Matt Caswell [Thu, 4 Jul 2019 09:21:53 +0000 (10:21 +0100)]
Prevent the use of RUN_ONCE inside the FIPS module
FIPS module code *always* runs within the scope of an associated
OPENSSL_CTX. When the module is loaded the OPENSSL_CTX gets created, and
when the module is unloaded the OPENSSL_CX gets freed. A module may be
loaded multiple times within the scope of different OPENSSL_CTX objects.
"Global" data should always be stored within the OPENSSL_CTX. In this
way it will always get cleaned up properly when the module is unloaded.
All current code within the FIPS module works this way. To avoid
"accidents" we disabled the RUN_ONCE code inside the FIPS module.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9308)
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9298)
Martin Peylo [Wed, 22 Aug 2018 09:48:22 +0000 (12:48 +0300)]
Adding Test.pm with workaround for Perl abs2rel bug
If SRCTOP != BLDTOP, and SRCTOP is given in relative form, e.g.
"./config ../openssl", then a bug in Perl's abs2rel may trigger that directory-
rewriting in __cwd results in wrong entries in %directories under certain
circumstances, e.g. when a test executes run(app(["openssl"]) after indir.
There should not be any need to go to a higher directory from BLDDIR or SRCDIR,
so it should be OK to use them in their absolute form, also resolving all
possible symlinks, right from the start.
Following the File::Spec::Functions bug description (reported to perl.org):
When abs2rel gets a path argument with ..s that are crossing over the ..s
trailing the base argument, the result is wrong.
Example
PATH: /home/goal/test/..
BASE: /home/goal/test/../../base
Good result: ../goal
Bad result: ../..
Bug verified with File::Spec versions
- 3.6301
- 3.74 (latest)
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7031)
Matt Caswell [Wed, 19 Jun 2019 14:20:03 +0000 (15:20 +0100)]
Add a dummy call to BN_rand_ex() in the FIPS provider
The previous commit made BIGNUM RAND operations available from within
the FIPS provider. We test this out by making a dummy call to check it
completes successfully.
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9193)
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9271)
Richard Levitte [Tue, 18 Jun 2019 09:39:13 +0000 (11:39 +0200)]
FIPS module: adapt for the changed error reporting methods
The FIPS module inner provider doesn't need to deal with error reason
strings or error library number, since it uses the outer provider's
error reporting upcalls. We therefore disable that code in
crypto/provider_core.c when building the FIPS module.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9174)
Richard Levitte [Tue, 18 Jun 2019 09:18:31 +0000 (11:18 +0200)]
Replumbing: re-implement error reporting for providers
The idea is that providers should only have to report a reason code.
The library code is considered to be libcrypto internal, and are
allocated dynamically and automatically for providers on creation.
We reserve the upper 8 bits of the reason code for internal OpenSSL
use. This allows our own providers to report errors in form of a
packed number that includes library number, function number and
reason number.
With this, a provider can potentially use any reason number it wants
from 1 to 16777216, although the current error semantics really only
allow 1 to 4095 (because only the lower 12 bits are currently
considered an actual reason code by the ERR subsystem).
A provider can provide a reason string table in form of an array of
ERR_STRING_DATA, with each item containing just the reason code and
the associated string, with the dispatch function numbered
OSSL_FUNC_PROVIDER_GET_REASON_STRINGS matching the type
OSSL_provider_get_reason_strings_fn.
If available, libcrypto will call that function on provider
activation.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9174)
Richard Levitte [Tue, 2 Jul 2019 12:57:36 +0000 (14:57 +0200)]
ossl_provider_upref to ossl_provider_up_ref
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)
Antoine Cœur [Tue, 2 Jul 2019 08:04:04 +0000 (16:04 +0800)]
Fix Typos
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)
Patrick Steuer [Mon, 1 Jul 2019 16:09:16 +0000 (18:09 +0200)]
test/p_test.c: silence -Wstringop-overflow
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/9282)
Antoine Cœur [Sun, 30 Jun 2019 18:02:06 +0000 (02:02 +0800)]
Fix Typos
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)
Patrick Steuer [Wed, 26 Jun 2019 21:41:35 +0000 (23:41 +0200)]
s390x assembly pack: update OPENSSL_s390xcap(3)
Add description of capability vector's pcc and kma parts.
