Ensure all EVP calls have their returns checked where appropriate

There are lots of calls to EVP functions from within libssl There were
various places where we should probably check the return value but don't.
This adds these checks.

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

Use better defaults for TSA.

Use SHA256 for TSA and setted permitted digests to a sensible value.

Based on PR#4141

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

Add support for signer_digest option in TS.

Based on PR#2145

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

Make GOST ciphersuites require TLSv1

PR#4141

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

aes/asm/vpaes-ppc.pl: eliminate overhung stores in misaligned cases.

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

e_os.h: bump minimal _WIN32_WINNT.

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

Rename RSA_eay_xxx to rsa_ossl_xxx

Final part of flushing out SSLEay API's.

Reviewed-by: Tim Hudson <tjh@openssl.org>

make update

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

Check error return from sysconf in secure memory code

We use the sysconf function to provide details about the page size in the
secure memory code. This function can return -1 on error so we should check
for this before proceeding.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Add comment explaining why we don't check a return value

A call to X509_verify_cert() is used to build a chain of certs for the
server to send back to the client. It isn't *actually* used for verifying
the cert at all - just building the chain. Therefore the return value is
ignored.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Remove an NULL ptr deref in an error path

The |passwd| variable in the code can be NULL if it goes to the err label.
Therefore we cannot call strlen on it without first checking that it is non
NULL.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

bn/asm/ppc64-mont.pl: adapt for little-endian.

The problem remained unnoticed so far, because it's never called by default.
You have to craft OPENSSL_ppccap environment variable to trigger the problem.

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

crypto/sec_mem.c: fix anonymous mmap on legacy systems.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

bn/asm/s390x.S: improve performance on z196 and z13 by up to 26%. [even z10 is couple percent faster]. Triggered by RT#4128, but solves the problem by real modulo-scheduling.

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

aesni-sha256-x86_64.pl: fix crash on AMD Jaguar.

It was also found that stich performs suboptimally on AMD Jaguar, hence
execution is limited to XOP-capable and Intel processors.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Explicit OpenSSL_version_num() implementation prototype

Reviewed-by: Tim Hudson <tjh@openssl.org>

Document new functions

Reviewed-by: Tim Hudson <tjh@openssl.org>

Use accessors for X509_print_ex().

Print certificate details using accessor functions.

Since X509_CERT_AUX_print is only used in one place and can't
be used by applications (it uses an internal X509_CERT_AUX structure)
this has been removed and replaced by a function X509_aux_print which
takes an X509 pointer instead.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Add new X509 accessors

Reviewed-by: Tim Hudson <tjh@openssl.org>

Update and clarify ciphers documentation.

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

add -psk option to ciphers command

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

add -tls1_2,-tls1_1 options to ciphers command

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

Add "TLSv1.0" cipher alias.

This adds a TLSv1.0 cipher alias for ciphersuites requiring
at least TLSv1.0: currently only PSK ciphersuites using SHA256
or SHA384 MAC (SSLv3 only supports SHA1 and MD5 MAC).

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

Don't alow TLS v1.0 ciphersuites for SSLv3

This disables some ciphersuites which aren't supported in SSL v3:
specifically PSK ciphersuites which use SHA256 or SHA384 for the MAC.

Thanks to the Open Crypto Audit Project for identifying this issue.

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

Use SSL_TLSV1 only if at least TLS v1.0 is needed.

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

RT2667: Add IRC support to -starttls

Reviewed-by: Tim Hudson <tjh@openssl.org>

RT32671: wrong multiple errs TS_check_status_info

Reviewed-by: Tim Hudson <tjh@openssl.org>
Signed-off-by: Rich Salz <rsalz@openssl.org>

absent identity hint should be NULL

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

modes/asm/ghash-armv4.pl: extend Apple fix to all clang cases.
Triggered by RT#3989.

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

Fix SSL_use_certificate_chain_file

The new function SSL_use_certificate_chain_file was always crashing in
the internal function use_certificate_chain_file because it would pass a
NULL value for SSL_CTX *, but use_certificate_chain_file would
unconditionally try to dereference it.

Reviewed-by: Stephen Henson <steve@openssl.org>

Remove redundant check from tls1_get_curvelist

The function tls1_get_curvelist() has an explicit check to see if s->cert
is NULL or not. However the check appears *after* calling the tls1_suiteb
macro which derefs s->cert. In reality s->cert can never be NULL because
it is created in SSL_new(). If the malloc fails then the SSL_new call fails
and no SSL object is created.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Continue malloc standardisation in engines

Continuing from previous work standardise use of malloc in the engine code.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Continue standardisation of malloc handling in apps

continue on from previous commits but in the apps directory

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Continue standardising malloc style for libcrypto

Continuing from previous commit ensure our style is consistent for malloc
return checks.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Standardise our style for checking malloc failures

if we have a malloc |x = OPENSSL_malloc(...)| sometimes we check |x|
for NULL and sometimes we treat it as a boolean |if(!x) ...|. Standardise
the approach in libssl.

