Bug#401843: /usr/bin/gpg-agent: gpg-agent sometimes get stuck and stops working
Package: gnupg-agent Version: 2.0.0-5.1 Severity: grave File: /usr/bin/gpg-agent Justification: renders package unusable after some time, gpg-agent sometimes get stuck, and does not work: gpg used with --use-agent option just says: gpg: problem with the agent - disabling agent use It's usually get stuck once a day rougly, implying an X session restart which is most than painful. neither verbose, nor --debug-all seems to give any valuable from the gpg command line, and I don't really know how to get into a situation where it gets stuck to debug it :| though I have one stuck instance running atm, is there any tests I can perform to know what's going on wrong ? I've checked $GPG_AGENT_INFO are correct, the socket is here, and gpg-agent is running the right pid, and fuser says that gpg-agent is still owning the socket. -- System Information: Debian Release: 4.0 APT prefers unstable APT policy: (500, 'unstable'), (500, 'testing'), (1, 'experimental') Architecture: amd64 (x86_64) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-2-amd64 Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Versions of packages gnupg-agent depends on: ii libc62.3.6.ds1-8 GNU C Library: Shared libraries ii libgcrypt11 1.2.3-2 LGPL Crypto library - runtime libr ii libgpg-error01.4-2 library for common error values an ii libpth20 2.0.7-4 The GNU Portable Threads Versions of packages gnupg-agent recommends: ii gnupg 1.4.5-3GNU privacy guard - a free PGP rep ii gnupg22.0.0-5.1 GNU privacy guard - a free PGP rep ii gpgsm 2.0.0-5.1 GNU privacy guard - S/MIME version ii pinentry-gtk2 [pinentry] 0.7.2-3GTK+-2-based PIN or pass-phrase en -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#401843: /usr/bin/gpg-agent: gpg-agent sometimes get stuck and stops working
reassign 401843 pinentry-gtk2
thanks
On Wed, Dec 06, 2006 at 12:09:43PM +0100, Pierre Habouzit wrote:
Package: gnupg-agent
Version: 2.0.0-5.1
Severity: grave
File: /usr/bin/gpg-agent
Justification: renders package unusable
after some time, gpg-agent sometimes get stuck, and does not work: gpg
used with --use-agent option just says:
gpg: problem with the agent - disabling agent use
It's usually get stuck once a day rougly, implying an X session
restart which is most than painful.
neither verbose, nor --debug-all seems to give any valuable from the
gpg command line, and I don't really know how to get into a situation
where it gets stuck to debug it :|
though I have one stuck instance running atm, is there any tests I can
perform to know what's going on wrong ? I've checked $GPG_AGENT_INFO are
correct, the socket is here, and gpg-agent is running the right pid, and
fuser says that gpg-agent is still owning the socket.
okay, further testing shows in fact that pinentry-gtk-2 is the culprit
in fact. after some time, for a reason I just can't understand,
pinentry-gtk-2 just refuses to fire, and says that it cannot open
display: 0.0. the interesting part of the strace is:
getuid()= 1003
uname({sys=Linux, node=mad, ...}) = 0
uname({sys=Linux, node=mad, ...}) = 0
socket(PF_FILE, SOCK_STREAM, 0) = 3
uname({sys=Linux, node=mad, ...}) = 0
uname({sys=Linux, node=mad, ...}) = 0
connect(3, {sa_family=AF_FILE, path=/tmp/.X11-unix/X0}, 19) = 0
uname({sys=Linux, node=mad, ...}) = 0
fcntl(3, F_SETFD, FD_CLOEXEC) = 0
access(/tmp/.gdmSNNFKT, R_OK) = -1 ENOENT (No such file or
directory)
writev(3, [{l\0\v\0\0\0\0\0\0\0\0\0, 12}], 1) = 12
fcntl(3, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK)= 0
read(3, \0\26\v\0\0\0\6\0, 8) = 8
read(3, No protocol specified\n\0\0, 24) = 24
write(2, Xlib: connection to \:0.0\ refused by server\r\nXlib: , 52) =
52
write(2, No protocol specified\n, 22) = 22
write(2, \r\n, 2) = 2
shutdown(3, 2 /* send and receive */) = 0
close(3)= 0
write(2, \n(pinentry-gtk-2:4007): Gtk-WARNING **: cannot open display:
:0.0\n, 66) = 66
munmap(0x2b347fd66000, 16384) = 0
exit_group(1) = ?
it tries to open gdm X Authentication cookie, sadly, it's not the
correct file name. I just don't know where that name comes from,
$XAUTHORITY is correct and in fact points to /tmp/.gdmY0V0JT and not
/tmp/.gdmSNNFKT.
--
·O· Pierre Habouzit
··O[EMAIL PROTECTED]
OOOhttp://www.madism.org
Bug#401843: /usr/bin/gpg-agent: gpg-agent sometimes get stuck and stops working
On Wed, Dec 06, 2006 at 12:09:43PM +0100, Pierre Habouzit wrote:
Package: gnupg-agent
Version: 2.0.0-5.1
Severity: grave
File: /usr/bin/gpg-agent
Justification: renders package unusable
after some time, gpg-agent sometimes get stuck, and does not work: gpg
used with --use-agent option just says:
gpg: problem with the agent - disabling agent use
It's usually get stuck once a day rougly, implying an X session
restart which is most than painful.
neither verbose, nor --debug-all seems to give any valuable from the
gpg command line, and I don't really know how to get into a situation
where it gets stuck to debug it :|
though I have one stuck instance running atm, is there any tests I can
perform to know what's going on wrong ? I've checked $GPG_AGENT_INFO are
correct, the socket is here, and gpg-agent is running the right pid, and
fuser says that gpg-agent is still owning the socket.
I've compared the strace of a stuck gpg-agent and a working one, the
stuck one seems to receive a SIGPIPE when it writes to stdout. and yes
my gpg-agent is launched with the switch --daemon:
broken trace:
fcntl(0, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(0, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(0, F_SETFL, O_RDWR|O_NONBLOCK)= 0
select(1, NULL, [0], NULL, {0, 0}) = 1 (out [0], left {0, 0})
write(0, OK closing connection, 21) = -1 EPIPE (Broken pipe)
--- SIGPIPE (Broken pipe) @ 0 (0) ---
fcntl(0, F_GETFL) = 0x802 (flags O_RDWR|O_NONBLOCK)
fcntl(0, F_SETFL, O_RDWR) = 0
close(0)= 0
good trace
fcntl(6, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(6, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(6, F_SETFL, O_RDWR|O_NONBLOCK)= 0
select(7, NULL, [6], NULL, {0, 0}) = 1 (out [6], left {0, 0})
write(6, OK closing connection, 21) = 21
fcntl(6, F_GETFL) = 0x802 (flags O_RDWR|O_NONBLOCK)
fcntl(6, F_SETFL, O_RDWR) = 0
fcntl(6, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(6, F_GETFL) = 0x2 (flags O_RDWR)
fcntl(6, F_SETFL, O_RDWR|O_NONBLOCK)= 0
select(7, NULL, [6], NULL, {0, 0}) = 1 (out [6], left {0, 0})
write(6, \n, 1) = 1
fcntl(6, F_GETFL) = 0x802 (flags O_RDWR|O_NONBLOCK)
fcntl(6, F_SETFL, O_RDWR) = 0
close(6)= 0
socket(PF_FILE, SOCK_STREAM, 0) = 6
=
I've strictly no idea how the broken gpg-agent got into that stupid
state though.
--
·O· Pierre Habouzit
··O[EMAIL PROTECTED]
OOOhttp://www.madism.org
