Now that we have an EVP namemap containing all aliases that providers
know about for any given algorithm, it is possible that an application
attempts to look up a digest or a cipher via EVP_get_digestbyname() or
EVP_get_cipherbyname() with an algorithm name that is unknown to the
legacy method database. Therefore we extend those functions to
additionally check the aliases in the namemap when searching for a
method in the event that our initial lookup attempt fails.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10324)
Because the algorithm to use is decided already when creating an
EVP_PKEY_CTX regardless of how it was created, it turns out that it's
unnecessary to provide the SIGNATURE method explicitly, and rather
always have it be fetched implicitly.
This means fewer changes for applications that want to use new
signature algorithms / implementations.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10303)
Because the algorithm to use is decided already when creating an
EVP_PKEY_CTX regardless of how it was created, it turns out that it's
unnecessary to provide the KEYEXCH method explicitly, and rather
always have it be fetched implicitly.
This means fewer changes for applications that want to use new key
exchange algorithms / implementations.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10305)
With provided algorithms, the library context is ever present, so of
course it should be specified alongside the algorithm name and
property query string.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10308)
There is a vagueness around how the provider data (algorithm name and
property query string) is initialized in the presence of an engine.
This change modifies this slightly so that the algorithm name for use
with providers is never set if the initilization was given an engine.
This makes it easier for other functions to simply check ctx->algorithm
to see if the context is meant to be used for strictly legacy stuff or
not.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10308)
Otherwise, should this function be called more than once on the same
EVP_PKEY_CTX, we get double free issues.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10292)
RSA-PSS keys use the same internal structure as RSA keys but do not
allow accessing it through EVP_PKEY_get0_RSA. This commit changes that
behavior.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10217)
This leaves minimal implementations of EVP_blake2b512 and EVP_blake2s256,
that are now only there to provide a name for implicit fetches.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9075)
It may be that the OSSL_PARAM array we used for getting parameter
values for a key had a few too many entries. These are detected by
their return_size == 0. Before making second export call, we prune
away these items so we only ask for parameters that exist.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10190)
Method data was passed down as provider to ossl_algorithm_do_all(),
which causes trouble as soon a it's non-NULL. Pass it via the data
structure instead.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9979)
The following new functions all do the same thing; they traverse
the set of names assigned to implementations of each algorithm type:
EVP_MD_names_do_all(), EVP_CIPHER_names_do_all(),
EVP_MAC_names_do_all(), EVP_KEYMGMT_names_do_all(),
EVP_KEYEXCH_names_do_all(), EVP_KDF_names_do_all(),
EVP_SIGNATURE_names_do_all()
We add a warning to the documentation of EVP_CIPHER_name() and
EVP_MD_name(), as they aren't suitable to use with multiple-name
implementation.
We also remove EVP_MAC_name() and evp_KDF_name(), as they serve no
useful purpose.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9979)
This adds the missing functions that should be common for all
fetchable EVP sub-APIs:
EVP_KEYMGMT_is_a(), EVP_KEYMGMT_do_all_provided(), EVP_KEYEXCH_is_a(),
EVP_KEYEXCH_do_all_provided(), EVP_KDF_is_a(), EVP_MD_is_a(),
EVP_SIGNATURE_do_all_provided(), EVP_SIGNATURE_is_a().
This also renames EVP_MD_do_all_ex(), EVP_CIPHER_do_all_ex(),
EVP_KDF_do_all_ex(), EVP_MAC_do_all_ex() to change '_ex'
to '_provided'.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9979)
This control command should never be used with provided methods, but
since this is publically available, someone might still make the
mistake. We make sure it returns 1 so as not to be overly
disruptive.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10163)
This works as much as possible EVP_PKEY_CTX_new_id(), except it takes
data that's relevant for providers, algorithm name and property query
string instead of NID and engine.
Additionally, if EVP_PKEY_CTX_new() or EVP_PKEY_CTX_new_id() was
called, the algorithm name in the EVP_PKEY context will be set to the
short name of the given NID (explicit or the one of the given
EVP_PKEY), thereby giving an easier transition from legacy methods to
provided methods.
The intent is that operations will use this information to fetch
provider methods implicitly as needed.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10184)
We need to pass the SSL3 Master Secret down to the provider code in order
for SSLv3 to work correctly.
