Richard Levitte [Thu, 7 May 2020 06:51:09 +0000 (08:51 +0200)]
PROV: Refactor the RSA SIGNATURE implementation for better param control
We want to catch errors in passed parameters early, which requires
kowledge of the ongoing operation. Fortunately, that's possible by
re-using the EVP_PKEY_OP macros in specific init functions.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Richard Levitte [Sat, 2 May 2020 11:33:24 +0000 (13:33 +0200)]
DER writer: Add the possibility to abandon empty SEQUENCEs
In some cases, a SEQUENCE that contains only optional fields may end
up empty. In some cases, this may be represented by dropping the
SEQUENCE entirely from the encoded DER.
To do this, we detect the case where WPACKET_FLAGS_ABANDON_ON_ZERO_LENGTH
is used, and adapt accordingly.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Richard Levitte [Sat, 2 May 2020 10:46:55 +0000 (12:46 +0200)]
RSA: Add a less loaded PSS-parameter structure
RSA_PSS_PARAMS carries with it a lot of baggage in form of X509_ALGOR
and ASN1_INTEGER, which we would rather avoid in our providers.
Therefore, we create a parallell structure - RSA_PSS_PARAMS_30 - that
contains the same information, but uses numeric identities (*) and C
integers (**). This makes it simpler to handle.
Note that neither this structure nor its contents are passed between
libcrypto and the providers. Instead, the numeric identities are
translated to and from names, which are then passed over that
boundary.
For future considerations, we might consider dropping RSA_PSS_PARAMS
entirely. For now, it's still reserved for EVP_PKEY_ASN1_METHOD code,
which RSA_PSS_PARAMS_30 is (almost entirely) reserved for use in our
providers.
(*) We use NIDs in this case, because we already have them and because
only algorithms that libcrypto knows about are permitted in PSS
restrictions. We could use any number series we want, as long as we
know for sure what they represent.
(**) That's for saltlen and for trailerfield, which are never expect
to surpass the set of numbers that fit in a regular 'int'.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Richard Levitte [Sat, 2 May 2020 10:41:39 +0000 (12:41 +0200)]
RSA: Add rsa_schemes.c, to store scheme data and translator functions
The scheme currently added is OAEP-PSSDigestAlgorithms codified.
The translator functions translate an EVP_MD into a NID, and a NID
into a name, to support the creation and parsing of OSSL_PARAM items.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Richard Levitte [Sat, 2 May 2020 09:22:23 +0000 (11:22 +0200)]
RSA: Add RSA key types
The support of restricted RSA key types (OAEP and PSS) was lacking, or
dependent on the presence of restriction parameters. For example,
this means that even though an RSA-PSS key may have been generated, it
may appear as a plain unrestricted RSA key if parameters weren't
present (which is the case when default restriction parameters are
used)
To make it clearer what an RSA key is intended for, and avoid
depending in an EVP_PKEY, we introduce RSA key types. This is done by
reserving a section of the RSA flags (4 bits, which allows a total of
16 different types).
This isn't terribly important for EVP_PKEY_ASN1_METHOD code, as that
has access to the wrapping EVP_PKEY. This is very important for
provider code, which has no access to the wrapping EVP_PKEY.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11710)
Rename OSSL_CMP_CTX_set1_clCert() to OSSL_CMP_CTX_set1_cert()
Also update documentation and example code in openssl-cmp.pod.in
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Add -reqin_new_tid option to apps/cmp.c and OSSL_CMP_MSG_update_transactionID()
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Improve CMP documentation regarding use of untrusted certs
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Chunk 11 of CMP contribution to OpenSSL: CMP command-line interface
Certificate Management Protocol (CMP, RFC 4210) extension to OpenSSL
Also includes CRMF (RFC 4211) and HTTP transfer (RFC 6712).
Adds the CMP and CRMF API to libcrypto and the "cmp" app to the CLI.
