• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 98 subscribers
  • Views 942 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.

AM3352: Changing OPP causes MPU to hang

Kanae
Guru 10105 points
Part Number: AM3352

Dear Sitara support team,

In a AM335x design, my customer are facing an issue during the field test.

[Issue]
-CPU was facing hang-up at changing frequency mode
: scaling_available_governors = ondemand
> A CPU frequency always fluctuates by a CPU load.
-It was happened at the field test for over a week, not happened at short time operation.
-CPU was not facing hang-up at changing frequency mode
: scaling_available_governors = performance
> A CPU frequency will be the biggest 1GHz fixing.

[Verified items]
- Timing and sequense for the Voltage setting ang the Frequency setting is totally correct.
OPP: high ⇒ low:
It's changed to a frequency which corresponds in OPP changed first
and the voltage is changed after that.
OPP: low ⇒ high:
It's changed to the voltage which corresponds in OPP changed first
and the frequency is changed after that.
- When the frequency is chaned from 300MHz to 600MHz,
the voltage drop is occured about 40mV.

[Questions]
1. Did you have like this issue caused by changing frequency of AM335x so far?
2. When the frequency is chaned from low to high, does it generate a voltage drop which caused CPU hang-up?
3. Is you have any check points for this issue on your experience, please let me know.

Best regards,
Kanae

  • 0
    TI__Guru**** 393215 points
    Hi,

    What software are you running? What is the processor part number? Please provide accurate details of the use case.
  • 0 in reply to Biser Gatchev-XID
    Guru 10105 points
    Hi Biser,

    Thank you for quick reply.
    Here are informations you asked.

    -Part number： AM3352
    -Software:Linux OS
    Base Distribution: Debian
    Linux Kernel 3.13.4
    -Details of the use case
    1. CPU frequecy switched from 300MHz to 1000MHz, from 300MHz to 800MHz,
    from 300MHz to 600MHz at randam for over a week and CPU faced the hang-up.
    2. When CPU frequecy kept 1000MHz for over a week and CPU did not face the hang-up.
    3. It also did not face the hang-up for a short time.
    4. CPU frequency switching sequence: Ex. 300MHz ->1000MHz->300MHz
    cpu0_set target->scaling up-(Y)->set voltage->set freq.->scaling down?-(Y)->set voltage->
    cpu0_set target->scaling up-(N)---------------->set freq.->scaling down?-(N)---------------->
    This means that timing and sequense for the Voltage setting ang the Frequency setting is totally correct.
    5. When the voltage is changed less than 1000mV at CPU frequency's 1000MHz on purpose,
    it of course makes CPU hang-up.
    Normally, when the frequency is chaned from 300MHz to 600MHz, the voltage drop is just occured about 40mV.

    [Questions]
    1. Did you have like this issue caused by changing frequency of AM335x so far?
    2. When the frequency is chaned from low to high, does it generate a voltage drop which caused CPU hang-up?
    3. Is you have any check points for this issue on your experience, please let me know.

    Best regards,
    Kanae
  • 0 in reply to Kanae
    TI__Guru**** 393215 points
    Sorry, the Debian distribution is not supported by TI. This forum supports only the TI released Processor SDK: www.ti.com/.../PROCESSOR-SDK-AM335X
  • 0 in reply to Biser Gatchev-XID
    Guru 10105 points

    Hi Biser,

    Thank you for your reply.
    And I am sorry not to write the proper part number: AM3352BZCZA100.

    According to your comment, do you guess from your experience that there were no such problems
    which occurred as AM335x; hardware and this kind of  problems dependent on software?
    I understand it is difficult to say, however any your comments would be helpful to study this issue.

    Best regards,
    Kanae

  • 0 in reply to Kanae
    TI__Guru**** 393215 points
    I don't remember having a similar issue reported on this forum. On your question #2, this is definitely something you should check, however this is highly dependent on PCB design and proper decoupling values and placement.
  • 0 in reply to Biser Gatchev-XID
    Guru 10105 points
    Hi Biser,
    Thank you for prompt reply.
    I have already ask my customer to check the voltage drop at a long time operation.
    And also my customer checked thier software.
    I really apprecite your support.
    When I know the solution, I will report it here.

    Best regards,
    Kanae