[Trac #318] Wireless firmware load failures

[Zarro Boogs per Child]

#318: Wireless firmware load failures
---------------------+------------------------------------------------------
 Reporter:  mfoster  |        Owner:  mfoster 
     Type:  defect   |       Status:  assigned
 Priority:  blocker  |    Milestone:  BTest-1 
Component:  distro   |   Resolution:          
 Keywords:           |  
---------------------+------------------------------------------------------
Changes (by mfoster):

  * status:  new => assigned

Old description:

> Dave Woodhouse has seen occasional failures where the wireless firmware
> will fail to load properly when booting from NAND Flash.  This also
> occurs when attempting to update the wireless firmware.  To help chase
> this bug, the firmware update process was used as an easier way to induce
> the problem.
>
> We've caught four USB protocol analyzer traces.
>
> 1) Successful upgrade of firmware from 31.02 to 31.04:
>
> http://www.talix.com/Successful_Upgrade.ufo
>
> 2) Successful downgrade of firmware from 31.04 to 31.02:
>
> http://www.talix.com/Successful_Downgrade.ufo
>
> 3) Unsuccessful upgrade of firmware from 31.02 to 31.04:
>
> http://www.talix.com/Unsuccessful_Upgrade_Fail_Mode_1
>
> [Failure mode 1 means that we see a single 'R' erase block error from the
> wireless firmware update tool, then a series of ...s.]
>
> 4) Unsuccessful downgrade offirmware from 31.04 to 31.02:
>
> http://www.talix.com/Unsuccessful_Downgrade_Fail_Mode_2
>
> [Failure mode 2 means that we will have a "hard" erase error, where the
> wireless firmware update tool will print four 'R's, then error out.
>
> When booting from USB, it appears that these failures do not occur. We're
> uncertain of precisely what USB bus condition is causing the failures.
>
> We would sincerely appreciate assistance from Marvell and AMD on these
> USB traces.

New description:

 Dave Woodhouse has seen occasional failures where the wireless firmware
 will fail to load properly when booting from NAND Flash.  This also occurs
 when attempting to update the wireless firmware.  To help chase this bug,
 the firmware update process was used as an easier way to induce the
 problem.

 We've caught four USB protocol analyzer traces.

 1) Successful upgrade of firmware from 31.02 to 31.04:

 http://www.talix.com/Successful_Upgrade.ufo

 2) Successful downgrade of firmware from 31.04 to 31.02:

 http://www.talix.com/Successful_Downgrade.ufo

 3) Unsuccessful upgrade of firmware from 31.02 to 31.04:

 http://www.talix.com/Unsuccessful_Upgrade_Fail_Mode_1.ufo

 [Failure mode 1 means that we see a single 'R' erase block error from the
 wireless firmware update tool, then a series of ...s.]

 4) Unsuccessful downgrade offirmware from 31.04 to 31.02:

 http://www.talix.com/Unsuccessful_Downgrade_Fail_Mode_2.ufo

 [Failure mode 2 means that we will have a "hard" erase error, where the
 wireless firmware update tool will print four 'R's, then error out.

 When booting from USB, it appears that these failures do not occur. We're
 uncertain of precisely what USB bus condition is causing the failures.

 We would sincerely appreciate assistance from Marvell and AMD on these USB
 traces.

-- 
Ticket URL: <http://dev.laptop.org/ticket/318#comment:1>
One Laptop Per Child <http://laptop.org/>