On Sun, 23 Mar 2014, Dâniel Fraga wrote:
> On Sun, 23 Mar 2014 11:56:50 -0400 (EDT)
> Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
>
> > > [ 28.336116] usbhid 2-1.6:1.0: disconnect by usbfs
> > > [ 28.336133] usb 2-1.6: unlink qh1-1c01/ffff8802153fb480 start 0 [1/3 us]
> > > [ 28.353165] usb 2-1.6: manual set_interface for iface 0, alt 0
> > > [ 28.353168] usb 2-1.6: ep 81: release intr @ 0+8 (0.0+1) [1/3 us] mask 1c01
> > > [ 28.354041] usbhid 2-1.6:1.0: usb_probe_interface
> > > [ 28.354044] usbhid 2-1.6:1.0: usb_probe_interface - got id
> > > [ 28.356155] input: Razer Razer DeathAdder as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.6/2-1.6:1.0/input/input10
> > > [ 28.356226] usb 2-1.6: ep 81: reserve intr @ 0+8 (0.0+1) [1/3 us] mask 1c01
> > > [ 28.356237] usb 2-1.6: link qh1-1c01/ffff8800d55d6300 start 0 [1/3 us]
> >
> > Is this the sort of thing you're seeing? The "disconnect by usbfs"
> > means that some program on your computer is causing this to happen.
> > The program unbinds the usbhid driver from the mouse and then rebinds
> > it.
>
> Hmm. The only program I use for the mouse is razerd daemon to
> switch off the leds.
>
> > If you want, I can post a patch that will print out the name of the
> > process doing this. Did this happen when you were starting up KDE?
>
> Yes, it just happens when starting KDE. but I'd like to test
> with your patch, so we can be sure what process is causing this.
Okay, the patch is at the end of this email.
> > I wasn't able to learn anything from the usbmon data. It just shows
> > the mouse being used. Did one of those disconnection events occur
> > while you were collecting the usbmon trace?
>
> Yes! This is what I found interesting, because usbmon keeps
> reporting mouse activity while the mouse pointer is stuck. Maybe it's a
> KDE only issue, *not* kernel related? I reported here because with 3.12
> kernel it doesn't happens (only with 3.13 and above so...). But maybe
> it's a bug in KDE which only shows with 3.13 kernel.
Or a bug in the input layer. If you load the gpm console mouse
manager and switch to a VT console after starting KDE, does the
console mouse pointer work?
> > Can you post an equivalent dmesg log with the f1125f81feca commit
> > reverted?
>
> Forget it. I reboot 2 times without the f1125f81eca commit. Only
> in the second reboot the bug appeared. So you're right. The commit
> f1125f81eca isn't the problem.
>
> It seems this bug is ocasional. That explains why the bisect is
> worthless in this case, since I can't be sure what commit really
> triggers the bug.
>
> What I can say for sure is that with 3.12 kernel the bug
> doesn't happen, while with 3.13 and above kernel the bug _sometimes_
> happens.
Then can you post an equivalent dmesg log for the 3.12 kernel?
Alan Stern
Index: usb-3.14/drivers/usb/core/devio.c
===================================================================
--- usb-3.14.orig/drivers/usb/core/devio.c
+++ usb-3.14/drivers/usb/core/devio.c
@@ -1860,6 +1860,7 @@ static int proc_ioctl(struct dev_state *
/* disconnect kernel driver from interface */
case USBDEVFS_DISCONNECT:
+ dev_info(&intf->dev, "DISCONNECT by %s\n", current->comm);
if (intf->dev.driver) {
driver = to_usb_driver(intf->dev.driver);
dev_dbg(&intf->dev, "disconnect by usbfs\n");
@@ -1979,6 +1980,7 @@ static int proc_disconnect_claim(struct
if (!intf)
return -EINVAL;
+ dev_info(&intf->dev, "DISCONNECT_CLAIM by %s\n", current->comm);
if (intf->dev.driver) {
struct usb_driver *driver = to_usb_driver(intf->dev.driver);
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
