freebsd-skq/UPDATING

1517 lines
60 KiB
Plaintext
Raw Normal View History

Updating Information for FreeBSD current users
This file is maintained and copyrighted by M. Warner Losh
<imp@village.org>. See end of file for further details. For commonly
done items, please see the COMMON ITEMS: section later in the file.
Items affecting the ports and packages system can be found in
/usr/ports/UPDATING. Please read that file before running
portupgrade.
NOTE TO PEOPLE WHO THINK THAT FreeBSD 8.x IS SLOW:
FreeBSD 8.x has many debugging features turned on, in
both the kernel and userland. These features attempt to detect
incorrect use of system primitives, and encourage loud failure
through extra sanity checking and fail stop semantics. They
also substantially impact system performance. If you want to
do performance measurement, benchmarking, and optimization,
you'll want to turn them off. This includes various WITNESS-
related kernel options, INVARIANTS, malloc debugging flags
in userland, and various verbose features in the kernel. Many
developers choose to disable these features on build machines
to maximize performance. (To disable malloc debugging, run
ln -s aj /etc/malloc.conf.)
Permit buiding kernels with options VIMAGE, restricted to only a single active network stack instance. Turning on options VIMAGE at compile time yields the following changes relative to default kernel build: 1) V_ accessor macros for virtualized variables resolve to structure fields via base pointers, instead of being resolved as fields in global structs or plain global variables. As an example, V_ifnet becomes: options VIMAGE: ((struct vnet_net *) vnet_net)->_ifnet default build: vnet_net_0._ifnet options VIMAGE_GLOBALS: ifnet 2) INIT_VNET_* macros will declare and set up base pointers to be used by V_ accessor macros, instead of resolving to whitespace: INIT_VNET_NET(ifp->if_vnet); becomes struct vnet_net *vnet_net = (ifp->if_vnet)->mod_data[VNET_MOD_NET]; 3) Memory for vnet modules registered via vnet_mod_register() is now allocated at run time in sys/kern/kern_vimage.c, instead of per vnet module structs being declared as globals. If required, vnet modules can now request the framework to provide them with allocated bzeroed memory by filling in the vmi_size field in their vmi_modinfo structures. 4) structs socket, ifnet, inpcbinfo, tcpcb and syncache_head are extended to hold a pointer to the parent vnet. options VIMAGE builds will fill in those fields as required. 5) curvnet is introduced as a new global variable in options VIMAGE builds, always pointing to the default and only struct vnet. 6) struct sysctl_oid has been extended with additional two fields to store major and minor virtualization module identifiers, oid_v_subs and oid_v_mod. SYSCTL_V_* family of macros will fill in those fields accordingly, and store the offset in the appropriate vnet container struct in oid_arg1. In sysctl handlers dealing with virtualized sysctls, the SYSCTL_RESOLVE_V_ARG1() macro will compute the address of the target variable and make it available in arg1 variable for further processing. Unused fields in structs vnet_inet, vnet_inet6 and vnet_ipfw have been deleted. Reviewed by: bz, rwatson Approved by: julian (mentor)
2009-04-30 13:36:26 +00:00
20090430:
The layout of the following structs has changed: sysctl_oid,
socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
panics may be experienced. World rebuild is required for
correctly checking networking state from userland.
Bump __FreeBSD_version to 800085.
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
20090429:
MLDv2 and Source-Specific Multicast (SSM) have been merged
to the IPv6 stack. VIMAGE hooks are in but not yet used.
The implementation of SSM within FreeBSD's IPv6 stack closely
follows the IPv4 implementation.
For kernel developers:
* The most important changes are that the ip6_output() and
ip6_input() paths no longer take the IN6_MULTI_LOCK,
and this lock has been downgraded to a non-recursive mutex.
* As with the changes to the IPv4 stack to support SSM, filtering
of inbound multicast traffic must now be performed by transport
protocols within the IPv6 stack. This does not apply to TCP and
SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
* The KPIs used by IPv6 multicast are similar to those used by
the IPv4 stack, with the following differences:
* im6o_mc_filter() is analogous to imo_multicast_filter().
* The legacy KAME entry points in6_joingroup and in6_leavegroup()
are shimmed to in6_mc_join() and in6_mc_leave() respectively.
* IN6_LOOKUP_MULTI() has been deprecated and removed.
* IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
for MLDv1 have an additional 'timer' argument which is used to
jitter the initial membership report for the solicited-node
multicast membership on-link.
* This is not strictly needed for MLDv2, which already jitters
its report transmissions. However, the 'timer' argument is
preserved in case MLDv1 is active on the interface.
* The KAME linked-list based IPv6 membership implementation has
been refactored to use a vector similar to that used by the IPv4
stack.
Code which maintains a list of its own multicast memberships
internally, e.g. carp, has been updated to reflect the new
semantics.
* There is a known Lock Order Reversal (LOR) due to in6_setscope()
acquiring the IF_AFDATA_LOCK and being called within ip6_output().
Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
implementation constraint which needs to be addressed in HEAD.
For application developers:
* The changes are broadly similar to those made for the IPv4
stack.
* The use of IPv4 and IPv6 multicast socket options on the same
socket, using mapped addresses, HAS NOT been tested or supported.
* There are a number of issues with the implementation of various
IPv6 multicast APIs which need to be resolved in the API surface
before the implementation is fully compatible with KAME userland
use, and these are mostly to do with interface index treatment.
* The literature available discusses the use of either the delta / ASM
API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
using setsourcefilter(3)/getsourcefilter(3). For more information
please refer to RFC 3768, 'Socket Interface Extensions for
Multicast Source Filters'.
* Applications which use the published RFC 3678 APIs should be fine.
For systems administrators:
* The mtest(8) utility has been refactored to support IPv6, in
addition to IPv4. Interface addresses are no longer accepted
as arguments, their names must be used instead. The utility
will map the interface name to its first IPv4 address as
returned by getifaddrs(3).
* The ifmcstat(8) utility has also been updated to print the MLDv2
endpoint state and source filter lists via sysctl(3).
* The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
loopback of IPv6 multicast datagrams by default; it defaults to 1
to preserve the existing behaviour. Disabling multicast loopback is
recommended for optimal system performance.
* The IPv6 MROUTING code has been changed to examine this sysctl
instead of attempting to perform a group lookup before looping
back forwarded datagrams.
Bump __FreeBSD_version to 800084.
20090422:
Implement low-level Bluetooth HCI API.
Bump __FreeBSD_version to 800083.
20090419:
The layout of struct malloc_type, used by modules to register new
memory allocation types, has changed. Most modules will need to
be rebuilt or panics may be experienced.
