#12032 NORM Not Tri: XO-1.5s in critical low-battery conditions do not cleanly shutdown or restart

[Zarro Boogs per Child]

#12032: XO-1.5s in critical low-battery conditions do not cleanly shutdown or
restart
---------------------------------+------------------------------------------
 Reporter:  greenfeld            |                 Owner:  rsmith                           
     Type:  defect               |                Status:  new                              
 Priority:  normal               |             Milestone:  Not Triaged                      
Component:  embedded controller  |               Version:  Development build as of this date
 Keywords:                       |           Next_action:  diagnose                         
 Verified:  0                    |   Deployment_affected:                                   
Blockedby:                       |              Blocking:                                   
---------------------------------+------------------------------------------
 I am doing runin testing with Q3C07 with the battery discharge test turned
 on.  This testing causes a lot of kernel hangs.

 I have a strong hunch that XO-1.5s which die with the discharge test
 active are misbehaving instead of having the embedded controller detect
 the critical battery shutdown and power off the XO.

 This cases the XO to stay turned on the EC is supposed to turn it off
 *without* Linux turning the XO off first.

 When this was seen once in person instead of observed remotely, the XO-1.5
 in question went into reboot loops with the AC adapter plugged in.
 Removing the AC adapter caused the XO to turn on with the display lit but
 it did not get further than that.  A battery pull to reset the EC was
 required to restore proper operation.

 This usually shows up as a series of reboots where "Forth+" is printed
 repeatedly:

 {{{
 [ 4251.337237] will schedule 98 mmcqd/2
 [ 4251.340865] done CS
 [ 4251.342970] will schedule 430 rtcwake
 [ 4251.346667] done CS
 [ 4251.348778] done.
 +
 Forthmacs
 Type 'i' to interrupt stand-init sequence
 Unknown value in TS tag
 USB2 devices:
 USB1 devices:
 OLPC D5, 1 GiB memory installed, 4 GB internal storage, S/N SHC01901E0B
 OpenFirmware  Q3C07   EC Firmware Ver:2.2.10   2012-07-28 00:36:21 UTC

 Type the Esc key to interrupt automatic startup
 Boot device: /sd/disk at 3:\boot\olpc.fth  Arguments:
 Boot device: /pci/sd at c/disk at 3:\boot\vmlinuz  Arguments:
 +
 Forthmacs
 Type 'i' to interrupt stand-init sequence
 Unknown value in TS tag
 USB2 devices:
 USB1 devices:
 OLPC D5, 1 GiB memory installed, 4 GB internal storage, S/N SHC01901E0B
 OpenFirmware  Q3C07   EC Firmware Ver:2.2.10   2012-07-28 00:36:32 UTC

 Type the Esc key to interrupt automatic startup
 Boot device: /sd/disk at 3:\boot\olpc.fth  Arguments:
 +
 Forthmacs
 Type 'i' to interrupt stand-init sequence
 Unknown value in TS tag
 USB2 devices:
 USB1 devices:
 OLPC D5, 1 GiB memory installed, 4 GB internal storage, S/N SHC01901E0B
 OpenFirmware  Q3C07   EC Firmware Ver:2.2.10   2012-07-28 00:36:40 UTC

 Type the Esc key to interrupt automatic startup
 Boot device: /sd/disk at 3:\boot\olpc.fth  Arguments:
 +
 Forth+
 Forth+
 Forth+
 Forth+
 (...)
 }}}

 One system right now isn't getting that far, and is repeatedly logging
 "Fort+" multiple times every second, so far thousands of times.

 A third XO-1.5 currently in this condition simply decided it had invalid
 firmware:

 {{{
 [ 4404.397737] will schedule 81 mmcqd/2
 [ 4404.401408] done CS
 [ 4404.403552] will schedule 2693 rtcwake
 [ 4404.407388] done CS
 [ 4404.409542] done.
 [ 4429.487272] PM: Syncing filesystems ... +
 Forthmacs
 Type 'i' to interrupt stand-init sequence
 Unknown value in TS tag
 USB2 devices:
 USB1 devices:
 OLPC D7, 1 GiB memory installed, 4 GB internal storage, S/N SHC10000001
 OpenFirmware  Q3C07   EC Firmware Ver:2.2.10   2012-07-27 21:48:52 UTC

 Type the Esc key to interrupt automatic startup
 Boot device: /sd/disk at 3:\boot\olpc.fth  Arguments:
 SDHCI: Error: ISR = 8000 ESR = 20 Data CRC,
 Command reg: 123a Mode reg: 37 Arg reg: 1303a
 Recent commands (decimal): 18 18 18 18 18 18 18 18
 Stopping
 Invalid Firmware image
 OLPC D7, 1 GiB memory installed, 4 GB internal storage, S/N SHC10000001
 OpenFirmware  Q3C07   EC Firmware Ver:2.2.10   2012-07-27 21:48:55 UTC

 Powering off in 30 seconds
 }}}

 No motherboard revision or SKU dependance seems to be required to cause
 this.  Sometimes the battery charge LED never turns on after the XO dies.

 Expected: XO-1.5s should be able to power off safely in low-power
 conditions which Linux does not power off the XO (even potentially in
 runin).  After this happens the EC should be able to restart the XO with
 external power attached.

 Actual: Currently this does not appear to be the case.

-- 
Ticket URL: <http://dev.laptop.org/ticket/12032>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system