On Thu, May 16, 2002, David Brownell <[EMAIL PROTECTED]> wrote: > However, for the inevitable buggy device drivers, this guarantees > the cleanup would never be done in_interrupt().
I think you still misunderstand the guarantees the reference counting code provide. You can call usb_inc_dev_use and usb_dec_dev_use in either a thread or an interrupt context. And since usb_dec_dev_use will call deallocate() (and some other code), deallocate() can be called in either a thread or an interrupt context. The fact that it mostly happens in a thread context is to be considered dumb luck. A quick recap: The reference counting code makes no guaratees in which context the count will hit 0, thusly no guarantees that deallocate() will be in a thread context (before your patch atleast). Don't confuse your implementation of sohci_free_dev (which needs to be called in a thread context) from the requirements of the hardware (triggering the start of an unlink). JE _______________________________________________________________ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