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9258)
Patrick Steuer [Wed, 26 Jun 2019 23:07:54 +0000 (01:07 +0200)]
s390x assembly pack: add support for pcc and kma instructions
Signed-off-by: Patrick Steuer <patrick.steuer@de.ibm.com> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9258)
Matt Caswell [Thu, 20 Jun 2019 10:48:50 +0000 (11:48 +0100)]
Add a nid 2 algorithm name mapping capability
Providers that link against libcrypto can just use OBJ_nid2sn() to look
up the name of an algorithm given a NID. However that doesn't work for the
FIPS provider because OBJ_nid2sn() is not available there (due to the
reliance of the code on ASN.1 types). Therefore we provider a new function
to do this mapping. For providers linking against libcrypto the new function
just wraps OBJ_nid2sn(). For the FIPS provider it has a look up for all the
NIDs known there.
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9035)
Matt Caswell [Tue, 28 May 2019 10:26:17 +0000 (11:26 +0100)]
Call RAND_DRBG_bytes from inside the FIPS provider
Insert a dummy call to RAND_DRBG_bytes from inside the FIPS provider to
demonstrate that it is possible to use the RAND code from inside the
module. This is temporary and will be removed once real uses of the RAND
code are available inside the module.
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com>
(Merged from https://github.com/openssl/openssl/pull/9035)
Matt Caswell [Mon, 24 Jun 2019 16:34:14 +0000 (17:34 +0100)]
Add documentation for EVP_CIPHER_fetch
We extend the EVP_MD_fetch documentation to be more generic and to also
cover EVP_CIPHER_fetch. We expect this to be further expanded with other
"fetch" functions in the future.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9233)
man: clarify the 'random number generator must be seeded' requirement
The manual pages require for some API functions that the 'random number
generator must be seeded' before calling the function. Initially, this
was meant literally, i.e. the OpenSSL CSPRNG had to be seeded manually
before calling these functions.
Since version 1.1.1, the CSPRNG is seeded automatically on first use,
so it's not the responsibility of the programmer anymore. Still, he
needs to be aware that the seeding might fail.
Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9257)
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9260)
Matt Caswell [Mon, 24 Jun 2019 15:07:30 +0000 (16:07 +0100)]
Move the public SIV mode functions from public headers to internal ones
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)
Benjamin Kaduk [Thu, 13 Jun 2019 19:26:12 +0000 (12:26 -0700)]
Move 'shared_sigalgs' from cert_st to ssl_st
It was only ever in cert_st because ssl_st was a public structure
and could not be modified without breaking the API. However, both
structures are now opaque, and thus we can freely change their layout
without breaking applications. In this case, keeping the shared
sigalgs in the SSL object prevents complications wherein they would
inadvertently get cleared during SSL_set_SSL_CTX() (e.g., as run
during a cert_cb).
Fixes #9099
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9157)
While this change did prevent the sigalgs from getting inadvertently
clobbered by SSL_set_SSL_CTX(), it also caused the sigalgs to not be
set when the cert_cb runs. This, in turn, caused significant breakage,
such as SSL_check_chain() failing to find any valid chain. An alternate
approach to fixing the issue from #7244 will follow.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9157)
Benjamin Kaduk [Thu, 13 Jun 2019 19:02:03 +0000 (12:02 -0700)]
Add regression test for #9099
Augment the cert_cb sslapitest to include a run that uses
SSL_check_chain() to inspect the certificate prior to installing
it on the SSL object. If the check shows the certificate as not
valid in that context, we do not install a certificate at all, so
the handshake will fail later on in processing (tls_choose_sigalg()),
exposing the indicated regression.
