Re: PATCH: (as177) Add class_device_unregister_wait() and platform_device_unregister_wait() to the driver model core

[Steve Youngs <sryoungs@bigpond.net.au>]

[-- Attachment #1: Type: text/plain, Size: 1383 bytes --]

* Valdis Kletnieks <Valdis.Kletnieks@vt.edu> writes:

  > On Mon, 26 Jan 2004 15:06:48 +1000, Steve Youngs <sryoungs@bigpond.net.au>  said:
  >> > A boolean is just a one-bit reference count. If the maximum number of
  >> > simultaneous 'users' for a given module is one, then a boolean will work.
  >> > If there is potential for more than one simultaneous user then you need
  >> > more bits.
  >> 
  >> Why?  A module is either being used or it isn't, the number of uses
  >> shouldn't even come into it.

  > OK. There's 2 users of the module.  The first one exits.  How does
  > it (or anything else) know that it's NOT safe to just clear the
  > in-use bit and clean it up?

Because the 2nd user is still using the module so its in-use bit
should still be set.  Remember that when the module was first loaded
it registered a function with the kernel for testing whether the
module is in use.

I must be overlooking something because I see the answer so clearly.
Maybe if someone could give me a real world example of a situation
where it'd be hard/impossible/unsafe to unload a module and I'll see
if my ideas can be applied.



-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|              Ashes to ashes, dust to dust.               |
|      The proof of the pudding, is under the crust.       |
|------------------------------<sryoungs@bigpond.net.au>---|

[-- Attachment #2: Type: application/pgp-signature, Size: 256 bytes --]