Re: What's the purpose of subsys_interface in linux/device.h?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Dec 16, 2016 at 8:57 AM, Perr Zhang <strongbox8@xxxxxxxx> wrote:
>  ---- On Fri, 16 Dec 2016 00:26:24 +0800 Alexander Kapshuk <alexander.kapshuk@xxxxxxxxx> wrote ----
>  >
>  > Having looked at the pieces of source code below, it appears that
>  > subsys_interface is the means to expose specific functionality of a
>  > subsystem/class of devices, as the commentary below states.
>  > include/linux/device.h:331,350
>  > /**
>  >  * struct subsys_interface - interfaces to device functions
>  >  * @name:       name of the device function
>  >  * @subsys:     subsytem of the devices to attach to
>  >  * @node:       the list of functions registered at the subsystem
>  >  * @add_dev:    device hookup to device function handler
>  >  * @remove_dev: device hookup to device function handler
>  >  *
>  >  * Simple interfaces attached to a subsystem. Multiple interfaces can
>  >  * attach to a subsystem and its devices. Unlike drivers, they do not
>  >  * exclusively claim or control devices. Interfaces usually represent
>  >  * a specific functionality of a subsystem/class of devices.
>  >  */
>  > struct subsys_interface {
>  > const char *name;
>  > struct bus_type *subsys;
>  > struct list_head node;
>  > int (*add_dev)(struct device *dev, struct subsys_interface *sif);
>  > void (*remove_dev)(struct device *dev, struct subsys_interface *sif);
>  > };
>  >
>  > Here's sample usage:
>  > arch/tile/kernel/sysfs.c:211,216
>  > static struct subsys_interface hv_stats_interface = {
>  > .name = "hv_stats",
>  > .subsys = &cpu_subsys,
>  > .add_dev = hv_stats_device_add,
>  > .remove_dev = hv_stats_device_remove,
>  > };
>  >
>  > And the add_dev method, whose main purpose seems to be creating a
>  > sysfs file for a particular device:
>  > arch/tile/kernel/sysfs.c:189,199
>  > static int hv_stats_device_add(struct device *dev, struct subsys_interface *sif)
>  > {
>  > int err, cpu = dev->id;
>  >
>  > if (!cpu_online(cpu))
>  > return 0;
>  >
>  > err = sysfs_create_file(&dev->kobj, &dev_attr_hv_stats.attr);
>  >
>  > return err;
>  > }
>  >
>  > Here's an example of the hv_stats_interface being registered with the subsystem:
>  > arch/tile/kernel/sysfs.c:261
>  > err = subsys_interface_register(&hv_stats_interface);
>  >
>  > Hopefully, other members of this list will offer a better informed
>  > explanation than that given here.
>
> Sorry for sending email repeatedly.
> I have also read the code:
> device_add() ->  bus_add_device() -> device_add_attrs(bus, dev) -> device_create_file(dev, &bus->dev_attrs[i])
> Since the attribute file/functionality could be added through bus->dev_attrs, why not to do it in this way?
>

My understanding is that an interface is like a protocol that has been
put in place to enable a particular set of functions within a given
subsystem or device model.

I am sending a copy of this email to the list as well, so you can
hopefully get a better informed reply to your query.

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@xxxxxxxxxxxxxxxxx
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies



[Index of Archives]     [Newbies FAQ]     [Linux Kernel Mentors]     [Linux Kernel Development]     [IETF Annouce]     [Git]     [Networking]     [Security]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux SCSI]     [Linux ACPI]
  Powered by Linux