On Sun 2016-11-13 01:45:45 +0900, Ian Jackson wrote:
> Package: gnupg
> Version: 2.1.15-4
>
> I just tried to use caff. I found that it seemed to hang, every
> time. I tried debugging it with strace.
>
> I discovered that caff had run a gpg-connect-agent process. It was
> calling connect() to connect to /run/user/1000/gnupg/S.dirmngr. The
> corresponding listening socket in question was owned by a dirmngr
> process.
>
> I straced the dirnmgr process and got:
>
> root(ian)@zealot:~> strace -fp14006
> strace: Process 14006 attached with 3 threads
> [pid 17554] futex(0x7f459d683200,
> FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, NULL, ffffffff) = -1
> EAGAIN (Resource temporarily unavailable)
> [pid 17554] getsockopt(1, SOL_SOCKET, SO_PEERCRED, {pid=17553, uid=1000,
> gid=1000}, [12]) = 0
> [pid 17554] futex(0x7f459d683200, FUTEX_WAKE_PRIVATE, 1) = 0
> ...
>
> And then caff unwedged.
>
> This all happened to me twice. I may be able to repro it.
>
> I suspect thread mishandling by dirmngr.
This looks like #840680 to me, which was resolved in 2.1.15-6.
Please test a more recent version of dirmngr, and if it resolves your
problems, please merge this bug report into 840680.
Regards,
--dkg
signature.asc
Description: PGP signature
