On Mon, Feb 04, 2008 at 03:23:14PM -0500, Alan Stern wrote: > On Mon, 4 Feb 2008, Oliver Neukum wrote: > > > Am Montag, 4. Februar 2008 16:36:51 schrieb Alan Stern: > > > > > That's because you don't bind usbfs to a device through sysfs. You > > > bind it by running a program that calls the USBDEVFS_CLAIMINTERFACE > > > ioctl. > > > > True, and unavoidable given the temporal order of usbfs and sysfs. Though > > looking at it now, isn't that a misdesign? IMHO we should avoid it in > > usbfs2. > > I don't know -- it depends on some fundamental design decisions about > how usbfs2 will work, and I haven't tried to follow the progress in > that area. > > For example, usbfs essentially binds an interface to an open file. Any > task with a descriptor for that open file can do I/O through the > interface's endpoints (although there may be some question as to which > tasks receive the signal when an async I/O completes). > > usbfs2 doesn't have to work that way. Instead interfaces could be > bound to usbfs2 itself, and then usbfs2 would need to have some way of > deciding which tasks are allowed to use the interface's endpoints. > > But maybe this has already been settled -- like I said, I don't know > what's happening with usbfs2.
usbfs2 has been on my back burner for a while now. I've been fairly busy with USB 3.0, and I'd like to throughly test the in-kernel AIO core changes that Zach Brown proposed before I move on with the design. If someone else wants to take over the usbfs2 design while I work on aio, I would appreciate it. Sarah Sharp - To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
