• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 98 subscribers
  • Views 481 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.

wake up from sleep source

DavidVescovi
Genius 5470 points

I have been doing some low power suspend work with the A8_1_01_xx 35xx BSP.
About 2/3 of the time it goes into off mode suspend correctly.
I have the debugsuspend flag set.

About 1/3 of the time, or so, it wakes right back up after entering off and I get a "Wake do to SYSINTR -1"

According to Madhvi's suggestion in another post, I commented out:
     //if (!OALPowerWakeSource(sysIntr))
        //    continue;


Now I see Wake do to SYSINTR 17.
Which translates to IRQ 37 or the GPTIMER1 ..the OS timer.

Why is this? Shouldn't this interrupt be masked off before entering "off"?

  • 0
    TI__Expert 5320 points

    Hi David

    Look at this post:

    http://e2e.ti.com/support/embedded/f/353/p/103563/364704.aspx#364704

    See my suggested changes for PrcmSuspend function. You dont need to do the changes that you have shown in the post - that is only to figure out which non-wakeup source interrupt is waking the system.

    There is a finite time between interrupts being turned off (using co-proc registers), thereby making it a single threaded system, and GPTIMER1 being stopped in PrcmSuspend function in BSP. During this small interval time, timer match interrupt could happen but not handled due to single threaded. Hence the issue that you observe.

    FYI - this fix will be officially made available in 01.02.00 release.

    -Madhvi

  • 0 in reply to Madhvi
    Genius 5470 points

    Thanks Madhvi,

    Just the information I was looking for. I only made the changes to determine the interrupt source.

    Your comment and my observations confirms my suspicions about an extra unhandled timer interrupt.

    I will make some patches and retest.

    Thanks for the confirmation w... ill be looking for the next release.

     

    DV

  • 0 in reply to DavidVescovi
    Genius 5470 points

    I made the changes Madhvi suggested and indeed it does seem the sleep entry is more robust.

    I no longer get immediate wake-ups. Unfortunately now I occasionally get into a state where it will enter sleep but can not be woken up with the power key (or any other wake up source for that matter). I am beginning to think it is probably better to have the nuisance wake up over not being able to wake it at all.

    Have you observed this behavior also?

  • 0 in reply to DavidVescovi
    TI__Expert 5320 points

    David

    The issue where system doesnt resume at times is a known issue with our BSP and is not related to the GPTIMER fix that I suggested earlier. We have fixed a few issues related to suspend resume in 01.02.00 release - and with these fixes, we havent seen the resume issue so far. Once the release is on the web, please integrate it and let us know if you still see issues.

    -Madhvi

  • 0 in reply to Madhvi
    Genius 5470 points

    Madhvi,

    Much appreciate the straight feed back.

    DV