libunwind-devel
[Top][All Lists]
Advanced

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

Re: [Libunwind-devel] Segfault in dwarf_readu32 (x86-64)


From: Peter Wu
Subject: Re: [Libunwind-devel] Segfault in dwarf_readu32 (x86-64)
Date: Tue, 25 Nov 2014 10:28:10 +0100
User-agent: KMail/4.14.3 (Linux/3.17.0-rc4-custom-00168-g7ec62d4; KDE/4.14.3; x86_64; ; )

On Tuesday 25 November 2014 12:36:11 Arun Sharma wrote:
> On Sun, Nov 23, 2014 at 5:54 PM, Peter Wu <address@hidden> wrote:
> > I am using libunwind in gobject-list[1] to get a backtrace at certain
> > points. It generally works fine, but there is one condition that
> > triggers a segfault and I have no idea how to tackle that.
> >
> 
> Thanks for the bug report.
> 
> Suggest running with UNW_DEBUG_LEVEL=n to get more info about what happened.
> 
> Instructions here:
> https://lists.nongnu.org/archive/html/libunwind-devel/2014-02/msg00004.html

I already figured out what was wrong. In essence the .eh_frame section
is malformed due to a bug in the gold linker:
https://sourceware.org/bugzilla/show_bug.cgi?id=17639

In dwarf_extract_proc_info_from_fde the FDE length is read, but not
checked whether it is actually within a segment. It should do so and
return an error.

Do you think that you can reproduce the error with the details from the
binutils bug report and the test program from my previous mail? Or would
you like me to send you the details from UNW_DEBUG_LEVEL?
-- 
Kind regards,
Peter
https://lekensteyn.nl




reply via email to

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