Saturday, March 9, 2013


We've gone down the path of CyanogenMod for the Nexus 4.  The Android development community has an apparent average age of about 14, and all the arrogance and haughty superior attitude of, say, your typical Los Alamos National Laboratory staff member in the mid-80's, but at least they know how to bang out code. It's kind of refreshing actually after the ponderous interactions with "Fuck you, we're Google".

The two primary outstanding N4 bugs that we've been tracking, and seeing zero progress from Google are
  1. Nexus 4 stops responding to ARP requests when screen is off with wifi on, and
  2. Bluetooth kills wifi on the Nexus 4.
We've reported both of these on the CyanogenMod Jira bug tracking site. Let's see if the CM folks can make better progress on them than Google.  Actually, if they do *anything* they'll be making better progress than Google, so the bar really isn't set all that high.

We've been favorably impressed with CM 10.1 M2 experimental release, as well as the recent nightly builds, so we're holding out some hope that a more agile, more energized bunch than the Google behemoth can finally get these issues fixed.



  1. I think I'd like to add a third bug. Amusingly the 1st doesn't affect me at home; my Debian Linux server runs DHCP for the house and a static lease for the phone appears to also solve the ARP problem.

    However, MY Nexus 4 will sometimes (every couple of days) not come back from deep sleep, needing a power cycle to recover. Often after I've changed locations but sometimes just after a prolonged sleep (eg. overnight & left on charge)

    /proc/last_kmsg suggests the last thing it's doing is powering off the HPD circuitry, so I don't know if this figures the HDMI driver, or whatever happens next.

    However, there was some suggestion that it might be the WiFi power optimisation feature, so I'm experimenting with that turned off. So far, so good.

    1. Interesting, thanks. Good to hear that your Nexus 4 does not that the "suspend wifi dropout" issue at home. I'm using the workaround described here:

      Re: the other bug where bluetooth kills wifi, someone posted a comment on my Cynaogen bug report for this that it only happens with wifi routers using the 2.4 GHz band, and not with the 5GHz band. You can see the comment here:

    2. The bluetooth bug sort of makes sense, insofar as they're both using 2.4GHz.

      It's kind of moot, since SOMETHING in JellyBean Bluetooth crashes my car's (BMW 2008 3-series) built-in handsfree (locks up the entire radio / telematics unit, in fact) if I try to interact with it.

      Gingerbread (Exynos Galaxy S2) worked just fine, but CM10 on the S2 and stock Nexus 4 don't, at all. I initially blamed CM10, but it looks more fundamental than that.

  2. Hey, 12:33.

    I have an old HTC Thunderbolt running Gingerbread 2.3, and bluetooth and wifi coexist well on it, never a problem. BT and wifi might use overlapping frequency bands, but that should not prevent them from coexisting by using different frequencies.