libunwind-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Libunwind-devel] SIGSEGV Segmentation Fault in access_mem


From: Dave Watson
Subject: Re: [Libunwind-devel] SIGSEGV Segmentation Fault in access_mem
Date: Tue, 20 Jun 2017 09:28:29 -0700
User-agent: Mutt/1.6.0 (2016-04-01)

On 06/20/17 03:15 PM, Ali Nakipoglu wrote:
> Hi,
> 
> Hope everyone very well.
> 
> Im working on a memory tracking tool that gets linked at run-time via
> LD_PRELOAD. It's using libunwind to get stack trace information. With most
> of the test applications its working fine but, there is one that libunwind
> always crashes with the stack trace:
> 
> #0  0x00007ffff7bbfe74 in access_mem () from /build-debug/Intercept.so
> #1  0x00007ffff7bbf724 in _Ux86_64_step () from /build-debug/Intercept.so
> #2  0x00007ffff7bbc2b8 in Intercept::trace() () at
> /src/Intercept/Intercept.cpp:185
> #3  0x00007ffff7bbc4c3 in calloc () at src/Intercept/Intercept.cpp:252
> #4  0x00007fffe35f1b33 in ?? () from /usr/lib64/nvidia/libGL.so.1
> #5  0x00007fffcb50302e in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #6  0x00007fffcb50c11b in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #7  0x00007fffcb50d1b8 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #8  0x00007fffcb510c63 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #9  0x00007fffcb6686f3 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #10 0x00007fffcb63be54 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #11 0x00007fffcb63d647 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> #12 0x00007fffcb655f24 in ?? () from
> /usr/lib64/nvidia/libnvidia-glcore.so.375.39
> 
> Libunwind 1.2.1.
> 
> Is there a way to bypass it or its is a bug? Can CONSERVATIVE_CHECKS 
> (https://github.com/pathscale/libunwind/blob/1fcbfc649837d2a28e1901986437ca6bf9c4f4d4/src/x86_64/Gstep.c#L61)
> build flag solve this?

You can try CONSERVATIVE_CHECKS.  There is also a known issue if the
memory is mapped, but without PROT_READ.  A potential fix is here:

https://github.com/libunwind/libunwind/pull/7



reply via email to

[Prev in Thread] Current Thread [Next in Thread]