Fixes#10182
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10186)
This leaves minimal implementations of EVP_md2, EVP_md4, EVP_md5 and
EVP_mdc2, that are now only there to provide a name for implicit fetches.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10164)
If we remove these, the functions EVP_get_digestbyname() and
EVP_get_cipherbyname() will stop working entirely, and it's too early
to criple them yet.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10176)
EVP_Cipher() would return whatever ctx->cipher->ccipher() returned
with no regard for historical semantics.
We change this to first look if there is a ctx->cipher->ccipher(), and
in that case we treat the implementation as one with a custom cipher,
and "translate" it's return value like this: 0 => -1, 1 => outl, where
|outl| is the output length.
If there is no ctx->cipher->ccipher, we treat the implementation as
one without a custom cipher, call ctx->cipher->cupdate or
ctx->cipher->cfinal depending on input, and return whatever they
return (0 or 1).
Furthermore, we add a small hack in EVP_CIPHER_flags() to check if the
cipher is a provided one, and add EVP_CIPH_FLAG_CUSTOM_CIPHER to the
flags to be returned if there is a cipher->ccipher. That way,
provided implementations never have to set that flag themselves, all
they need to do is to include a OSSL_FUNC_CIPHER_CIPHER function.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10137)
Prior to OpenSSL 3.0 EVP_Digest[Sign|Verify|Update were just macros for
EVP_DigestUpdate. They are now separate functions. Unfortunately some
code assumes that EVP_Digest[Sign|Verify]Update is interchangeable with
EVP_DigestUpdate. For example the dgst app uses an MD bio which always
calls EVP_DigestUpdate(). However the dgst app supports signing instead
of digesting and may initialise with EVP_DigestSignInit_ex() instead of
just EVP_DigestInit().
We now detect these differences and redirect to the correct function
where appropriate.
Fixes#10114
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10116)
There is no need for us to be diving inside the EVP_MD_CTX in the
implementation of an MD BIO. We can just use public APIs. By doing this
certain calls (such as getting the MD out of the BIO were not working
correctly) where providers are in use.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10116)
From providers/{common,default,legacy}/ to providers/implementations/
However, providers/common/digests/digest_common.c stays where it is,
because it's support code rather than an implementation.
To better support all kinds of implementations with common code, we
add the library providers/libcommon.a. Code that ends up in this
library must be FIPS agnostic.
While we're moving things around, though, we move digestscommon.h
from providers/common/include/internal to providers/common/include/prov,
thereby starting on a provider specific include structure, which
follows the line of thoughts of the recent header file reorganization.
We modify the affected '#include "internal/something.h"' to
'#include "prov/something.h"'.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10088)
We put almost everything in these internal static libraries:
libcommon Block building code that can be used by all
our implementations, legacy and non-legacy
alike.
libimplementations All non-legacy algorithm implementations and
only them. All the code that ends up here is
agnostic to the definitions of FIPS_MODE.
liblegacy All legacy implementations.
libnonfips Support code for the algorithm implementations.
Built with FIPS_MODE undefined. Any code that
checks that FIPS_MODE isn't defined must end
up in this library.
libfips Support code for the algorithm implementations.
Built with FIPS_MODE defined. Any code that
checks that FIPS_MODE is defined must end up
in this library.
The FIPS provider module is built from providers/fips/*.c and linked
with libimplementations, libcommon and libfips.
The Legacy provider module is built from providers/legacy/*.c and
linked with liblegacy, libcommon and libcrypto.
If module building is disabled, the object files from liblegacy and
libcommon are added to libcrypto and the Legacy provider becomes a
built-in provider.
The Default provider module is built-in, so it ends up being linked
with libimplementations, libcommon and libnonfips. For libcrypto in
form of static library, the object files from those other libraries
are simply being added to libcrypto.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10088)
Also added blanks lines after declarations in a couple of places.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9916)
Even thought the underlying calls might return something other than 0
or 1, EVP_CIPHER_CTX_ctrl() and EVP_MD_CTX_ctrl() were made to only
return those values regardless. That behaviour was recently lost, so
we need to restore it.
