_gcry_ath_mutex_lock: Assertion

[David Goulet]

On 04 Aug (11:37:18), Werner Koch wrote:
> On Wed, 30 Jul 2014 18:01, dgoulet at ev0ke.net said:
> 
> > I see that in libgcrypt 1.6, it has been removed so is there a way to
> > avoid the issue for < 1.6 without defining NDEBUG?
> 
> You mean the problem is in 1.5.3 but not in 1.6 ?

Yes, sorry for the confusion. The problem is in 1.5.3 and not 1.6.x.

> What version of
> libgpg-error is used.  I need the exact versions and whether it has been
> patched by Debian or you are using upstream versions.

I'm using the Debian package and the exact version of the package is:

	libgpg-error-dev 1.13-0.2

Debian does *not* apply patches on top, I checked.

https://packages.debian.org/jessie/libgpg-error-dev

> 
> Can I look at the OTR code? 

Sure, the pastebin link in the previous email shows the call stack from
libotr which basically gives that:

otrl_message_receiving
  -> otrl_proto_accept_data
    -> gcry_md_read (_gcry_md_read)
      -> md_final
        -> md_digest_length
          assert mutext lock here

This happens when two threads at the same time do a
otrl_message_receiving() call and end up in that code path. Top of the
calling chain starts here:

https://bugs.otr.im/projects/libotr/repository/revisions/master/entry/src/message.c#L922

Let me know if you need more info! :)

Thanks!
David

> 
> 
> Salam-Shalom,
> 
>    Werner
> 
> -- 
> Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.
> 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 603 bytes
Desc: Digital signature
URL: </pipermail/attachments/20140804/08ecec59/attachment-0001.sig>