Currently it fails, since we have not yet set the shared sigalgs
by the time the cert_cb runs.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9157)
Keep the externally visible macro BIO_FLAGS_UPLINK in bio.h
and rename the internally used macro to BIO_FLAGS_UPLINK_INTERNAL.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7307)
Pauli [Mon, 24 Jun 2019 15:37:07 +0000 (01:37 +1000)]
The params_test uses wrong size for BN check.
Before the return size was included in the OSSL_PARAM structure, the size
global was updated via a pointer. The size global variable has been removed
and the in structure return size used instead.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9248)
Shane Lontis [Tue, 4 Jun 2019 22:24:16 +0000 (08:24 +1000)]
EC only uses approved curves in FIPS mode.
Once there are buildable fips tests, some tests that are data driven
from files will need to be modified to exclude non approved curves in
fips mode.
These changes were tested by temporarily adding #define FIPS_MODE 1 to
all the modified source files.
Reviewed-by: Nicola Tuveri <nic.tuv@gmail.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9081)
Patrick Steuer [Mon, 24 Jun 2019 22:10:20 +0000 (00:10 +0200)]
Fix s390x build errors and warnings
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/9236)
OSSL_TRACE: enhance documentation and fix doc-nit errors
- Add the following macros to the NAME section:
- with synopsis
OSSL_TRACE_CANCEL, OSSL_TRACE, OSSL_TRACE_ENABLED
- without synopsis
OSSL_TRACEV (helper macro, not intended for public use)
OSSL_TRACE[3-8] (omitted on purpose)
- Revise the NOTES section
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9224)
Reviewed-by: Matthias St. Pierre <Matthias.St.Pierre@ncp-e.com> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9228)
Bernd Edlinger [Mon, 17 Jun 2019 14:57:25 +0000 (16:57 +0200)]
Handle CTRL-C as non-redoable abort signal
This is a bit annoying, if for instance "openssl genrsa -aes128"
tries to read a 4+ character size password, but CTRL-C does no longer
work after a RETURN key, since the flag UI_FLAG_REDOABLE is set by
UI_set_result_ex, together with the error "You must type in 4 to 1023 characters".
Thus remove the REDOABLE flag to allow CTRL-C to work.
[extended tests]
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9170)
Pauli [Thu, 20 Jun 2019 23:27:48 +0000 (09:27 +1000)]
Remove OPENSSL_memcmp.
After avoiding OPENSSL_memcmp for EC curve comparison, there are no remaining
uses in the source code. The function is only defined in an internal header
and thus should be safe to remove for 3.0.0.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/9207)
It turned out that the internal trace cleanup handler was added too
late, so it would be executed before OPENSSL_cleanup().
This results in address errors, as the trace code that's executed in
OPENSSL_cleanup() itself tries to reach for data that's been freed at
that point.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9196)
Changed a bad reference: test/Makefile -> test/build.info
Add ../apps/include as include path to include opt.h used by test/testutil.h.
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9154)
Miquel Ruiz [Mon, 22 Oct 2018 21:34:57 +0000 (22:34 +0100)]
Add SSL_shutdown to SSL_get_error's documentation
SSL_shutdown can fail if called during initialization, and in such case, it'll
add an error to the error queue. This adds SSL_shutdown to the list of functions
that should preceed the call to SSL_get_error.
CLA: trivial
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/7465)
Rebecca Cran [Wed, 12 Jun 2019 20:03:36 +0000 (14:03 -0600)]
Fix UEFI build on FreeBSD by not including system headers
CLA: trivial
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9149)
Matt Caswell [Tue, 18 Jun 2019 17:37:38 +0000 (18:37 +0100)]
Provide an ability to deregister thread stop handlers
If a provider gets unloaded then any thread stop handlers that it had
registered will be left hanging. We should clean them up before tearing
down the provider.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9186)
Richard Levitte [Wed, 19 Jun 2019 06:37:19 +0000 (08:37 +0200)]
Add extra error text in the property parser
With properties being specified in all kinds of places, including hard
coded in providers, it's not always easy to figure out exactly what
string was incorrect when the parser would just say something like
'parse failed' with no more details.