Fixes#10106
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/10108)
This leaves minimal implementations of EVP_md5_sha1, which is now only
there to provide a name for implicit fetches.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9076)
This leaves minimal implementations of EVP_sha* and EVP_shake*, which
is now only there to provide a name for implicit fetches.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10059)
This modifies the treatment of algorithm name strings to allow
multiple names separated with colons.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/8985)
EVP_DigestInit_ex forced following of the legacy path if ctx->pctx is
set (meaning we've actually been called via EVP_DigestSignInit_ex).
There is some code in the legacy path that calls the
EVP_PKEY_CTRL_DIGESTINIT ctrl on the pctx. Not going down the legacy path
if ctx->pctx is set means that ctrl message will neve get sent. However,
it turns out that all algs that understand that ctrl also set the
EVP_MD_CTX_FLAG_NO_INIT flag which forces legacy anyway. Therefore the
ctx->pctx check is not required and can be removed.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10082)
Otherwise a mem leak can occur since EVP_MD_free() calls
EVP_MD_CTX_reset() which then clears the contents of the ctx.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10013)
This resets the fields of the EVP_MD_CTX and means we can no longer
make calls using the EVP_MD_CTX, such as to query parameters.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10013)
If doing an EVP_DigestSign*() or EVP_DigestVerify*() operation we use
the embedded pctx for communication with the provider. Any MD params need
to use that ctx instead.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10013)
This flag is still relevant even for non-legacy code so we should check
it where appropriate.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/10013)
So far, these two funtions have depended on legacy EVP_CIPHER
implementations to be able to do their work. This change adapts them
to work with provided implementations as well, in one of two possible
ways:
1. If the implementation's set_asn1_parameters or get_asn1_parameters
function pointers are non-NULL, this is a legacy implementation,
and that function is called.
2. Otherwise, if the cipher doesn't have EVP_CIPH_FLAG_CUSTOM_ASN1
set, the default AlgorithmIdentifier parameter code in libcrypto
is executed.
3. Otherwise, if the cipher is a provided implementation, the ASN1
type structure is converted to a DER blob which is then passed to
the implementation as a parameter (param_to_asn1) or the DER blob
is retrieved from the implementation as a parameter and converted
locally to a ASN1_TYPE (asn1_to_param).
With this, the old flag EVP_CIPH_FLAG_DEFAULT_ASN1 has become
irrelevant and is simply ignored.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10008)
All instances of EVP_*_CTX_gettable_params functions have been renamed
to EVP_*_gettable_ctx_params. Except for the EVP_MD ones which were changed
already.
These functions do not take EVP_*_CTX arguments so their prior naming was
misleading.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10052)
Apart from public and internal header files, there is a third type called
local header files, which are located next to source files in the source
directory. Currently, they have different suffixes like
'*_lcl.h', '*_local.h', or '*_int.h'
This commit changes the different suffixes to '*_local.h' uniformly.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)
Currently, there are two different directories which contain internal
header files of libcrypto which are meant to be shared internally:
While header files in 'include/internal' are intended to be shared
between libcrypto and libssl, the files in 'crypto/include/internal'
are intended to be shared inside libcrypto only.
To make things complicated, the include search path is set up in such
a way that the directive #include "internal/file.h" could refer to
a file in either of these two directoroes. This makes it necessary
in some cases to add a '_int.h' suffix to some files to resolve this
ambiguity:
#include "internal/file.h" # located in 'include/internal'
#include "internal/file_int.h" # located in 'crypto/include/internal'
This commit moves the private crypto headers from
'crypto/include/internal' to 'include/crypto'
As a result, the include directives become unambiguous
#include "internal/file.h" # located in 'include/internal'
#include "crypto/file.h" # located in 'include/crypto'
hence the superfluous '_int.h' suffixes can be stripped.
The files 'store_int.h' and 'store.h' need to be treated specially;
they are joined into a single file.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9333)
using PCC and KDSA instructions.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/10004)
EVP_MD_CTX_gettable_params() and EVP_MD_CTX_settable_params() were
confusingly named because they did not take an EVP_MD_CTX parameter.
In addition we add the functions EVP_MD_gettable_ctx_params() and
EVP_MD_settable_ctx_params() which do the same thing but are passed
an EVP_MD object instead.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9998)
With multiple names, it's no longer viable to just grab the "first" in
the set and use that to find the legacy NID. Instead, all names for
an algorithm must be checked, and if we encounter more than one NID
asssociated with those names, we consider it an error and make that
method unloadable.
