On Thu, Apr 02, 2015 at 01:48:35PM -0500, Liu, Bin wrote: > Felipe, > > > -----Original Message----- > > From: Balbi, Felipe > > Sent: Thursday, April 02, 2015 1:22 PM > > To: Liu, Bin > > Cc: linux-usb@xxxxxxxxxxxxxxx; Balbi, Felipe > > Subject: Re: [PATCH v2] usb: musb: only set test mode once > > > > On Thu, Apr 02, 2015 at 11:44:30AM -0500, Liu, Bin wrote: > > > Felipe, > > > > > > > -----Original Message----- > > > > From: Liu, Bin > > > > Sent: Tuesday, March 24, 2015 3:09 PM > > > > To: linux-usb@xxxxxxxxxxxxxxx > > > > Cc: Balbi, Felipe; Liu, Bin > > > > Subject: [PATCH v2] usb: musb: only set test mode once > > > > > > > > The MUSB test mode register can only be set once, otherwise the > > > > result is undefined. > > > > > > > > This prevents the debugfs testmode entry to set the register more > > > > than once which causes test failure. > > > > > > > > Signed-off-by: Bin Liu <b-liu@xxxxxx> > > > > > > I see you have not picked up this patch yet. Can you please hold on > > > it? > > > > > > I noticed only the USB2.0 Specs mentioned about USB reset to exit test > > > mode, but not in the MUSB data manual. > > > > > > I am guessing running multiple 'test packet' command causing incorrect > > > test packets is due to the driver loading the test packets into EP0 > > > FIFO multiple times. I simply masked this then the analyzer trace > > > shows the test packets are correct even after ran 'test packet' > > > command multiple times. > > > > > > I will have to hook up a scope to measure the signal to ensure the > > > DP/DM signals are still correct after run each test mode command > > > multiple times for MUSB. If so, then the patch we need is to ensure > > > not loading the test packet to FIFO multiple times. > > > > MUSB still needs to obey what USB 2.0 specification says and if it says we must > > reset, then we must reset ;-) > > This is right. But the Specs only says the hub reset is required when > *exit* test mode, it did not say a reset is required when switching > between test modes. to enter a new test mode, you must first exit the previous one ;-) > I really don't care about reset or not, it is just about board reboot, > not a big deal. The only thing I want to do is prevent the *** user > using the current driver loading test packets to FIFO multiple times > which breaks the test. You don't need to reboot the board at all. When USB spec says 'reset', it's referring to a USB bus reset. See if my device-reset tool helps you out [1] [1] https://gitorious.org/usb/usb-tools/source/373b417a9f0ded25312d0626486f6e7a2d660c32:device-reset.c -- balbi
Attachment:
signature.asc
Description: Digital signature
