OpenBSD src changes summary for 2016-12-07 ==========================================
distrib/sets lib/libc lib/libssl regress/usr.bin regress/usr.sbin share/man sys/arch/amd64/conf sys/dev/pci sys/dev/pv usr.bin/mandoc usr.bin/tmux usr.sbin/dhcrelay usr.sbin/syspatch == distrib =========================================================== 01/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib sets ~ lists/comp/mi > sync (deraadt@) ~ lists/comp/mi > sync (deraadt@) ~ lists/comp/mi > sync (deraadt@) == lib =============================================================== 02/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libc ~ net/linkaddr.c > Fix a typo, decrement rem, don't increment for single digit hex bytes. > From Henri Kemppainen (millert@) libssl ~ ssl_asn1.c > Ensure that we zero memory used to hold the ASN.1 encoded session, since > this contains the session master key. > ok deraadt@ doug@ (jsing@) ~ s3_srvr.c > Ensure that we zero memory that contiansthe ASN.1 encoded session, since > this contains the session master key. > ok deraadt@ doug@ (jsing@) ~ s3_clnt.c > Convert ssl3_send_client_kex_dhe() to CBB. > ok doug@ (jsing@) ~ man/SSL_load_client_CA_file.3 > Add documentation for SSL_add_file_cert_subjects_to_stack(3) and > SSL_add_dir_cert_subjects_to_stack(3), written from scratch. > Both functions are listed in ssl(3) and <openssl/ssl.h> and recommended > for the use by browsers in source code comments, so they are clearly > public interfaces. > Mention deduplication. > Purge some duplicate text and improve some wording while here. > Two additional cross references instead of the useless ssl(3). > Add HISTORY, AUTHORS, and BUGS. > It is depressing that BUGS (purely from code inspection) became > longer than the DESCRIPTION. (schwarze@) ~ man/Makefile + man/SSL_dup_CA_list.3 > New SSL_dup_CA_list(3) manual written from scratch. > Mentioned in ssl(3) and <openssl/ssl.h>, so it is public. (schwarze@) ~ man/Makefile + man/SSL_dup.3 > Write new SSL_dup(3) manual from scratch. This function is listed > in ssl(3) and <openssl/ssl.h>, so it is clearly public. (schwarze@) ~ man/Makefile + man/SSL_copy_session_id.3 > Write SSL_copy_session_id(3) manual from scratch. This function is > listed in ssl(3) and <openssl/ssl.h>, so it's clearly a public interface. > We might wish to merge the improved code from OpenSSL 1.1.0, > but that's major bump, so i'm documenting the BUGS for now. (schwarze@) ~ man/SSL_read.3 > add missing .Nm SSL_peek in the NAME section (schwarze@) ~ man/Makefile + man/SSL_renegotiate.3 > Write SSL_renegotiate(3) manual from scratch; this function is listed > in ssl(3) and <openssl/ssl.h>, so it's clearly a public interface. > More could probably be said, the code looks somewhat mysterious to me, > but i think this stub is already better than nothing. (schwarze@) == regress =========================================================== 03/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress usr.bin ~ mandoc/roff/args/roff.out_ascii ~ mandoc/roff/args/roff.out_lint > enable commented parts of the test, they seem to work now (schwarze@) ~ mandoc/roff/args/roff.in > enable commented parts of the test, they seem to work now (schwarze@) ~ mandoc/roff/args/man.in ~ mandoc/roff/args/man.out_ascii ~ mandoc/roff/args/man.out_lint > comment out non-working parts of the test with .ig rather than \", > this is not supposed to test whitespace at the end of comments; > noticed by bluhm@ who sent a different patch (schwarze@) ~ mandoc/char/space/Makefile + mandoc/char/space/trailing-mdoc.in + mandoc/char/space/trailing-mdoc.out_ascii + mandoc/char/space/trailing-mdoc.out_lint > test different kinds of the "WARNING: whitespace at end of input line"; > bluhm@ noticed that this area can grow regressions (schwarze@) usr.sbin ~ arp/arptest1.ok ~ arp/arptest2.ok ~ arp/arptest3.ok ~ arp/arptest4.ok ~ arp/arptest5.ok ~ arp/arptest6.ok > Spacing in arp -a output has changed, adapt test. (bluhm@) == share ============================================================= 04/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man + man4/xbf.4 > xbf(4) manual page (mikeb@) ~ man4/pvbus.4 ~ man4/Makefile > Add xbf(4) (mikeb@) ~ man4/ix.4 > Mention X552, don't mention X557 which is a PHY (mikeb@) ~ man4/pvbus.4 ~ man4/xbf.4 > tweak previous; (jmc@) == sys =============================================================== 