On Fri, Nov 15, 2013 at 09:18:34PM +0000, Paul Zimmerman wrote: > > From: Felipe Balbi [mailto:balbi@xxxxxx] > > Sent: Friday, November 15, 2013 9:31 AM > > To: Greg KH > > Cc: Paul Zimmerman; balbi@xxxxxx; linux-usb@xxxxxxxxxxxxxxx > > Subject: Re: [PATCH v2] Move DWC2 driver out of staging > > > > On Fri, Nov 15, 2013 at 01:33:05PM +0900, Greg KH wrote: > > > On Thu, Nov 14, 2013 at 02:50:24PM -0800, Paul Zimmerman wrote: > > > > DWC2 driver should be in good enough shape to move out of staging > > > > > > > > Signed-off-by: Paul Zimmerman <paulz@xxxxxxxxxxxx> > > > > --- > > > > Greg, is this the proper method for moving a driver out of staging? > > > > > > Yes, that's the way to do it. > > > > > > But, are all of the TODO entries really done? I don't want to have that > > > file in drivers/usb/dwc2/ as that doesn't make much sense, right? > > > > I would suggest running some extensive week-long stress tests with some > > known devices just to make sure. > > > > You could use g_zero running on a PLX card and usbtest to test your > > host. Or take some mass storage device and run some long tests with it. > > > > You shouldn't see any failures for weeks of stress tests. It's also > > important to exercise control transfers as much as you can, since a bug > > there would prevent enumeration. > > > > It should be very simple to write up some quick tool using libusbx to do > > most of it. > > > > I wrote a few myself and use them every now and again to test a bunch of > > things. They're all in a branch on gitorious [1]. note that I have a > > copy of testusb.c, but don't use that, it's best to use the one shipping > > with the kernel. > > > > [1] https://gitorious.org/usb/usb-tools/ > > Looks interesting. Did you find many SW or HW bugs using this? SW mostly. The only HW bug we found needed special packet generation with Lecroy's analyzer ;-) -- balbi
Attachment:
signature.asc
Description: Digital signature