libunwind-devel
[Top][All Lists]
Advanced

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

Re: [Libunwind-devel] libunwind test failure


From: Prabhat Verma
Subject: Re: [Libunwind-devel] libunwind test failure
Date: Fri, 5 Apr 2013 19:36:03 +0000

Thank you for the quick reply!

Let me dig into it a bit and try your suggestions…

 

Thanks and Regards,

Prabhat

 

From: address@hidden [mailto:address@hidden On Behalf Of Arun Sharma
Sent: Friday, April 05, 2013 1:51 PM
To: Prabhat Verma
Cc: address@hidden
Subject: Re: [Libunwind-devel] libunwind test failure

 

On Fri, Apr 5, 2013 at 8:31 AM, Prabhat Verma <address@hidden> wrote:

child's exit status 255 

FAILURE: detected 1 errors 
FAIL: test-ptrace 

 

Could you run:

cd tests

./test-ptrace -c -t ./test-ptrace-misc # less verbose

./test-ptrace -v -c -t ./test-ptrace-misc 

 

The -v flag should tell you where things are going wrong. For eg: on my gcc-4.1.2/glibc-2.5 based system I get a failure in:

================

 

FAILURE: unw_step() returned -8 for ip=31f74094f8 (start ip=31f74094f8)

unwind failed with ret=-8

 

00000031f74094f8 _dl_lookup_symbol_x+0x3f8        (sp=00007fff06ab5f58)

        proc=00000031f7409100-00000031f7409db7

        handler=0 lsda=0

================

 

I then disassembled /lib64/ld-linux.so.2 to find the exact location:

 

_dl_lookup_symbol_x:

  ...

   0x00000031f74094f7 <+1015>:  leaveq

   0x00000031f74094f8 <+1016>:  retq


readelf -wf on the binary shows:

00000550 00000024 00000554 FDE cie=00000000 pc=31f7409100..31f7409db7

  DW_CFA_advance_loc: 1 to 31f7409101

  DW_CFA_def_cfa_offset: 16

  DW_CFA_offset: r6 at cfa-16

  DW_CFA_advance_loc: 3 to 31f7409104

  DW_CFA_def_cfa_reg: r6

  DW_CFA_advance_loc: 6 to 31f740910a

  DW_CFA_offset: r13 at cfa-40

  DW_CFA_offset: r14 at cfa-32

  DW_CFA_offset: r15 at cfa-24

  DW_CFA_advance_loc: 16 to 31f740911a

  DW_CFA_offset: r3 at cfa-56

  DW_CFA_offset: r12 at cfa-48

  DW_CFA_nop

  DW_CFA_nop

  DW_CFA_nop

 

i.e. missing unwind info around 31f74094f8.

I remember people having better luck when compiling everything with -fasynchronous-unwind-tables.

 

 -Arun


reply via email to

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