I recently got a new laptop (HP 2510p) and have been playing with
suspend/resume. Initially the dmesg after resume was fairly clean, but
that was with the bluetooth device disabled (I initially disabled it in
Windows).
After enabling it I get repeated warnings on resume starting with:
<snip>
usb 1-1: reset full speed USB device using uhci_hcd and address 3
hci_usb 1-1:1.0: no reset_resume for driver hci_usb?
hci_usb 1-1:1.1: no reset_resume for driver hci_usb?
usb_endpoint usbdev1.3_ep83: parent 1-1:1.1 is sleeping, will not add
------------[ cut here ]------------
WARNING: at drivers/base/power/main.c:82 device_pm_add+0x77/0xd3()
</snip>
The last one starts with:
<snip>
bluetooth hci0: parent 1-1:1.0 is sleeping, will not add
------------[ cut here ]------------
WARNING: at drivers/base/power/main.c:82 device_pm_add+0x77/0xd3()
</snip>
The call traces for these all start with "warn_on_slowpath".
And then there is also the following error. In the attached dmesg this is
at the end of the dmesg, but I've also seen it a lot earlier (before 'PM:
Syncing filesystems ... done.' even). I'm not completely sure if this is
also bluetooth related, but I don't remember seeing it before enabling
the bluetooth device.
<snip>
irq 19: nobody cared (try booting with the "irqpoll" option)
Pid: 0, comm: swapper Tainted: G W 2.6.27-rc3 #52
Call Trace:
<IRQ> [<ffffffff8026bdcf>] __report_bad_irq+0x38/0x7c
[...]
</snip>
The attached full dmesg is from start of boot and includes one suspend to
mem / resume cycle. Kernel config attached as well.
The bluetooth device is (from lsusb):
Bus 001 Device 003: ID 03f0:171d Hewlett-Packard Wireless (Bluetooth +
WLAN) Interface [Integrated Module]
As I've not tested this with earlier kernels I have no idea if this is a
regression or not.
Cheers,
FJP
> I recently got a new laptop (HP 2510p) and have been playing with
> suspend/resume. Initially the dmesg after resume was fairly clean, but
> that was with the bluetooth device disabled (I initially disabled it in
> Windows).
avoid hci_usb. It is racy as hell. btusb actually works.
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
On Wednesday 20 August 2008, Rafael wrote:
> On Wednesday, 20 of August 2008, Frans Pop wrote:
> > I recently got a new laptop (HP 2510p) and have been playing with
> > suspend/resume. Initially the dmesg after resume was fairly clean,
> > but that was with the bluetooth device disabled (I initially disabled
> > it in Windows).
> >
> > After enabling it I get repeated warnings on resume starting with:
> > <snip>
> > usb 1-1: reset full speed USB device using uhci_hcd and address 3
> > hci_usb 1-1:1.0: no reset_resume for driver hci_usb?
> > hci_usb 1-1:1.1: no reset_resume for driver hci_usb?
> > usb_endpoint usbdev1.3_ep83: parent 1-1:1.1 is sleeping, will not add
> > ------------[ cut here ]------------
> > WARNING: at drivers/base/power/main.c:82 device_pm_add+0x77/0xd3()
> > </snip>
>
> Yeah.
>
> http://bugzilla.kernel.org/show_bug.cgi?id=11263#c11
>
> The patch is in the Greg's tree, AFAICS.
Thanks! I'll keep an eye on that and check if the IRQ problem is solved
too after that patch gets in. If not, I'll try to narrow that one down a
bit further.
On Wednesday, 20 of August 2008, Frans Pop wrote:
> I recently got a new laptop (HP 2510p) and have been playing with
> suspend/resume. Initially the dmesg after resume was fairly clean, but
> that was with the bluetooth device disabled (I initially disabled it in
> Windows).
>
> After enabling it I get repeated warnings on resume starting with:
> <snip>
> usb 1-1: reset full speed USB device using uhci_hcd and address 3
> hci_usb 1-1:1.0: no reset_resume for driver hci_usb?
> hci_usb 1-1:1.1: no reset_resume for driver hci_usb?
> usb_endpoint usbdev1.3_ep83: parent 1-1:1.1 is sleeping, will not add
> ------------[ cut here ]------------
> WARNING: at drivers/base/power/main.c:82 device_pm_add+0x77/0xd3()
> </snip>
Yeah.
http://bugzilla.kernel.org/show_bug.cgi?id=11263#c11
The patch is in the Greg's tree, AFAICS.
Thanks,
Rafael