Please tell your email client to wrap lines of text after 72 columns or so. On Thu, 23 Jul 2015, Vasudevan, Krishna PrasathX K wrote: > Hi, > > This mail is for RFC regarding the persist resume of some USB 2.0 devices, > > Problem Summary: Problem has been observed for some USB 2.0 devices > while resuming from sleep. When the USB �persist� feature is enabled > through sysfs it is expected to retain its previous mount point > across sleep and resume states, it works fine for most of the USB 2.0 > mass storage devices, but for some USB 2.0 pendrives such as > (Transcend Jet flash 16 GB & Alcor micro corporation) it seems to > fail and re-enumeration happens resulting in creation of new device > data structure. This problem seems to be similar to the problem faced > with USB 3.0 devices which is mentioned in below link > > http://marc.info/?l=linux-usb&m=140566728011240&w=2 > > > > In the above mentioned link the problem is observed with USB 3.0 > devices, but in this case it is observed with USB 2.0 mass storage > devices.While resuming from sleep, a USB disconnect message and re > enumeration messages are seen resulting in the reset of �persist � > variable in sysfs. The link above refers to a problem involving link training. USB-2.0 devices don't perform link training in the same way as SuperSpeed devices, so they don't suffer from the same problem. > In the above mentioned link a fix is proposed and that patch was > merged in mainline kernel, but the patch restricts the timeout only > to USB 3.0 devices. I am not sure why it was restricted to USB 3.0 > devices, but the same issue seems to appear for USB 2.0 devices too. No, it isn't the same problem. It may be a similar problem, but it can't be the same, for the reason mentioned above. > The below patch removes the restriction for USB 3.0 devices and makes > the time out applicable to all USB devices. > > Signed-off-by: vasudevan,krishna prasath <krishna.prasathx.k.vasudevan@xxxxxxxxx> > > --- > > drivers/usb/core/hub.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c > > index d2bd9d7..b2b709d 100644 > > --- a/drivers/usb/core/hub.c > > +++ b/drivers/usb/core/hub.c > > @@ -3320,7 +3320,7 @@ int usb_port_resume(struct usb_device *udev, pm_message_t msg) > > > > clear_bit(port1, hub->busy_bits); > > > > - if (udev->persist_enabled && hub_is_superspeed(hub->hdev)) > > + if (udev->persist_enabled) > > status = wait_for_ss_port_enable(udev, hub, &port1, &portchange, > > &portstatus); If you are going to do this then you also need to rename the wait_for_ss_port_enable function and update the comment preceding the function's definition. It now applies to all ports, not just to SuperSpeed (SS) ports. > RESULT: > > This patch has been tested with USB 2.0 devices(Transcend & Alcor > micro corp.) that were facing this issue and after applying this > patch, both the USB 2.0 devices seems to work fine without any issues > for all tested sleep-resume cycles . Please make this changes before you resubmit the patch. Alan Stern -- 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
