Print

[erich]

Hi,

Please could someone explain the process to modify uboot to boot without the jtag attached. I understand from another post it is something to do with the serial console output. what are the repercussions of doing this? (I suspect not being able to debug the boot processes)

[Confusticated]

modify uboot to boot without the jtag attached

I don't understand your post, this is the default condition, nothing requires modification.

[erich]

I have 2 factory installed Debian Dreamplugs both of which require a Jtag unit (not even connected to USB) just to boot - once booted it can be detached to boot other DP - this is obviously not very convenient. I have seen other posts (one guy described using a DP in his car with jtag attached) with the same problem but no one has a detailed fix.

[Confusticated]

Please post u-boot version and printenv output

EDIT: Possibly useful test...
Power up the Dreamplug with the JTAG Serial Console (UART) connected and the JTAG Interface disconnected, post any output.

And the Dreamplug Serial Number...

[erich]

DP still boots if jtag->serial connected and jtag->jtag disconnected but not the other way round

I hope this is the info what you meant-

Marvell>> version

U-Boot 2011.06-02334-g8f495d9-dirty (May 31 2011 - 02:06:26)
Marvell-DreamPlug


Marvell>> printenv
bootcmd=setenv ethact egiga0; ${x_bootcmd_ethernet}; setenv ethact egiga1; ${x_bootcmd_ethernet}; ${x_bootcmd_usb}; ${x_bootcmd_kernel}; setenv bootargs ${x_bootargs} ${x_bootargs_root}; bootm 0x6400000;
bootdelay=3
baudrate=115200
x_bootcmd_ethernet=ping 192.168.2.1
x_bootcmd_usb=usb start
x_bootcmd_kernel=fatload usb 0 0x6400000 uImage
x_bootargs=console=ttyS0,115200
x_bootargs_root=root=/dev/sda2 rootdelay=10
ethact=egiga0
ethaddr=F0:AD:4E:00:B5:08
eth1addr=F0:AD:4E:00:B5:09
stdin=serial
stdout=serial
stderr=serial

Environment size: 524/4092 bytes

[Confusticated]

still boots if jtag->serial connected

OK I have another test if you wish to try it...

Power up the Dreamplug into Linux with jtag->serial connected (and being monitored with minicom, cu, screen, Putty,  whatever you prefer), this is to ensure that the USB\Serial detects the correct baud rate and retains it (do not exit your console monitor application at any time during this test).
Shutdown linux ('poweroff' at the console prompt) and power down the Dreamplug.
Disconnect the JTAG->serial, power the Dreamplug back up (which should 'fail to boot').
Now with everything live, reconnect the jtag->serial, and strike the enter key several times.

If the 'Marvell>>' u-boot prompt appears, we have a result!
I will explain why later if it works

[erich]

I am now losing the will to live .

It appears (and this may change tomorrow ) that if the jtag is not connect at all it now boots correctly ( I tried this 3 times to be sure). If jtag-> serial connected and USB connected it boots normally if jtag->serial connected but USB disconnected it stops are Marvell>> prompt

If Jtag->serial and Jtag->jtag connected but USB not connected it hangs until USB connected then starts.

I have now restarted the DP with nothing except Ethernet. I tried tried to rsh to it - no joy (No route to host), but I can access the lighttpd server running. Then I connect the jtag->serial. I can then attach a second rsh session. I am wondering this is a booting order issue with rsh.  

Arrrgghhhh - another reboot to retest last test. So nothing but Ethernet. Same result  with rsh. Attach jtag->serial and USB ; this time Marvell>> prompt - it is not consistent.

[Confusticated]

if jtag->serial connected but USB disconnected it stops

If Jtag->serial and Jtag->jtag connected but USB not connected it hangs

The above is the expected behaviour, if you don't want to disconnect the JTAG Unit from the Dreamplug  (performing a test away from a PC, say) then
connect the JTAG Unit USB lead to the Dreamplug also (to provide power to the Unit).

http://www.newit.co.uk/forum/index.php/topic,2525.msg7069.html#msg7069

tried to rsh to it- no joy (No route to host), but I can access the lighttpd server running

'rsh'...I assume you mean 'ssh'...

If there is 'No route to host' for ssh, then why is there a route for HTTP ?
What is it that your web browser is doing that is different (are you using the correct IP for 'ssh') ?

[erich]

OK I think I understand the jtag stuff now

Yes it is ssh

I tried tried to rsh to it - no joy (No route to host), but I can access the lighttpd server running. Then I connect the jtag->serial. I can then attach a second rsh session. I am wondering this is a booting order issue with rsh. 

Both ssh and http connection are on the same ip address

And I have no idea why the following is happening...

Arrrgghhhh - another reboot to retest last test. So nothing but Ethernet. Same result  with rsh. Attach jtag->serial and USB ; this time Marvell>> prompt - it is not consistent.

I reran this test again - it booted properly, web server served static pages but not dynamic ones until I had connected via rsh (jtag not attached). Then dynamic pages worked.  coincidence? The access logs don't give any clues - so now for some more reboots
 
Thanks for your help

[erich]

It would seem patience is golden - eventually the the dynamic pages are displayed.

Though it does not really explain some of the other goings on:
Such as getting stuck at the Marvell>> prompt
No ssh access until jtag connected - sometimes
A few odd unexplained crashes/reboots along the way

Have I missed something?

If the 'Marvell>>' u-boot prompt appears, we have a result!
I will explain why later if it works