Re: [PATCH] remove BKL from drivers' release functions

From: Alan Cox (alan@lxorguk.ukuu.org.uk)
Date: Wed Nov 28 2001 - 20:47:47 EST


> Does everyone agree that we need to get the BKL out of common areas like
> this? For starters, what about adding a pair of spinlocks for block

Ideally the BKL itself should die.

> devices and character devices to take the place of the BKL in
> serializing opens? Or, should we make it the driver's responsibility
> completely?

It needs to be the drivers job, to be documented as such and to be
implemented properly in some drivers. In paticular there are some extremely
interesting closedown races in existing drivers where it goes

                CPU1 CPU2
        release
        [do slow thing]
                                        open
                                        ioctl
                                                setting stuff up
        slow thing done
        trash the chip setup
        turn the chip off
        return
                                        ouch bang splat Oops!!!

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



This archive was generated by hypermail 2b29 : Fri Nov 30 2001 - 21:00:33 EST