933589 - [bluedroid] Error message shows "GKI_LINUX( 178): ##### ERROR : timer_thread: tick delayed"

Status: RESOLVED INVALID

(Firefox OS Graveyard :: Bluetooth, defect)

Description

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Error message shows "GKI_LINUX(  178): ##### ERROR : timer_thread: tick delayed". Not sure it's related to mozglue? But I don't see this error on Android 4.3.

Comment 1

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Not sure why this error message shows up a lot. I need to check bluedroid GKI stuff again.

Comment 2

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

I turned on bluedroid bt stack with full-log options on Nexus 4 which run with Android 4.3.
But I don't see any error message "tick delayed > 5 slots (%d,%d) -- cpu overload" from gki_ulinux.c on Android. 
 
However, even doing Bluetooth discovery on Nexus 4 with b2g, I can see these error messages.


        if (__unlikely(timeout_ns < ((GKI_TICKS_TO_MS(1) * 1000000) * 0.1)))
        {
            timeout.tv_nsec = (GKI_TICKS_TO_MS(1) * 1000000) * 0.1;

            /* Print error message if tick really got delayed
               (more than 5 ticks) */
            if (timeout_ns < GKI_TICKS_TO_MS(-5) * 1000000)
            {
                GKI_ERROR_LOG("tick delayed > 5 slots (%d,%d) -- cpu overload ? ",
                        timeout_ns, GKI_TICKS_TO_MS(-5) * 1000000);
            }
        }
I will try non-debug build. Maybe it's related to debug build?

Comment 3

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

I don't have much time for this bug until bleutooth profiles get done. Pending to last sprint. Although I don't have idea about it.

Comment 4

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

It seems on Android KK, bluedroid library runs on Bluetooth java application, with only bluetooth permission, which causes scheduler based on RT policy fail to acquire. This seems can be fixed from bluedroid based on Comment 32. (https://bugzilla.mozilla.org/show_bug.cgi?id=1002917#c32)