[OLPC-devel] Re: [One Laptop Per Child] #53: LinuxBIOS sometimes hangs after "Jumping to LinuxBIOS"

[Zarro Boogs per Child]

#53: LinuxBIOS sometimes hangs after "Jumping to LinuxBIOS"
-------------------------------+--------------------------------------------
 Reporter:  wmb at firmworks.com  |        Owner:  rminnich  
     Type:  defect             |       Status:  new       
 Priority:  blocker            |    Milestone:  rev1 alpha
Component:  linuxbios          |   Resolution:            
 Keywords:                     |  
-------------------------------+--------------------------------------------
Old description:

> I have run afoul of this problem on two separate occasions.  It would
> happen over and over - just the first few lines on the serial port, up to
> "Jumping to LinuxBIOS", and then nothing.
>
> The first time it happened, eventually I gave up and went to sleep.  The
> next morning,
> it still happened, but I managed to fix it by unplugging all the USB
> devices.  It worked without the USB devices, and when I plugged them back
> in, it continued to work.    On that occasion, I was attempting to run
> LinuxBIOS from a PROMjet ROM emulator plugged into the PLCC socket, with
> InsydeBIOS in the SPI FLASH.
>
> The second time was several days later.  By that time I had LinuxBIOS in
> the SPI FLASH, but I was still mainly running LinuxBIOS from the PROMjet.
> This second time,
> I tried unplugging the USB devices, but that didn't help.  I also tried
> unplugging the PROMjet and running the LinuxBIOS that was in SPI FLASH.
> That didn't work either.  So I switched to another board and used it for
> awhile.  The next day, I tried the "failing" board again and it worked,
> both from SPI FLASH and from the PROMjet.

New description:

 I've got a hopefully consistent report which seems to indicate
 hardware problems.

 After happily flashing a new LinuxBIOS version on my test-A board (to
 PLCC), I rebooted the board and surprise: no graphics output.

 Plugged in a serial cable, and watched LinuxBIOS fail to boot:

 LinuxBIOS-1.1.8.0Fallback Wed Aug 30 08:56:50 BRT 2006 starting...
 done cpuRegInit
                Ram1.00
 Ram2.00
 computed msr.hi 10075012
 Ram3
 Ram4
 Copying LinuxBIOS to ram.
 Jumping to LinuxBIOS.

 (hang)

 OK, so I removed the PLCC, and tried to boot from the SPI flash (which was
 working reliably before the "accident"), and surprise:

 LinuxBIOS-1.1.8.0Fallback Thu Aug 24 10:32:22 BRT 2006 starting...
 done cpuRegInit
                Ram1.00
 Ram2.00
 computed msr.hi 10075012
 Ram3
 Ram4
 Copying LinuxBIOS to ram.
 Jumping to LinuxBIOS.

 (hang, again!)

 By reading this Trac entry I noticed that Jim fixed such problem,
 on two different occasions, by unplugging the board from power for 30 or
 so minutes ("ghost register" theory).

 Tom suggested on IRC that removing the battery could also help.

 So my attempts to bring it back to life were:

 - Remove from power plug for 1+ hours
 - Remove battery
 - Interchange between PLCC / SPI flash

 With no success at all.

 So I was left with a pre-A board. Attempted to use the same PLCC chip on
 it and _SURPRISE_, it works:

 LinuxBIOS-1.1.8.0Fallback Wed Aug 30 08:56:50 BRT 2006 starting...
 done cpuRegInit
                Ram1.00
 Ram2.00
 computed msr.hi 10075012
 Ram3
 Ram4
 Copying LinuxBIOS to ram.
 Jumping to LinuxBIOS.
 LinuxBIOS-1.1.8.0Fallback Thu Aug 31 16:14:15 BRT 2006 booting...
 end 5ea1e600, start 0
 32-bit delta 768
 calibrate_tsc 32-bit result is 768
 clocks_per_usec: 768
 Enumerating buses...

 Now, my suspicion is that somehow the RAM in the test-A board
 is flaky.

 Why? How? NO CLUE.

 I'll try to power on the board tomorrow again.

-- 
Ticket URL: <http://dev.laptop.org/ticket/53#comment:18>
One Laptop Per Child <http://laptop.org/>