Re: [linux-usb-devel] [PATCH] USB: add usbfs stubs for suspend, resume, etc.
Oliver Neukum <[email protected]>
| Newsgroups | gmane.linux.usb.general,gmane.linux.usb.devel |
|---|---|
| Organization | Novell |
| Message-ID | <[email protected]> |
Am Montag, 3. Dezember 2007 23:09:07 schrieb Alan Stern: > On Mon, 3 Dec 2007, Oliver Neukum wrote: > > Am Montag, 3. Dezember 2007 21:43:11 schrieb Alan Stern: > > > THis patch (as1022) adds stub methods for suspend/resume/reset_resume > > > and pre-rreset/post-reset to the usbfs_driver. There isn't much they > > > can do, since there's no way to inform a user task about any of these > > > > In this case, shouldn't resume/reset_resume/post_reset report > > errors? > > Well, no. That would be just as bad as not having the stubs in the > first place. Yes, you are running up against a limitation in the system. > Not only is it impossible to tell user-level drivers about suspend, > etc. events, it's also not possible to tell them that they have been -ENODEV. Last time I looked usbfs had a disconnect() method. > unbound! So I wanted to be conservative and keep the kernel's current > behavior: The suspend/resume/reset happens and the user-level driver is > blissfully igorant of it. In theory, blissfully ignoring suspend/resume must work. Anything that involves a reset cannot. User space must be prepared with a surprise unplug. Surprise state change asks too much. Regards Oliver - 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