summaryrefslogtreecommitdiff
path: root/src (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
* fix a leak reported by Ben L (bobsayshilol () live ! co ! uk)tb2018-11-101-1/+3
|
* Fix a leak reported by Ben L bobsayshilol () live ! co ! uk.tb2018-11-101-1/+3
|
* fix a leak pointed out by Ben L (bobsayshi () live ! co ! uk)tb2018-11-101-4/+8
|
* Avoid a double allocation and memory leak.jsing2018-11-101-4/+2
| | | | Reported by Ben L <bobsayshilol at live dot co dot uk>
* Stop keeping track of sigalgs by guessing it from digest and pkey,beck2018-11-108-92/+102
| | | | | | just keep the sigalg around so we can remember what we actually decided to use. ok jsing@
* More regress all the way to exporter_masterbeck2018-11-101-4/+44
|
* Use TLS13_HS_{CLIENT,SERVER} instead of using a redundant _SEND{,S}.tb2018-11-101-30/+28
| | | | ok jsing
* Fix last of the empty hash nonsensebeck2018-11-102-32/+6
| | | | ok jsing@
* Update key schedule regress to match API changes.jsing2018-11-091-21/+12
|
* Fix the TLSv1.3 key schedule implementation.jsing2018-11-092-66/+95
| | | | | | | | | | | | When the RFC refers to ("") for key derivation, it is referring to the transcript hash of an empty string, not an empty string. Rename tls13_secrets_new() to tls13_secrets_create(), make it take an EVP_MD * and calculate the hash of an empty string so that we have it available for the "derived" and other steps. Merge tls13_secrets_init() into the same function, remove the EVP_MD * from other functions and use the empty string hash at the appropriate places. ok beck@ tb@
* Use "send" and "recv" consistently instead of mixing them with "read"tb2018-11-091-98/+108
| | | | | | and "write". Use self-documenting C99 initializers. ok bcook, jsing
* Initialize priv_key and pub_key on first use instead of at the top.tb2018-11-091-6/+4
| | | | | | While there, eliminate a flag that was only used once. ok beck jsing mestre
* Initialize priv_key and pub_key on first use instead of at the top.tb2018-11-091-4/+4
| | | | ok beck jsing mestre
* The Botan library from ports an be configured to use OpenSSL orbluhm2018-11-092-1/+22
| | | | | | | | LibreSSL as crypto provider. When we run their regression tests, we are actually testing our library. This is far from perfect. A lot of LibreSSL features have not been implemented as Botan provider. Even if provider openssl is specified, botan-test runs a lot of non-openssl tests. This can be improved later.
* Avoid dereferencing eckey before checking it for NULL.tb2018-11-091-5/+6
| | | | | | CID 184282 ok beck jsing mestre
* remove the not yet implemented "handshake" subdirectorytb2018-11-091-2/+1
|
* Add subdirectires with SUBDIR += instead of a single assignment withtb2018-11-093-67/+65
| | | | line continuations.
* Remove ethers(5) YP support bits from libc as it makes it difficult tobrynet2018-11-092-70/+8
| | | | | | | | effectively use pledge(2) in some programs. approval from many, thanks! idea by & ok deraadt@
* Ensure we free the handshake transcript upon session resumption.jsing2018-11-091-1/+4
| | | | | | Found the hard way by jmc@ ok tb@
* The cert subdir is testing all combinations of certificate validation.bluhm2018-11-0911-48/+244
| | | | | | Having the three libraries, client and server certificates, missing or invalid CA or certificates, and enforcing peer certificate results in 1944 new test cases.
* Ensure we only choose sigalgs from our prefernce list, not the whole listbeck2018-11-094-10/+19
| | | | ok jsing@
* Add the ability to have a separate priority list for sigalgs.beck2018-11-094-12/+43
| | | | | Add a priority list for tls 1.2 ok jsing@
* Correct defines for writer tests in connect/accept loops.jsing2018-11-091-3/+3
| | | | ok tb@
* Correct function naming for tls13_handshake_advance_state_machine().jsing2018-11-091-4/+2
| | | | ok tb@
* Avoid leak: free existing SRTP connection profiles beforetb2018-11-091-1/+2
| | | | | | setting it. From Ben L <bobsayshilol () live ! co ! uk>.
