cvs commit: src/sys/dev/netif/mii_layer

[Matthew Dillon]

:>   Modified files:
:>     sys/dev/netif/mii_layer acphy.c amphy.c brgphy.c dcphy.c 
:>                             e1000phy.c exphy.c inphy.c lxtphy.c 
:>                             mii_physubr.c miivar.h mlphy.c 
:>                             nsgphy.c nsphy.c pnaphy.c pnphy.c 
:>                             qsphy.c rlphy.c tlphy.c ukphy.c 
:>                             xmphy.c 
:>   Log:
:>   timeout/untimeout ==> callout_*
:
:The critical part of this is IMO that we need both mii_softc_init and
:mii_sofc_deinit or so. Not all driver call mii_phy_auto_stop and therefore
:might result in dangling timeouts.
:
:Joerg

    The question is: do the drivers ensure (directly or indirectly) that
    no timeout is pending before detaching ?  If they don't then we
    need mii_softc_deinit().  If they do then we don't.

    all we have left is ndis.  Ndis could be a problem because the existing
    pointer is being overloaded onto a windows structure.  We can't just
    change the windows structure (though it *might* be possible to extend
    it).

					-Matt
					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>