SAFE(4) | MidnightBSD Kernel Interfaces Manual | SAFE(4) |
safe
— SafeNet
crypto accelerator
To compile this driver into the kernel, place the following lines in your kernel configuration file:
device crypto
device cryptodev
device safe
Alternatively, to load the driver as a module at boot time, place the following line in loader.conf(5):
safe_load="YES"
sysctl hw.safe.debug |
sysctl hw.safe.dump |
sysctl
hw.safe.rnginterval |
sysctl
hw.safe.rngbufsize |
sysctl
hw.safe.rngmaxalarm |
The safe
driver supports cards containing
SafeNet crypto accelerator chips.
The safe
driver registers itself to
accelerate DES, Triple-DES, AES, MD5-HMAC, SHA1-HMAC, and NULL operations
for ipsec(4) and
crypto(4).
On all models, the driver registers itself to provide random data to the random(4) subsystem. Periodically the driver will poll the hardware RNG and retrieve data for use by the system. If the driver detects that the hardware RNG is resonating with any local signal, it will reset the oscillators that generate random data. Three sysctl(8) settings control this procedure: hw.safe.rnginterval specifies the time, in seconds, between polling operations, hw.safe.rngbufsize specifies the number of 32-bit words to retrieve on each poll, and hw.safe.rngmaxalarm specifies the threshold for resetting the oscillators.
When the driver is compiled with
SAFE_DEBUG
defined, two
sysctl(8) variables are
provided for debugging purposes: hw.safe.debug can be
set to a non-zero value to enable debugging messages to be sent to the
console for each cryptographic operation, hw.safe.dump
is a write-only variable that can be used to force driver state to be sent
to the console. Set this variable to
“ring
” to dump the current state of
the descriptor ring, to “dma
” to dump
the hardware DMA registers, or to
“int
” to dump the hardware interrupt
registers.
The safe
driver supports cards containing
any of the following chips:
crypt(3), crypto(4), intro(4), ipsec(4), random(4), crypto(9)
Public key support is not implemented.
April 1, 2006 | midnightbsd-3.1 |