* Avoid leaking memory that was already allocated in ASN1_item_new().tb2018-11-091-5/+1
| | | | From Ben L <bobsayshilol () live ! co ! uk>
* Fix a buffer overrun in asn1_parse2().tb2018-11-091-4/+7
| | | | | From Ben L bobsayshilol () live ! co ! uk Similar fixes in BoringSSL and OpensSSL.
* In verbose mode netcat reports to stderr when the listen systembluhm2018-11-091-16/+29
| | | | | | call has finished. This allows to write race free scripts as they can check that the server is up and running. OK sthen@ tb@
* Add const to the data argument of ASN1_set{,_int}_octetstring().tb2018-11-092-6/+6
| | | | | | From Ben L bobsayshilol () live ! co ! uk ok jsing
* Add header guards and hidden declarations.jsing2018-11-092-2/+19
|
* Add header guards and hidden declarations.jsing2018-11-091-1/+10
|
* Reimplement the sigalgs processing code into a new implementationbeck2018-11-099-253/+340
| | | | | that will be usable with TLS 1.3 with less eye bleed. ok jsing@ tb@
* First skeleton of the TLS 1.3 state machine. Based on RFC 8446 andtb2018-11-083-4/+607
| | | | | | inspired by s2n's state machine. Lots of help and input from jsing. ok beck, jsing
* KNFbeck2018-11-081-3/+4
|
* Clean up and simplify the handshake transcript code.jsing2018-11-0810-166/+144
| | | | | | | This provides a cleaner, simpler and more readable API, with code that uses a BUF_MEM instead of a BIO. ok beck@ ("hurry up") and tb@.
* Add missing NULL checks on allocation, style(9) and consistently usejsing2018-11-081-35/+34
| | | | goto err instead of handrolling.
* Add missing BN_free() calls.jsing2018-11-081-1/+3
| | | | From Ben L <bobsayshilol at live dot co dot uk>.
* Use ASN1_TYPE_new()/ASN1_TYPE_free() to avoid leaking memory.jsing2018-11-081-12/+17
| | | | From Ben L <bobsayshilol at live dot co dot uk>.
* Stop pretending that a cert member in a SSL and SSL_CTX can be NULL.jsing2018-11-088-121/+18
| | | | ok beck@ tb@
* Move #include <openssl/evp.h> to the header.tb2018-11-082-3/+4
| | | | discussed with beck and jsing
* Ensure the handshake transcript is cleaned up.jsing2018-11-082-3/+20
| | | | | | | | | Add a check at the completion of the client/server handshake to ensure that the handshake transcript has been freed. Fix the case where a server asks the client for a certificate, but it does not have one, resulting in the handshake transcript being left around post-handshake. ok bcook@ tb@
* Remove some function prototypes that should have been removed in thejsing2018-11-081-10/+2
| | | | | | previous clean up. Spotted by bcook@
* LibreSSL 2.9.0bcook2018-11-081-3/+3
|
* zap last remains of malloc.conf; prompted by and ok jmc@otto2018-11-081-6/+6
|
* Use in-place (un)wrapping in the keywrap tests.tb2018-11-071-5/+8
|
* Add in key_schedule regress tests to regress buildbeck2018-11-071-1/+2
|
* Add a self test for each SSL library by connecting client withbluhm2018-11-075-23/+88
| | | | | | server. Check that the highest available TLS version is selected. LibreSSL TLS 1.3 check is disabled until the feature becomes available.
* Add initial TLS 1.3 key schedule support with basic regress testsbeck2018-11-075-2/+685
| | | | ok jsing@ tb@
* Add interop test with OpenSSL 1.1. TLS 1.3 should be used automaticallybluhm2018-11-075-10/+55
| | | | | when it becomes available in LibreSSL. thanks to sthen@ for the new OpenSSL port
* Use memmove() instead of memcpy() to get rid of the need fortb2018-11-071-3/+3
| | | | | | | | | non-overlapping *in and *out buffers as we're already implementing the "in place (un)wrapping" algorithms as given in RFC 3394. This removes a gratuitous API difference to OpenSSLin these undocumented functions. Found while working on wycheproof regress tests. ok beck jsing
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-28 11:34:29 +0000