Adding extra data to the error, showing exactly what string is
incorrect, helps a bit. At the very least, this gives anyone
interested something to grep for.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9190)
Rich Salz [Tue, 18 Jun 2019 16:06:17 +0000 (12:06 -0400)]
Change ERR_add_error_[v]data to append
The "add error data" functions now append to the current error.
Add a test for this.
Cleanup some of the ERR_put functions.
In the FIPS module, always append "(in the FIPS module)" to any errors.
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de> Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9181)
Richard Levitte [Tue, 18 Jun 2019 14:23:06 +0000 (16:23 +0200)]
crypto/modes/build.conf: Fix MODES asm mistakes
The old rule in Configure was that if the asm source had a file name
with 'ghash-' as part of the name, GHASH_ASM should be defined. Since
none of the aarch64 asm files has such a name, that macro shouldn't
have been defined.
Fixes #9173
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9178)
Matt Caswell [Tue, 18 Jun 2019 10:45:26 +0000 (11:45 +0100)]
Following the previous 2 commits also move ecpointformats out of session
The previous 2 commits moved supported groups and ciphers out of the
session object to avoid race conditions. We now also move ecpointformats
for consistency. There does not seem to be a race condition with access
to this data since it is only ever set in a non-resumption handshake.
However, there is no reason for it to be in the session.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9162)
Matt Caswell [Fri, 14 Jun 2019 13:06:55 +0000 (14:06 +0100)]
Fix a race condition in ciphers handling
Similarly to the previous commit we were storing the peer offered list
of ciphers in the session. In practice there is no need for this
information to be avilable from one resumption to the next since this
list is specific to a particular handshake. Since the session object is
supposed to be immutable we should not be updating it once we have decided
to resume. The solution is to remove the session list out of the session
object.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9162)
Matt Caswell [Fri, 14 Jun 2019 11:46:13 +0000 (12:46 +0100)]
Fix a race condition in supported groups handling
In TLSv1.3 the supported groups can be negotiated each time a handshake
occurs, regardless of whether we are resuming or not. We should not store
the supported groups information in the session because session objects
can be shared between multiple threads and we can end up with race
conditions. For most users this won't be seen because, by default, we
use stateless tickets in TLSv1.3 which don't get shared. However if you
use SSL_OP_NO_TICKET (to get stateful tickets in TLSv1.3) then this can
happen.
The answer is to move the supported the supported group information into
the SSL object instead.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/9162)
Bernd Edlinger [Fri, 14 Jun 2019 12:48:37 +0000 (14:48 +0200)]
Fix error handling at openssl_strerror_r
When bufsize == 0, openssl_strerror_r should return 0 (if _GNU_SOURCE is defined),
to be consistent with non-_GNU_SOURCE variants, which exhibit the same behavior.
Fix a few cases, where the return value of openssl_strerror_r was ignored.
Reviewed-by: Paul Dale <paul.dale@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/9163)
Matt Caswell [Mon, 27 May 2019 15:31:27 +0000 (16:31 +0100)]
Tell the FIPS provider about thread stop events
The RAND code needs to know about threads stopping in order to cleanup
local thread data. Therefore we add a callback for libcrypto to tell
providers about such events.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9040)
Matt Caswell [Fri, 24 May 2019 10:45:48 +0000 (11:45 +0100)]
Convert thread stop handling into a publish/subscribe model
In later commits this will allow providers to subscribe to thread stop
events. We will need this in the FIPS module. We also make thread stop
handling OPENSSL_CTX aware (different OPENSSL_CTXs may have different
thread local data that needs cleaning up).
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9040)