Bump __FreeBSD_version to 800081.
20090415:
Anticipate overflowing inp_flags - add inp_flags2.
This changes most offsets in inpcb, so checking v4 connection
state will require a world rebuild.
Bump __FreeBSD_version to 800080.
20090415:
Add an llentry to struct route and struct route_in6. Modules
embedding a struct route will need to be recompiled.
Bump __FreeBSD_version to 800079.
20090414:
The size of rt_metrics_lite and by extension rtentry has changed.
Networking administration apps will need to be recompiled.
The route command now supports show as an alias for get, weighting
of routes, sticky and nostick flags to alter the behavior of stateful
load balancing.
Bump __FreeBSD_version to 800078.
2009-04-12 19:42:25 +00:00
20090408:
Do not use Giant for kbdmux(4) locking. This is wrong and
apparently causing more problems than it solves. This will
re-open the issue where interrupt handlers may race with
kbdmux(4) in polling mode. Typical symptoms include (but
not limited to) duplicated and/or missing characters when
low level console functions (such as gets) are used while
interrupts are enabled (for example geli password prompt,
mountroot prompt etc.). Disabling kbdmux(4) may help.
20090407:
The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
kernel modules referencing any of the above need to be recompiled.
Bump __FreeBSD_version to 800075.
20090320:
GEOM_PART has become the default partition slicer for storage devices,
replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
introduces some changes:
MSDOS/EBR: the devices created from MSDOS extended partition entries
(EBR) can be named differently than with GEOM_MBR and are now symlinks
to devices with offset-based names. fstabs may need to be modified.
BSD: the "geometry does not match label" warning is harmless in most
cases but it points to problems in file system misalignment with
disk geometry. The "c" partition is now implicit, covers the whole
top-level drive and cannot be (mis)used by users.
General: Kernel dumps are now not allowed to be written to devices
whose partition types indicate they are meant to be used for file
systems (or, in case of MSDOS partitions, as something else than
the "386BSD" type).
Most of these changes date approximately from 200812.
20090319:
The uscanner(4) driver has been removed from the kernel. This follows
Linux removing theirs in 2.6 and making libusb the default interface
(supported by sane).
Introduce a number of changes to the MROUTING code. This is purely a forwarding plane cleanup; no control plane code is involved. Summary: * Split IPv4 and IPv6 MROUTING support. The static compile-time kernel option remains the same, however, the modules may now be built for IPv4 and IPv6 separately as ip_mroute_mod and ip6_mroute_mod. * Clean up the IPv4 multicast forwarding code to use BSD queue and hash table constructs. Don't build our own timer abstractions when ratecheck() and timevalclear() etc will do. * Expose the multicast forwarding cache (MFC) and virtual interface table (VIF) as sysctls, to reduce netstat's dependence on libkvm for this information for running kernels. * bandwidth meters however still require libkvm. * Make the MFC hash table size a boot/load-time tunable ULONG, net.inet.ip.mfchashsize (defaults to 256). * Remove unused members from struct vif and struct mfc. * Kill RSVP support, as no current RSVP implementation uses it. These stubs could be moved to raw_ip.c. * Don't share locks or initialization between IPv4 and IPv6. * Don't use a static struct route_in6 in ip6_mroute.c. The v6 code is still using a cached struct route_in6, this is moved to mif6 for the time being. * More cleanup remains to be merged from ip_mroute.c to ip6_mroute.c. v4 path tested using ports/net/mcast-tools. v6 changes are mostly mechanical locking and *have not* been tested. As these changes partially break some kernel ABIs, they will not be MFCed. There is a lot more work to be done here. Reviewed by: Pavlin Radoslavov
2009-03-19 01:43:03 +00:00
20090319:
The multicast forwarding code has been cleaned up. netstat(1)
only relies on KVM now for printing bandwidth upcall meters.
The IPv4 and IPv6 modules are split into ip_mroute_mod and
ip6_mroute_mod respectively. The config(5) options for statically
compiling this code remain the same, i.e. 'options MROUTING'.
20090315:
Support for the IFF_NEEDSGIANT network interface flag has been
removed, which means that non-MPSAFE network device drivers are no
longer supported. In particular, if_ar, if_sr, and network device
drivers from the old (legacy) USB stack can no longer be built or
used.
20090313:
POSIX.1 Native Language Support (NLS) has been enabled in libc and
a bunch of new language catalog files have also been added.
This means that some common libc messages are now localized and
they depend on the LC_MESSAGES environmental variable.
20090313:
The k8temp(4) driver has been renamed to amdtemp(4) since
support for K10 and K11 CPU families was added.
20090309:
IGMPv3 and Source-Specific Multicast (SSM) have been merged
to the IPv4 stack. VIMAGE hooks are in but not yet used.
For kernel developers, the most important changes are that the
ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
and this lock has been downgraded to a non-recursive mutex.
Transport protocols (UDP, Raw IP) are now responsible for filtering
inbound multicast traffic according to group membership and source
filters. The imo_multicast_filter() KPI exists for this purpose.
Transports which do not use multicast (SCTP, TCP) already reject
multicast by default. Forwarding and receive performance may improve
as a mutex acquisition is no longer needed in the ip_input()
low-level input path. in_addmulti() and in_delmulti() are shimmed
to new KPIs which exist to support SSM in-kernel.
For application developers, it is recommended that loopback of
multicast datagrams be disabled for best performance, as this
will still cause the lock to be taken for each looped-back
datagram transmission. The net.inet.ip.mcast.loop sysctl may
be tuned to 0 to disable loopback by default; it defaults to 1
to preserve the existing behaviour.
For systems administrators, to obtain best performance with
multicast reception and multiple groups, it is always recommended
that a card with a suitably precise hash filter is used. Hash
collisions will still result in the lock being taken within the
transport protocol input path to check group membership.
If deploying FreeBSD in an environment with IGMP snooping switches,
it is recommended that the net.inet.igmp.sendlocal sysctl remain
enabled; this forces 224.0.0.0/24 group membership to be announced
via IGMP.
The size of 'struct igmpstat' has changed; netstat needs to be
recompiled to reflect this.
Bump __FreeBSD_version to 800070.
20090309:
libusb20.so.1 is now installed as libusb.so.1 and the ports system
updated to use it. This requires a buildworld/installworld in order to
update the library and dependencies (usbconfig, etc). Its advisable to
rebuild all ports which uses libusb. More specific directions are given
in the ports collection UPDATING file. Any /etc/libmap.conf entries for
2009-03-09 22:43:00 +00:00
libusb are no longer required and can be removed.
