OpenBSD src changes summary for 2016-12-06
==========================================

distrib/sets                            distrib/sgi
etc/netstart                            lib/libc
lib/libcrypto                           lib/libssl
regress/sbin                            sys/arch/mips64/include
sys/dev/pci                             sys/dev/pv
sys/net                                 sys/net80211
usr.bin/ssh                             usr.sbin/syspatch

== distrib =========================================================== 01/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib

sets

  ~ lists/comp/mi                         

  > sync (deraadt@)

sgi

  ~ ramdisk/Makefile                      

  > Use makefs to build bsd.rd; build & boot tested by visa (natano@)

== etc =============================================================== 02/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/etc

netstart

  ~ netstart                              

  > Do not lose the default route when netstart(8) is run a second time on
  > the interface pointed to by the default route.
  > Since the kernel no longer keep routes with dangling address pointer,
  > netstart(8) has to re-add the default route when the corresponding ifa
  > has been deleted and re-created.
  > deraadt@ points out that even if the previous semantic was not necessarily
  > better, a script like netstart(8) cannot totally fix the default route
  > problem.
  > Regression reported by and fix tested by Hrvoje Popovski.
  > ksh foo checked by halex@ (mpi@)

== lib =============================================================== 03/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib

libc

  ~ net/linkaddr.c                        

  > CVE-2016-6559: fix potential buffer overflow(s) in link_ntoa(3).
  > A specially crafted struct sockaddr_dl argument can trigger a stack
  > overflow of a static buffer in libc.  An attacker may be able to
  > use this to write to arbitrary locations in the data segment.
  > From FreeBSD (glebius); OK deraadt@ mestre@ (millert@)

libcrypto

  ~ man/BIO_f_buffer.3                    ~ man/BIO_s_bio.3
  ~ man/BUF_MEM_new.3                     ~ man/crypto.3
  ~ man/lh_stats.3                        

  > replace all references to BIO(3) with references to BIO_new(3) (schwarze@)

  - man/BIO.3                             ~ man/BIO_new.3
  ~ man/BIO_push.3                        ~ man/Makefile

  > We don't want section 3 manual pages with names that do not correspond
  > to functions, so delete the BIO(3) manual page and merge its content
  > into BIO_new(3) and BIO_push(3).
  > Sort the content of BIO_new(3) into a logical order
  > and improve the wording in various ways.
  > Add the required cross references to BIO_push(3). (schwarze@)

  ~ man/BIO_new.3                         

  > tweak previous; (jmc@)

  ~ man/crypto.3                          

  > Add Copyright and license.
  > Mention AES, Chacha20, ECDSA, OCSP, OPENSSL_config.
  > Improve many cross references.
  > Convert the table into some real text and polish some wording.
  > Stop talking about SSL.
  > Drop vague references to undocumented internals (objects, stack, txt_db).
  > Delete verbiage that says nothing or is duplicate.
  > Ultimately, the content of this page should be merged into
  > OPENSSL_init_crypto(3), but we don't have that function yet,
  > so the merge has to wait. (schwarze@)

  ~ man/BIO_ctrl.3                        ~ man/BIO_f_base64.3
  ~ man/BIO_f_cipher.3                    ~ man/BIO_f_md.3
  ~ man/BIO_f_null.3                      ~ man/BIO_find_type.3
  ~ man/BIO_read.3                        ~ man/BIO_s_accept.3
  ~ man/BIO_s_connect.3                   ~ man/BIO_s_fd.3
  ~ man/BIO_s_file.3                      ~ man/BIO_s_mem.3
  ~ man/BIO_s_null.3                      ~ man/BIO_s_socket.3
  ~ man/BIO_set_callback.3                ~ man/BIO_should_retry.3
  ~ man/PEM_read_bio_PrivateKey.3         ~ man/PEM_write_bio_PKCS7_stream.3
  ~ man/d2i_PKCS8PrivateKey_bio.3         ~ man/i2d_PKCS7_bio_stream.3

