> > Why would that be different from read or write ?
>
> What do you mean? If you mean: "of course the ioctl callback
> can be called during disconnect, just like for other reads and
> writes", well I agree. I didn't write the driver, I'm trying to fix
> it so it works with SMP; this looks like an obvious race in the
> driver (ioctl during disconnect), but I would like to be sure.
> Maybe the driver author knows something I don't...
A look through usbdevfs shows that there's no locking.
All imaginable race are possible. In terms of locking you are on your own.
And yes, simply taking the same locks as in read and write should do the job.
You need to do this anyway as the syscall may be reentered.
Regards
Oliver
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
