Am Mittwoch, 31. Januar 2007 14:47 schrieb Jiri Kosina:
> On Wed, 31 Jan 2007, Oliver Neukum wrote:
>
> > > coming from the URB submission which happened later - either the flags
> > > are cleared, or wake_up() will eventually happen (you are right that
> > > it might possibly be wake_up() corresponding to submission of URB
> > > submitted later, but does that matter?).
> > Yes, as you cannot make sure it'll come before the timeout strikes, nor
> > can you make sure the condition will be true when a second URB finishes.
> > This logic can fail under load.
>
> Actually, this is a current semantics of the usbhid_wait_io() function -
> it doesn't wait for a given URB to be delivered (which is what you seem to
> expect from it), but it rather waits until the given HID device completes
> pending I/O. Do you find this semantics broken in some way?
Yes, it is broken. Hiddev allows multiple openings of the same device.
You cannot just mix IO requests of different readers. You are allowing
a second process to delay IO completion of the first process.
Regards
Oliver
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