This ensures that all methods that do have an internal NID associated
will get that NID in their structure, thereby ensuring that other
parts of libcrypto that haven't gone away from using NIDs for
comparison will continue to work as expected.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9969)
It either makes the flow of control simpler and more obvious, or it is
just a "cleanup" so that the editing scripts will find and fixup things.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9441)
With some provider implementations, there are underlying ciphers,
digests and macs. For some of them, the name was retrieved from the
method, but since the methods do not store those any more, we add
different mechanics.
For code that needs to pass on the name of a cipher or diges via
parameters, we simply locally store the name that was used when
fetching said cipher or digest. This will ensure that any underlying
code that needs to fetch that same cipher or digest does so with the
exact same name instead of any random name from the set of names
associated with the algorithm.
For code that needs to check what kind of algorithm was passed, we
provide EVP_{type}_is_a(), that returns true if the given method has
the given name as one of its names.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9897)
Multiple names per implementation is already supported in the namemap,
but hasn't been used yet. However, as soon as we have multiple names,
we will get an issue with what name should be saved in the method.
The solution is to not save the name itself, but rather the number
it's associated with. This number is supposed to be unique for each
set of names, and we assume that algorithm names are globally unique,
i.e. there can be no name overlap between different algorithm types.
Incidently, it was also found that the 'get' function used by
ossl_construct_method() doesn't need all the parameters it was given;
most of what it needs, it can now get through the data structure given
by the caller of ossl_construct_method(). As a consequence,
ossl_construct_method() itself doesn't need all the parameters it was
given either.
There are some added internal functions that are expected to disappear
as soon as legacy code is removed, such as evp_first_name() and
ossl_namemap_num2name().
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9897)
This caused a SEGV inside tls13_enc() when using chacha_poly.
The tls code assigns the iv_length to a size_t (even though it is an int).
This is actually really bad since it could be -1, which will then trash the iv buffer.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9890)
The output C code was made to use ERR_func_error_string() to see if a
string table was already loaded or not. Since this function returns
NULL always, this check became useless.
Change it to use ERR_reason_error_string() instead, as there's no
reason to believe we will get rid of reason strings, ever.
To top it off, we rebuild all affected C sources.
Fixes#9756
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9756)
It is undefined behaviour to send NULL as either the src, or dest params
in memcpy.
In pkey_kdf.c we had a check to ensure that the src address is non-NULL.
However in some situations it is possible that the dest address could also
be NULL. Specifically in the case where the datalen is 0 and we are using
a newly allocated BUF_MEM.
We add a check of datalen to avoid the undefined behaviour.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9868)
Make sure we pass the provider side ctx and not the libcrypto side ctx.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9865)
An EVP_PKEY can be used for multiple different algorithm operations.
Only one can be used at a time, so we move those into a union.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9753)
We add new functions for getting parameters and discovering the gettable
and settable parameters. We also make EVP_PKEY_CTX_get_signature_md() a
function and implement it in terms of the new functions.
This enables applications to discover the set of parameters that are
supported for a given algorithm implementation.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9753)
This makes EVP_PKEY_sign and EVP_PKEY_sign_init provider aware. It
also introduces the new type EVP_SIGNATURE to represent signature
algorithms. This also automatically makes the EVP_Sign* APIs provider
aware because they use EVP_Digest* (which is already provider aware)
and EVP_PKEY_sign(_init) under the covers.
At this stage there are no signature algorithms in any providers. That
will come in the following commits.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9753)
BUF_MEM_grow() returns the passed length, but also zero on error. If
the passed length was zero, an extra check to see if a returned zero
was an error or not is needed.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
pkey_kdf_ctrl_str() has to do the same kind of special treatment as
pkey_kdf_ctrl() does.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
Undo the caching scheme, pass through most controls as parameters, except
for SEED and INFO, where we keep supporting adding data through additional
ctrl calls by collecting the data, and only passing it to the EVP_KDF
before calling its derive function.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9662)
If the passed string length is zero, the function computes the string length
from the passed string.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9760)
This function re-implements EVP_CIPHER_meth_free(), but has a name that
isn't encumbered by legacy EVP_CIPHER construction functionality.
