SIGSEGV before main ? and gcc-debuginfo

[Marco Atzeri]

On 7/15/2015 3:11 PM, Corinna Vinschen wrote:
> Hi Marco,
>
>
> I seem to have missed this...
>
> On Jul 11 09:22, Marco Atzeri wrote:
>> trying to debug latest release candidate of netcdf
>> I hit this SIGSEGV before reaching main
>>
>> $ gdb .libs/ctest.exe
>> GNU gdb (GDB) Cygwin 7.9.1-1
>> .....
>> Reading symbols from .libs/ctest.exe...done.
>> (gdb) break main
>> Breakpoint 1 at 0x100401171: file
>> /pub/devel/netcdf/prova/netcdf-c-4.4.0-rc2/ncdump/ctest.c, line 32.
>> (gdb) run
>> Starting program: /cygdrive/e/cyg_pub/devel/netcdf/prova/netcdf-c-4.4.0-rc2_build64/ncdump/.libs/ctest.exe
>> [New Thread 9476.0x2a88]
>> [New Thread 9476.0x1fa0]
>> [New Thread 9476.0x9a8]
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> ___chkstk_ms () at
>> /usr/src/debug/gcc-4.9.3-1/libgcc/config/i386/cygwin.S:146
>> 146             orq     $0x0, (%rcx)            /* probe there */
>>
>> I thought it was a stack issue and I raised from 2M to 8M
>
> It *is* a stack issue of some sort, otherwise it wouldn't have called
> ___chkstk_ms, which is the guard page handler.
>
> Did you find out in the meantime?

No. I put at side the issue and reported as upstream regression
http://www.unidata.ucar.edu/mailing_lists/archives/netcdfgroup/2015/msg00192.html

As the crash is on initialization of a test program
is not too high on my TODO.

>> SizeOfStackReserve      0000000000200000
>>
>> SizeOfStackReserve      00000000007a1200
>>
>> with no success.
>>
>> May be it is not a netcdf bug ? Ideas what to look for ?
>
> If it's prior to main, it must be some constructor.  Other than that,
> no idea.
>

>
> Hmm, the relative path might break finding the debuginfo.  You
> should be able to help gdb by calling
>
>    (gdb) dir /usr/src/debug/gcc-4.9.3-1/libgcc/config/i386
>
>

thanks. I will try



> Corinna
>

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple