I wrote a very simple Android application(AndroidGLSurfaceViewActivity.apk).mainly use opengl use muti-thread to draw frame on the screen
I tried to run on the freescale platform ( i.mx6q , 1GB Ram , android 4.0.4 , hdmi 1080p output )
It can't run more than one hour will cause the system to crash. (Kernel panic or rcu_preempt_state detected stalls on CPUs/tasks).
but it can run all night on the Android Emulator and Samsung Galaxy Note 8.0
PS : Android apk and program source code and detailed error messages, see attached file
Original Attachment has been moved to: rcu_preempt_state-detected-stalls-on-CPUstasks.txt.zip
Original Attachment has been moved to: AndroidGLSurfaceViewActivity.apk.zip
Original Attachment has been moved to: AndroidGLSurfaceViewActivity.7z.zip
Original Attachment has been moved to: Kernel-panic--not-syncing-Fatal-exception.txt.zip
已解决! 转到解答。
Hi Aaron, please apply the attached two patches, from our test, the issue can be fixed.
Hi DraganOstojic,
We are also facing same issue in i.MX 8MQ CPU, Board will hang if WDOG disable if WDOG is enabled board will reset without any error print.
Please let me know if issue found
BR,
MK
Hi Qiang, I use Android ICS BSP 13.4.1 + 13.4.1.04 patches.
I use a custom board that is a similar to imx6q sabresd board. However there are some changes in BSP because we use a different display, touch screen and both cameras are different. We also disabled a lot of services such as geo-location, bluetooth, wireless etc. We removed most of pre-installed applications and a launcher app so that we can run system with only our application (kiosk mode).
First our own application starts. After that, I manually start AndroidGLSurfaceViewActivity.apk. Aftet that I stop our own application. I have to do it that way because we removed launcher app so our app starts by default.
I built AndroidGLSurfaceViewActivity.apk from the sources provided in the link.
Let me know what you want me to try. I have sabresd and all original source code before we moved to the new board. I can run whatever software or hardware you suggest.
Hi Qiang, I couldn't find JB4.2.2_GA1.1.0 on the Freescale site (currently only jb4.3.1 and kitkat4.4.2 are posted) so I tried android_jb4.3_1.1.0-ga directly downloaded and built as per instructions with the p13 applied (again as per Freescale post in the forum). I ran on the sabre-sd board. Unfortunately, I still get an exception "rcu_sched_state detected stalls".
This happened in less than 24h run. I attached a log file for your reference.
Is there anything else I could do or try?
Hi Dragan, the JB4.3_1.1.0 kernel hasn't included the patch "0027-ENGR00306276-iMX6-Add-workaround-for-ARM-errata-7613.patch" too. This patch had been merged in KK 4.4.2 BSP.
And "0025-ENGR00295714-GPT-Status-register-bits-are-cleared-in.patch" was not included in JB4.3_1.1.0 BSP too.
Hi Qiang_FSL LeonardoSandovalGonzalez
does Android 4.4.2 BSP fix out these problems??
i found that FSL release the "android_jb4.3_1.1.1-ga_source.tar" patch on fsl official imx6 website,
As what I understand, this jb4.3_1.1.1 patch is just for FSL Android 4.3 version.
while, what about FSL Android 4.4.2 BSP version??does this version need to do the patch??
FSL didnot release any patch about Android 4.4.2 on the official imx6 site ,
could i think that FSL Android 4.4.2 BSP have been merged all these patch?
AND, Could I consider that Android 4.4.2 BSP didnothave schedule problem。
AND,do I need to update GPU patch on Android 4.4.2 BSP?? something like p13 GPU patch. does android 4.4.2 merge with it??
Hi Qiang, I confirm that KK 4.4.2 doesn't have multi-core issue. Test that was previously failing on R13.4.1 is still running on KK 4.4.2 after 5d of continuous run. The best I got on R13.4.1 after applying all recommended patches was 3d.
Thank you for all your great help.
Dragan
Hi Qiang, I upgraded to p13 and let it run overnight. Unfortunately, an exception was generated and log was printed on the screen (see attached). It's original rcu_preempt_state detected stall. Interestingly, test app AndroidGLSurfaceViewActivity continues to run and system is responsive.
Now, it appears that in upgrade p13, patch 0027-ENGR00306276-iMX6-Add-workaround-for-ARM-errata-7613.patch is disabled by subsequent patch 0030-disable-errata-794072-and-761320-enable-errata-74223. Should I re-enable 0027 or you have some other suggestion?
This patches seems work for me.
our product had been running for 3 days without hang with these patches applied.
but these patches is not on freescale git, I wonder where these two patch come from?
Hi , All
it could running over one week at now when i'm added args "maxcpus=0" in bootloader ( disable kernel symmetric multiprocessing (SMP) feature )
so. that mean kernel indeed have scheduling problem , right ?