We also refactor most of EVP_CIPHER_meth_new() into an internal
evp_cipher_new() that's used when creating fetched methods.
EVP_CIPHER_meth_new() and EVP_CIPHER_meth_free() are rewritten in terms of
evp_cipher_new() and EVP_CIPHER_free(). This means that at any time, we can
deprecate all the EVP_CIPHER_meth_ functions with no harmful consequence.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9758)
This function re-implements EVP_MD_meth_free(), but has a name that
isn't encumbered by legacy EVP_MD construction functionality.
We also refactor most of EVP_MD_meth_new() into an internal
evp_md_new() that's used when creating fetched methods.
EVP_MD_meth_new() and EVP_MD_meth_free() are rewritten in terms of
evp_md_new() and EVP_MD_free(). This means that at any time, we can
deprecate all the EVP_MD_meth_ functions with no harmful consequence.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9758)
KEYMGMT methods were attached to other methods after those were fully
created and registered, thereby creating a potential data race, if two
threads tried to create the exact same method at the same time.
Instead of this, we change the method creating function to take an
extra data parameter, passed all the way from the public fetching
function. In the case of EVP_KEYEXCH, we pass all the necessary data
that evp_keyexch_from_dispatch() needs to be able to fetch the
appropriate KEYMGMT method on the fly.
Fixes#9592
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9678)
Use the header file internal/cryptlib.h instead.
Remove checks for OPENSSL_NO_ASM and I386_ONLY
in cryptlib.c, to match the checks in other
places where OPENSSL_ia32cap_P is used and
assumed to be initialized.
Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/9688)
ENGINE_get_name() actually returns more of a long description of the
engine, whilst ENGINE_get_id() returns a shorter id.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9720)
Make sure references to ENGINE functions are appropriately guarded.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9720)
Don't modify the cipher/md we just fetched - it could be shared by multiple
threads.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9590)
Moved the relevant ciphers into default and restructed headers to allow the move.
This removed most of the cases of #ifdef NO_XXX (which are now specified in build.info)
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9482)
The EVP_PKEY MAC implementations had a diversity of controls that were
really the same thing. We did reproduce that for the provider based
MACs, but are changing our minds on this. Instead of that, we now use
one parameter name for passing the name of the underlying ciphers or
digests to a MAC implementation, "cipher" and "digest", and one
parameter name for passing the output size of the MAC, "size".
Then we leave it to the EVP_PKEY->EVP_MAC bridge to translate "md"
to "digest", and "digestsize" to "size".
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9667)
A few declarations and static functions / arrays that are no longer
used were left behind when MACs moved to be implemented by providers.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9667)
The aes code has been refactored into generic and algorithn specific parts,
so that most of the code can be shared.
The cipher related files have been broken up into smaller parts.
Add chunked variant of mode ciphers - aria uses this (many other ciphers will use this new code instead of the
generic code used by aes).
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9451)
If ossl_method_store_add() gets called with a method that already exists
(i.e. the store has one with matching provider, nid and properties), that
method should not be stored. We do this check inside ossl_method_store_add()
because it has all the locking required to do so safely.
Fixes#9561
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9650)
Because this function affects the reference count on failure (the call
to impl_free() does this), it may as well handle incrementing it as
well to indicate the extra reference in the method store.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9650)
Add Cleanups for gcm - based on the changes to ccm.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
(Merged from https://github.com/openssl/openssl/pull/9280)
The macros are defined in include/openssl/core_names.h and follow the
naming standard OSSL_{OPNAME}_NAME_{ALGONAME}, where {OPNAME} is the
name of the operation (such as MAC) and {ALGONAME} is the name of the
algorithm. Example: OSSL_MAC_NAME_HMAC
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9635)
Recently, we added dispatched functions to get parameter descriptions,
and those for operation context parameters ended up being called
something_gettable_ctx_params and something_settable_ctx_params.
The corresponding dispatched functions to actually perform parameter
transfers were previously called something_ctx_get_params and
something_ctx_set_params, which doesn't quite match, so we rename them
to something_get_ctx_params and something_set_ctx_params.
An argument in favor of this name change is English, where you'd
rather say something like "set the context parameters".
