Do not use a dynamically allocated buffer to store the structure used to pass random bytes to the kernel. The maximum size is known in advance, so we can use a buffer allocated on the stack. We can thus get rid of the xmalloc wrapper, which was only used for that allocation.
|7 years ago|
|lib||7 years ago|
|m4||8 years ago|
|man||7 years ago|
|src||7 years ago|
|.gitignore||8 years ago|
|AUTHORS||7 years ago|
|COPYING||8 years ago|
|Makefile.am||7 years ago|
|NEWS||7 years ago|
|README.md||7 years ago|
|configure.ac||7 years ago|
Scdtools - Tools for Scdaemon and OpenPGP smartcards
Scdrand - Extract random numbers from a smartcard
Scdrand obtains up to 256 bytes of random data from a ISO7816-compliant smartcard and adds them to the kernel entropy pool.
It uses Scdaemon to send a
GET CHALLENGE to the smartcard. As per
ISO 7816-4, the smartcard replies with random bytes, which scdrand then
sends to the Linux kernel entropy pool through the RNDADDENTROPY ioctl
Since that call requires root privileges, scdrand is installed with
the setuid bit set. To mitigate risks associated with running as root,
the program forks itself at startup and only the child retains the root
privileges needed to write to
/dev/random. The parent process, in
charge of communicating with the smartcard, then runs with the
privileges on the calling user.
The most basic usage of scdrand is as follows:
$ scdrand 32
The program will request 32 random bytes from the card, feed them to the entropy pool, then quits. You can request from 1 to 256 bytes.
-l, --loop option, scdrand will not quit, but will
periodically request random bytes from the card and feed them to the
pool. With the
-t, --threshold option, scdrand will attempt to feed
the pool only if the available entropy in the pool falls below a
specified threshold. So in the following example:
$ scdrand -l -i 1 -t 512 256
scdrand will check the state of the entropy pool every second; if
there is less than 512 bits of entropy available, it will feed the pool
with 256 random bytes. The program will quit upon receiving a
Scdtotp - Generate time-based OTP from an OpenPGP smartcard
Scdtotp uses an OpenPGP smartcard as a poor man’s one-time password generator token. It generates time-based one-time password (TOTP) as per RFC 6238, based on a key it expects to find in the private data object of the inserted OpenPGP smartcard.
Contrary to a true password generator token, the key cannot remain only on the smartcard, it has to be sent to the computer so that scdtotp can derive the password from it. Thus it cannot provide the same level of security.
The key must be stored as an
otpauth:// URI as specified in Google
where the secret parameter is the Base32-encoded key. This format allows to specify also the non-secret parameters of the TOTP algorithm:
- the HMAC algorithm to use:
&algorithm=mac, where mac can be
- the time period:
&period=N, where N is expressed in seconds (30 seconds by default);
- the number of digits to output:
&digits=N(defaults to 6).
privatedo Ncommand of the GnuPG’s card editor may be used to store the URI into the Nth private DO of the OpenPGP smartcard.
Once the URI is on the card and the card is inserted in the card reader, simply calling scdtotp will print the one-time password for the current time window on standard output.
Scdtools is distributed under the terms of the GNU General Public License, version 3 or higher. The full license is included in the COPYING file of the source distribution.
Homepage and repository
The project is located at http://www.incenp.org/dvlpt/scdrand.html. The latest source code is available in a Git repository at git://git.incenp.org/scdrand.git and may also be browsed at http://www.incenp.org/gitweb/?p=scdrand.git;a=summary.