On Wed, 2005-16-11 at 17:14 +0100, Thomas Graf wrote: > * jamal <[EMAIL PROTECTED]> 2005-11-15 21:16 > > 3) There is a kernel dev->operstate_kernel which is accessible via > > user space in the same manner IFF_UP flags are set etc. > > Just be careful about synchronization, that's one issue with the > currently proposed implementation. The callers cannot be made > responsible to synchronize this, therefore the set-and-test before > adding to the linkwatch queue must be made atomic.
I think if we are in agreement it is time for a patch then we can see if there are synchronization issues. The suggestion is: user space doesnt set any operational state but can set an admin state to select the operational mode (something along the lines of IFF_OPMODE) and this will be transfered to the kernel state no different than IFF_UP is. > The latest patch looks a lot better, even though I still don't see > what we gainin exchange for the slightly reduced flexibility. What > worries me in particular are checks like operstate < STATE_X && > operstate > STATE_Y which are fine for the moment but once we add > new states we have to make sure that all checks remain correct which > could mean that certain logically valid combinations cannot be > implemented anymore. > In my review of Stefan's patch i made a similar comment on using the inequalities and i think he is fine with not doing so. > I think it would help me a lot if I'd see the actual gain from > using a state variable over state flags. I am not sure if i followed, but we are close enough that a single patch now needs to be put out which unifies the ideas. cheers, jamal - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