This only change the libcrypto <-> provider interface.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9612)
For information processing.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8877)
Now that all our MACs have moved to the default provider, we let it
take over completely
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8877)
Quite a few adaptations are needed, most prominently the added code
to allow provider based MACs.
As part of this, all the old information functions are gone, except
for EVP_MAC_name(). Some of them will reappear later, for example
EVP_MAC_do_all() in some form.
MACs by EVP_PKEY was particularly difficult to deal with, as they
need to allocate and deallocate EVP_MAC_CTXs "under the hood", and
thereby implicitly fetch the corresponding EVP_MAC. This means that
EVP_MACs can't be constant in a EVP_MAC_CTX, as their reference count
may need to be incremented and decremented as part of the allocation
or deallocation of the EVP_MAC_CTX. It may be that other provider
based EVP operation types may need to be handled in a similar manner.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8877)
These functions were missing for a completes API:
EVP_MD_get_params(), EVP_CIPHER_get_params(), EVP_CIPHER_CTX_set_params(),
and EVP_CIPHER_CTX_get_params
Additionally, we also add all the corresponding parameter descriptor
returning functions, along the correspoding provider dispatches:
EVP_MD_gettable_params(), EVP_MD_CTX_settable_params(),
EVP_MD_CTX_gettable_params(), EVP_CIPHER_gettable_params(),
EVP_CIPHER_CTX_settable_params(), and EVP_CIPHER_CTX_gettable_params()
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9576)
Also clean up EVP_MD_CTX_ctrl(), which did use these interfaces, but
development since allows for more elegant code.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9391)
If an implict EVP_CIPHER_fetch fails then ctx->cipher should not be set
otherwise strange things will happen when trying to free the ctx.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9531)
The code has been modularized so that it can be shared by algorithms.
A fixed size IV is now used instead of being allocated.
The IV is not set into the low level struct now until the update (it uses an
iv_state for this purpose).
Hardware specific methods have been added to a PROV_GCM_HW object.
The S390 code has been changed to just contain methods that can be accessed in
a modular way. There are equivalent generic methods also for the other
platforms.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Patrick Steuer <patrick.steuer@de.ibm.com>
(Merged from https://github.com/openssl/openssl/pull/9231)
Also added EVP_CTRL_RET_UNSUPPORTED define (so magic numbers can be removed)
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9464)
Extends the existing provider documentation with information about the
CIPHER operation. This is primarily for provider authors.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9473)
The biggest part in this was to move the key->param builder from EVP
to the DH ASN.1 method, and to implement the KEYMGMT support in the
provider DH.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9394)
This function is used to traverse all algorithm implementations for a
given operation type, and execute the given function for each of them.
For each algorithm implementation, a method is created and passed to
the given function, and then freed after that function's return. If
the caller wishes to keep the method for longer, they must call the
appropriate up_ref function on the method, and they must also make
sure to free the passed methods at some point.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9356)
They will do the same as usual for non-provider algorithms
implementations, but can handle provider implementations as well.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9356)
This will be useful for information display, as well as for code that
want to check the name of an algorithm. This can eventually replace
all NID checks.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9356)
This function clears the cache of provider key references, and is used
in evp_keymgmt_export_to_provider() when the internal key is dirty, as
well as by EVP_PKEY_free_it().
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)
This affects all its callers: EVP_PKEY_CTX_new(), EVP_PKEY_CTX_new_id().
They are now possible to called with "zero" values, i.e.:
EVP_PKEY_CTX *pctx = EVP_PKEY_CTX_new(NULL, NULL);
or
EVP_PKEY_CTX *pctx = EVP_PKEY_CTX_new_id(0, NULL);
This is suitable for provider use, as the key functionality is tied
with its keys, and the operation time is determined by the init
functions the EVP_PKEY_CTX is used with.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)
This function is used to transport private key materia from whatever
is already attached to the EVP_PKEY to the new provider, using key
data export and import functionality.
If a legacy lower level key has been assigned to the EVP_PKEY, we use
its data to create a provider side key, and thereby have a bridge
between old style public key types and the EVP_PKEY on providers.
If successful, this function returns a reference to the appropriate
provider side data for the key.
