libtpms/man/man3/TPMLIB_RegisterCallbacks.pod
Corey Bryant a0098eda2d Initial import of project
This is the initial import of the libtpms library.  The libtpms library
provides software emulation of a Trusted Platform Module (TPM).  It is
intended to be used by applications when a hardware TPM is not adequate.
For example, a hypervisor can use libtpms to emulate an independent TPM
for each of it's virtual machine guests.  The library provides a high-
level API for developers to integrate the emulated TPM support into
their application.

The code was originally written by Kenneth Goldman <kgoldman@us.ibm.com>
and Stefan Berger <stefanb@us.ibm.com>.

The code is licensed under the Modified BSD License.

Signed-off-by: Corey Bryant <coreyb@linux.vnet.ibm.com>
2013-10-31 15:33:22 -05:00

273 lines
8.6 KiB
Plaintext

=head1 NAME
TPMLIB_RegisterCallbacks - Register callbacks for implementing customized
behavior of certain functions
=head1 LIBRARY
TPM library (libtpms, -ltpms)
=head1 SYNOPSIS
B<#include <libtpms/tpm_types.h>>
B<#include <libtpms/tpm_library.h>>
B<#include <libtpms/tpm_error.h>>
B<TPM_RESULT TPMLIB_RegisterCallbacks(struct tpmlibrary_callbacks *);>
=head1 DESCRIPTION
The B<TPMLIB_RegisterCallbacks()> functions allows to register several
callback functions with libtpms that enable a user to implement customized
behavior of several library-internal functions. This feature will typically
be used if the behavior of the provided internal functions is not as needed.
An example would be that libtpms writes all data into files with certain names.
If, however, the data needs to be written into a special type of storage
the user will register callbacks with the library that are invoked when
the TPM needs to write, read or delete data from storage and the user may
then implement custom behavior in these functions.
The following shows the data structure used for registering the callbacks.
struct libtpms_callbacks {
int sizeOfStruct;
TPM_RESULT (*tpm_nvram_init)(void);
TPM_RESULT (*tpm_nvram_loaddata)(unsigned char **data,
uint32_t *length,
uint32_t tpm_number,
const char *name);
TPM_RESULT (*tpm_nvram_storedata)(const unsigned char *data,
uint32_t length,
uint32_t tpm_number,
const char *name);
TPM_RESULT (*tpm_nvram_deletename)(uint32_t tpm_number,
const char *name,
TPM_BOOL mustExist);
TPM_RESULT (*tpm_io_init)(void);
TPM_RESULT (*tpm_io_getlocality)(TPM_MODIFIER_INDICATOR *localityModifer,
uint32_t tpm_number);
TPM_RESULT (*tpm_io_getphysicalpresence)(TPM_BOOL *physicalPresence,
uint32_t tpm_number);
};
Currently 7 callbacks are supported. If a callback pointer in the above
structure is set to NULL the default library-internal implementation
of that function will be used.
If one of the callbacks in either the I<tpm_nvram> or I<tpm_io> group is
set, then all of the callbacks in the respective group should
be implemented.
=over 4
=item B<tpm_nvram_init>
This function is called before any access to persitent storage is done. It
allows the user to perform initialization of access to persitent storage.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation requires that the environment variable
I<TPM_PATH> is set and points to a directory where the TPM's state
can be written to. If the variable is not set, it will return B<TPM_FAIL>
and the initialization of the TPM in B<TPMLIB_MainInit()> will fail.
=item B<tpm_nvram_loaddata>
This function is called when the TPM wants to load state from persistent
storage. The implementing function must allocate a buffer (I<data>)
and return it to the TPM along with the length of the buffer (I<length>).
The I<tpm_number> is always 0 and can be ignored.
The I<name> parameter is either one of B<TPM_SAVESTATE_NAME>,
B<TPM_VOLATILESTATE_NAME>, or B<TPM_PERMANENT_ALL_NAME> and indicates
which one of the 3 types of state is supposed to be loaded.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation writes the TPM's state into files in a directory
where the I<TPM_PATH> environment variable pointed to when
B<TPMLIB_MainInit()> was executed. Failure to write the TPM's state into
files will put the TPM into failure mode.
=item B<tpm_nvram_storedata>
