259A_shutdown(void)
 {
 	/* Put the i8259A into a quiescent state that
 	 * the kernel initialization code can get it
···
 	 */
 	outb(0xff, PIC_MASTER_IMR);	/* mask all of 8259A-1 */
 	outb(0xff, PIC_SLAVE_IMR);	/* mask all of 8259A-1 */
-	return 0;
 }
 
-static struct sysdev_class i8259_sysdev_class = {
-	.name = "i8259",
+static struct syscore_ops i8259_syscore_ops = {
 	.suspend = i8259A_suspend,
 	.resume = i8259A_resume,
 	.shutdown = i8259A_shutdown,
-};
-
-static struct sys_device device_i8259A = {
-	.id	= 0,
-	.cls	= &i8259_sysdev_class,
 };
 
 static void mask_8259A(void)
···
 
 struct legacy_pic *legacy_pic = &default_legacy_pic;
 
-static int __init i8259A_init_sysfs(void)
+static int __init i8259A_init_ops(void)
 {
-	int error;
+	if (legacy_pic == &default_legacy_pic)
+		register_syscore_ops(&i8259_syscore_ops);
 
-	if (legacy_pic != &default_legacy_pic)
-		return 0;
-
-	error = sysdev_class_register(&i8259_sysdev_class);
-	if (!error)
-		error = sysdev_register(&device_i8259A);
-	return error;
+	return 0;
 }
 
-device_initcall(i8259A_init_sysfs);
+device_initcall(i8259A_init_ops);

arch/x86/kernel/microcode_core.c

···
 #include <linux/cpu.h>
 #include <linux/fs.h>
 #include <linux/mm.h>
+#include <linux/syscore_ops.h>
 
 #include <asm/microcode.h>
 #include <asm/processor.h>
···
 	return 0;
 }
 
-static int mc_sysdev_resume(struct sys_device *dev)
-{
-	int cpu = dev->id;
-	struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
-
-	if (!cpu_online(cpu))
-		return 0;
-
-	/*
-	 * All non-bootup cpus are still disabled,
-	 * so only CPU 0 will apply ucode here.
-	 *
-	 * Moreover, there can be no concurrent
-	 * updates from any other places at this point.
-	 */
-	WARN_ON(cpu != 0);
-
-	if (uci->valid && uci->mc)
-		microcode_ops->apply_microcode(cpu);
-
-	return 0;
-}
-
 static struct sysdev_driver mc_sysdev_driver = {
 	.add			= mc_sysdev_add,
 	.remove			= mc_sysdev_remove,
-	.resume			= mc_sysdev_resume,
+};
+
+/**
+ * mc_bp_resume - Update boot CPU microcode during resume.
+ */
+static void mc_bp_resume(void)
+{
+	int cpu = smp_processor_id();
+	struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
+
+	if (uci->valid && uci->mc)
+		microcode_ops->apply_microcode(cpu);
+}
+
+static struct syscore_ops mc_syscore_ops = {
+	.resume			= mc_bp_resume,
 };
 
 static __cpuinit int
···
 	if (error)
 		return error;
 
+	register_syscore_ops(&mc_syscore_ops);
 	register_hotcpu_notifier(&mc_cpu_notifier);
 
 	pr_info("Microcode Update Driver: v" MICROCODE_VERSION

arch/x86/kernel/pci-gart_64.c

···
 #include <linux/kdebug.h>
 #include <linux/scatterlist.h>
 #include <linux/iommu-helper.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/io.h>
 #include <linux/gfp.h>
 #include <asm/atomic.h>
···
 	aperture_alloc = aper_alloc;
 }
 
-static void gart_fixup_northbridges(struct sys_device *dev)
+static void gart_fixup_northbridges(void)
 {
 	int i;
 
···
 	}
 }
 
-static int gart_resume(struct sys_device *dev)
+static void gart_resume(void)
 {
 	pr_info("PCI-DMA: Resuming GART IOMMU\n");
 
-	gart_fixup_northbridges(dev);
+	gart_fixup_northbridges();
 
 	enable_gart_translations();
-
-	return 0;
 }
 
-static int gart_suspend(struct sys_device *dev, pm_message_t state)
-{
-	return 0;
-}
-
-static struct sysdev_class gart_sysdev_class = {
-	.name		= "gart",
-	.suspend	= gart_suspend,
+static struct syscore_ops gart_syscore_ops = {
 	.resume		= gart_resume,
 
-};
-
-static struct sys_device device_gart = {
-	.cls		= &gart_sysdev_class,
 };
 
 /*
···
 	unsigned aper_base, new_aper_base;
 	struct pci_dev *dev;
 	void *gatt;
-	int i, error;
+	int i;
 
 	pr_info("PCI-DMA: Disabling AGP.\n");
 
···
 
 	agp_gatt_table = gatt;
 
-	error = sysdev_class_register(&gart_sysdev_class);
-	if (!error)
-		error = sysdev_register(&device_gart);
-	if (error)
-		panic("Could not register gart_sysdev -- "
-		      "would corrupt data on next suspend");
+	register_syscore_ops(&gart_syscore_ops);
 
 	flush_gart();
 

arch/x86/oprofile/nmi_int.c

···
 #include <linux/notifier.h>
 #include <linux/smp.h>
 #include <linux/oprofile.h>
-#include <linux/sysdev.h>
+#include <linux/syscore_ops.h>
 #include <linux/slab.h>
 #include <linux/moduleparam.h>
 #include <linux/kdebug.h>
···
 
 #ifdef CONFIG_PM
 
-static int nmi_suspend(struct sys_device *dev, pm_message_t state)
+static int nmi_suspend(void)
 {
 	/* Only one CPU left, just stop that one */
 	if (nmi_enabled == 1)
···
 	return 0;
 }
 
-static int nmi_resume(struct sys_device *dev)
+static void nmi_resume(void)
 {
 	if (nmi_enabled == 1)
 		nmi_cpu_start(NULL);
-	return 0;
 }
 
-static struct sysdev_class oprofile_sysclass = {
-	.name		= "oprofile",
+static struct syscore_ops oprofile_syscore_ops = {
 	.resume		= nmi_resume,
 	.suspend	= nmi_suspend,
 };
 
-static struct sys_device device_oprofile = {
-	.id	= 0,
-	.cls	= &oprofile_sysclass,
-};
-
-static int __init init_sysfs(void)
+static void __init init_suspend_resume(void)
 {
-	int error;
-
-	error = sysdev_class_register(&oprofile_sysclass);
-	if (error)
-		return error;
-
-	error = sysdev_register(&device_oprofile);
-	if (error)
-		sysdev_class_unregister(&oprofile_sysclass);
-
-	return error;
+	register_syscore_ops(&oprofile_syscore_ops);
 }
 
-static void exit_sysfs(void)
+static void exit_suspend_resume(void)
 {
-	sysdev_unregister(&device_oprofile);
-	sysdev_class_unregister(&oprofile_sysclass);
+	unregister_syscore_ops(&oprofile_syscore_ops);
 }
 
 #else
 
-static inline int  init_sysfs(void) { return 0; }
-static inline void exit_sysfs(void) { }
+static inline void init_suspend_resume(void) { }
+static inline void exit_suspend_resume(void) { }
 
 #endif /* CONFIG_PM */
 
···
 
 	mux_init(ops);
 
-	ret = init_sysfs();
-	if (ret)
-		return ret;
+	init_suspend_resume();
 
 	printk(KERN_INFO "oprofile: using NMI interrupt.\n");
 	return 0;
···
 
 void op_nmi_exit(void)
 {
-	exit_sysfs();
+	exit_suspend_resume();
 }