On Thu, May 16, 2002, David Brownell <[EMAIL PROTECTED]> wrote: > > So I finally hunkered down and got uhci-hcd working. I was surprised how > > easy it was to make the modifications for hcd.c. > > Great! I'll have a look at that one too ... gotta upgrade a bunch of > systems first though. Maybe I'll be able to test with 2.5.16 ... :)
Excellent, the more testing the better. > > There are some other messages that get printed such as: > > > > hcd.c: usb_hcd_pci_remove 00:10.0, count != 1 > > > > but it appears from the code that this is just a soft warning that all > > of the references to the bus aren't freed yet. This isn't a bug. > > It's an err(), as noted in the documentation for that method, so I'd > look at that a bit more. It means your driver could be rmmodded > but something (who knows what?) hasn't dropped all references > to that device. Minimally, you've got a memory leak. Worst case, > you've got some device driver thinking it can call into a device on > that bus ... oops! Oh boy, not another reference counting discussion. Look at the code. With your recent patch to move deallocate() and usb_put_bus(), thankfully there is no problem with a device driver doing any operations on that device. rmmod'ing uhci-hcd is still safe, FWIW. And that's not necessarily a memory leak. I won't guarantee that there are no leaks since the code hasn't received extensive testing yet, but I think you're misunderstanding this particular scenario. 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