20090302:
A workaround is committed to allow the creation of System V shared
memory segment of size > 2 GB on the 64-bit architectures.
Due to a limitation of the existing ABI, the shm_segsz member
of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
wrong for large segments. Note that limits must be explicitely
raised to allow such segments to be created.
20090301:
The layout of struct ifnet has changed, requiring a rebuild of all
network device driver modules.
2009-02-27 17:32:49 +00:00
20090227:
The /dev handling for the new USB stack has changed, a
buildworld/installworld is required for libusb20.
20090223:
The new USB2 stack has now been permanently moved in and all kernel and
module names reverted to their previous values (eg, usb, ehci, ohci,
2009-02-23 19:28:29 +00:00
ums, ...). The old usb stack can be compiled in by prefixing the name
with the letter 'o', the old usb modules have been removed.
Updating entry 20090216 for xorg and 20090215 for libmap may still
apply.
20090217:
The rc.conf(5) option if_up_delay has been renamed to
defaultroute_delay to better reflect its purpose. If you have
customized this setting in /etc/rc.conf you need to update it to
use the new name.
20090216:
xorg 7.4 wants to configure its input devices via hald which does not
yet work with USB2. If the keyboard/mouse does not work in xorg then
add
Option "AllowEmptyInput" "off"
to your ServerLayout section. This will cause X to use the configured
2009-02-17 10:49:36 +00:00
kbd and mouse sections from your xorg.conf.
20090215:
The GENERIC kernels for all architectures now default to the new USB2
stack. No kernel config options or code have been removed so if a
problem arises please report it and optionally revert to the old USB
stack. If you are loading USB kernel modules or have a custom kernel
that includes GENERIC then ensure that usb names are also changed over,
2009-02-17 10:49:36 +00:00
eg uftdi -> usb2_serial_ftdi.
Older programs linked against the ports libusb 0.1 need to be
redirected to the new stack's libusb20. /etc/libmap.conf can
be used for this:
# Map old usb library to new one for usb2 stack
libusb-0.1.so.8 libusb20.so.1
20090203:
The ichsmb(4) driver has been changed to require SMBus slave
addresses be left-justified (xxxxxxx0b) rather than right-justified.
All of the other SMBus controller drivers require left-justified
slave addresses, so this change makes all the drivers provide the
same interface.
20090201:
INET6 statistics (struct ip6stat) was updated.
netstat(1) needs to be recompiled.
20090119:
NTFS has been removed from GENERIC kernel on amd64 to match
GENERIC on i386. Should not cause any issues since mount_ntfs(8)
will load ntfs.ko module automatically when NTFS support is
actually needed, unless ntfs.ko is not installed or security
level prohibits loading kernel modules. If either is the case,
"options NTFS" has to be added into kernel config.
20090115:
TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
800061. User space tools that rely on the size of struct tcpcb in
tcp_var.h (e.g. sockstat) need to be recompiled.
20081225:
ng_tty(4) module updated to match the new TTY subsystem.
Due to API change, user-level applications must be updated.
New API support added to mpd5 CVS and expected to be present
in next mpd5.3 release.
20081219:
2008-12-19 23:12:14 +00:00
With __FreeBSD_version 800060 the makefs tool is part of
the base system (it was a port).
20081216:
2009-02-17 10:50:18 +00:00
The afdata and ifnet locks have been changed from mutexes to
rwlocks, network modules will need to be re-compiled.
20081214:
__FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
2009-02-17 10:50:18 +00:00
The new code reduced struct rtentry{} by 16 bytes on 32-bit
architecture and 40 bytes on 64-bit architecture. The userland
applications "arp" and "ndp" have been updated accordingly.
The output from "netstat -r" shows only routing entries and
none of the L2 information.
20081130:
__FreeBSD_version 800057 marks the switchover from the
binary ath hal to source code. Users must add the line:
2008-12-01 23:09:58 +00:00
options AH_SUPPORT_AR5416
to their kernel config files when specifying:
device ath_hal
The ath_hal module no longer exists; the code is now compiled
together with the driver in the ath module. It is now
possible to tailor chip support (i.e. reduce the set of chips
and thereby the code size); consult ath_hal(4) for details.
20081121:
__FreeBSD_version 800054 adds memory barriers to
<machine/atomic.h>, new interfaces to ifnet to facilitate
multiple hardware transmit queues for cards that support
them, and a lock-less ring-buffer implementation to
enable drivers to more efficiently manage queueing of
packets.
20081117:
A new version of ZFS (version 13) has been merged to -HEAD.
This version has zpool attribute "listsnapshots" off by
default, which means "zfs list" does not show snapshots,
and is the same as Solaris behavior.
20081028:
dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
20081009:
The uhci, ohci, ehci and slhci USB Host controller drivers have
been put into separate modules. If you load the usb module
separately through loader.conf you will need to load the
appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
controller add the following to loader.conf:
uhci_load="YES"
ehci_load="YES"
20081009:
The ABI used by the PMC toolset has changed. Please keep
userland (libpmc(3)) and the kernel module (hwpmc(4)) in
sync.
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
20080820:
The TTY subsystem of the kernel has been replaced by a new
implementation, which provides better scalability and an
improved driver model. Most common drivers have been migrated to
the new TTY subsystem, while others have not. The following
drivers have not yet been ported to the new TTY layer:
PCI/ISA:
cy, digi, rc, rp, sio
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
USB:
ubser, ucycom
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
Line disciplines:
ng_h4, ng_tty, ppp, sl, snp
Adding these drivers to your kernel configuration file shall
cause compilation to fail.
2008-09-03 08:30:17 +00:00
20080818:
ntpd has been upgraded to 4.2.4p5.
20080801:
OpenSSH has been upgraded to 5.1p1.
For many years, FreeBSD's version of OpenSSH preferred DSA
over RSA for host and user authentication keys. With this
upgrade, we've switched to the vendor's default of RSA over
DSA. This may cause upgraded clients to warn about unknown
host keys even for previously known hosts. Users should
follow the usual procedure for verifying host keys before
accepting the RSA key.
This can be circumvented by setting the "HostKeyAlgorithms"
option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
command line.
Please note that the sequence of keys offered for
authentication has been changed as well. You may want to
specify IdentityFile in a different order to revert this
behavior.
20080713:
The sio(4) driver has been removed from the i386 and amd64
kernel configuration files. This means uart(4) is now the
default serial port driver on those platforms as well.
2008-07-25 09:30:53 +00:00
To prevent collisions with the sio(4) driver, the uart(4) driver
uses different names for its device nodes. This means the
onboard serial port will now most likely be called "ttyu0"
instead of "ttyd0". You may need to reconfigure applications to
use the new device names.
