Re: PCMCIA/network behaviour broken

Alan Cox (alan@lxorguk.ukuu.org.uk)
Mon, 31 May 1999 22:01:48 +0100 (BST)


> Given that (I'll check the driver in a sec), what's the correct way to
> deal with this problem? Should pcmcia-cs explicitly shut down the

pcmcia-cs should down the interface but it has to down the interface in
a non interrupt context. Either via the pcmcia daemon or a kernel thread.
Thats my guess as to what is going on

> interface before removing the card structures and modules? Or should the
> drivers in question force their own associated interfaces down as they are
> being removed? Or is it purely a kernel issue?

Basically when an interface is "down" the kernel ensures no references are
ever made to it unsafely. That means it can be unloaded

> What I'm wondering is, what changed between 2.0 and 2.2 that caused this
> behaviour change? Given what I know of the networking code, the 2.2
> behaviour is "correct" in the naive (non-PCMCIA) case, so what was it in
> 2.0 that made it work?

In 2.0 downing an interface from an interrupt kind of mostly worked. Thats
why Im guessing this is the actual problem

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/