Re: small regression in mediatree/for_v3.7-3 - media_build

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



--- On Sat, 3/11/12, VDR User <user.vdr@xxxxxxxxx> wrote:

> On Tue, Aug 14, 2012 at 4:51 PM,
> Antti Palosaari <crope@xxxxxx>
> wrote:
> >> There seems to be a small regression on
> mediatree/for_v3.7-3
> >> - dmesg/klog get flooded with these:
> >>
> >> [201145.140260] dvb_frontend_poll: 15 callbacks
> suppressed
> >> [201145.586405] usb_urb_complete: 88 callbacks
> suppressed
> >> [201150.587308] usb_urb_complete: 3456 callbacks
> suppressed
> >>
> >> [201468.630197] usb_urb_complete: 3315 callbacks
> suppressed
> >> [201473.632978] usb_urb_complete: 3529 callbacks
> suppressed
> >> [201478.635400] usb_urb_complete: 3574 callbacks
> suppressed
> >>
> >> It seems to be every 5 seconds, but I think that's
> just klog skipping
> >> repeats and collapsing duplicate entries. This does
> not happen the last time
> >> I tried playing with the TV stick :-).
> >
> > That's because you has dynamic debugs enabled!
> > modprobe dvb_core; echo -n 'module dvb_core +p' >
> > /sys/kernel/debug/dynamic_debug/control
> > modprobe dvb_usbv2; echo -n 'module dvb_usbv2 +p' >
> > /sys/kernel/debug/dynamic_debug/control
> >
> > If you don't add dvb_core and dvb_usbv2 modules to
> > /sys/kernel/debug/dynamic_debug/control you will not
> see those.
> 
> I'm getting massive amounts of "dvb_frontend_poll: 20
> callbacks
> suppressed" messages in dmesg also and I definitely did not
> put
> dvb_core or anything else in
> /sys/kernel/debug/dynamic_debug/control.
> For that matter I don't even have a
> /sys/kernel/debug/dynamic_debug/control file.
> 
> > I have added ratelimited version for those few debugs
> that are flooded
> > normally. This suppressed is coming from ratelimit - it
> does not print all
> > those similar debugs.
> 
> I'm using kernel 3.6.3 with media_build from Oct. 21, 2012.
> How I can
> disable those messages? I'd rather not see hundreds,
> possibly
> thousands or millions of those messages. :)

okay, if you had followed the threads further down, you would probably see that I patched the host side copy of device.h, i.e:

/lib/modules/`uname -r `/build/include/linux/device.h

to get around this, since media_build does not use a full dev kernel tree, but uses the host-side copy of the kernel headers.

So you need to apply the patch (attached) to that. i.e.
  cd /lib/modules/`uname -r `/build/include/linux/
  patch -p3 < /tmp/patch

Mind you this messes up your kernel-dev headers. In my case, because that file is part of the fedora distro kernel-devel package, and as soon as I get a new kernel the whole thing is thrown away, so I don't care. But if you compile your own kernel (and hence have your own kernel-devel headers which you want to keep in clean state), you might want to take note about this.

Hope this helps.

Attachment: patch
Description: Binary data


[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]
  Powered by Linux