When using the serial port as a boot console, be sure to update
/boot/device.hints and /etc/ttys before booting the new kernel.
If you forget to do so, you can still manually specify the hints
at the loader prompt:
set hint.uart.0.at="isa"
set hint.uart.0.port="0x3F8"
set hint.uart.0.flags="0x10"
set hint.uart.0.irq="4"
boot -s
2008-06-09 21:33:57 +00:00
20080609:
The gpt(8) utility has been removed. Use gpart(8) to partition
disks instead.
20080603:
The version that Linuxulator emulates was changed from 2.4.2
to 2.6.16. If you experience any problems with Linux binaries
2008-07-25 09:30:53 +00:00
please try to set sysctl compat.linux.osrelease to 2.4.2 and
if it fixes the problem contact emulation mailing list.
20080525:
ISDN4BSD (I4B) was removed from the src tree. You may need to
update a your kernel configuration and remove relevant entries.
20080509:
I have checked in code to support multiple routing tables.
2008-07-25 09:13:18 +00:00
See the man pages setfib(1) and setfib(2).
This is a hopefully backwards compatible version,
but to make use of it you need to compile your kernel
with options ROUTETABLES=2 (or more up to 16).
2008-04-20 21:25:37 +00:00
20080420:
The 802.11 wireless support was redone to enable multi-bss
operation on devices that are capable. The underlying device
is no longer used directly but instead wlanX devices are
cloned with ifconfig. This requires changes to rc.conf files.
For example, change:
ifconfig_ath0="WPA DHCP"
to
wlans_ath0=wlan0
ifconfig_wlan0="WPA DHCP"
2008-04-29 19:55:18 +00:00
see rc.conf(5) for more details. In addition, mergemaster of
/etc/rc.d is highly recommended. Simultaneous update of userland
and kernel wouldn't hurt either.
2008-04-20 21:25:37 +00:00
As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
modules were merged into the base wlan module. All references
to these modules (e.g. in kernel config files) must be removed.
20080408:
psm(4) has gained write(2) support in native operation level.
Arbitrary commands can be written to /dev/psm%d and status can
be read back from it. Therefore, an application is responsible
for status validation and error recovery. It is a no-op in
other operation levels.
20080312:
Support for KSE threading has been removed from the kernel. To
run legacy applications linked against KSE libmap.conf may
be used. The following libmap.conf may be used to ensure
compatibility with any prior release:
libpthread.so.1 libthr.so.1
libpthread.so.2 libthr.so.2
libkse.so.3 libthr.so.3
20080301:
The layout of struct vmspace has changed. This affects libkvm
and any executables that link against libkvm and use the
kvm_getprocs() function. In particular, but not exclusively,
it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
The effects are minimal, but it's advisable to upgrade world
nonetheless.
20080229:
The latest em driver no longer has support in it for the
82575 adapter, this is now moved to the igb driver. The
split was done to make new features that are incompatible
with older hardware easier to do.
20080220:
The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
likewise the kernel option is now GEOM_LINUX_LVM.
20080211:
The default NFS mount mode has changed from UDP to TCP for
increased reliability. If you rely on (insecurely) NFS
mounting across a firewall you may need to update your
firewall rules.
2008-02-08 21:24:58 +00:00
20080208:
Belatedly note the addition of m_collapse for compacting
mbuf chains.
Our fts(3) API, as inherited from 4.4BSD, suffers from integer fields in FTS and FTSENT structs being too narrow. In addition, the narrow types creep from there into fts.c. As a result, fts(3) consumers, e.g., find(1) or rm(1), can't handle file trees an ordinary user can create, which can have security implications. To fix the historic implementation of fts(3), OpenBSD and NetBSD have already changed <fts.h> in somewhat incompatible ways, so we are free to do so, too. This change is a superset of changes from the other BSDs with a few more improvements. It doesn't touch fts(3) functionality; it just extends integer types used by it to match modern reality and the C standard. Here are its points: o For C object sizes, use size_t unless it's 100% certain that the object will be really small. (Note that fts(3) can construct pathnames _much_ longer than PATH_MAX for its consumers.) o Avoid the short types because on modern platforms using them results in larger and slower code. Change shorts to ints as follows: - For variables than count simple, limited things like states, use plain vanilla `int' as it's the type of choice in C. - For a limited number of bit flags use `unsigned' because signed bit-wise operations are implementation-defined, i.e., unportable, in C. o For things that should be at least 64 bits wide, use long long and not int64_t, as the latter is an optional type. See FTSENT.fts_number aka FTS.fts_bignum. Extending fts_number `to satisfy future needs' is pointless because there is fts_pointer, which can be used to link to arbitrary data from an FTSENT. However, there already are fts(3) consumers that require fts_number, or fts_bignum, have at least 64 bits in it, so we must allow for them. o For the tree depth, use `long'. This is a trade-off between making this field too wide and allowing for 64-bit inode numbers and/or chain-mounted filesystems. On the one hand, `long' is almost enough for 32-bit filesystems on a 32-bit platform (our ino_t is uint32_t now). On the other hand, platforms with a 64-bit (or wider) `long' will be ready for 64-bit inode numbers, as well as for several 32-bit filesystems mounted one under another. Note that fts_level has to be signed because -1 is a magic value for it, FTS_ROOTPARENTLEVEL. o For the `nlinks' local var in fts_build(), use `long'. The logic in fts_build() requires that `nlinks' be signed, but our nlink_t currently is uint16_t. Therefore let's make the signed var wide enough to be able to represent 2^16-1 in pure C99, and even 2^32-1 on a 64-bit platform. Perhaps the logic should be changed just to use nlink_t, but it can be done later w/o breaking fts(3) ABI any more because `nlinks' is just a local var. This commit also inludes supporting stuff for the fts change: o Preserve the old versions of fts(3) functions through libc symbol versioning because the old versions appeared in all our former releases. o Bump __FreeBSD_version just in case. There is a small chance that some ill-written 3-rd party apps may fail to build or work correctly if compiled after this change. o Update the fts(3) manpage accordingly. In particular, remove references to fts_bignum, which was a FreeBSD-specific hack to work around the too narrow types of FTSENT members. Now fts_number is at least 64 bits wide (long long) and fts_bignum is an undocumented alias for fts_number kept around for compatibility reasons. According to Google Code Search, the only big consumers of fts_bignum are in our own source tree, so they can be fixed easily to use fts_number. o Mention the change in src/UPDATING. PR: bin/104458 Approved by: re (quite a while ago) Discussed with: deischen (the symbol versioning part) Reviewed by: -arch (mostly silence); das (generally OK, but we didn't agree on some types used; assuming that no objections on -arch let me to stick to my opinion)
2008-01-26 17:09:40 +00:00
20080126:
The fts(3) structures have been changed to use adequate
integer types for their members and so to be able to cope
with huge file trees. The old fts(3) ABI is preserved
through symbol versioning in libc, so third-party binaries
using fts(3) should still work, although they will not take
advantage of the extended types. At the same time, some
third-party software might fail to build after this change
due to unportable assumptions made in its source code about
fts(3) structure members. Such software should be fixed
by its vendor or, in the worst case, in the ports tree.
