Pullup ticket #4471 - requested by obache
security/openssl: security update

Revisions pulled up:
- security/openssl/Makefile                                    1.195
- security/openssl/PLIST.common                                1.19
- security/openssl/distinfo                                    1.108

-------------------------------------------------------------------
  Module Name: pkgsrc
  Committed By: obache
  Date: Thu Aug  7 01:29:45 UTC 2014

  Modified Files:
  pkgsrc/security/openssl: Makefile PLIST.common distinfo

  Log Message:
  Update openssl to 1.0.1i.

    Changes between 1.0.1h and 1.0.1i [6 Aug 2014]

    *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
        SRP code can be overrun an internal buffer. Add sanity check that
        g, A, B < N to SRP code.

        Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
        Group for discovering this issue.
        (CVE-2014-3512)
        [Steve Henson]

    *) A flaw in the OpenSSL SSL/TLS server code causes the server to
  negotiate      TLS 1.0 instead of higher protocol versions when the
  ClientHello message      is badly fragmented. This allows a
  man-in-the-middle attacker to force a      downgrade to TLS 1.0 even if
  both the server and the client support a      higher protocol version, by
  modifying the client's TLS records.

        Thanks to David Benjamin and Adam Langley (Google) for discovering and
        researching this issue.
        (CVE-2014-3511)
        [David Benjamin]

    *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are =
  subject
        to a denial of service attack. A malicious server can crash the client
        with a null pointer dereference (read) by specifying an anonymous
  (EC)DH      ciphersuite and sending carefully crafted handshake messages.

        Thanks to Felix Grvbert (Google) for discovering and researching this
        issue.
        (CVE-2014-3510)
        [Emilia K=E4=B3=B0er]

    *) By sending carefully crafted DTLS packets an attacker could cause
  openssl      to leak memory. This can be exploited through a Denial of
  Service attack.      Thanks to Adam Langley for discovering and researching
  this issue.      (CVE-2014-3507)
        [Adam Langley]

    *) An attacker can force openssl to consume large amounts of memory =
  whilst
        processing DTLS handshake messages. This can be exploited through a
        Denial of Service attack.
        Thanks to Adam Langley for discovering and researching this issue.
        (CVE-2014-3506)
        [Adam Langley]

    *) An attacker can force an error condition which causes openssl to crash
        whilst processing DTLS packets due to memory being freed twice. This
        can be exploited through a Denial of Service attack.
        Thanks to Adam Langley and Wan-Teh Chang for discovering and
  researching      this issue.
        (CVE-2014-3505)
        [Adam Langley]

    *) If a multithreaded client connects to a malicious server using a
  resumed      session and the server sends an ec point format extension it
  could write      up to 255 bytes to freed memory.

        Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching
  this      issue.
        (CVE-2014-3509)
        [Gabor Tyukasz]

    *) A malicious server can crash an OpenSSL client with a null pointer
        dereference (read) by specifying an SRP ciphersuite even though it was
  not      properly negotiated with the client. This can be exploited through
  a      Denial of Service attack.

        Thanks to Joonas Kuorilehto and Riku Hietam=E4=AB=A9 (Codenomicon) for
        discovering and researching this issue.
        (CVE-2014-5139)
        [Steve Henson]

    *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
        X509_name_oneline, X509_name_print_ex et al. to leak some information
        from the stack. Applications may be affected if they echo pretty
  printing      output to the attacker.

        Thanks to Ivan Fratric (Google) for discovering this issue.
        (CVE-2014-3508)
        [Emilia K=E4=B3=B0er, and Steve Henson]

    *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
        for corner cases. (Certain input points at infinity could lead to
        bogus results, with non-infinity inputs mapped to infinity too.)
        [Bodo Moeller]

  To generate a diff of this commit:
  cvs rdiff -u -r1.194 -r1.195 pkgsrc/security/openssl/Makefile
  cvs rdiff -u -r1.18 -r1.19 pkgsrc/security/openssl/PLIST.common
  cvs rdiff -u -r1.107 -r1.108 pkgsrc/security/openssl/distinfo

(spz)