Adds extensive documentation and tests.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Preliminary fix of memory leak in try_decode_PKCS12() - full fix is in #11733
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Improve description of algorithm NIDs in doc/man3/OSSL_CMP_CTX_new.pod
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Reflect constifications of 62dcd2aa in doc/man3/OSSL_CRMF_MSG_get0_tmpl.pod
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11470)
Richard Levitte [Tue, 12 May 2020 06:46:23 +0000 (08:46 +0200)]
PROV: Add a proper provider context structure for OpenSSL providers
The provider context structure is made to include the following information:
- The core provider handle (first argument to the provider init
function). This handle is meant to be used in all upcalls that need
it.
- A library context, used for any libcrypto calls that need it, done in
the provider itself.
Regarding the library context, that's generally only needed if the
provider makes any libcrypto calls, i.e. is linked with libcrypto. That
happens to be the case for all OpenSSL providers, but is applicable for
other providers that use libcrypto internally as well.
The normal thing to do for a provider init function is to create its own
library context. For a provider that's meant to become a dynamically
loadable module, this is what MUST be done.
However, we do not do that in the default provider; it uses the library
context associated with the core provider handle instead. This is
permissible, although generally discouraged, as long as the provider in
question is guaranteed to be built-in, into libcrypto or into the
application that uses it.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11803)
I used echo `echo "<paste>" ` | sed "s/ //g"
to get everything in one line, put that again
into the clipboard and
then start sage, type N=0x<paste><CR>
>> P seems to be a prime, and is indeed a factor of N.
sage: is_pseudoprime(Q)
False
sage: gcd(N,Q)
1
sage: ecm(Q)
Found composite factor of 3 digits: 675
Composite cofactor ... has 1231 digits.
Q has a small factor. The large cofactor
is way too large to be factorized (today).
>> surprise, this is a sage artefact.
>> is_prime is supposed to tell if Qcorrect
>> is a provable prime, but these numbers are
>> too large for a proof.
sage: help(Qcorrect)
class Rational
...
>> oops, it is of course not a rational number.
sage: Qcorrect = Integer(N/P)
class Integer
...
>> okay now it is an integer.
sage: is_prime(Qcorrect)
>> takes way too long: press CTRL-C
sage: is_pseudoprime(Qcorrect)
True
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/11783)
Tomas Mraz [Mon, 11 May 2020 15:15:40 +0000 (17:15 +0200)]
Replace misleading error message when loading PEM
The error message "short header" when the end line
of PEM data cannot be identified is misleading.
Replace it with already existing "bad end line" error.
Fixes #8815
Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11793)
Richard Levitte [Mon, 11 May 2020 09:10:41 +0000 (11:10 +0200)]
CORE: Attach the provider context to the provider late
There are concerns that if |prov->provctx| is populated early,
sensitive information may leak from the provider. Therefore, we use a
temporary variable, and only assign it to |prov->provctx| when the
provider init function has returned successfully.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11777)
Richard Levitte [Sat, 9 May 2020 08:11:14 +0000 (10:11 +0200)]
Fix some misunderstandings in our providers' main modules
This started with adding forward declarations of all provider side
interface functions, and fixing all compiler errors.
Furthermore, diminish the faulty assumption that the provider context
is and always will be just a library context. That means adding a
teardown function in all providers that aren't necessarily built into
libcrypto.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11777)
Richard Levitte [Mon, 11 May 2020 07:14:11 +0000 (09:14 +0200)]
Fix d2i_PrivateKey_ex() to work as documented
d2i_PrivateKey(), and thereby d2i_PrivateKey_ex(), is documented to
return keys of the type given as first argument |type|, unconditionally.
Most specifically, the manual says this:
> An error occurs if the decoded key does not match type.
However, when faced of a PKCS#8 wrapped key, |type| was ignored, which
may lead to unexpected results.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11787)
Beat Bolli [Thu, 7 May 2020 12:19:22 +0000 (14:19 +0200)]
doc: fix two invalid <B> tags
Signed-off-by: Beat Bolli <dev@drbeat.li> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11759)
Extract HTTP server code from apps/ocsp.c to apps/lib/http_server.c
Also adds apps/include/http_server.h.
