summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2015-02-08Move a few typedef up in this file in order to be able to use them in env_md_stMiod Vallat
and get rid of a silly FIXME comment. ok doug@ jsing@
2015-02-08Check memory allocation results in EVP_PBE_alg_add_type().Miod Vallat
ok doug@ jsing@
2015-02-08Check memory allocation results, as well as stack pushes.Miod Vallat
Also fix a memory leak in one of the error paths of SMIME_read_ASN1(), spotted by doug@ tweaks&ok doug@ jsing@
2015-02-08Rely upon enc_flags rather than the tls version, to upgrade SHA1+MD5 toMiod Vallat
SHA256 in ssl_get_algorithm2(). From OpenSSL HEAD; ok jsing@
2015-02-08Lob a KNF grenade into the ecdsa code.Joel Sing
2015-02-07Don't leak addresses in error messages.Miod Vallat
2015-02-07Remove useless variables and use the values directly.Doug Hogan
From OpenSSL commit 3d47c1d331fdc7574d2275cda1a630ccdb624b08. ok miod@, jsing@
2015-02-07Don't support very old versions of Netscape (is there any other kind?).Doug Hogan
Apparently "very old" Netscape versions illegally included empty content and a detached signature. OpenSSL removed the #if 0 that protected these users and added a new button OPENSSL_DONT_SUPPORT_OLD_NETSCAPE. It appears to be off by default to keep the hopes and dreams of very old Netscape users alive. We decided to be rebels and disable support. If you installed your browser from floppy disks, it's time to upgrade! Based on OpenSSL commit: 02a938c953b3e1ced71d9a832de1618f907eb96d ok tedu@, miod@, jsing@
2015-02-07Delete a lot of #if 0 code in libressl.Doug Hogan
There are a few instances where #if 1 is removed but the code remains. Based on the following OpenSSL commits. Some of the commits weren't strictly deletions so they are going to be split up into separate commits. 6f91b017bbb7140f816721141ac156d1b828a6b3 3d47c1d331fdc7574d2275cda1a630ccdb624b08 dfb56425b68314b2b57e17c82c1df42e7a015132 c8fa2356a00cbaada8963f739e5570298311a060 f16a64d11f55c01f56baa62ebf1dec7f8fe718cb 9ccc00ef6ea65567622e40c49aca43f2c6d79cdb 02a938c953b3e1ced71d9a832de1618f907eb96d 75d0ebef2aef7a2c77b27575b8da898e22f3ccd5 d6fbb194095312f4722c81c9362dbd0de66cb656 6f1a93ad111c7dfe36a09a976c4c009079b19ea1 1a5adcfb5edfe23908b350f8757df405b0f5f71f 8de24b792743d11e1d5a0dcd336a49368750c577 a2b18e657ea1a932d125154f4e13ab2258796d90 8e964419603d2478dfb391c66e7ccb2dcc9776b4 32dfde107636ac9bc62a5b3233fe2a54dbc27008 input + ok jsing@, miod@, tedu@
2015-02-07Convert several of the server side handshake functions to the new handshakeJoel Sing
message handling routines. ok miod@
2015-02-07Crank major for libcrypto since symbols have been removed.Joel Sing
Requested by deraadt@
2015-02-07Don't allow tag number 31 in CBB_add_asn1().Doug Hogan
Tag 31 is invalid for a short form identifier octet (single byte). KNF a little more. Based on BoringSSL commit 5ba305643f55d37a3e45e8388a36d50c1b2d4ff5 ok miod@
2015-02-07Crank libssl major due to the recent changes and removals.Joel Sing
ok beck@ deraadt@ miod@
2015-02-07Clean up the {get,put}_cipher_by_char() implementations. Also useJoel Sing
ssl3_get_cipher_by_value() in other parts of the code where it simplifies things. ok doug@
2015-02-07Only call free in CBB_init().Doug Hogan
CBB_init_fixed() should not call free because it can lead to use after free or double free bugs. The caller should be responsible for creating and destroying the buffer. From BoringSSL commit a84f06fc1eee6ea25ce040675fbad72c532afece miod agrees with the reasoning ok jsing@, beck@
2015-02-07Provide a SSL_CIPHER_get_by_value() function that allows a cipher to beJoel Sing
retrieved via its cipher suite value. A corresponding SSL_CIPHER_by_value() function returns the cipher suite value for a given SSL_CIPHER. These functions should mean that software does not need to resort to put_cipher_by_char()/get_cipher_by_char() in order to locate a cipher. Begrudgingly also provide a SSL_CIPHER_get_by_id() function that locates a cipher via the internal cipher identifier. Unfortunately these have already been leaked outside the library via SSL_CIPHER_by_id() and the various SSL3_CK_* and TLS1_CK_* defines in the ssl3.h/tls1.h headers. ok beck@ miod@
2015-02-07Combine c_allc.c and c_alld.c into c_all.c - there is not much point havingJoel Sing
this split across files, especially when two of them have less code than license text. ok bcook@ beck@ doug@ miod@
2015-02-07Fix typo and ASN.1 tag number range comment in bytestring.h.Doug Hogan
CBS_get_asn1() and CBS_get_any_asn1_element() only support the single byte ASN.1 identifier octets (aka short form tags). Tag number 31 is the start of the multi-byte long form per X.690 section 8.1.2.4. From BoringSSL commit 2683af70e73f116e14db2bca6290fa4a010a2ee4 ok miod@
2015-02-06KNF bytestring files.Doug Hogan
I checked that this doesn't change anything. Compiled with clang using -Wno-pointer-sign -g0 to reduce the differences. Only difference in the asm is due to assert(0) line number changes in bs_cbs.c and bs_cbb.c. miod is ok with the general process.