  > make sure that all the BIO pages point back to the master page BIO_new(3)
  > (schwarze@)

  ~ man/EC_GFp_simple_method.3            ~ man/EC_GROUP_copy.3
  ~ man/EC_GROUP_new.3                    ~ man/EC_KEY_new.3
  ~ man/EC_POINT_add.3                    ~ man/EC_POINT_new.3
  ~ man/OCSP_REQUEST_new.3                ~ man/OCSP_cert_to_id.3
  ~ man/OCSP_request_add1_nonce.3         ~ man/OCSP_resp_find_status.3
  ~ man/OCSP_response_status.3            ~ man/OCSP_sendreq_new.3
  ~ man/d2i_ECPKParameters.3              ~ man/d2i_PrivateKey.3

  > Delete most references to crypto(3).  That page isn't useful enough
  > to be pointed to from random individual pages. (schwarze@)

  ~ man/X509V3_get_d2i.3                  

  > We don't support Certificate Transparency Extensions and
  > we don't define preprocessor constants for the related NIDs,
  > so delete the subsection discussing them. (schwarze@)

  ~ man/evp.3                             

  > Copyright and license (schwarze@)

  ~ man/x509.3                            

  > Copyright and license (schwarze@)

libssl

  ~ man/DTLSv1_listen.3                   ~ man/SSL_accept.3
  ~ man/SSL_connect.3                     ~ man/SSL_do_handshake.3
  ~ man/SSL_get_fd.3                      ~ man/SSL_get_rbio.3
  ~ man/SSL_read.3                        ~ man/SSL_set_bio.3
  ~ man/SSL_set_fd.3                      ~ man/SSL_shutdown.3
  ~ man/SSL_write.3                       

  > replace all references to BIO(3) with references to BIO_new(3) (schwarze@)

  ~ s3_both.c                             ~ s3_clnt.c
  ~ s3_lib.c                              ~ s3_srvr.c
  ~ ssl_locl.h                            

  > Convert certificate handshake message generation to CBB, with some clean
  > up and restructure.
  > This also adds CBB based variants of the ssl3_handshake_msg_{start,finish}
  > functions - for the time being these use a CBB to build the messages, then
  > copy back into the init_buf.
  > ok doug@ (jsing@)

  ~ d1_both.c                             ~ d1_clnt.c
  ~ d1_srvr.c                             ~ ssl_locl.h

  > Now that ssl3_send_{client,server}_certificate() are using the common
  > handshake functions, we can remove more copied code from DTLS. (jsing@)

  ~ s3_clnt.c                             

  > Convert ssl3_send_client_kex_rsa() to CBB.
  > ok doug@ (jsing@)

  ~ man/Makefile                          + man/SSL_SESSION_new.3

  > New manual page SSL_SESSION_new(3) written from scratch.
  > The function prototype is listed in ssl(3) and <openssl/ssl.h>, so
  > it's clearly a public interface, but OpenSSL has no documentation
  > about it whatsoever. (schwarze@)

  ~ man/SSL_SESSION_free.3                

  > Add Copyright and license.
  > Garbage collect empty RETURN VALUES section.
  > Delete useless cross reference to ssl(3).
  > Add cross reference to SSL_SESSION_new(3). (schwarze@)

  ~ man/SSL_CTX_get_ex_new_index.3        ~ man/SSL_CTX_set_tmp_rsa_callback.3
  ~ man/SSL_SESSION_get_ex_new_index.3    ~ man/SSL_SESSION_get_time.3
  ~ man/SSL_get_ex_data_X509_STORE_CTX_idx.3
  ~ man/SSL_get_ex_new_index.3            

  > Copyright and license for the remaining manuals in this directory
  > (schwarze@)