This is used so far by apps/ocsp.c and is going to be used for apps/cmp.c
and will be helpful also for any future app acting as HTTP server.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11736)
Fix bio_wait() in crypto/bio/bio_lib.c in case OPENSSL_NO_SOCK
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11736)
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11736)
Fix bug in OSSL_CMP_SRV_process_request() on transaction renewal
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11736)
As described in https://github.com/openssl/openssl/issues/9187, the
loading of PEM certificates sometimes fails if a line of base64
content has the length of a multiple of 254.
The problem is in get_header_and_data(). When such a line with a
length of 254 (or a multiple) has been read, the next read will
only read a newline. Due to this get_header_and_data() expects to be
in the header not in the data area. This commit fixes that by checking
if lines have been read completely or only partially. In case of a
previous partial read, a newline will be ignored.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11741)
Rich Salz [Thu, 7 May 2020 11:44:01 +0000 (13:44 +0200)]
Fix issues reported by markdownlint
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11739)
Rich Salz [Thu, 7 May 2020 11:42:14 +0000 (13:42 +0200)]
travis: enable markdownlint checks
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/11739)
Richard Levitte [Wed, 6 May 2020 18:48:25 +0000 (20:48 +0200)]
EVP: when setting the operation to EVP_PKEY_OP_UNDEFINED, clean up!
There were a few instances where we set the EVP_PKEY_CTX operation to
EVP_PKEY_OP_UNDEFINED, but forgot to clean up first. After the
operation is made undefined, there's no way to know what should be
cleaned away, so that must be done first, in all spots.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/11750)
Shane Lontis [Thu, 7 May 2020 06:08:18 +0000 (16:08 +1000)]
Remove legacy FIPS_mode functions
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11680)
Shane Lontis [Thu, 7 May 2020 05:31:05 +0000 (15:31 +1000)]
Remove gen_get_params & gen_gettable_params from keygen operation
EVP_PKEY_CTX_gettable_params() was missing code for the keygen operation.
After adding it it was noticed that it is probably not required for this type, so instead
the gen_get_params and gen_gettable_params have been remnoved from the provider interface.
gen_get_params was only implemented for ec to get the curve name. This seems redundant
since normally you would set parameters into the keygen_init() and then generate a key.
Normally you would expect to extract data from the key - not the object that we just set up
to do the keygen.
Added a simple settable and gettable test into a test that does keygen.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11683)
Richard Levitte [Tue, 5 May 2020 14:53:43 +0000 (16:53 +0200)]
Configure: avoid perl regexp bugs
It seems that in older perl versions '(?P' doesn't interact very well
with '(?|' or '(?:'.
Since we make extensive use of '(?P' in build.info parsing, we avoid
combining that with '(?|' and '(?:' when parsing build.info variables,
and end up parsing variable modifier twice (first generally, and then
parse that result into the modifier components).
Fixes #11694
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11737)
Matt Caswell [Fri, 1 May 2020 11:24:57 +0000 (12:24 +0100)]
Extend test_ssl_get_shared_ciphers
Ensure we test scenarios where a FIPS peer is communication with a
non-FIPS peer. Check that a FIPS client doesn't offer ciphersuites it
doesn't have, and that a FIPS server only chooses ciphersuites it can
support.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11700)
Matt Caswell [Fri, 1 May 2020 08:17:40 +0000 (09:17 +0100)]
Don't offer or accept ciphersuites that we can't support
We were not correctly detecting whether TLSv1.3 ciphersuites could
actually be supported by the available provider implementations. For
example a FIPS client would still offer CHACHA20-POLY1305 based
ciphersuites even though it couldn't actually use them. Similarly on
the server would try to use CHACHA20-POLY1305 and then fail the
handshake.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11700)
Matt Caswell [Thu, 30 Apr 2020 15:08:57 +0000 (16:08 +0100)]
Centralise Environment Variables for the tests
The test_includes test was failing if OPENSSL_CONF_INCLUDE happened to
be set in the user's environment. To ensure that no tests accidentally
use this or other enviroment variables from the user's environment we
automatically set them centrally for all tests.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11691)
Tomas Mraz [Tue, 5 May 2020 07:52:25 +0000 (09:52 +0200)]
The synthesized OPENSSL_VERSION_NUMBER must be long
(to keep API compatibility with older releases)
Fixes #11716
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/11732)
The documentation for ``EVP_default_properties_is_fips_enabled()`` uses
``fips=yes`` in one place and ``fips=true`` in another place. Stick to
``fips=yes`` like everywhere else.