2015-02-06Remove accidental, commented out code.Doug Hogan
This was to test a patch for upstream.
2015-02-06Unifdef NETSCAPE_HANG_BUG.Joel Sing
If you're still using a buggy version of Netscape from 2000, for HTTPS with client certificates, it is probably a good time to find a new browser. "kill it softly... with napalm and kisses" miod@
2015-02-06Add additional checks to ssl3_send_client_key_exchange() that ensuresJoel Sing
ephemeral keys exist for SSL_kDHE and SSL_kECDHE. This would have prevented CVE-2014-3572. ok doug@
2015-02-06Import BoringSSL's crypto bytestring and crypto bytebuilder APIs.Doug Hogan
This is imported with as few changes as possible for the initial commit. I removed OPENSSL_EXPORT, replaced OPENSSL_malloc() etc with malloc() and changed a few header includes. BoringSSL has this as part of their public API. We're leaving it internal to libssl for now. Based on BoringSSL's CBB/CBS API as of commit c5cc15b4f5b1d6e9b9112cb8d30205a638aa2c54. input + ok jsing@, miod@
2015-02-06Bring back the horrible API that is get_cipher_by_char/put_cipher_by_char.Joel Sing
This API was intended to be an internal only, however like many things in OpenSSL, it is exposed externally and parts of the software ecosystem are now using it since there is no real alternative within the public API. ok doug@, tedu@ and reluctantly miod@
2015-02-06Rename SSL_CTX_use_certificate_chain() to SSL_CTX_use_certificate_chain_mem().Reyk Floeter
As discussed with beck@ jsing@ and others OK beck@
2015-02-05Declare the x509_(mem|file|dir)_lookup symbols as static because theyReyk Floeter
shouldn't be used directly. They aren't part of the API; each module (file, dir, mem) provides an actual function to export the now-static object. OK miod@
2015-01-28Fix a number of issues relating to algorithms in signatures, MostlyBob Beck
from OpenSSL with a hint of boring and some things done here. Addresses CVE-2014-8275 for OpenSSL fully ok miod@ doug@
2015-01-26Add AEAD as a "MAC alias" so that it is possible to identify/select ciphersJoel Sing
that use AEAD instead of a MAC. This allows for TLSv1.2 AEAD ciphers (effectively the only ciphers that are still considered to be secure) to be selected using TLSv1.2+AEAD as a cipher string. ok bcook@ doug@ miod@
2015-01-23Ensure that a ServerKeyExchange message is received if the selected cipherJoel Sing
suite uses ephemeral keys. This avoids an issue where an ECHDE cipher suite can effectively be downgraded to ECDH, if the server omits the ServerKeyExchange message and has provided a certificate with an ECC public key. Issue reported to OpenSSL by Karthikeyan Bhargavan. Based on OpenSSL. Fixes CVE-2014-3572. ok beck@
2015-01-22Use field names in struct initialisers.Joel Sing
No change to generated assembly.
2015-01-22Support CA verification in chroot'ed processes without direct fileReyk Floeter
access to the certificates. SSL_CTX_load_verify_mem() is a frontend to the new X509_STORE_load_mem() function that allows to load the CA chain from a memory buffer that is holding the PEM-encoded files. This function allows to handle the verification in privsep'ed code. Adopted for LibreSSL based on older code from relayd (by pyr@ and myself) With feedback and OK bluhm@
2015-01-22Add X509_STORE_load_mem() to load certificates from a memory bufferReyk Floeter
instead of disk. OpenSSL didn't provide a built-in API from loading certificates in a chroot'ed process that doesn't have direct access to the files. X509_STORE_load_mem() provides a new backend that will be used by libssl and libtls to implement such privsep-friendly functionality. Adopted for LibreSSL based on older code from relayd (by pyr@ and myself) With feedback and OK bluhm@
2015-01-22Assume that the size of a pointer will not change at runtime.Brent Cook
Change the runtime check for whether a long is smaller than a pointer to a compile-time check. Replace the silly hash for LLP64 platforms. ok tedu@
2015-01-21Fix DTLS memory leak (CVE-2015-0206).Doug Hogan
There were four bugs fixed by this patch: * dtls1_buffer_record() now frees rdata->rbuf.buf on error. Since s->s3->rbuf was memset, rdata->rbuf is the only pointer left which points to the old rbuf. On error, rdata is freed so there will not be any way of freeing this memory unless we do it here. * Changed the return code of dtls1_buffer_record() to differentiate between queue full (0) and error (-1). See below as this differs from upstream. * Handle errors if calls to dtls1_buffer_record() fail with -1. Previously, it did not check the return value. * Changed the way receipts are recorded. Previously, it was recorded when processed successfully (whether buffered or not) in dtls1_process_record(). Now, it records when it is handled in dtls1_get_record(): either when it is entered into the queue to buffer for the next epoch or when it is processed