picker() is type agnostic, but its output consumer is not. Or rather
it doesn't work if picker() picks nothing when consumer expects
array. So ensure array is returned when array is expected.
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5770)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5770)

The VMS C RTL has setbuf() working for short pointers only, probably
the FILE pointer will always be in P0 (the lower 4GB).  Fortunately,
this only generates a warning about possible data loss (doesn't apply
in this case) that we can simply turn off.
Reviewed-by: NTim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5789)

Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5776)

Instead of invoking the fuzz test programs once for every corpora
file, we invoke them once for each directory of corpora files.  This
dramatically reduces the number of program invikations, as well as the
time 99-test_fuzz.t takes to complete.

fuzz/test-corpus.c was enhanced to handle directories as well as
regular files.
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5776)

This wasn't a good solution, too many things depend on the quotes being
there consistently.

This reverts commit 49cd47ea.

Fixes #5772
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5773)

... to compute s390x aes function code from keylength.
Signed-off-by: NPatrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5250)

Signed-off-by: NPatrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5250)

Signed-off-by: NPatrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5250)

Signed-off-by: NPatrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5250)

Signed-off-by: NPatrick Steuer <patrick.steuer@de.ibm.com>
Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5250)

Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5780)

This commit removes the contribution of a user that we cannot trace to
gain their consent for the licence change.

After this commit the various IS_*() macros in the auto-generated file
conf_def.h may incorrectly return true if the supplied character has its
most significant bit set. The IS_*() macros should be able to correctly
handle 8-bit characters. Note that UTF-8 support is not a requirement.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5767)

If a server has been configured to use an ECDSA certificate, we should
allow it regardless of whether the server's own supported groups list
includes the certificate's group.

Fixes #2033
Reviewed-by: NRich Salz <rsalz@openssl.org>
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5601)

This reverts commit 37a38595.

These tests should now be fixed by commit e6e9170d.
Reviewed-by: NBen Kaduk <kaduk@mit.edu>
(Merged from https://github.com/openssl/openssl/pull/5765)

Based on the description in https://github.com/openssl/openssl/pull/5757,
this re-implements the "allow NULL to be passed" behavior of a number of
xxx_free routines.  I also fixed up some egregious formatting errors
that were nearby.
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5761)

ts/ts_rsp_sign.c: change to OPENSSL_gmtime.
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5720)

When the "certificate purpose" is checked and KeyUsage extension is present,
either 'digitalSignature' or 'nonRepudiation' is accepted.

Manual page corrected to reflect the above.
Signed-off-by: NKonstantin Shemyak <konstantin@shemyak.com>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5727)

Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5729)

Apparently applications rely on RAND_load_file's ability to work with
non-regular files, customarily with /dev/urandom, so that the ban was
not exactly appropriate.
Reviewed-by: NRich Salz <rsalz@openssl.org>
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: NPaul Dale <paul.dale@oracle.com>
Reviewed-by: NTim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5737)

This is quick-n-dirty ad-hoc solution, the problem asks for more
elegant one...
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5755)

The previous commit causes some tests to hang so we temporarily disable them.
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5757)

This removes some code because we cannot trace the original contributor
to get their agreement for the licence change (original commit e03ddfae).

After this change there will be numerous failures in the test cases until
someone rewrites the missing code.

All *_free functions should accept a NULL parameter. After this change
the following *_free functions will fail if a NULL parameter is passed:

BIO_ACCEPT_free()
BIO_CONNECT_free()
BN_BLINDING_free()
BN_CTX_free()
BN_MONT_CTX_free()
BN_RECP_CTX_free()
BUF_MEM_free()
COMP_CTX_free()
ERR_STATE_free()
TXT_DB_free()
X509_STORE_free()
ssl3_free()
ssl_cert_free()
SSL_SESSION_free()
SSL_free()

[skip ci]
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5757)

The alpn_selected value in the session should be NULL before we first
populate it if this is a new session. We assert to make sure it is.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5715)

If we don't have OID data for an object then we should fail if we
are asked to encode the ASN.1 for that OID.

Fixes #5723
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5725)

Reviewed-by: NRichard Levitte <levitte@openssl.org>

Reviewed-by: NMatt Caswell <matt@openssl.org>

Comparison was effectively reduced to least significant bits.

CVE-2018-0733
Reviewed-by: NMatt Caswell <matt@openssl.org>

Reviewed-by: NRich Salz <rsalz@openssl.org>

Constructed types with a recursive definition (such as can be found in
PKCS7) could eventually exceed the stack given malicious input with
excessive recursion. Therefore we limit the stack depth.

CVE-2018-0739

Credit to OSSFuzz for finding this issue.
Reviewed-by: NRich Salz <rsalz@openssl.org>

Reviewed-by: NMatt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5756)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>

openssl x509 -in server.pem -signkey privkey.pem -out server.pem
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5746)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5744)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5741)

Reviewed-by: NRich Salz <rsalz@openssl.org>
Reviewed-by: NAndy Polyakov <appro@openssl.org>
Reviewed-by: NBernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/5675)

'nmake clean' was leaving some artefacts behind.
Reviewed-by: NRichard Levitte <levitte@openssl.org>

myDSO_dsobyaddr and myDSO_free are only used in a narrow block of
code, and can therefore be made local to that block.  Otherwise, some
compilers may warn that they are unused.
Reviewed-by: NAndy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5733)

These errors were hidden because compiling this file didn't get the
macros derived from the dso_scheme attribute, and therefore, some code
never got compiled.
Reviewed-by: NAndy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5733)

Reviewed-by: NAndy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5733)