FreeBSD_version 800015 marks this change for the unlikely
case that a portable fix is impossible.
20080123:
To upgrade to -current after this date, you must be running
FreeBSD not older than 6.0-RELEASE. Upgrading to -current
from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
20071128:
The ADAPTIVE_GIANT kernel option has been retired because its
functionality is the default now.
20071118:
The AT keyboard emulation of sunkbd(4) has been turned on
by default. In order to make the special symbols of the Sun
keyboards driven by sunkbd(4) work under X these now have
to be configured the same way as Sun USB keyboards driven
by ukbd(4) (which also does AT keyboard emulation), f.e.:
Option "XkbLayout" "us"
Option "XkbRules" "xorg"
Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
20071024:
It has been decided that it is desirable to provide ABI
backwards compatibility to the FreeBSD 4/5/6 versions of the
PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
broken with the introduction of PCI domain support (see the
20070930 entry). Unfortunately, this required the ABI of
PCIOCGETCONF to be broken again in order to be able to
provide backwards compatibility to the old version of that
IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
again. As for prominent ports this affects neither pciutils
nor xorg-server this time, the hal port needs to be rebuilt
however.
20071020:
The misnamed kthread_create() and friends have been renamed
to kproc_create() etc. Many of the callers already
used kproc_start()..
I will return kthread_create() and friends in a while
with implementations that actually create threads, not procs.
Renaming corresponds with version 800002.
20071010:
RELENG_7 branched.
20071009:
Setting WITHOUT_LIBPTHREAD now means WITHOUT_LIBKSE and
WITHOUT_LIBTHR are set.
20070930:
The PCI code has been made aware of PCI domains. This means that
the location strings as used by pciconf(8) etc are now in the
following format: pci<domain>:<bus>:<device>[:<function>]. It
also means that consumers of <sys/pciio.h> potentially need to
be recompiled; this includes the hal and xorg-server ports.
20070928:
The caching daemon (cached) was renamed to nscd. nscd.conf
configuration file should be used instead of cached.conf and
nscd_enable, nscd_pidfile and nscd_flags options should be used
instead of cached_enable, cached_pidfile and cached_flags in
rc.conf.
20070921:
The getfacl(1) utility now prints owning user and group name
instead of owning uid and gid in the three line comment header.
This is the same behavior as getfacl(1) on Solaris and Linux.
20070704:
The new IPsec code is now compiled in using the IPSEC option. The
IPSEC option now requires "device crypto" be defined in your kernel
configuration. The FAST_IPSEC kernel option is now deprecated.
20070702:
The packet filter (pf) code has been updated to OpenBSD 4.1 Please
note the changed syntax - keep state is now on by default. Also
note the fact that ftp-proxy(8) has been changed from bottom up and
has been moved from libexec to usr/sbin. Changes in the ALTQ
handling also affect users of IPFW's ALTQ capabilities.
20070701:
Remove KAME IPsec in favor of FAST_IPSEC, which is now the
only IPsec supported by FreeBSD. The new IPsec stack
supports both IPv4 and IPv6. The kernel option will change
after the code changes have settled in. For now the kernel
option IPSEC is deprecated and FAST_IPSEC is the only option, that
will change after some settling time.
20070701:
The wicontrol(8) utility has been removed from the base system. wi(4)
cards should be configured using ifconfig(8), see the man page for more
information.
20070612:
The i386/amd64 GENERIC kernel now defaults to the nfe(4) driver
instead of the nve(4) driver. Please update your configuration
accordingly.
20070612:
By default, /etc/rc.d/sendmail no longer rebuilds the aliases
database if it is missing or older than the aliases file. If
desired, set the new rc.conf option sendmail_rebuild_aliases
to "YES" to restore that functionality.
Import rewrite of IPv4 socket multicast layer to support source-specific and protocol-independent host mode multicast. The code is written to accomodate IPv6, IGMPv3 and MLDv2 with only a little additional work. This change only pertains to FreeBSD's use as a multicast end-station and does not concern multicast routing; for an IGMPv3/MLDv2 router implementation, consider the XORP project. The work is based on Wilbert de Graaf's IGMPv3 code drop for FreeBSD 4.6, which is available at: http://www.kloosterhof.com/wilbert/igmpv3.html Summary * IPv4 multicast socket processing is now moved out of ip_output.c into a new module, in_mcast.c. * The in_mcast.c module implements the IPv4 legacy any-source API in terms of the protocol-independent source-specific API. * Source filters are lazy allocated as the common case does not use them. They are part of per inpcb state and are covered by the inpcb lock. * struct ip_mreqn is now supported to allow applications to specify multicast joins by interface index in the legacy IPv4 any-source API. * In UDP, an incoming multicast datagram only requires that the source port matches the 4-tuple if the socket was already bound by source port. An unbound socket SHOULD be able to receive multicasts sent from an ephemeral source port. * The UDP socket multicast filter mode defaults to exclusive, that is, sources present in the per-socket list will be blocked from delivery. * The RFC 3678 userland functions have been added to libc: setsourcefilter, getsourcefilter, setipv4sourcefilter, getipv4sourcefilter. * Definitions for IGMPv3 are merged but not yet used. * struct sockaddr_storage is now referenced from <netinet/in.h>. It is therefore defined there if not already declared in the same way as for the C99 types. * The RFC 1724 hack (specify 0.0.0.0/8 addresses to IP_MULTICAST_IF which are then interpreted as interface indexes) is now deprecated. * A patch for the Rhyolite.com routed in the FreeBSD base system is available in the -net archives. This only affects individuals running RIPv1 or RIPv2 via point-to-point and/or unnumbered interfaces. * Make IPv6 detach path similar to IPv4's in code flow; functionally same. * Bump __FreeBSD_version to 700048; see UPDATING. This work was financially supported by another FreeBSD committer. Obtained from: p4://bms_netdev Submitted by: Wilbert de Graaf (original work) Reviewed by: rwatson (locking), silence from fenner, net@ (but with encouragement)
2007-06-12 16:24:56 +00:00
20070612:
The IPv4 multicast socket code has been considerably modified, and
moved to the file sys/netinet/in_mcast.c. Initial support for the
RFC 3678 Source-Specific Multicast Socket API has been added to
the IPv4 network stack.