This can be used by any operation that wants to use this key.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)
The idea with the key management "operation" is to support the
following set of functionality:
- Key domain parameter generation
- Key domain parameter import
- Key domain parameter export
- Key generation
- Key import
- Key export
- Key loading (HSM / hidden key support)
With that set of function, we can support handling domain parameters
on one provider, key handling on another, and key usage on a third,
with transparent export / import of applicable data. Of course, if a
provider doesn't offer export / import functionality, then all
operations surrounding a key must be performed with the same
provider.
This method also avoids having to do anything special with legacy
assignment of libcrypto key structures, i.e. EVP_PKEY_assign_RSA().
They will simply be used as keys to be exported from whenever they are
used with provider based operations.
This change only adds the EVP_KEYMGMT API and the libcrypto <->
provider interface. Further changes will integrate them into existing
libcrypto functionality.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9312)
The recent move of the DH code into the default provider broke no-dh. This
adds back in various missing guards.
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9399)
This also adds the ability to set arbitrary parameters on key exchange
algorithms. The ability to pad the output is one such parameter for DH.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)
The default provider now has support for PKCS#3 Diffie-Hellman so we
switch libcrypto to using providers for that algorithm.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)
We introduce a new EVP_KEYEXCH type to represent key exchange algorithms
and refactor the existing code to use it where available.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9266)
Custom aes ciphers will be placed into multiple new files
(instead of the monolithic setup used in the e_aes.c legacy code)
so it makes sense to have a header for the platform specific
code that needs to be shared between files.
modes_lcl.h has also moved to modes_int.h to allow sharing with the
provider source.
Code that will be common to AEAD ciphers has also been added. These
will be used by seperate PR's for GCM, CCM & OCB.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9301)
Move the KDF code for CMS DH key agreement into an EVP_KDF object.
There are 2 specifications for X9.42 KDF. This implementation uses DER for
otherinfo which embeds the KDF loop counter inside the DER object.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8898)
Common pattern is that the routines to increment the reference count
are called something_up_ref, not something_upref. Adapt
ossl_provider_upref() accordingly.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9293)
CLA: trivial
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9288)
The maximum key length for rc5 is 2040 bits so we should not attempt to
use keys longer than this.
Issue found by OSS-Fuzz and Guido Vranken.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8834)
We were attempting to cache a method after we failed to create it
which leads to an assertion failure.
Fixes#9264
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9269)
CLA: trivial
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9275)
All the other upref functions are spelled as "up_ref". These new functions
should be consistent.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9233)
SIV mode is accessible via EVP. There should be no reason to make the low
level SIV functions from the modes directory part of the public API. Since
these functions do not exist in 1.1.1 we are still able to make this change.
This also reduces the list of newly added undocumented symbols from
issue #9095.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9232)
Algorithms may have multiple names, as seen in the legacy names
database. We need to support that as well.
This implementations modifies ossl_namemap to support multiple names
for the same identifier.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8967)
This feature is enabled by default outside of FIPS builds
which ban such actions completely.
Encryption is always disallowed and will generate an error.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9112)
Instead of referencing the return size from the OSSL_PARAM structure, make the
size a field within the structure.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9135)
The existing code used PKCS5 specifications.
SP800-132 adds the following additional constraints for:
- the range of the key length.
- the minimum iteration count (1000 recommended).
- salt length (at least 128 bits).
These additional constraints may cause errors (in scrypt, and
some PKCS5 related test vectors). To disable the new
constraints use the new ctrl string "pkcs5".
For backwards compatability, the checks are only enabled by
default for fips mode.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8868)
This is still required currently by engines and digestsign/digestverify.
This PR contains merged in code from Richard Levitte's PR #9126.
[extended tests]
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9103)
Because the operation identity wasn't integrated with the created
methods, the following code would give unexpected results:
EVP_MD *md = EVP_MD_fetch(NULL, "MD5", NULL);
EVP_CIPHER *cipher = EVP_CIPHER_fetch(NULL, "MD5", NULL);
if (md != NULL)
printf("MD5 is a digest\n");
if (cipher != NULL)
printf("MD5 is a cipher\n");
The message is that MD5 is both a digest and a cipher.