This function is called when the TPM wants to store state to persistent
storage. The I<data> and I<length> parameters provide the data to be
stored and the number of bytes. The implementing function must not
free the I<data> buffer.
The I<tpm_number> is always 0 and can be ignored.
The I<name> parameter is either one of B<TPM_SAVESTATE_NAME>,
B<TPM_VOLATILESTATE_NAME>, or B<TPM_PERMANENT_ALL_NAME> and indicates
which one of the 3 types of state is supposed to be stored.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation reads the TPM's state from files in a directory
where the I<TPM_PATH> environment variable pointed to when
B<TPMLIB_MainInit()> was executed. Failure to read the TPM's state from
files may put the TPM into failure mode.
=item B<tpm_nvram_deletename>
This function is called when the TPM wants to delete state on persistent
storage.
The I<tpm_number> is always 0 and can be ignored.
The I<name> parameter is either one of B<TPM_SAVESTATE_NAME>,
B<TPM_VOLATILESTATE_NAME>, or B<TPM_PERMANENT_ALL_NAME> and indicates
which one of the 3 types of state is supposed to be deleted.
The I<mustExist> parameter indicates wheteher the given data must exist
and the implementing function should return B<TPM_FAIL> if the data did
not exist.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation deletes the TPM's state files in a directory
where the I<TPM_PATH> environment variable pointed to when
B<TPMLIB_MainInit()> was executed. Failure to delete the TPM's state
files may put the TPM into failure mode.
=item B<tpm_io_init>
This function is called to initialize the IO subsystem of the TPM.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation simply returns B<TPM_SUCCESS>.
=item B<tpm_io_getlocality>
This function is called when the TPM needs to determine the locality
under which a command is supposed to be executed. The implementing function
should return the number of the locality by writing it into the
B<localityModifier> pointer.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation returns 0 as the locality.
=item B<tpm_io_getphysicalpresence>
This function is called when the TPM needs to determine whether physical
presence has been asserted. The implementing function should write either
B<TRUE> or B<FALSE> into the physicalPresence pointer.
Upon success this function should return B<TPM_SUCCESS>, a failure code
otherwise.
The default implementation returns B<FALSE> for physical presence.
=back
=head1 RETURN VALUE
Upon successful completion, B<TPMLIB_MainInit()> returns B<TPM_SUCCESS>,
an error value otherwise.
=head1 ERRORS
=over 4
=item B<TPM_SUCCESS>
The function completed sucessfully.
=item B<TPM_FAIL>
General failure.
=back
For a complete list of TPM error codes please consult the include file
B<libtpms/tpm_error.h>
=head1 EXAMPLE
#include <libtpms/tpm_types.h>
#include <libtpms/tpm_library.h>
#include <libtpms/tpm_error.h>
static TPM_MODIFIER_INDICATOR locality;
static TPM_RESULT mytpm_io_init(void)
{
return TPM_SUCCESS;
}
static TPM_RESULT tpm_io_getlocality(TPM_MODIFIER_INDICATOR *locModif)
{
*locModif = locality;
return TPM_SUCCESS:
}
static TPM_RESULT mytpm_io_getphysicalpresence(TPM_BOOL *phyPres)
{
*physicalPresence = FALSE;
return TPM_SUCCESS;
}
int main(void) {
TPM_RESULT res;
unsigned char *respbuffer;
uint32_t resp_size;
uint32_t respbufsize;
unsigned char *command;
uint32_t command_size;
struct libtpms_callbacks cbs = {
.sizeOfStruct = sizeof(struct libtpms_callbacks),
.tpm_nvram_init = NULL,
.tpm_nvram_loaddata = NULL,
.tpm_nvram_storedata = NULL,
.tpm_nvram_deletename = NULL,
.tpm_io_init = mytpm_io_init,
.tpm_io_getlocality = mytpm_io_getlocality,
.tpm_io_getphysicalpresence = mytpm_io_getphysicalpresence,
};
[...]
if (TPMLIB_RegisterCallbacks(cbs) != TPM_SUCCESS) {
fprintf(stderr, "Could not register the callbacks.\n");
return 1;
}
if (TPMLIB_MainInit()) != TPM_SUCCESS) {
fprintf(stderr, "Could not start the TPM.\n");
return 1;
}
[...]
/* build TPM command */
[...]
res = TPMLIB_Process(&respbuffer, &resp_size,
&respbufsize,
command, command_size);
[...]
TPMLIB_Terminate();
return 0;
}
=head1 SEE ALSO
B<TPMLIB_Process>(3), B<TPMLIB_MainInit>(3), B<TPMLIB_Terminate>(3),
B<TPMLIB_DecodeBlobs>(3)
=cut