Teleconference information for software status meeting (today, 21:00 EDT Boston

Chris Ball cjb at
Tue Aug 14 22:29:14 EDT 2007

System software meeting minutes, 2007-08-14:

Joel Stanley, Bernando Innocenti, Mitch, Bryan, Luna, Richard Smith,
Jim, Jordan, Andres, Scott, Marcelo, Javier, Chris Ball


Wireless resume problems:
  We have firmware from Javier that appears to work, but we don't
  understand the root cause.  cjb will run suspend/resume tests overnight.

  (The short version -- the bug is exhibited when the wireless module
  sees traffic inbetween the host sending a HOSTSLEEP command and the
  module returning an ACK.  The bandaid fix is to disable the radio
  inbetween the two.  There still exists an unresolved race condition,
  probably in either the wireless hardware or firmware.)

Joel Stanley, Power measurement:
  520mW savings by turning the DCON/video off all the way, still
Chris Ball, Tinderbox:
  Added power measurement, suspend/resume tests, memory use tracking.
Richard Smith, C-build status update:
  pre-C build has happened, 12 machines.  Remainder of the build is
  waiting on batteries to clear customs; supposedly happened a few hours
  ago, so we'll get real C build machines today.  Audio doesn't work;
  probably the same lack of firmware recognition that happened at B4.

  Scott:  Is the write-protect (WP) bit being set in the mfg-data at Quanta?
  Richard:  Will check, think so.
Scott, activation:
  Waiting on a real activation server from Ivan, and then can get real
  crypto.  Worked out the key format with Ivan, Michael and Mitch.
  Jim:  Are the machines' manufacturing data going to our server now?
  Richard:  No, that happens right before shipping.

Scott, upgrades:
  Has a specification, waiting for feedback, then implement.

Bugs with action items:

#1396 -- We're waiting on the problem with keyboard repeat events.
* AI: Richard to get us a release firmware with this patched.

#1500 -- Need public key for mfg-data.
* AI: Jim to talk to Noah/Ivan to understand why we still don't
      have the public keys generated and signing working yet.

#1547 -- Decide frame buffer depth.
* AI: cjb/bernie to performance-test color depth change.

#1748 -- USB key upgrades currently disabled in USB autoreinstall
* AI: cscott to enable them.

#2184 -- OHM idleness detection isn't working
* AI: jg and cjb to meet and decide what OHM should do for Trial-3,
      mail the result to Richard Hughes and devel at .

#2237 -- Need to find the best DCON write combination for power-savings
* AI: Joel to do so and pass on to Andres to implement.

- Chris.
Chris Ball   <cjb at>

More information about the Devel mailing list