Strict multicast and broadcast reception is now the default for
UDP/IPv4 sockets; the net.inet.udp.strict_mcast_mship sysctl variable
has now been removed.
The RFC 1724 hack for interface selection has been removed; the use
of the Linux-derived ip_mreqn structure with IP_MULTICAST_IF has
been added to replace it. Consumers such as routed will soon be
updated to reflect this.
These changes affect users who are running routed(8) or rdisc(8)
from the FreeBSD base system on point-to-point or unnumbered
interfaces.
2007-06-11 04:06:50 +00:00
20070610:
The net80211 layer has changed significantly and all wireless
drivers that depend on it need to be recompiled. Further these
changes require that any program that interacts with the wireless
support in the kernel be recompiled; this includes: ifconfig,
wpa_supplicant, hostapd, and wlanstats. Users must also, for
the moment, kldload the wlan_scan_sta and/or wlan_scan_ap modules
if they use modules for wireless support. These modules implement
scanning support for station and ap modes, respectively. Failure
to load the appropriate module before marking a wireless interface
up will result in a message to the console and the device not
operating properly.
20070610:
The pam_nologin(8) module ceases to provide an authentication
function and starts providing an account management function.
Consequent changes to /etc/pam.d should be brought in using
mergemaster(8). Third-party files in /usr/local/etc/pam.d may
need manual editing as follows. Locate this line (or similar):
auth required pam_nologin.so no_warn
and change it according to this example:
account required pam_nologin.so no_warn
That is, the first word needs to be changed from "auth" to
"account". The new line can be moved to the account section
within the file for clarity. Not updating pam.conf(5) files
will result in nologin(5) ignored by the respective services.
20070529:
The ether_ioctl() function has been synchronized with ioctl(2)
and ifnet.if_ioctl. Due to that, the size of one of its arguments
has changed on 64-bit architectures. All kernel modules using
ether_ioctl() need to be rebuilt on such architectures.
20070516:
Improved INCLUDE_CONFIG_FILE support has been introduced to the
config(8) utility. In order to take advantage of this new
functionality, you are expected to recompile and install
src/usr.sbin/config. If you don't rebuild config(8), and your
kernel configuration depends on INCLUDE_CONFIG_FILE, the kernel
build will be broken because of a missing "kernconfstring"
symbol.
20070513:
Symbol versioning is enabled by default. To disable it, use
option WITHOUT_SYMVER. It is not advisable to attempt to
disable symbol versioning once it is enabled; your installworld
will break because a symbol version-less libc will get installed
before the install tools. As a result, the old install tools,
which previously had symbol dependencies to FBSD_1.0, will fail
because the freshly installed libc will not have them.
The default threading library (providing "libpthread") has been
changed to libthr. If you wish to have libkse as your default,
use option DEFAULT_THREAD_LIB=libkse for the buildworld.
20070423:
2007-06-12 17:33:56 +00:00
The ABI breakage in sendmail(8)'s libmilter has been repaired
so it is no longer necessary to recompile mail filters (aka,
milters). If you recompiled mail filters after the 20070408
note, it is not necessary to recompile them again.
20070417:
The new trunk(4) driver has been renamed to lagg(4) as it better
reflects its purpose. ifconfig will need to be recompiled.
20070408:
sendmail(8) has been updated to version 8.14.1. Mail filters
(aka, milters) compiled against the libmilter included in the
base operating system should be recompiled.
20070302:
Firmwares for ipw(4) and iwi(4) are now included in the base tree.
In order to use them one must agree to the respective LICENSE in
share/doc/legal and define legal.intel_<name>.license_ack=1 via
loader.conf(5) or kenv(1). Make sure to deinstall the now
deprecated modules from the respective firmware ports.
20070228:
The name resolution/mapping functions addr2ascii(3) and ascii2addr(3)
were removed from FreeBSD's libc. These originally came from INRIA
IPv6. Nothing in FreeBSD ever used them. They may be regarded as
deprecated in previous releases.
The AF_LINK support for getnameinfo(3) was merged from NetBSD to
replace it as a more portable (and re-entrant) API.
2007-02-24 19:45:09 +00:00
20070224:
2008-07-25 09:30:53 +00:00
To support interrupt filtering a modification to the newbus API
has occurred, ABI was broken and __FreeBSD_version was bumped
2008-07-25 09:30:53 +00:00
to 700031. Please make sure that your kernel and modules are in
sync. For more info:
http://docs.freebsd.org/cgi/mid.cgi?20070221233124.GA13941
2007-02-24 19:45:09 +00:00
20070224:
The IPv6 multicast forwarding code may now be loaded into GENERIC
kernels by loading the ip_mroute.ko module. This is built into the
module unless WITHOUT_INET6 or WITHOUT_INET6_SUPPORT options are
set; see src.conf(5) for more information.
20070214:
The output of netstat -r has changed. Without -n, we now only
print a "network name" without the prefix length if the network
address and mask exactly match a Class A/B/C network, and an entry
exists in the nsswitch "networks" map.
With -n, we print the full unabbreviated CIDR network prefix in
the form "a.b.c.d/p". 0.0.0.0/0 is always printed as "default".
This change is in preparation for changes such as equal-cost
multipath, and to more generally assist operational deployment
of FreeBSD as a modern IPv4 router.
20070210:
PIM has been turned on by default in the IPv4 multicast
routing code. The kernel option 'PIM' has now been removed.
PIM is now built by default if option 'MROUTING' is specified.
It may now be loaded into GENERIC kernels by loading the
ip_mroute.ko module.
20070207:
Support for IPIP tunnels (VIFF_TUNNEL) in IPv4 multicast routing
has been removed. Its functionality may be achieved by explicitly
configuring gif(4) interfaces and using the 'phyint' keyword in
mrouted.conf.
XORP does not support source-routed IPv4 multicast tunnels nor the
integrated IPIP tunneling, therefore it is not affected by this
change. The __FreeBSD_version macro has been bumped to 700030.
20061221:
Support for PCI Message Signalled Interrupts has been
re-enabled in the bge driver, only for those chips which are
believed to support it properly. If there are any problems,
MSI can be disabled completely by setting the
'hw.pci.enable_msi' and 'hw.pci.enable_msix' tunables to 0
in the loader.
