* scd/apdu.c (reader_table_lock, apdu_init): New.
(new_reader_slot): Serialize by reader_table_lock.
* scd/app.c (lock_app, unlock_app, app_new_register): Fix error code
usage.
(initialize_module_command): Call apdu_init.
* scd/scdaemon.c (main): Handle error for initialize_module_command.
--
This is a long standing bug. There are two different things; The
serialization of allocating a new SLOT, and the serialization of using
the SLOT. The latter was implemented in new_reader_slot by lock_slot.
However, the former was not done. Thus, there was a possible race where
a same SLOT is allocated to multiple threads.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* scd/app.c (lock_app): Rename from lock_reader and use internal field
of APP.
(unlock_app): Likewise.
(app_dump_state): Use APP.
(application_notify_card_reset): Remove.
(check_conflict): Change API for APP, instead of SLOT.
(check_application_conflict): Likewise.
(release_application_internal): New.
(app_reset): New.
(app_new_register): New.
(select_application): Change API for APP, instead of SLOT.
(deallocate_app, release_application): Modify for manage link.
(report_change): New.
(scd_update_reader_status_file): Moved from command.c and
use APP list, instead of VREADER.
(initialize_module_command): Moved from command.c.
* scd/command.c (TEST_CARD_REMOVAL): Remove.
(IS_LOCKED): Simplify.
(vreader_table): Remove.
(vreader_slot, update_card_removed): Remove.
(do_reset): Call app_reset.
(get_current_reader): Remove.
(open_card): Add SCAN arg.
(cmd_serialno): No retry, since retry is done in lower layer in apdu.c.
No do_reset, since it is done in lower layer.
Add clearing card_removed flag.
(cmd_disconnect): Call apdu_disconnect.
(send_client_notifications): Modify for APP.
(update_reader_status_file): Remove.
--
APP is the abstraction of the card application. For management of
cards, it is better to focus on the APP instead of the physical reader.
This change makes support of multiple card/token easier.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/util.h: Provide replacement for GPGRT_ATTR_ macros when using
libgpg-error < 1.20.
* common/mischelp.h: Ditto.
* common/types.h: Ditto.
--
Given that libgpg-error is a dependency of all GnuPG related libraries
it is better to define such macros at only one place instead of having
similar macros at a lot of places. For now we need repalcement
macros, though.
The vreader table makes the code more clear by explicitly talking
about APDU slots and reader indices. It also accommodates for future
extensions.
* scd/scdaemon.h (server_control_s): Remove READER_SLOT.
* scd/scdaemon.c (scd_init_default_ctrl): Do not init READER_SLOT.
* scd/app.c (check_application_conflict): Add arg SLOT.
* scd/command.c (slot_status_s): Rename to vreader_s.
(server_local_s): Add field VREADER_IDX as replacement for
the READER_SLOT in server_control_s. Change all users.
(slot_table): Rename to vreader_table. Change all users.
(vreader_slot): New.
(do_reset, cmd_apdu): Map vreader to apdu slot.
(get_reader_slot): Rename to get_current_reader. Return -1 on error.
(open_card): Map vreader toapdu slot. Pass slot to
check_application_conflict.
(scd_command_handler): Init VREADER_IDX.
(update_reader_status_file): Reset SLOT field on error.
We better do this once and for all instead of cluttering all future
commits with diffs of trailing white spaces. In the majority of cases
blank or single lines are affected and thus this change won't disturb
a git blame too much. For future commits the pre-commit scripts
checks that this won't happen again.
* gpg-agent.c: New option --write-env-file.
* gpg-agent.c (handle_connections): Make sure that the signals we
are handling are not blocked.Block signals while creating new
threads.
* estream.c: Use HAVE_CONFIG_H and not USE_CONFIG_H!
(es_func_fd_read, es_func_fd_write): Protect against EINTR.
* gpg-agent.texi (Agent UPDATESTARTUPTTY): New.
* scdaemon.c (handle_connections): Make sure that the signals we
are handling are not blocked.Block signals while creating new
threads.
(handle_connections): Include the file descriptor into the name of
the thread.
* acinclude.m4: Ditto.
* scdaemon.texi: Removed OpenSC specific options.
* app-p15.c: New. Basic support for pkcs15 cards without OpenSC.
There are quite a couple of things missing but at least I can use
my old TCOS cards from the Aegypten-1 development for signing.
* app.c (select_application): Detect pkcs15 applications.
* Makefile.am (scdaemon_SOURCES): Removed card.c, card-common.h
and card-p15.c because they are now obsolete. Added app-p15.c.
Removed all OpenSC stuff.
* command.c (do_reset, open_card, cmd_serialno, cmd_learn)
(cmd_readcert, cmd_readkey, cmd_pksign, cmd_pkdecrypt): Removed
all special cases for the old card.c based mechanisms.
* scdaemon.c, apdu.c: Removed all special cases for OpenSC.
stable version 0.1.x.
* scdaemon.texi (Card applications): New section.
* scdaemon.c (main): New option --disable-application.
* app.c (is_app_allowed): New.
(select_application): Use it to check for disabled applications.
* ccid-driver.h (CCID_DRIVER_ERR_ABORTED): New.
* ccid-driver.c (ccid_open_reader): Support the stable 0.1 version
of libusb.
(ccid_get_atr): Handle short messages.
* apdu.c (my_rapdu_get_status): Implemented.
* Makefile.am (pkglib_PROGRAMS): Install it here.
* apdu.c (writen, readn): New.
(open_pcsc_reader, pcsc_send_apdu, close_pcsc_reader): Use the
pcsc-wrapper if we are using Pth.
* scdaemon.c, scdaemon.h: New option --pcsc-ccid.
* ccid-driver.c, ccid-driver.h: New but far from being useful.
* Makefile.am: Add above.
* apdu.c: Add support for that ccid driver.
* scdaemon.c, scdaemon.h: New option --disable-opensc.
* card.c (card_open): Implement it.
* apdu.c (open_osc_reader, osc_send_apdu): New.
(apdu_open_reader) [HAVE_OPENSC]: Use the opensc driver if not
disabled.
(error_string) [HAVE_OPENSC]: Use sc_strerror.
(send_apdu) [HAVE_OPENSC]: Call osc_apdu_send.
* card-common.h (struct p15_private_s): Forward declaration. Add
it to card_ctx_s.
* card.c (card_close): Make sure private data is released.
(card_enum_certs): New.
* card-p15.c (p15_release_private_data): New.
(init_private_data): New to work around an OpenSC weirdness.
(p15_enum_keypairs): Do an OpenSC get_objects only once.
(p15_enum_certs): New.
(card_p15_bind): Bind new function.
* command.c (cmd_learn): Return information about the certificates.
* card.c (card_open): set it here.
* card-p15.c (p15_prepare_key): Factored out common code from ...
(p15_sign, p15_decipher): here and made the decryption work the
regular way.
