On Tuesday, November 01, 2011, Linus Torvalds wrote:
> On Tue, Nov 1, 2011 at 7:47 AM, Linus Torvalds
> <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > Will try.
>
> Ok, the delay makes the USB port work for me. I still don't think it's
> the *right* approach, though, as per the previous mail. But it does
> make that port work.
>
> I also have to say that I think there is something *seriously* rotten
> in runtime power management. Maybe it's just this particular laptop,
> but look at what happens when I plug a device into this machine with
> everything set to be runtime power-managed...
>
> Some background:
>
> - "powertop" used to make everything power-managed
>
> - I tested with two different USB devices just to be sure
>
> - for the last device (which is a serial device, it shows up as
> ftdi_sio), in order to show you the crazy behavior, I did
>
> dmesg > before
> .. plug in device ..
> dmesg > after
> diff -u before after > plug-in
>
> - "lsusb -t" output with that device plugged in is:
>
> 1-1.3:1.0: No such file or directory
> /: Bus 04.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/2p, 5000M
> /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/2p, 480M
> |__ Port 1: Dev 4, If 0, Class=vend., Driver=ftdi_sio, 12M
> /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=ehci_hcd/3p, 480M
> |__ Port 1: Dev 2, If 0, Class=hub, Driver=hub/8p, 480M
> /: Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci_hcd/3p, 480M
> |__ Port 1: Dev 2, If 0, Class=hub, Driver=hub/6p, 480M
> |__ Port 3: Dev 3, If 0, Class=vend., Driver=, 12M
> |__ Port 6: Dev 4, If 0, Class='bInterfaceClass 0x0e not yet
> handled', Driver=uvcvideo, 480M
> |__ Port 6: Dev 4, If 1, Class='bInterfaceClass 0x0e not yet
> handled', Driver=uvcvideo, 480M
>
> just to show that the ftdi_sio thing is directly under the xhci root hub.
>
> With that said, appended is the crazy diff of dmesg output ("plug-in").
>
> NOTE! The absolute *only* thing I did was to plug in the one USB
> device into the xhci port. Comments/questions:
>
> - That e1000e thing seems to be just coincidence. It seems to happen
> every few minutes or so, it just happened to happen three seconds
> before the plug-in.
>
> - But that ehci_hcd thing is *not* pure coincidence. It consistently
> happens when I plug something into the xhci port. In fact, it seems to
> happen for *any* runtime PM event, because as far as I can tell, it
> also happens each time that e1000e thing happens. WTF?
>
> IOW, there seems to be some random cross-talk between devices, where
> that stupid ehci thing happens every time something else should
> happen. Notice in the trace below how the ehci spewage happens both at
> 1010 (e1000e power event) and at 1013 (xhci plug-in event).
>
> Ideas?
I seem to have seen that already.
Can you please attach the output of acpidump from your machine perchance?
Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
