gpg card status crash

[David Shaw]

On May 11, 2010, at 7:39 PM, Nicholas Cole wrote:

> On Tue, May 11, 2010 at 11:15 PM, David Shaw <dshaw at jabberwocky.com> wrote:
>> On May 11, 2010, at 5:12 PM, Nicholas Cole wrote:
>> 
>>>>> On gpg 1.4.10, the following command works fine:
>>>>> 
>>>>> gpg --card-status
>>>>> 
>>>>> The following command causes gpg to crash:
>>>>> 
>>>>> Macintosh-2:~ nicholas$ gpg  --status-fd 1 --card-status
>>>>> gpg(34195) malloc: *** mmap(size=6594112295125651456) failed (error code=12)
>>>>> *** error: can't allocate region
>>>>> *** set a breakpoint in malloc_error_break to debug
>>>> 
>>>> Need more information to even guess.  What OS is this?  (Your prompt is "Macintosh" - is this an OSX box?  If so, what version of OSX?)
>>>> What card reader do you use?  What version of the card is this?
>>> 
>>> Dear David,
>>> 
>>> Sorry for the poor report.  I'm running OS X 10.6.3, with version 2 of
>>> the card and with the GREMALTO USB Shell Token version 2.
>> 
>> Hmm.  I tried to duplicate this (same OSX, same card, but different card reader), but without any luck.  Can you try and get a backtrace?  To do that, run "gdb /the/path/to/gpg" then enter "break malloc_error_break", and finally "run --status-fd 1 --card-status".  When you get a prompt, enter "bt full".  Send the output of the bt full.
>> 
> 
> Dear David,
> 
> I have posted the full report from that at the end of this email.

Alas, no symbols.  Did you compile this yourself or get the binary from somewhere else?  If you got the binary from somewhere else, do you think you could try and compile it yourself to get the backtrace?

Clearly something is getting confused and making a really massive (6594112295125651456 bytes!) memory request, but it's not clear why without the backtrace.

Also, are you really sure you are running gpg 1.4.10 (what does "gpg --version" say)?

David