I will look into doing that. I think it would be preferable to add some functionality to mark all devices matching a certain VendorID rather than dumping the entire device table into `quirks.c', as I'd have to routinely update it every time we release a new product. In the meantime, if there's a way to disable LPM from within my software, I'd prefer to take that route as I could build, test, and ship a new version of my software to customers before end-of-week and not require them to build/use a custom kernel until the patch finds its way into production Ubuntu kernels. ============================================================ Matthew Giassa, MASc, BASc, EIT Security and Embedded Systems Specialist linkedin: https://ca.linkedin.com/in/giassa e-mail: matthew@xxxxxxxxxx website: www.giassa.net > -------- Original Message -------- > Subject: RE: USB xHCI regression after upgrading from kernel 3.19.0-51 > to 4.2.0-34.] > From: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> > Date: Wed, April 13, 2016 10:55 am > To: Matthew Giassa <matthew@xxxxxxxxxx> > Cc: Felipe Balbi <felipe.balbi@xxxxxxxxxxxxxxx>, Mathias Nyman > <mathias.nyman@xxxxxxxxxxxxxxx>, <linux-usb@xxxxxxxxxxxxxxx> > > > On Wed, 13 Apr 2016, Matthew Giassa wrote: > > > Hi Felipe, Mathias; > > > > I pulled a clean copy of v4.5 and applied the patch suggested by Felipe, > > and the issue is no longer present. I would like to disable this > > behavior via changes to my software instead of requiring customers to > > build a custom kernel. What would be the ideal solution to this issue? > > Would it be acceptable to wrap the lines of code with an if-else using a > > new kernel boot parameter? > > > > Mathias also provided a suggestion that I could put in to my code: > > > If that is the issue then adding pm_runtime_get_xxx() will increase the device > > > usage count before you know you are going to use a device. It prevents runtime suspend. > > > Then once you know the device is going to be idle, decrease the referece cound > > > with a pm_runtime_put(). > > > > May I please request a short code snippet on how I could get the `struct > > device*' handle, assuming I have a libusb_device_handle and the bus/dev > > number of the camera in use? I'm not very familiar with the core USB > > API, and would just need to know how to do the set/reset logic to > > prevent runtime suspend. > > It is possible to disable LPM for one particular device, based on the > vendor and product IDs. To do so would require a kernel patch, but > this patch would be accepted for the mainline kernel and therefore > would be available to everyone using these devices, not just the people > running your software. > > To write such a patch yourself, edit the drivers/core/usb/quirks.c > file and copy one of the entries for the Blackmagic Design devices, > changing the ID numbers as appropriate. The USB_QUIRK_NO_LPM flag is > the important factor. > > 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