20061214:
Support for PCI Message Signalled Interrupts has been
disabled again in the bge driver. Many revisions of the
hardware fail to support it properly. Support can be
re-enabled by removing the #define of BGE_DISABLE_MSI in
"src/sys/dev/bge/if_bge.c".
20061214:
Support for PCI Message Signalled Interrupts has been added
to the bge driver. If there are any problems, MSI can be
disabled completely by setting the 'hw.pci.enable_msi' and
'hw.pci.enable_msix' tunables to 0 in the loader.
20061205:
2008-07-25 09:30:53 +00:00
The removal of several facets of the experimental Threading
system from the kernel means that the proc and thread structures
have changed quite a bit. I suggest all kernel modules that might
reference these structures be recompiled.. Especially the
linux module.
Welcome to Once-a-year Sound Mega-Commit. Enjoy numerous updates and fixes in every sense. General ------- - Multichannel safe, endian safe, format safe * Large part of critical pcm filters such as vchan.c, feeder_rate.c, feeder_volume.c, feeder_fmt.c and feeder.c has been rewritten so that using them does not cause the pcm data to be converted to 16bit little endian. * Macrosses for accessing pcm data safely are defined within sound.h in the form of PCM_READ_* / PCM_WRITE_* * Currently, most of them are probably limited for mono/stereo handling, but the future addition of true multichannel will be much easier. - Low latency operation * Well, this require lot more works to do not just within sound driver, but we're heading towards right direction. Buffer/block sizing within channel.c is rewritten to calculate precise allocation for various combination of sample/data/rate size. As a result, applying correct SNDCTL_DSP_POLICY value will achive expected latency behaviour simmilar to what commercial 4front driver do. * Signal handling fix. ctrl+c of "cat /dev/zero > /dev/dsp" does not result long delay. * Eliminate sound truncation if the sound data is too small. DIY: 1) Download / extract http://people.freebsd.org/~ariff/lowlatency/shortfiles.tar.gz 2) Do a comparison between "cat state*.au > /dev/dsp" and "for x in state*.au ; do cat $x > /dev/dsp ; done" - there should be no "perceivable" differences. Double close for PR kern/31445. CAVEAT: Low latency come with (unbearable) price especially for poorly written applications. Applications that trying to act smarter by requesting (wrong) blocksize/blockcount will suffer the most. Fixup samples/patches can be found at: http://people.freebsd.org/~ariff/ports/ - Switch minimum/maximum sampling rate limit to "1" and "2016000" (48k * 42) due to closer compatibility with 4front driver. Discussed with: marcus@ (long time ago?) - All driver specific sysctls in the form of "hw.snd.pcm%d.*" have been moved to their own dev sysctl nodes, notably: hw.snd.pcm%d.vchans -> dev.pcm.%d.vchans Bump __FreeBSD_version. Driver specific --------------- - Ditto for sysctls. - snd_atiixp, snd_es137x, snd_via8233, snd_hda * Numerous cleanups and fixes. * _EXPERIMENTAL_ polling mode support using simple callout_* mechanisme. This was intended for pure debugging and latency measurement, but proven good enough in few unexpected and rare cases (such as problematic shared IRQ with GIANT devices - USB). Polling can be enabled/disabled through dev.pcm.0.polling. Disabled by default. - snd_ich * Fix possible overflow during speed calibration. Delay final initialization (pcm_setstatus) after calibration finished. PR: kern/100169 Tested by: Kevin Overman <oberman@es.net> * Inverted EAPD for few Nec VersaPro. PR: kern/104715 Submitted by: KAWATA Masahiko <kawata@mta.biglobe.ne.jp> Thanks to various people, notably Joel Dahl, Yuriy Tsibizov, Kevin Oberman, those at #freebsd-azalia @ freenode and others for testing. Joel Dahl will do the manpage update.
2006-11-26 12:24:06 +00:00
20061126:
Sound infrastructure has been updated with various fixes and
improvements. Most of the changes are pretty much transparent,
with exceptions of followings:
1) All sound driver specific sysctls (hw.snd.pcm%d.*) have been
moved to their own dev sysctl nodes, for example:
hw.snd.pcm0.vchans -> dev.pcm.0.vchans
2) /dev/dspr%d.%d has been deprecated. Each channel now has its
own chardev in the form of "dsp%d.<function>%d", where <function>
is p = playback, r = record and v = virtual, respectively. Users
are encouraged to use these devs instead of (old) "/dev/dsp%d.%d".
This does not affect those who are using "/dev/dsp".
20061122:
geom(4)'s gmirror(8) class metadata structure has been
rev'd from v3 to v4. If you update across this point and
your metadata is converted for you, you will not be easily
able to downgrade since the /boot/kernel.old/geom_mirror.ko
kernel module will be unable to read the v4 metadata. You
can resolve this by doing from the loader(8) prompt:
set vfs.root.mountfrom="ufs:/dev/XXX"
where XXX is the root slice of one of the disks that composed
the mirror (i.e.: /dev/ad0s1a). You can then rebuild
the array the same way you built it originally.
20061122:
The following binaries have been disconnected from the build:
mount_devfs, mount_ext2fs, mount_fdescfs, mount_procfs, mount_linprocfs,
and mount_std. The functionality of these programs has been
moved into the mount program. For example, to mount a devfs
filesystem, instead of using mount_devfs, use: "mount -t devfs".
This does not affect entries in /etc/fstab, since entries in
/etc/fstab are always processed with "mount -t fstype".
20061113:
Support for PCI Message Signalled Interrupts on i386 and amd64
has been added to the kernel and various drivers will soon be
updated to use MSI when it is available. If there are any problems,
MSI can be disabled completely by setting the 'hw.pci.enable_msi'
and 'hw.pci.enable_msix' tunables to 0 in the loader.
20061110:
The MUTEX_PROFILING option has been renamed to LOCK_PROFILING.
The lockmgr object layout has been changed as a result of having
a lock_object embedded in it. As a consequence all file system
2008-07-25 09:30:53 +00:00
kernel modules must be re-compiled. The mutex profiling man page
has not yet been updated to reflect this change.
20061026:
KSE in the kernel has now been made optional and turned on by
default. Use 'nooption KSE' in your kernel config to turn it
off. All kernel modules *must* be recompiled after this change.
There-after, modules from a KSE kernel should be compatible with
modules from a NOKSE kernel due to the temporary padding fields
added to 'struct proc'.
20060929:
mrouted and its utilities have been removed from the base system.
20060927:
Some ioctl(2) command codes have changed. Full backward ABI
compatibility is provided if the "options COMPAT_FREEBSD6" is
present in the kernel configuration file. Make sure to add
this option to your kernel config file, or recompile X.Org
and the rest of ports; otherwise they may refuse to work.