Signed-off-by: Christian Heimes <christian@python.org> Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/11723)
Strengthen X509_STORE_CTX_print_verify_cb() to print expected host etc.
Add X509_VERIFY_PARAM_get0_host(), X509_VERIFY_PARAM_get0_email(),
and X509_VERIFY_PARAM_get1_ip_asc() to support this,
as well as the internal helper function ipaddr_to_asc(), which
is used also for simplifying other IP address output functions.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com>
(Merged from https://github.com/openssl/openssl/pull/11693)
Shane Lontis [Fri, 1 May 2020 07:09:01 +0000 (17:09 +1000)]
Fix aix compile error in cmp_ctx_test.c
Errors were of the form 1506-226 (S) The ":" operator is not allowed between "int" and "char*".
I think it is valid syntax the way it was written, But just rewrote so it compiled.
The aix compiler must be looking at the type of blah() when doing test ? (blah(), NULL) : X.
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11698)
Shane Lontis [Sat, 2 May 2020 03:51:35 +0000 (13:51 +1000)]
Fix incorrect default keysize for CAST ofb and cfb modes.
Fixes #11459
It was incorrectly using 8 bytes instead of 16 as the default.
This was verified by expanding the macros used in e_cast.c.
The issue occurs if EVP_CIPHER_CTX_set_key_length() is not called.
evp_test.c hides this issue as it always calls EVP_CIPHER_CTX_set_key_length() before
using EVP_CipherInit_ex(...., key, ..).
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11707)
Richard Levitte [Fri, 1 May 2020 16:06:18 +0000 (18:06 +0200)]
WPACKET: don't write DER length when we don't want to
With endfirst writing, it could be that we want to abandon any zero
length sub-packet. That's what WPACKET_FLAGS_ABANDON_ON_ZERO_LENGTH
was supposed to make happen, but the DER length writing code didn't
look at that flag. Now it does.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/11703)
Shane Lontis [Fri, 1 May 2020 05:15:02 +0000 (15:15 +1000)]
Add solaris assembler fixes for legacy provider
The legacy provider contains assembler references. Most code is automagically pulled in from the libcrypto - but the platform specific assembler functions will not be visible in the symbol table. Copying BNASM and DESASM into liblegacy seems to be a better solution than exposing platform specific function in libcrypto.num.
Added a missing call in the des_cbc code for sparc.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11697)
Benjamin Kaduk [Tue, 17 Mar 2020 17:08:11 +0000 (10:08 -0700)]
Add test for SSL_new_session_ticket()
Run a normal handshake and then request some extra tickets,
checking that the new_session_cb is called the expected number of
times. Since the tickets are generated in the same way as other
tickets, there should not be a need to verify that these specific ones
can be used to resume.
Run the test with both zero and a non-zero number of tickets issued in the
initial handshake.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11416)
Benjamin Kaduk [Mon, 16 Mar 2020 18:25:58 +0000 (11:25 -0700)]
Add SSL_new_session_ticket() API
This API requests that the TLS stack generate a (TLS 1.3)
NewSessionTicket message the next time it is safe to do so (i.e., we do
not have other data pending write, which could be mid-record). For
efficiency, defer actually generating/writing the ticket until there
is other data to write, to avoid producing server-to-client traffic when
not needed.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/11416)