  ~ man/SSL_CTX_set_tmp_rsa_callback.3    

  > s/SSL_CTX_need_tmp_rsa/SSL_CTX_need_tmp_RSA/
  > because that's what <openssl/ssl.h> #defines.
  > That's likely a typo in the header file because all the other
  > functions are called *tmp_rsa*().  But it would be a bad idea to
  > fix such a bug in interfaces that are only provided for backward
  > compatibility in the first place, so i'm adjusting the manual to
  > be bug-compatible with the code, for now.
  > But, pretty please, for the next major bump, somebody go get Bob's
  > flensing knife and excise this part of the interface.  Like, export
  > ciphers?  Really? (schwarze@)

  ~ man/Makefile                          + man/SSL_SESSION_print.3

  > An SSL_SESSION_print(3) manual written from scratch.  It's clearly
  > a public interface since it's listed both in ssl(3) and in
  > <openssl/ssl.h>.  Nothing to pilfer from OpenSSL in this case...
  > (schwarze@)

== regress =========================================================== 04/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress

sbin

  ~ route/rttest14.ok                     ~ route/rttest16.ok

  > Update to match recent changes.
  > IPv6 multicast "All Nodes Address" and "All Routers Address" routes are
  > now attached to the corresponding interface.
  > ok bluhm@ (mpi@)

== sys =============================================================== 05/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys

arch/mips64/include

  ~ pmap.h                                

  > Get PTE shift using PTE_LOG. (visa@)

dev/pci

  ~ if_iwm.c                              

  > Make sure iwm(4) only calls ieee80211_mira_choose() while in RUN state.
  > Should fix "panic: timeout_add: not initialized" reported by danj@ (stsp@)

  ~ ixgbe_x550.c                          

  > Let X552 SFP+ boot without a module plugged in
  > Due to insufficient error handling X552 SFP+ card isn't configured
  > properly on boot when an SFP+ module is not plugged in.  As a side
  > effect the interface becomes completely unusable afterwards.
  > The issue was discovered and fix tested by Hrvoje Popovski, thanks!
  > (mikeb@)

  ~ if_ix.c                               

  > Improve error handling and don't fail if SFP module is not present;
  > tested by Hrvoje Popovski, thanks! (mikeb@)

  ~ ixgbe_x550.c                          

  > Don't forget to return an error (mikeb@)

  ~ pcidevs                               

  > X557 is the name of an external PHY not the PCI device
  > Rename 8086:15a[abce] back to X552, give 8086:15ad a proper moniker
  > X552/X557-AT and the yet-unreleased parts 8086:15c* are tentatively
  > named X553. (mikeb@)

  ~ pcidevs.h                             ~ pcidevs_data.h

  > regen (mikeb@)

dev/pv

  ~ pvbus.c                               

  > The kvop is not mandatory to query the type, fixes hostctl -t on vmm(4)
  > VMs.
  > OK mikeb@ (reyk@)

net

  ~ pf.c                                  

  > Use the sizeof udphdr not the sizeof a pointer to it in the af-to
  > specific part of pf_test_state_icmp().  This worked by accident on LP64
  > archs as the struct is eight bytes long.
  > ok mikeb@ bluhm@ krw@ jca@ (jsg@)

net80211

  ~ ieee80211_mira.c                      

  > avoid a potentially uninitialised return value
  > ok stsp@ (jsg@)

== usr.bin =========================================================== 06/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin

ssh

  ~ ssh.c                                 

  > make IdentityFile successfully load and use certificates that have no
  > corresponding bare public key. E.g. just a private id_rsa and certificate
  > id_rsa-cert.pub (and no id_rsa.pub).
  > bz#2617 ok dtucker@ (djm@)

== usr.sbin ========================================================== 07/07 ==

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin

syspatch

  ~ syspatch.sh                           

  > Tweak comment. (ajacoutot@)

  ~ syspatch.sh                           

  > Catch mtree(8) error; it's only run when applying patches and we already
  > check for read-only... in this case. (ajacoutot@)

  ~ syspatch.sh                           

  > Better matching and remove a useless use of sort(1). (ajacoutot@)

  ~ syspatch.sh                           

  > Missed in previous. (ajacoutot@)

===============================================================================
_______________________________________________
odc mailing list
[email protected]
http://www.squish.net/mailman/listinfo/odc

Reply via email to