Doug Wilson <doug.lkml@gmail.com> writes:
All,
I have been looking at sysfs cpu side code. Had a quick question.
On my x86 box: #ls /sys/devices/system/cpu/ I see 4 cpus plus a few more fields such as online/present/possible etc.
This come from drivers/base/cpu.c.
My question here is, in each of those cpu0-3 folders, there's file by name online which one can use to disable/enable cpu using cpumask. Could someone point me to that code, I could not really pin point to where exactly this attribute is defined. Am just exploring cpu hotplug feature.
drivers/base/core.c has this generic code for creating that file: if (device_supports_offline(dev) && !dev->offline_disabled) { error = device_create_file(dev, &dev_attr_online); if (error) goto err_remove_dev_groups; } where device_supports_offline is: static inline bool device_supports_offline(struct device *dev) { return dev->bus && dev->bus->offline && dev->bus->online; } and the cpu "bus" is (from drivers/base/cpu.c): struct bus_type cpu_subsys = { .name = "cpu", .dev_name = "cpu", .match = cpu_subsys_match, #ifdef CONFIG_HOTPLUG_CPU .online = cpu_subsys_online, .offline = cpu_subsys_offline, #endif }; The fields of the cpu device are initialized in drivers/base/cpu.c: int register_cpu(struct cpu *cpu, int num) { int error; cpu->node_id = cpu_to_node(num); memset(&cpu->dev, 0x00, sizeof(struct device)); cpu->dev.id = num; cpu->dev.bus = &cpu_subsys; cpu->dev.release = cpu_device_release; cpu->dev.offline_disabled = !cpu->hotpluggable; cpu->dev.offline = !cpu_online(num); cpu->dev.of_node = of_get_cpu_node(num, NULL); #ifdef CONFIG_GENERIC_CPU_AUTOPROBE cpu->dev.bus->uevent = cpu_uevent; #endif cpu->dev.groups = common_cpu_attr_groups; if (cpu->hotpluggable) cpu->dev.groups = hotplugable_cpu_attr_groups; error = device_register(&cpu->dev); if (!error) per_cpu(cpu_sys_devices, num) = &cpu->dev; if (!error) register_cpu_under_node(num, cpu_to_node(num)); return error; } So you get the "online" file in the cpuX folders when CONFIG_HOTPLUG_CPU is defined and cpu->hotpluggable is true. Which will exclude cpu0 in some cases. See arch/x86/kernel/topology.c for the x86 specific logic concerning that. I assume other platforms have similar exceptions Bjørn