* One sc_context has only a single reader driver.
* remove dynamic reader driver loading capabilities
* remove opensc-tool -R command
* change the internal API, we don't need to pass around a "driver data" pointer as it can be found directly from the context.
* check in ./configure for only a single enabled reader driver
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@4709 c6295689-39f2-0310-b995-f0e70906c6a9
From http://en.wikipedia.org/wiki/Malloc#Casting_and_type_safety
" Casting and type safety
malloc returns a void pointer (void *), which indicates that it is a
pointer to a region of unknown data type. One may "cast" (see type
conversion) this pointer to a specific type, as in
int *ptr = (int*)malloc(10 * sizeof (int));
When using C, this is considered bad practice; it is redundant under the
C standard. Moreover, putting in a cast may mask failure to include the
header stdlib.h, in which the prototype for malloc is found. In the
absence of a prototype for malloc, the C compiler will assume that
malloc returns an int, and will issue a warning in a context such as the
above, provided the error is not masked by a cast. On certain
architectures and data models (such as LP64 on 64 bit systems, where
long and pointers are 64 bit and int is 32 bit), this error can actually
result in undefined behavior, as the implicitly declared malloc returns
a 32 bit value whereas the actually defined function returns a 64 bit
value. Depending on calling conventions and memory layout, this may
result in stack smashing.
The returned pointer need not be explicitly cast to a more specific
pointer type, since ANSI C defines an implicit conversion between the
void pointer type and other pointers to objects. An explicit cast of
malloc's return value is sometimes performed because malloc originally
returned a char *, but this cast is unnecessary in standard C
code.[4][5] Omitting the cast, however, creates an incompatibility with
C++, which does require it.
The lack of a specific pointer type returned from malloc is type-unsafe
behaviour: malloc allocates based on byte count but not on type. This
distinguishes it from the C++ new operator that returns a pointer whose
type relies on the operand. (see C Type Safety). "
See also
http://www.opensc-project.org/pipermail/opensc-devel/2010-August/014586.html
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@4636 c6295689-39f2-0310-b995-f0e70906c6a9
* reduce to a few, supported functions.
* change all functions to take the debug level as parameter.
* use symbolic names for the debug levels.
* fix tools to pass "verbose"/"opt_debug" as ctx->debug.
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@4118 c6295689-39f2-0310-b995-f0e70906c6a9
- Remove slot abstraction from internal API and all reader drivers. CT-API (from where it all comes from) readers with multiple slots (if still found) can be presented as separate readers, OpenCT should remove the slot abstraction, PC/SC never knew about it. None of the tools knew how to use slots.
- Add sc_cancel (translates to SCardCancel)
- Re-implement sc_wait_for_event; support a blocking call.
- Replace the "int reader" API with "* sc_reader_t" style; add "Get reader by name" functionality.
- Remove "action" parameter from sc_disconnect_card() (was not used)
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@3931 c6295689-39f2-0310-b995-f0e70906c6a9
This is not the best solution, but focus on smallest code change.
Changes:
1. Add detect_readers() to reader opts, this adds new readers to the end
of the readers list until list is full.
2. Add sc_ctx_detect_readers() that calls readers' detect_readers().
3. Fixup pcsc_lock() so that it reconnect to the card and report proper
error so caller may be notified if session was lost.
4. Allow context to be created without readers.
5. Call sc_ctx_detect_readers() from PKCS#11 C_GetSlotList with NULL_PTR.
6. Allow no reader at detect_card, as reader my be removed.
7. Since I broke ABI, I updated the external module version requirement
to match OpenSC version. In the future a separate version should be
maintained for each interface, this should be unrelated to the package
version.
Alon
---
svn merge -r 3480:3505 https://www.opensc-project.org/svn/opensc/branches/alonbl/pnp
M src/tools/opensc-tool.c
M src/pkcs11/pkcs11-global.c
M src/pkcs11/slot.c
M src/libopensc/reader-pcsc.c
M src/libopensc/internal-winscard.h
M src/libopensc/ctx.c
M src/libopensc/reader-ctapi.c
M src/libopensc/libopensc.exports
M src/libopensc/reader-openct.c
M src/libopensc/opensc.h
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@3506 c6295689-39f2-0310-b995-f0e70906c6a9
- log APDUs only if DEBUG is defined (sensitive APDUs should
never be logged and we cannot know whether a APDU is sensitive
or not => enable APDU logging only in a non-production debug
build)
- remove OPENSC_DONT_LOG_SENSITIVE configure option as it's
needed anymore
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@2868 c6295689-39f2-0310-b995-f0e70906c6a9
use pkg-config for openct, openssl, pcsc.
do not compilke libp11. do not compile sslengines.
remove scdl. use libltdl instead.
use libassuan.m4 macro for m4 detection.
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@2528 c6295689-39f2-0310-b995-f0e70906c6a9
- added new parameters max_send_size and max_recv_size, roughly corresponding
to the old max_le (SC_APDU_CHOP_SIZE) parameter. You can now set this
chop limit per driver class (pcsc, openct, ctapi), which sets
driver->max_{send,recv}_size. This value is copied to
card->max_{send,recv}_size in sc_connect_card, and can be overridden
by the card driver.
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@1683 c6295689-39f2-0310-b995-f0e70906c6a9
building PC/SC, OpenCT or USBToken support,
use ifdef's directly in source.
- Because of above, add HAVE_PCSCLITE for winconfig.h
- Remove unnecessary includes for log.h, opensc.h and
errors.h in libopensc sources, they're already taken
care by internal.h.
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@1406 c6295689-39f2-0310-b995-f0e70906c6a9
are routed through sc_pin_cmd(), which builds the APDU and either passes
it to the card directly, or to the card reader along with a request to
read the PIN(s) from the reader's keypad.
Currently, entering PIN in the standard way (i.e. via the application)
should still work - I have verified GPK and eToken; Cryptoflex verify
should work as well. Anything else needs additional testing, and support
for keypad input in particular (I cannot test this at the moment for
lack of a suitable reader).
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@811 c6295689-39f2-0310-b995-f0e70906c6a9
- Changed call convention for reader finish()
- CT-API driver now frees its resources correctly
- Added year 2002 to some of the copyright statements
- sc_pkcs15_decipher() and sc_pkcs15_compute_signature()
now select only the parent DF of the private key file
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@463 c6295689-39f2-0310-b995-f0e70906c6a9
- Add header check for dlfcn.h
- Add internal functions sc_module_{open,close,get_address}
- Use environ instead __environ for scldap_search, should be
more portable.
- Fix compiler warnings noticed by Tru64 / AIX cc
git-svn-id: https://www.opensc-project.org/svnp/opensc/trunk@401 c6295689-39f2-0310-b995-f0e70906c6a9