05/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/amd64/conf ~ GENERIC > Add a commented out entry for xbf(4) > xbf(4) driver creates a shim on top of Xen I/O rings and exports > it as a SCSI adaptor with a single target for a disk or a cdrom. > The disks that are attached to it are managed by the sd(4) driver, > but thankfully OpenBSD users do not require any migration strategy > to move from emulated wd(4) disks to paravirtualized ones since > disklabel UIDs are used by default in /etc/fstab making referencing > particular partitions independent from the underlying device driver. > (mikeb@) dev/pci ~ if_iwn.c ~ if_iwnvar.h > Add mira support to iwn(4). > Please let me know about any regressions. > Tested by myself, tb@, okan@, jca@, naddy@, mlarkin@ > ok tb@ jca@ (stsp@) dev/pv ~ xen.c > Remove some leftovers from before the dynamic grant table code (mikeb@) ~ xenstore.c > Return proper error values from xs_{get,set}prop (mikeb@) ~ xenstore.c > Factor out the polling loop into a separate function (mikeb@) ~ xenstore.c ~ xenvar.h > Add a simple mechanism to poll for a change in the property value (mikeb@) + xbf.c > A driver for Xen Blkfront heavily based on vdsk(4) from kettenis@ (mikeb@) ~ files.pv > Add xbf(4) (mikeb@) ~ xbf.c > splbio isn't doing us any good here and prevents polling from working > (mikeb@) ~ xbf.c > Fall back to the minimum amount of ring pages > Some AWS volumes lack both max-ring-page-order and max-ring-pages > properties so we have to default to a minimum amount of 1 ring page. > Noticed and fix tested by reyk@ on EC2. (mikeb@) ~ xbf.c > Response status field is signed; adjust the debug message (mikeb@) ~ xbf.c > Add required padding to the response descriptor > Xen source code relies on the compiler to pad members of the structure > representing the descriptor layout in memory; we're however trying to > be more defensive and define packed structures. > Figured out the hard way with reyk@. (mikeb@) == usr.bin =========================================================== 06/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin mandoc ~ read.c > When reporting "whitespace at end of input line" on lines ending with > roff(7) comments, let the column number in the message point to the > end of the line rather than to the beginning of the comment. > Improvement suggested by bluhm@. (schwarze@) tmux ~ key-bindings.c ~ server-client.c > Make prefix work in all tables (except the prefix table). Users who want > to avoid this can set prefix to "none" and bind send-prefix > themselves. Allows C-b t be bound in the copy mode tables again, pointed > out by millert@. (nicm@) ~ tty.c > Only skip moving the cursor if it is already in the last position _on > the same line_, fixes redraw bug reported by patrick keshishian. (nicm@) ~ status.c > Do not clear the prompt when a message is shown, just leave it around and > return to it when the message is finished. (nicm@) == usr.sbin ========================================================== 07/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin dhcrelay ~ dhcpd.h ~ dhcrelay.c ~ dispatch.c > Rename function discover_interfaces into get_interface and change its > prototype to be more useful. > ok reyk@ (rzalamena@) ~ dhcrelay.c > pledge(2) dhcrelay with "stdio route". route is needed for interface > ioctls. > OK rzalamena@ (reyk@) ~ dhcrelay.c > If the BROADCAST flag is set on a BOOTREPLY, the RFC specifies that > we SHOULD forward the packet not only as L3 broadcast, but also as > L2 broadcast. Apparently that helps on older machines that can't > handle L2 unicast replies. > ok jca@ (patrick@) ~ dhcrelay.c > DHCP requests can be relayed through multiple relays. Currently we > drop requests that have already been relayed. To allow usage in the > middle of a chain, remove this check and only set giaddr if it has not > been set yet. This giaddr will be used by the DHCP server to identify > which subnet the client is connected to. > RFC 1542 specifies that we should increase the hop counter every time > we relay a request. If we receive a request whose hop counter exceeds > the value of 16 we must silently drop it. > ok jca@ (patrick@) syspatch ~ syspatch.sh > Cleanup syspatch directories that do not contain a tarball; it means it > was not applied properly. (ajacoutot@) =============================================================================== _______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
