Re: [PATCH 2/2] USB: ehci-platform: Support ehci reset after resume quirk

[Doug Anderson]

Alan,

On Mon, Dec 22, 2014 at 8:06 AM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
> On Mon, 22 Dec 2014, Wu Liang feng wrote:
>
>> The EHCI controller doesn't properly detect the case when
>
> "The" EHCI controller?  I don't know what EHCI controller you're
> talking about, but my controllers don't have any trouble detecting
> device removal during suspend.

This is specifically the EHCI controller on rk3288.  Not sure why Wu
Liang feng removed that part of the description.


>> a device is removed during suspend. Specifically,when we
>> resume from suspend the EHCI controller maintaining the
>> USB state (FLAG_CF is 1 Current Connect Status is 1) but
>> a USB device (like a USB camera on rk3288) may have been
>> disconnected actually.
>>
>> Let's add a quirk to force ehci to go into the
>> usb_root_hub_lost_power() path and reset after resume.
>> This should generally reset the whole controller and all
>> ports and initialize everything cleanly again, and bring
>> the devices back up.
>
> Isn't this solution too extreme?  What if the device was a flash
> storage drive and it wasn't unplugged during suspend?  This patch would
> force it to be removed, messing up any mounted filesystems, when there
> was no need.

I'm told by Julius (CCed, who knows the USB stack infinitely better
than I do) that you can work around this using "persist".  I would
imagine that anyone on a machine using hibernation would run into the
same problem, right?


> Can you find a better way to work around the problem?

We asked a lot about this and you can find a whole set of detailed
discussion at:

https://chromium-review.googlesource.com/#/c/232077/1/drivers/usb/host/ehci-platform.c


Since I don't know the USB subsystem particularly well, my review
isn't terribly meaningful, but just in case:

Reviewed-by: Doug Anderson <dianders@xxxxxxxxxxxx>


On rk3288-pinky I have tested that without this patch the EHCI
controller flips out when a USB webcam is plugged into the port and is
power cycled across suspend/resume.  With this patch the controller
properly unplugs / replugs the webcam.

Tested-by: Doug Anderson <dianders@xxxxxxxxxxxx>


-Doug
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/