Crash Test Scenario

[no_spam_for_me]

Inspired of my own experiences at 002Beta4

Ok, ok, ok, I'm also able to freeze the system [IMG]file:///C:/DOKUME%7E1/FUN/LOKALE%7E1/Temp/msohtml1/01/clip_image002.gif[/IMG]
It runs for ~7 hours without any freeze (last 2 hours stress test), but if I do math operations (e.g. integer), e.g. at AnTuTu between 8% and 35%, I'm able to freeze the system...
The easiest way to freeze the system is to use 'PassMark PerformanceTest' (https://play.google.com/store/apps/d...mark.pt_mobile) and there at CPU do 1, 2 or 3 times the integer test and the system freezes....
I never know before that this can depends on the kernel, I thought it's only a problem of the SoC (because at stock it isn't a problem)...
BUT after this I find out with "816/1992 interactive" at stock I'm also able to reproduce it at 'PassMark PerformanceTest'...
Only a quick shot because now I have to go to bed...
I will analyse it tomorrow...

and because I often read about freezes specially during the first 35% at AnTuTu and about soft restarts and reboots, I try to provoke the freeze and what should I say, nothings easier like that...

First try at fresh and clean installed Stock Rom (20140529) I only have to use AnTuTu (4.4.2) maybe 10 times (cancelling after 35% (after processing the math operations) and directly restarting the test).

BUT the system desn't really freezes, because I'm able to use ADB and so I do

adb shell dmesg > dmesg_s001_1.txt

and have a look at the end of "dmesg_s001_1.txt" and find this nearly the end

 :
<4>[  740.054762@0] Mali: ERR: /Work/s802/s8_44_0328/common/drivers/amlogic/../../../hardware/arm/gpu/mali/common/mali_group.c
<4>[  740.060072@0]            mali_group_bottom_half_pp()1643
<4>[  740.060072@0]            Mali PP: Unexpected interrupt 0x00000F19 from core Mali_PP_Broadcast, aborting job
<4>[  740.073577@0] 
<4>[  740.077394@0] Mali: ERR: /Work/s802/s8_44_0328/common/drivers/amlogic/../../../hardware/arm/gpu/mali/common/mali_pp.c
<4>[  740.085811@0]            mali_pp_stop_bus_wait() 150
<4>[  740.085811@0]            Mali PP: Failed to stop bus on Mali_PP5. Status: 0x00000005
<4>[  740.096980@0] 
<4>[  740.100700@0] Mali: ERR: /Work/s802/s8_44_0328/common/drivers/amlogic/../../../hardware/arm/gpu/mali/common/mali_mmu.c
<4>[  740.109299@0]            mali_mmu_raw_reset() 282
<4>[  740.109299@0]            Reset request failed, MMU status is 0x80000019
 :

During I wrote this (so some minutes later) I get a black screen so I redo the dmesg to "dmesg_s001_2.txt" and what should I say? Correct, I get much more 'Mali: ERR'.
The delta after the end of "dmesg_s001_1.txt" I wrote to "dmesg_s001_1+2_delta.txt" so maybe MINX (and/or one of there developer) will have a look at it, because I'm also able to provoke these freeze at 002Beta4....

(BTW: please don't ask me why this will happen during math operations...)

Maybe others find other ways to provoke a system crash and will post it, so I or others can check, if it is also possible to get a log and we can help MINIX to find a solution....

Greetings

[no_spam_for_me]

At stress test at 002beta2 with "816/1992 interactive" at that moment I want to finish the stress test I got a black screen and after ~ 2 seconds I get back the launcher without any running application (also at background), I done a dmesg, but I don't know if I was fast enough, but maybe MINIX is interesting in...

[no_spam_for_me]

Also freeze at AnTuTu wile the flying android-men (I think it is 2D) at 002Beta5 and after a wile I get a black screen and after a wile the freeze screen back and then after a wile a black screen........ and again and again and again....

Maybe minix is interested in the dmesg.txt...

BUT I begin to loose my interest in these fu*** amlogic SoC's (they sucks)... and begin to think about an Intel C1037U, something like this

http://www.aliexpress.com/store/product/build-your-own-pc-Barebone-System-Intel-HM76-Wireless-display-fanless-Alluminum-chassis-Four-Native-USB/800900_1847650084.html

or

http://www.aliexpress.com/item/2014-NEW-Intel-Celeron-C1037U-aluminum-fanless-dual-core-living-room-HTPC-Barebone-Mini-PC-with/1638834276.html

[Batfink32]

Yeah,I said this a couple of days ago in another thread but time to put mine on ****. The amlogic chips seem to have poor quality with the reboots etc.

Sent From My xx.004 Using Tapatalk

[no_spam_for_me]

Flashed 002beta6 "Erase flash" (Normal erase) and "Erase bootloader" + wait 10 minutes after first boot

5 runs are ok (21850 - 21950) BUT at sixth run 3D test didn't start and after a wile I get "Benchmarking has stop unexpectedly, please try again!"

After that after a short pause I've done 10 AnTuTu runs without any problem...

+ rooted (by finless .zip) 10 AnTuTu runs without any problem (but 100 - 200 less than untouched)
++ self-made "81600/1992000 interactive" 4-5 AnTuTu runs without any problem (but 0 - 100 less than untouched) BUT at the next run freeze the system totally at 3D test (no dmesg possible)...

=> it seams so that they have fixed the "Mali: ERR:"but maybe the totally freeze is because of overheating... who knows... In the next time I will do some stress test also at open case with table fan... we will see

[titooo7]

That's great feedback man

Sent from my atv_google_sdk using Tapatalk