Provided by: freebsd-manpages_12.2-1_all bug

NAME

     cmx — Omnikey CardMan 4040 smartcard reader device driver

SYNOPSIS

     device cmx

DEPRECATION NOTICE

     This driver is scheduled for removal prior to the release of FreeBSD 13.0

DESCRIPTION

     The cmx driver provides support for the PCCARD based Omnikey CardMan 4040 smartcard reader.  The driver
     provides a character device special file based Chip/Smart Card Interface Devices (CCID) interface.  The
     driver implements what the vendor calls the Synchronous API onto the smartcard reader device.

     Reading and writing is synchronous, meaning that a call to write(2) directly corresponds to a complete CCID
     command sent to the device, while the following read(2) will return the complete answer from the reader.
     There is no support for partial reads or writes.  There is no upper limit on CCID request or response
     sizes, but the complete CCID request must be sent to the driver in write(2) and the complete CCID response
     must fit into the buffer supplied to read(2).

     Non-blocking I/O, select(2) and poll(2) are supported and work as expected.  An open file descriptor will
     always be ready for writing, but only ready for reading if the device indicates that it has data available.

FILES

     /dev/cmxn                   Character device special file.
     /usr/ports/security/openct  OpenCT, a userspace smartcard daemon containing a CCID driver which directly
                                 supports cmx devices.

COMPATIBILITY

     Userland smartcard code written for the vendor's Linux drivers should work with the cmx driver without
     modification.

SEE ALSO

     pccard(4)

HISTORY

     The cmx driver first appeared in FreeBSD 7.1.

AUTHORS

     The cmx driver was written by Daniel Roethlisberger <daniel@roe.ch>, originally based on the Linux driver
     v1.1.0 by Omnikey GmbH www.omnikey.com. Early testing and bug fixes by Marcin Cieslak <saper@system.pl>.

BUGS

     The way the cmx driver talks to the CardMan 4040 is a bit rough.  Due to the complete lack of hardware
     documentation other than vendor drivers for other operating systems, the gory details of the device's I/O
     registers are not understood very well.  There may be error conditions which can only be solved by
     physically reinserting the reader.