Bug 356675

Summary:	callgrind test apk in android 5.0.2
Product:	[Developer tools] valgrind	Reporter:	winnie_hx <winnie_hx>
Component:	general	Assignee:	Josef Weidendorfer <josef.weidendorfer>
Status:	REPORTED ---
Severity:	crash	CC:	winnie_hx
Priority:	NOR
Version First Reported In:	3.10.0
Target Milestone:	---
Platform:	Android
OS:	Android 5.x
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:
Attachments:	this is test a apk in android5.0.2 by callgrind patch that uses signal handle to fix this problem attachment-20835-0.html

Description winnie_hx 2015-12-14 12:07:33 UTC

Created attachment 96060 [details]
this is test a apk in android5.0.2 by callgrind

We want to test hot functions in apk  at android 5.0.2 by callgrind. because there are oat files in android 5.0.2 .Oat file is a elf of android.So we think callgrind can test the hot functions in apk.
    However,when we test by callgrind,we find a bug in dex file(oat file): at 0x1599449A: ??? (in /data/dalvik-cache/arm/data@app@com.kiloo.subwaysurf-1@base.apk@classes.dex)

Comment 1 winnie_hx 2016-05-12 09:22:54 UTC

Created attachment 98929 [details]
patch that uses signal handle to fix this problem

add this patch can solve the signal problem that some apk may meet SIGSEGV .
since then  you can test apk by valgrind in android 5.0 or most

Comment 2 Josef Weidendorfer 2016-05-12 12:02:23 UTC

Can you provide the complete output of the crash?
As the patch is about VG core stuff, I assume the crash also happens with other VG tools?
If so, It seems better to re-assign this to "general" component.

Comment 3 winnie_hx 2016-05-13 13:38:44 UTC

Created attachment 98947 [details]
attachment-20835-0.html

Sorry,I think it's my fault,this problem must be in the core of valgrind  ,i have modified it to "general".And Thank you very much!

At 2016-05-12 20:02:23, "Josef Weidendorfer via KDE Bugzilla" <bugzilla_noreply@kde.org> wrote:
>https://bugs.kde.org/show_bug.cgi?id=356675
>
>--- Comment #2 from Josef Weidendorfer <josef.weidendorfer@gmx.de> ---
>Can you provide the complete output of the crash?
>As the patch is about VG core stuff, I assume the crash also happens with other
>VG tools?
>If so, It seems better to re-assign this to "general" component.
>
>-- 
>You are receiving this mail because:
>You are on the CC list for the bug.
>You reported the bug.