segfault in recent snapshots

Christopher Faylor cgf-use-the-mailinglist-please@cygwin.com
Sun Mar 3 18:06:00 GMT 2013


On Sun, Mar 03, 2013 at 12:02:43PM +0000, Henry S. Thompson wrote:
>Christopher Faylor writes:
>
>> On Sat, Mar 02, 2013 at 08:00:38PM +0000, Henry S. Thompson wrote:
>>>Snapshots of 31/1 and 1/3 are giving segfaults for psql (see previous
>>>thread?)
>>
>> "previous thread"?
>
>http://cygwin.1069669.n5.nabble.com/psql-crashes-with-snapshot-td96270.html

That's the thread where Marco said:

"strangely I noticed that psql of 9.2.2-1 does not suffer this problem,
and it seems due to cygpq.dll (rebase address ?)."

i.e., not a Cygwin snapshot problem.

>>>and locally compiled xemacs.
>>>
>>>Any way I can help debug?
>>
>> Yes.  As always, provide a simple test case which demonstrates the
>> problem.
>
>Sigh, not within my competence in this case.
>
>> And, if it is segfaulting then there is a stackdump file.
>> Post that.
>
>Exception: STATUS_ACCESS_VIOLATION at eip=003D0000
>eax=003D0000 ebx=00000000 ecx=7792389A edx=006F017C esi=61006C50 edi=0028CD78
>ebp=0028AC88 esp=0028AC4C program=C:\Cygwin\bin\psql.exe, pid 11572, thread main
>cs=0023 ds=002B es=002B fs=0053 gs=002B ss=002B
>Stack trace:
>Frame     Function  Args
>0028AC88  003D0000  (61273B28, 00000000, 00000000, 00000000)
>0028ACF8  61006CF5  (00000000, 0028CD78, 61006C50, 00000000)
>End of stack trace
>
>Exception: STATUS_ACCESS_VIOLATION at eip=003D0000
>eax=003D0000 ebx=00000000 ecx=7792389A edx=0112017C esi=61006C50 edi=0028CD78
>ebp=0028AC88 esp=0028AC4C program=C:\cygwin\usr\local\bin\xemacs-21.5-b33.exe, pid 10904, thread main
>cs=0023 ds=002B es=002B fs=0053 gs=002B ss=002B
>Stack trace:
>Frame     Function  Args
>0028AC88  003D0000  (61273B28, 00000000, 00000000, 00000000)
>0028ACF8  61006CF5  (00000000, 0028CD78, 61006C50, 00000000)
>End of stack trace
>
>I think there's probably more detail in the gdb stacktraces I posted
>subsequently. . .

The 003D0000 was in your gdb stacktrace but there is more useful detail,
and less superfluous cruft above.

This address indicates that the error is in user code, so this again
devolves to something that Marco would likely have to look into.  The
"003D0000" looks like a DLL which has not been properly rebased.

If you can duplicate this under gdb then "info shared" should tell you
which dll is associated with that address.

cgf

--
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



More information about the Cygwin mailing list