Updated with fixes > to code layout, typecasting, and use of internal functions. > From: Mike Murphy > Improved support for Xbox 360 wireless devices by enabling rumble and > LED control, improved general usability by adding input controls, and > added a sysfs interface for tuning driver behavior. Since the current stable version only supports rumble on wiredĬontrollers anyway, removing the force feedback code altogether mightīe the simplest resolution if nobody wants to maintain the code for The current stable xpad driver has exactly the same concurrency bug. Pick up this code, I suggest dropping my patches, with the note that I do not have time to continue working on this problem and am Versions) takes a mutex while holding the spinlock, which doesn't seem Rumble-handling code in the xpad driver (both the modified and stable

It turns out theįf_memoryless code is holding an IRQ spinlock before invoking the xpadĬode, so the urb submission should be atomic. GFP_KERNEL when rumble (force feedback) is utilized. I can confirm the presence of a concurrency bug in the modified xpadĬode, which directly results from submission of the USB urb using

X-MIMEDefang-Filter: lf$Revision: 1.188 $ X-Spam-Checker-Version: SpamAssassin 3.2.4-osdl_revision_1.47_

X-Spam-Status: No, hits=-3.513 required=5 tests=AWL, BAYES_00, Subject: input: drivers/input/xpad.c: improve Xbox 360 wirelessĬontent-Type: text/plain charset=ANSI_X3.4-1968 Received: from localhost.localdomain (localhost ) (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) ' rhost-flags-OK-OK-OK-OK)īy with ESMTP id S1752217AbZIRTtq (ORCPT

Xbox 360 wireless driver