Partially fixes#9106
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9109)
If a EVP_MD_CTX holds a reference to a previously given engine, and
the type of its digest isn't the same as the one given in the new
call, drop that engine reference, allowing providers or other engines
to provide the new algorithm on an equal basis.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9077)
Move digest code into the relevant providers (fips, default, legacy).
The headers are temporarily moved to be internal, and will be moved
into providers after all external references are resolved. The deprecated
digest code can not be removed until EVP_PKEY (signing) is supported by
providers. EVP_MD data can also not yet be cleaned up for the same reasons.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8763)
In all legacy code ctx->cipher is dereferenced without checks, so it
makes no sense to jump there is ctx->cipher is NULL as it will just lead
to a crash. Catch it separately and return an error.
This is simlar to the fix in d2c2e49eab
Signed-off-by: Simo Sorce <simo@redhat.com>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9002)
CLA: trivial
Fixes#8904
Commit 48feaceb53 ("Remove the possibility to disable the UI module
entirely", 2017-07-03) made the BUFSIZ references in "evp_key.c"
unconditional, by deleting the preprocessing directive "#ifndef
OPENSSL_NO_UI". This breaks the build when compiling OpenSSL for edk2
(OPENSSL_SYS_UEFI), because edk2's <stdio.h> doesn't #define BUFSIZ.
Provide a fallback definition, like we do in "crypto/ui/ui_util.c" (from
commit 984d6c6052, "Fix no-stdio build", 2015-09-29).
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8922)
67c81ec311 forgot about s390x
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8971)
X963 KDF is used for CMS ec keyagree Recipient Info.
The X963 KDF that is used by CMS EC Key Agreement has been moved
into a EVP_KDF object. This KDF is almost identical to the the SSKDF
hash variant, so it has been implemented inside the SSKDF code with
its own method table.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8902)
Now that the legacy NID isn't used as a main index for fetched
algorithms, the legacy NID was just transported around unnecessarily.
This is removed, and the legacy NID is simply set by EVP_{API}_fetch()
after the construction process is done.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8878)
We didn't deal very well with names that didn't have pre-defined NIDs,
as the NID zero travelled through the full process and resulted in an
inaccessible method. By consequence, we need to refactor the method
construction callbacks to rely more on algorithm names.
We must, however, still store the legacy NID with the method, for the
sake of other code that depend on it (for example, CMS).
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8878)
This avoids using the ASN1_OBJECT database, which is bloated for the
purpose of a simple number <-> name database.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8878)
This change allows to pass the authentication tag after specifying
the AAD in CCM mode. This is already true for the other two supported
AEAD modes (GCM and OCB) and it seems appropriate to match the
behavior.
GCM and OCB also support to set the tag at any point before the call
to `EVP_*Final`, but this won't work for CCM due to a restriction
imposed by section 2.6 of RFC3610: The tag must be set before
actually decrypting data.
This commit also adds a test case for setting the tag after supplying
plaintext length and AAD.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7243)
If ctx->cipher->cupdate/ctx->cipher->cfinal failed, 'soutl' is left
uninitialized.
This patch incorporates the same logic as present in EVP_DecryptUpdate and
EVP_DecryptFinal_ex: only branch on 'soutl' if the preceding call succeeded.
Bug found by OSS-Fuzz.
Signed-off-by: Guido Vranken <guidovranken@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8874)
Various core and property related code files used global data. We should
store all of that in an OPENSSL_CTX instead.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8857)
Introduce a macro that allows all structure alignment tricks to be rolled up
into a single place.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8845)
OSSL_provider_init() gets another output parameter, holding a pointer
to a provider side context. It's entirely up to the provider to
define the context and what it's being used for. This pointer is
passed back to other provider functions, typically the provider global
get_params and set_params functions, and also the diverse algorithm
context creators, and of course, the teardown function.
With this, a provider can be instantiated more than once, or be
re-loaded as the case may be, while maintaining instance state.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8848)
The OP_cipher_final function takes a return output size and an output
buffer size argument. The oneshot OP_cipher_cipher function should do
the same.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8849)
Make sure we use the the correct key length in EVP_CIPHER_CTX_rand_key().
Now that ciphers may come from providers we need to make sure we ask the
provider for the value if appropriate.
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/8794)
We also lay the ground work for various of other the basic AES ciphers.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/8700)
