#6053 NORM Future : Clock drift during suspend/resume.

[Zarro Boogs per Child]

#6053: Clock drift during suspend/resume.
---------------------+------------------------------------------------------
  Reporter:  cscott  |       Owner:  JordanCrouse  
      Type:  defect  |      Status:  new           
  Priority:  normal  |   Milestone:  Future Release
 Component:  kernel  |     Version:                
Resolution:          |    Keywords:                
  Verified:  0       |    Blocking:                
 Blockedby:          |  
---------------------+------------------------------------------------------

Comment(by nealmcb):

 Here is some data using build 650 on a G1G1 machine.  I followed the
 directions in #1396 to turn on suspend during ebook mode (`touch
 /home/olpc/ebook-enable-sleep` to turn it on, then relaunch ebook).

 Then from the journal I resumed a pdf file which I had previously
 downloaded, and logged in via ssh so I could easily cut-and-paste some
 data.  I ran ntpdate against my local ntp server.   Then I used the "down"
 game key to scroll down, which induces the laptop to go to sleep after 5
 seconds or so, then used "enter" in the ssh window to wake it back up.  I
 repeated the ntpdate and sleep several times, and observed the time
 jumping both backwards and forwards by up to 0.6 seconds, as documented
 below.

 {{{
 -bash-3.2# ntpdate -q 192.168.1.15  # show it was initially in sync
 server 192.168.1.15, stratum 3, offset 0.001929, delay 0.02710
 21 Jan 17:40:01 ntpdate[1800]: adjust time server 192.168.1.15 offset
 0.001929 sec
 -bash-3.2#
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.440471, delay 0.02705
 21 Jan 17:40:46 ntpdate[1809]: adjust time server 192.168.1.15 offset
 0.440471 sec
 -bash-3.2#
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 1.010302, delay 0.02705
 21 Jan 17:41:19 ntpdate[1816]: step time server 192.168.1.15 offset
 1.010302 sec
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.772156, delay 0.02698
 21 Jan 17:41:35 ntpdate[1823]: step time server 192.168.1.15 offset
 0.772156 sec
 -bash-3.2#
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.590046, delay 0.02707
 21 Jan 17:41:52 ntpdate[1830]: step time server 192.168.1.15 offset
 0.590046 sec
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 1.033643, delay 0.02708
 21 Jan 17:42:08 ntpdate[1837]: step time server 192.168.1.15 offset
 1.033643 sec
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.415472, delay 0.02718
 21 Jan 17:42:20 ntpdate[1844]: adjust time server 192.168.1.15 offset
 0.415472 sec
 }}}

 Then I ran it without suspending several times in a row just to
 demonstrate a stable clock:

 {{{
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.416191, delay 0.02708
 21 Jan 17:42:26 ntpdate[1845]: adjust time server 192.168.1.15 offset
 0.416191 sec
 -bash-3.2# ntpdate -q 192.168.1.15
 server 192.168.1.15, stratum 3, offset 0.416773, delay 0.02705
 21 Jan 17:42:31 ntpdate[1846]: adjust time server 192.168.1.15 offset
 0.416773 sec
 }}}

-- 
Ticket URL: <http://dev.laptop.org/ticket/6053#comment:3>
One Laptop Per Child <http://dev.laptop.org>
OLPC bug tracking system