path: root/contrib/isc-dhcp/RELNOTES
diff options
authorDavid E. O'Brien <obrien@FreeBSD.org>1999-02-11 12:40:22 +0000
committerDavid E. O'Brien <obrien@FreeBSD.org>1999-02-11 12:40:22 +0000
commitb02401bdd6455578fb36dc5eed002b83d88e4954 (patch)
tree1ab69d6be8758743eebe6650b959ea559178112b /contrib/isc-dhcp/RELNOTES
parentcd2d014aab8332077aba9b0f6e0d06678fbe1aff (diff)
Virgin import of ISC-DHCP v2.0b1pl11vendor/isc-dhcp/2.0b1-pl.11
Notes: svn path=/vendor/isc-dhcp/dist/; revision=43892 svn path=/vendor/isc-dhcp/2.0b1-pl.11/; revision=43894; tag=vendor/isc-dhcp/2.0b1-pl.11
Diffstat (limited to 'contrib/isc-dhcp/RELNOTES')
1 files changed, 217 insertions, 12 deletions
diff --git a/contrib/isc-dhcp/RELNOTES b/contrib/isc-dhcp/RELNOTES
index 2cad409c8986..c136fd7bac69 100644
--- a/contrib/isc-dhcp/RELNOTES
+++ b/contrib/isc-dhcp/RELNOTES
@@ -1,7 +1,7 @@
Internet Software Consortium
Dynamic Host Configuration Protocol Distribution
- Version 2, Beta 1, Patchlevel 6
- June 26, 1998
+ Version 2, Beta 1, Patchlevel 10
+ February 8, 1998
Release Notes
@@ -13,8 +13,8 @@ stable.
Version 1 of the ISC DHCP Distribution includes just a DHCP Server.
Version 1 has been in feature freeze since late 1996, and is quite
-stable. This is the release that we would expect most sites to run in
+stable. This is the release that we would expect very conservative
+sites to run in production, but it is no longer recommended.
Version 2 of the ISC DHCP Distribution adds a DHCP Client and a
DHCP/BOOTP Relay Agent to the DHCP Server that was offered in version
@@ -33,19 +33,224 @@ server:
addresses other than the one the server knows they should be
using are disciplined quickly.
-This version is now in Beta testing, and is planned for release in
-mid-1998. It has a number of new features, and is the release that we
-would expect sites that want some stability but need the new lease
-testing feature, or need a client or relay agent. Note that it is
-possible to run the Version 1 server with the Version 2 client.
+This version has been in a near feature freeze since January of 1998,
+has been in Beta test since then, and is planned for final release in
+mid-1999. It has a number of important features, and is the release
+that we would expect most sites to run. It is possible to run the
+Version 1 server with the Version 2 client at sites that want to be
+really conservative.
-Version 3 of the ISC DHCP Distribution will add Dynamic DNS Support,
+Version 3 of the ISC DHCP Distribution will add conditional behaviour,
+client classing, Dynamic DNS Support, DHCPv4 16-bit option codes,
asynchronous DNS query resolution, DHCP Authentication, and possibly
support for a DHCP Interserver Protocol and live querying of the DHCP
-database. This release is not expected to be stable in the near
-future, and is intended for sites that are in a position to
+database. Currently, only client classing and conditional behaviour
+have been implemented - the DNS code is waiting for an enhanced DNS
+resolver. The code has gone through a major internal restructuring
+which will help to support wider option codes, and possibly IPv6, as
+well as a more sensible memory allocation strategy. This release is
+running in producion at the ISC, but is not expected to be stable in
+the near future, and is intended for sites that are in a position to
experiment, or for sites that desperately need the new features.
+- Update top-level Makefile so that it exits correctly on errors in
+ submakes under bash/gnu make (dunno which is the culprit, and don't
+ really care).
+- Print a more helpful message if no free BPF devices are found.
+- Add support for specifying that the server is or is not
+ authoritative for a particular network segment.
+- Fix two stupid typos in lpf.c.
+- Print a more helpful message if we can't create an LPF socket or
+ can't attach a filter to it.
+- Correct the hopelessly outdated information about Linux at the top
+ of the README - many apologies to the Linux people who have had to
+ read that nonsense for the past couple of snapshots and have been
+ confused or annoyed by it. I simply hadn't read it recently, and
+ didn't realize how out-of-date it was.
+- Print a message if the client finds no broadcast interfaces to
+ configure.
+- Add support for use-lease-addr-for-default-route flag in server, so
+ that Windows machines can be made to ARP for all addresses.
+- Update README file to mention new Linux gotchas.
+- After finally understanding Brian Murrel's code (my fault, not his)
+ to get interface names from /proc/net/dev on Linux, fix what I broke
+ of his code and document it.
+- Use sendto rather than send for SOCK_PACKET sockets, because they
+ can't be connected, only bound. :'(
+- Fix up SOCK_PACKET creation so that the kernel doesn't complain
+ about it.