directly. Processing buffered records does not add a receipt because it needed one in order to get into the queue. The above bugs combined contributed to an eventual DoS through memory exhaustion. The memory leak came from dtls1_buffer_record()'s error handling. The error handling can be triggered by a duplicate record or malloc failure. It was possible to add duplicate records because they were not being dropped. The faulty receipts logic did not detect replays when dealing with records for the next epoch. Additionally, dtls1_buffer_record()'s return value was not checked so an attacker could send repeated replay records for the next epoch. Reported to OpenSSL by Chris Mueller. Patch based on OpenSSL commit 103b171d8fc282ef435f8de9afbf7782e312961f and BoringSSL commit 44e2709cd65fbd2172b9516c79e56f1875f60300. Our patch matches BoringSSL's commit. OpenSSL returns 0 when the queue is full or when malloc() or pitem_new() fails. They return -1 on error including !ssl3_setup_buffers() which is another failure to allocate memory. BoringSSL and LibreSSL changed the return code for dtls1_buffer_record() to be 1 on success, 0 when the queue is full and -1 on error. input + ok bcook@, jsing@
2015-01-19Add arc4random/getentropy shims for NetBSD.Brent Cook
The latest NetBSD (6.1.5) arc4random does not appear to reseed the CRNG state after a fork, so provide an override until the fork-safe version in CVS appears in a release. These are the same as the FreeBSD shims. ok deraadt@
2015-01-15back in september I did the large abstraction refactoring to allow theseTheo de Raadt
other systems to fit into the same mold, so add copyright
2015-01-12Fix a memory leak in bss_dgram.Doug Hogan
Free data->saved_message.data. Based on OpenSSL commit: 41cd41c4416f545a18ead37e09e437c75fa07c95 except this version sets a->ptr to NULL to avoid accidental reuse and handles malloc failing. ok beck@, input + ok miod@
2015-01-08Avoid a double-free in an error path.Doug Hogan
ok jsing@ beck@
2015-01-07mix in more virtual memory and process informationBrent Cook
2015-01-06add initial HP-UX getentropy/arc4random support.Brent Cook
patch from Kinichiro Inoguchi, tested on HP-UX 11.31 ok deraadt@
2015-01-03Fix incorrect OPENSSL_assert() usage.Doug Hogan
Instead of asserting, return an error code for I/O errors. This is based on OpenSSL commit 2521fcd8527008ceb3e4748f95b0ed4e2d70cfef. Added checks for two calloc()s while I'm here. ok miod@
2014-12-29don't leak timing info about padding errors by generating a fake keyTed Unangst
afterwards. openssl has a more complicated fix, but it's less intrusive for now to simply hoist the expensive part (fake key generation) up without sweating a branch or two. ok bcook jsing
2014-12-16Now that we have Camellia support in libcrypto, bring in the SHA256 flavour ofMiod Vallat
the Camellia ciphersuites for TLS 1.2 introduced in RFC 5932. From OpenSSL HEAD.
2014-12-15Add error handling for EVP_DigestInit_ex().Doug Hogan
A few EVP_DigestInit_ex() calls were left alone since reporting an error would change the public API. Changed internal ssl3_cbc_digest_record() to return a value due to the above change. It will also now set md_out_size=0 on failure. This is based on part of BoringSSL's commit to fix malloc crashes: https://boringssl.googlesource.com/boringssl/+/69a01608f33ab6fe2c3485d94aef1fe9eacf5364 ok miod@
2014-12-14unconditionally align SSL payloadsBrent Cook
Remove support for conditional payload alignment, since we would never want to turn it off. Also, consistently use size_t for calculating the alignment. ok miod@
2014-12-14Convert all of the straight forward client handshake handling code to useJoel Sing
the new handshake functions. ok miod@
2014-12-14Provide functions for starting, finishing and writing SSL handshakeJoel Sing
messages. This will allow for removal of repeated/duplicated code. Additionally, DTLS was written by wholesale copying of the SSL/TLS code, with some DTLS specifics being added to the duplicated code. Since these SSL handshake message functions know how to handle both SSL/TLS and DTLS, upon conversion the duplicate versions will become identical (or close to), at which point the DTLS versions can be removed and the SSL/TLS versions used for both protocols. Partially based on similar changes in OpenSSL. ok miod@
2014-12-14Remove trailing whitespace.Joel Sing
2014-12-14unifdef OPENSSL_NO_NEXTPROTONEG, which is one of the last standing #ifndefJoel Sing
mazes in libssl. NPN is being replaced by ALPN, however it is still going to be around for a while yet. ok miod@
2014-12-14unifdef NETSCAPE_HANG_BUG from the DTLS code. The code is not currentlyJoel Sing
enabled and I would hope that no one is using client certificates with DTLS and Netscape, assuming it even supported it... ok bcook@ miod@
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-01 16:43:52 +0000