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Remove redundant check from SSL_shutdown

The SSL object was being deref'd and then there was a later redundant check
to see if it is NULL. We assume all SSL_foo functions pass a non NULL SSL
object and do not check it.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Make the match for previous cflags a bit more strict

./Configure [target] --strict-warnings -Wno-pedantic-ms-format
would not add '-pedantic' because it matches '-Wno-pedantic-ms-format',
which was added first.

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

Replace L suffix with U

Reviewed-by: Andy Polyakov <appro@openssl.org>

Use uint32_t and int32_t for SSL_CIPHER structure.

Reviewed-by: Andy Polyakov <appro@openssl.org>

Fix compilation problems with SCTP

The SCTP code is not compiled by default. This fixes some compilation
problems in that code.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Remove some redundant assignments

We were setting |s->renegotiate| and |s->new_session| to 0 twice in
tls_finish_handshake. This is redundant so now we just do it once!

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Don't finish the handshake twice

We finish the handshake when we move into the TLS_ST_OK state. At various
points we were also unnecessarily finishing it when we were reading/writing
the Finished message. It's much simpler just to do it in TLS_ST_OK, so
remove the other calls.

Reviewed-by: Viktor Dukhovni <viktor@openssl.org>

Rebuild error source files.

Rebuild error source files: the new mkerr.pl functionality will now
pick up and translate static function names properly.

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

fix discrepancy

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

Read function names from C source files.

In mkerr.pl read parse functions names in C source files and use
them for translation and sanity checks.

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

Ensure the dtls1_get_*_methods work with DTLS_ANY_VERSION

The various dtls1_get*_methods did not handle the DTLS_ANY_VERSION case,
so this needed to be added.

Reviewed-by: Tim Hudson <tjh@openssl.org>

Minor EVP_SignInit_ex doc fix

EVP_SignInit_ex was missing from the NAME section of its man page so
typing "man EVP_SignInit_ex" failed to load the page.

Reviewed-by: Stephen Henson <steve@openssl.org>

Fix usage of BIO_ctrl_set_connected

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

Remove the state parameter from BIO_ctrl_set_connected

The actual implementation has the state of the connection being
controlled with the peer parameter, non-NULL meaning connected and
NULL meaning connected.

Reviewed-by: Andy Polyakov <appro@openssl.org>

BIO_s_datagram() ctrl doesn't support SEEK/TELL, so don't pretend it does

Reviewed-by: Andy Polyakov <appro@openssl.org>

Correct or add comments indicating what controls belong to what

Reviewed-by: Andy Polyakov <appro@openssl.org>

Remove PROXY controls that aren't used anywhere

Reviewed-by: Andy Polyakov <appro@openssl.org>

Remove dummy argument from BIO_get_bind_mode

Reviewed-by: Andy Polyakov <appro@openssl.org>

Document how BIO_get_conn_ip and BIO_get_conn_int_port actually work

No dummy arguments.

Reviewed-by: Andy Polyakov <appro@openssl.org>

Have BIO_get_conn_int_port use BIO_ctrl instead BIO_int_ctrl

BIO_int_ctrl isn't made for the purpose BIO_get_conn_int_port used it
for.

This also changes BIO_C_GET_CONNECT to actually return the port
instead of assigning it to a pointer that was never returned back to
the caller.

Reviewed-by: Andy Polyakov <appro@openssl.org>

Clarify the preferred way of creating patch files

Clarify that git format-patch output is preferred for creating patch files.

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

Remove a reachable assert from ssl3_write_bytes

A buggy application that call SSL_write with a different length after a
NBIO event could cause an OPENSSL_assert to be reached. The assert is not
actually necessary because there was an explicit check a little further
down that would catch this scenario. Therefore remove the assert an move
the check a little higher up.

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

Remove an OPENSSL_assert which could fail

An OPENSSL_assert was being used which could fail (e.g. on a malloc
failure).

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

Remove a trivially true OPENSSL_assert

This OPENSSL_assert in (d)tls1_hearbeat is trivially always going to be
true because it is testing the sum of values that have been set as
constants just a few lines above and nothing has changed them. Therefore
remove this.

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

Fixed typo in rsautl.pod

Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>

PR1279: Clean up CONTRIBUTING

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Various README updates

Close GH Issue 69
Close GH PR 457
Some other updates

By Rich Salz, Alessandro Ghedini, Steve Marquess, Collin Anderson

Reviewed-by: Kurt Roeckx <kurt@openssl.org>

Remove des_ver.h; broke build.

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

Replace "SSLeay" in API with OpenSSL

All instances of SSLeay (any combination of case) were replaced with
the case-equivalent OpenSSL.

