• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 97 subscribers
  • Views 600 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Linux did not start

Stefan Sals
Prodigy 60 points
Other Parts Discussed in Thread: OMAP-L138

Hi,

we have strong problems getting Linux to run on our custom board with an OMAP-L138.
-> U-Boot is working and loads kernel image from ftp-server
-> during U_Boot is running communikation via UART2 works
-> After "Uncompressing Linux... done, booting the kernel." nothing is to see

On our EVM (OMAP-L138 eXperimenter Kit) it is working with the same Kernel...
-> changing in bootargs console to ttsS0/1/2 did not solve the problem

Taking o look with the debugger told us:
start_kernel is called, but inserted printk() did not print on our board (works fine on the EVM)
-> if i let the software run, it hangs after a while on 0xffff00c0 -> data prefetch failure
-> if i disable the mmu (when halting the ARM in the first line of start_kernel) with the CCS
    it runs until a kernal panic (seen only in debugger be comparing PC with map-file)

Sending a sign by writing directly to the UART2 in the beginning of start kernel:
-> work if mmu is disabled as above with CCS
-> causes a hang (PC is at 0xFFFF00??, i'm not shure now where)

We have tried with PSP3.20.00.14 and 3.21.00.04 now...

Our gretest Questions:
- Why is there no debug output on our board, but it works with the other one?
  (With this output it should be easier to find the next problems...)
- Why did our board hang when MMU is enabled?

Any hints would be fine, thanks!

regards, stefan

 

 

  • 0
    Prodigy 60 points

    Hi,

    here is some additional info:
    adding earlyprintk=xxx (it seems to be equal whats entered for xxx) we got soemthing like that

    ¬˜WÑyMFk,ž5`}òkSl¸]16‚‰¯ØsÑq¡‚ÒˆPÐS@        'f˜`Á
    P½éܑ‰yq-aM™¡’‹dS“S.ˆ“n³ÅÉaiӓl²ìёÉr‹¯S/œ²îÏö‹“d¤=¡
    &}}¡íòÓSS®Ñl·Áyry‘]!M‘}bkS®^¾ó\yyay]Òk®›Ÿ1báÖ–"âi

    on the console, but it stops aftere round about 15 lines (i think when kernel gots panic)

    But on the other board calling "printk(KERN_NOTICE "text");" in start_kernel (first line after variable setup)
    will print out "text" without this option.

    Because of the equal kernel i think now that there must be a difference in U-Boot (which we had to customize),
    but dont't know what it could be....

    ... we use the same uart for the console, we just changed something to include our flash and some pinmux (not from the uart).

    stefan

  • 0 in reply to Stefan Sals
    Prodigy 60 points

    Hi,

    we found out with an Oszilloskop on the serial port that baudrate is about 100k,

    means that the PLL runs on 528M instead of 600M.

    (and Serial port depends on this clock)

    So the diffierence is somewhere in u-boot or ais-config. We Changed DDR2 Timing to 150M

    instead of 132M before (132*4 = 528M) and now we see all the output from the kernel...

     

    We did not know why PLL config is changed on the startup of the kernel, we' ll search for it...

     

    Hope this informations helps someone..

    regards,

    Stefan