+- Fix incorrect tests in linux client script:
+ [ $relmajor == 2 ] -> [ $relmajor -eq 2 ]
+- Make typedefs for u8, u16 and u32 types. These are Linux kernel
+ internal data types which are unfortunately exposed in the linux
+ packetfilter header file.
+- Don't include <net/ethernet.h> in lpf.c - it defines things we're
+ already correctly defining elsewhere, and doesn't define any useful
+ new stuff.
+- Finally fix client PREINIT bug that causes interfaces not specified
+ on the command line to be preinitialized. If no interfaces are
+ specified on the command line, all interfaces are still
+ preinitialized.
+- Fix socket API fallback setup code, which was causing Linux servers
+ and clients to loop endlessly on select when run as daemons.
+- Add support for Linux 2.2 version number (treated the same as Linux
+ 2.1, for now).
+- Correct apparent error in DHCPREQUEST destination address handling
+ when in INIT-REBOOT state.
+- Do not set BROADCAST flag if we have a valid IP address.
+- Remove hard-coded filenames and use system-specific manifest
+ constants.
+- Add entry and exit hooks to Linux dhclient-script (should be added
+ to all operating systems once tested).
+- Test for linux major and minor version so as to correctly invoke
+ network configuration programs.
+- Add support for Linux's gratuitous name change of bpf_insn structure
+ (can't pollute precious Linux sources with the "Berkeley" word, I
+ guess.
+- Correct USE_BPF_{SEND,RECEIVE} ifdefs for if_reinitialize_*
+ functions.
+- Ensure that we have ifreq structure before initializing interface -
+ if an interface was specified on the command line on Linux, this was
+ not the case.
+- Get rid of references to enstamp structure in lpf.c. Correctly
+ declare and initialize sock_fprog structure (aka bpf_filter
+ structure on non-Linux machines).
+- Define ssize_t on Ultrix.
+- Generalize FDDI support.
+- Fix potential core dump in interface discovery code.
+- Put explicit release versions on startup messages.
+- Add support for Linux Packet Filter (thanks to Brian Murrell, Interlinx).
+- Add support for FDDI hardware type.
+- Fix a long-standing bug in DLPI support where the ethertype was
+ being set incorrectly (thanks to Gong Wei, CCENet).
+- Don't use DLPI RAW mode on Solaris.
+- In the client, when a lease expires, the interface to which that
+ lease is attached is unconfigured. On systems that use the socket
+ API, the interface needs to then be reconfigured with the
+ address so that it can be used to get a new address.
+- Add fallback support for Linux. This fixes a problem with the
+ relay agent when relaying over non-broadcast links, and may also fix
+ some obscure problems with unicasting DHCPACKs in both the server and
+ relay agent.
+- When allocating leases, if the oldest lease is abandoned, try to
+ find a younger-but-still-expired lease rather than reclaiming the
+ abandoned lease.
+- Add more documentation to README.
+- The absence of the /etc/dhclient.conf file is no longer considered
+ an error.
+- The dhcp client's lease file name can be specified on the command
+ line.
+- The DHCP client should no longer zap interfaces that it has not been
+ directed to configure.
+- If a client starts up in the init-reboot state, the xid will be a
+ "random" number rather than always being zero, as was previously the
+ case.
+- In addition to comparing transaction IDs, compare hardware addresses
+ in response packets to verify that they are ours.
+- Rewrite the client lease database after 20 leases have been written.
+- Fix the exponential backoff code.
+- Add a Y2k comment to indicate that something suspicious-looking is
+ in fact _not_ a problem.
+- Use mkstemp if possible.
+- Add missing fi in various client scripts.
+- Use "search" instead of "domain" in linux resolv.conf files.
+- Specify a hop count in all route command on solaris.
+- If an allocation fails, don't try to zero out the allocation buffer
+ we didn't get.
+- Support subnets that are subsets of other subnets - that is, for
+ example, and This is useful in fairly
+ obscure circumstances.
+- Don't set the lease end time if it's already expired.
+- Don't define INADDR_LOOPBACK on FreeBSD if it's already defined in a
+ system header.
+- Use the broadcast address in the relay agent if we are using the BSD
+ socket API.
+- Allow host declarations without names.
+- Allow the server identifier option to be specified.
+- Don't dump hostnames into the lease file if they contain
+ non-printable characters.
+- Copy the entire client hardware address buffer that the client sends
+ to the output packet, not just the portion of it that's supposedly
+ significant according to the hardware address length field. This
+ is done for the benefit of certain Microsoft clients.
+- Don't send a second ICMP echo request if we receive two DHCPDISCOVER
+ messages in quick succession. This prevents a rather annoying
+ timing race in configuring some Win95 clients.
+- Fix up dhcp-options man page to make it more readable. Note that
+ netbios-name-server is the same thing as WINS.
- Define some extra DLPI support flags that make DLPI work much better