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

Remove some SSLv2 references

There were a few remaining references to SSLv2 support which are no longer
relevant now that it has been removed.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Add SRP and PSK to disallowed CertificateRequest ciphersuites

There was a discrepancy between what ciphersuites we allowed to send a
CertificateRequest, and what ciphersuites we allowed to receive one. So
add PSK and SRP to the disallowed ones.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove the inline keyword

Some functions were marked as inline in statem_srvr.c where they probably
didn't need to be, so remove it.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove superfluous check

|tls_process_finished| was checking that |peer_finish_md_len| was
non-negative. However neither |tls1_final_finish_mac| or
|ssl3_final_finish_mac| can ever return a negative value, so the check is
superfluous.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Change SSL_state_string return strings to start with a T

Due the rest of the state machine changes it makes sense to change the
SSL_state_string return strings from 3* to T*. They are not SSL3 specific

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Change snprintf to memcpy

There was a few uses of snprintf in the DTLS SCTP code which made more
sense to be a memcpy.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Add ossl_statem prefix to various functions

Add the ossl_statem prefix to various funtions to avoid name clashes.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Convert enums to typedefs

Various enums were introduced as part of the state machine rewrite. As a
matter of style it is preferred for these to be typedefs.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Make dtls1_link_min_mtu static

The function dtls1_link_min_mtu() was only used within d1_lib.c so make
it static.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix a bogus clang warning

Clang with --strict-warnings was complaining about an uninitalised
variable. In reality it will never be used uninitialised but clang can't
figure out the logic, so just init it anyway to silence the warning.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix empty NewSessionTicket processing

Rebasing the state machine code introduced a problem with empty
NewSessionTicket processing. The return value from the
tls_process_new_session_ticket() is supposed to be an enum, but a bare
integer was being used. Unfortunately this is valid C so the compiler
doesn't pick it up.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix uninitialised variable

Fix another instance of |al| being unitialised in certain error scenarios.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

make update

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix various error codes

A number of error codes were wrong due to a rebase of the state machine
code.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Move in_handshake into STATEM

The SSL variable |in_handshake| seems misplaced. It would be better to have
it in the STATEM structure.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix unitialised variable problem

tls_process_client_hello() failed to initialise the |al| variable in some
(error) scenarios. This could cause issues with creating the alert.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix DTLSv1_listen following state machine changes

Adding the new state machine broke the DTLSv1_listen code because
calling SSL_in_before() was erroneously returning true after DTLSv1_listen
had successfully completed. This change ensures that SSL_in_before returns
false.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Add a function to get the info_callback

Remove repeated blocks of checking SSL and then SSL_CTX for the
info_callback.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove the old state defines

Remove all the defines for the old state machines states. Mapping old to new
is probably going to cause more problems than it solves so it is probably
better to just remove them.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Change SUB_STATE_RETURN into a typedef

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Rename STATEM to OSSL_STATEM

Add an OSSL_ prefix onto STATEM

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove SSL_state and SSL_set_state

SSL_state has been replaced by SSL_get_state and SSL_set_state is no longer
supported.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Move |no_cert_verify| into state machine

The |no_cert_verify| should be in the state machine structure not in SSL

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove extraneous parens

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Change statem prefix to ossl_statem

Change various state machine functions to use the prefix ossl_statem
instead.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Don't depend on SSL structure internals

The macros SSL_in_connect_init() and SSL_in_accept_init() inadvertently
depended on SSL structure internals. This fixes it to use public API calls
instead.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Change HANDSHAKE_STATE to OSSL_HANDSHAKE_STATE

Rename the enum HANDSHAKE_STATE to OSSL_HANDSHAKE_STATE to ensure there are
no namespace clashes, and convert it into a typedef.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix some client side transition logic

Fixed some issues in the logic for determining whether an SKE should be
expected or not. In particular only allow an SKE for RSA if its export and
the key size is not allowed. Also fix the ephemeral ciphersuite checks and
add in a missing call to ssl3_check_cert_and_algorithm().

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Minor documentation tweak

Update the return type for SSL_state in the documentation.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Update CHANGES

Update the CHANGES file for the state machine rewrite

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove a call to SSL_set_state from s_server

s_server was (ab)using SSL_set_state to force a renegotiation. This is a
bad way to do things and does not work with the new state machine code, so
we need to do it a different way.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Fix a comment

Fix an out of date reference to old state machine code in a comment

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Remove some unused variables

The next_state variable is no longer needed in the new state machine.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

Add a state machine README

Add some documentation on the thinking behind the state machine.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

make update

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>

More state machine reorg

Move some function definitions around within the state machine to make sure
they are in the correct files. Also create a statem_locl.h header for stuff
entirely local to the state machine code and move various definitions into
it.

Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>