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.

OMAP non-funtional Warm Reset

Other Parts Discussed in Thread: TPS65920, OMAP3530

Hello,

My system consists of OMAP3530 and TPS65920.

I have a problem using warm reset. Once triggered by hardware, OMAP hangs, and will stay this way until switching the system on and off again.

The warm reset line in my layout is connected directly from the TPS65920 to the OMAP on the designated pins, with a 1.8v pull-up resistor. A closer look with an oscilloscope after the trigger shows that the warm reset line is high, with short '0' pulses (all this while the OMAP hangs). This happens on all of my boards, at u-boot level, which means warm reset is unusable for me.

Any known reason for this behavior?

Thanks,

Matan

  • Hello Matan,

    OMAP supports a different boot sequence from warm reset with power on reset. Please refer to the section Table 25-5. Booting Configuration Pins After a Warm Reset in the TRM (spruf98m.pdf). Please make sure that a boot strap is in your boot source set by sys_boot pins.

  • Hi Kazunobu,

    My sys_boot pins are set as follows: (1010101) [6:0].

    Which means I'm set to boot from NAND, which stores my bootstrap. The board boots perfectly on POR, but it hangs on warm reset. Anything else that may cause this problem?

     

    Thanks,

    Matan

  • Hello Matan,

    Can you use a JTAG emulator to read register values after the failure ? If you can, please check tracing vector values described in the wiki to debug the issue.

  • I just faced this  problem.  Since I'm running on a Gumstix Overo, my boot pins are set :  (peripheral boot) 0x01111, which would mean that we look to MMC first and then NAND for power on reset.   Now since I develop on my SD card, I've never had thought to replace the MLO which was in NAND flash. 

    So when I make my watchdog fire in Linux, or in the u-boot, warm reset would cause a freeze rather than a reset.  However, if I re-flashed the MLO which was in nand, NOW my watchdog warm reset goes back through the reboot cycle!  The warm reset causes the OMAP to look to nand first.  I haven't looked at the code in the X-Loader yet to see if the x-loader is testing to check if there's an SD and jumping to the SD instead in that case.

    I'm just relieved that I have a method of getting the watchdog to function as I intend.

    I just read the above post (look in table 25-5 ((THANKS TI)))  hence the edit.