[pvrusb2] Module ehci-hcd may prevent pvrusb2 reconnect

[roger]

On Thu, 2008-03-27 at 23:16 -0500, Mike Isely wrote:
> On Thu, 27 Mar 2008, roger wrote:
> 
> > Found a bug, if the pvrusb2 device is disconnected via power cable and
> > then plugged back in with the ehci-hcd module still loaded, the device
> > may fail initializing and the system logger will only show:
> > 
> > Mar 27 19:59:07 pvrusb2: unregistered device video0 [mpeg]
> > Mar 27 19:59:07 pvrusb2: unregistered device radio0 [mpeg]
> > Mar 27 19:59:07 msp3400 1-0040: chip reset failed
> > 
> > Or, you may only get a one or two line notice stating the pvrusb2 has
> > been plugged in (with no device initialization).
> > 
> > 
> > A work around, compile ehci-hcd as module and manually reload
> > (rmmd/modprobe) the ehci-hcd module.  Unloading the pvrusb2 module or
> > any of it's dependents (msp3400) doesn't seem to be need for the pvrusb2
> > to initialize.  Matter of fact, just removing the ehci-hcd module (ie.
> > rmmod ehci-hcd) without even loading the ehci-hcd module will
> > automatically allow the pvrusb2 device to initialize.
> > 
> > This may have been a long standing issue for the past year or more.
> > 
> 
> Really?  I'm surprised.  This is the first report of a problem like this 
> that I've seen.
> 
> Reloading the USB host controller driver is going to have the same 
> effect as re-plugging in everything that was attached to the 
> controller's USB ports.  So of course the pvrusb2 device will be 
> re-enumerated by this action.

Well, this is the "The Enhanced Host Controller Interface (EHCI) is
standard for USB 2.0 (high speed) (480 Mbit/sec, 60 Mbyte/sec) host
controller".

If I'm not mistaken, the ehci-hid driver provides "high speed" rates
above the hid 1.0 controller driver?

Even though removing usb controller re-enumerate attached devices, the
controller drivers should not prevent a device from being accessed as in
my condition.  (Sounds to me like a USB driver bug.)

> When a pvrusb2 device is plugged in, you should ALSO see a message like 
> this:
> 
> usb 5-2: new high speed USB device using ehci_hcd and address 6
> 
> (with different address values of course)

You're correct.  This is the line I was referring too.. for some reason,
couldn't get my device to spawn this line, but it usually does show
*this* line and nothing about msp3400 failing.

> That message comes directly from the USB core not the pvrusb2 driver.  
> If it doesn't appear, then the USB core is simply failing to spot the 
> newly plugged in device - and there is very very little that a USB 
> device could do to prevent itself from being at least seen by the USB 
> controller (detection is done via a resistor connected between ground 
> and one of the USB data wires).  Said another way, if you are having 
> this problem, then (1) it can't be from the pvrusb2 driver, and (2) it 
> can't be from the pvrusb2 hardware either.  Sounds more likely that 
> you've got general USB issues here.
> 
> I routinely power cycle plugged in devices here.  Reinitialization works 
> every time.  Are you SURE about this?
> 
>   -Mike

Again, you're correct.  I was merrily stating it here.  Since I don't
work on USB dev, figured this would be the best place for me to file for
now and this is probably more like FAQ issue here in case anybody else
sees this issue and starts wondering why the device isn't initializing.


--
Roger
http://www.eskimo.com/~roger/index.html
Key fingerprint = 8977 A252 2623 F567 70CD 1261 640F C963 1005 1D61

Thu Mar 27 21:10:52 AKDT 2008