2006-09-28 13:00:50 +00:00
20060924:
tcpslice has been removed from the base system.
20060913:
The sizes of struct tcpcb (and struct xtcpcb) have changed due to
the rewrite of TCP syncookies. Tools like netstat, sockstat, and
systat needs to be rebuilt.
20060903:
libpcap updated to v0.9.4 and tcpdump to v3.9.4
20060816:
The IPFIREWALL_FORWARD_EXTENDED option is gone and the behaviour
for IPFIREWALL_FORWARD is now as it was before when it was first
committed and for years after. The behaviour is now ON.
20060725:
enigma(1)/crypt(1) utility has been changed on 64 bit architectures.
Now it can decrypt files created from different architectures.
Unfortunately, it is no longer able to decrypt a cipher text
generated with an older version on 64 bit architectures.
If you have such a file, you need old utility to decrypt it.
20060709:
The interface version of the i4b kernel part has changed. So
after updating the kernel sources and compiling a new kernel,
the i4b user space tools in "/usr/src/usr.sbin/i4b" must also
be rebuilt, and vice versa.
2008-07-25 09:30:53 +00:00
20060627:
The XBOX kernel now defaults to the nfe(4) driver instead of
the nve(4) driver. Please update your configuration
accordingly.
20060514:
The i386-only lnc(4) driver for the AMD Am7900 LANCE and Am79C9xx
PCnet family of NICs has been removed. The new le(4) driver serves
as an equivalent but cross-platform replacement with the pcn(4)
driver still providing performance-optimized support for the subset
of AMD Am79C971 PCnet-FAST and greater chips as before.
20060511:
The machdep.* sysctls and the adjkerntz utility have been
modified a bit. The new adjkerntz utility uses the new
sysctl names and sysctlbyname() calls, so it may be impossible
to run an old /sbin/adjkerntz utility in single-user mode
with a new kernel. Replace the `adjkerntz -i' step before
`make installworld' with:
/usr/obj/usr/src/sbin/adjkerntz/adjkerntz -i
and proceed as usual with the rest of the installworld-stage
steps. Otherwise, you risk installing binaries with their
timestamp set several hours in the future, especially if
you are running with local time set to GMT+X hours.
20060412:
The ip6fw utility has been removed. The behavior provided by
ip6fw has been in ipfw2 for a good while and the rc.d scripts
2006-05-14 20:54:45 +00:00
have been updated to deal with it. There are some rules that
might not migrate cleanly. Use rc.firewall6 as a template to
rewrite rules.
20060428:
The puc(4) driver has been overhauled. The ebus(4) and sbus(4)
attachments have been removed. Make sure to configure scc(4)
on sparc64. Note also that by default puc(4) will use uart(4)
and not sio(4) for serial ports because interrupt handling has
been optimized for multi-port serial cards and only uart(4)
implements the interface to support it.
20060330:
The scc(4) driver replaces puc(4) for Serial Communications
Controllers (SCCs) like the Siemens SAB82532 and the Zilog
Z8530. On sparc64, it is advised to add scc(4) to the kernel
configuration to make sure that the serial ports remain
functional.
20060317:
Most world/kernel related NO_* build options changed names.
New knobs have common prefixes WITHOUT_*/WITH_* (modelled
after FreeBSD ports) and should be set in /etc/src.conf
(the src.conf(5) manpage is provided). Full backwards
compatibility is maintained for the time being though it's
highly recommended to start moving old options out of the
system-wide /etc/make.conf file into the new /etc/src.conf
while also properly renaming them. More conversions will
likely follow. Posting to current@:
http://lists.freebsd.org/pipermail/freebsd-current/2006-March/061725.html
20060305:
The NETSMBCRYPTO kernel option has been retired because its
functionality is always included in NETSMB and smbfs.ko now.
20060303:
The TDFX_LINUX kernel option was retired and replaced by the
tdfx_linux device. The latter can be loaded as the 3dfx_linux.ko
kernel module. Loading it alone should suffice to get 3dfx support
for Linux apps because it will pull in 3dfx.ko and linux.ko through
its dependencies.
20060204:
The 'audit' group was added to support the new auditing functionality
in the base system. Be sure to follow the directions for updating,
including the requirement to run mergemaster -p.
20060201:
The kernel ABI to file system modules was changed on i386.
Please make sure that your kernel and modules are in sync.
2008-07-25 09:30:53 +00:00
20060118:
This actually occured some time ago, but installing the kernel
now also installs a bunch of symbol files for the kernel modules.
This increases the size of /boot/kernel to about 67Mbytes. You
will need twice this if you will eventually back this up to kernel.old
on your next install.
If you have a shortage of room in your root partition, you should add
-DINSTALL_NODEBUG to your make arguments or add INSTALL_NODEBUG="yes"
to your /etc/make.conf.
20060113:
libc's malloc implementation has been replaced. This change has the
potential to uncover application bugs that previously went unnoticed.
See the malloc(3) manual page for more details.
20060112:
The generic netgraph(4) cookie has been changed. If you upgrade
kernel passing this point, you also need to upgrade userland
and netgraph(4) utilities like ports/net/mpd or ports/net/mpd4.
20060106:
si(4)'s device files now contain the unit number.
Uses of {cua,tty}A[0-9a-f] should be replaced by {cua,tty}A0[0-9a-f].
20060106:
The kernel ABI was mostly destroyed due to a change in the size
of struct lock_object which is nested in other structures such
as mutexes which are nested in all sorts of other structures.
Make sure your kernel and modules are in sync.
20051231:
The page coloring algorithm in the VM subsystem was converted
from tuning with kernel options to autotuning. Please remove
any PQ_* option except PQ_NOOPT from your kernel config.
20051211:
The net80211-related tools in the tools/tools/ath directory
have been moved to tools/tools/net80211 and renamed with a
"wlan" prefix. Scripts that use them should be adjusted
accordingly.
20051202:
Scripts in the local_startup directories (as defined in
/etc/defaults/rc.conf) that have the new rc.d semantics will
now be run as part of the base system rcorder. If there are
errors or problems with one of these local scripts, it could
cause boot problems. If you encounter such problems, boot in
single user mode, remove that script from the */rc.d directory.
Please report the problem to the port's maintainer, and the
freebsd-ports@freebsd.org mailing list.
20051129:
The nodev mount option was deprecated in RELENG_6 (where it
was a no-op), and is now unsupported. If you have nodev or dev listed
in /etc/fstab, remove it, otherwise it will result in a mount error.
20051129:
ABI between ipfw(4) and ipfw(8) has been changed. You need
to rebuild ipfw(8) when rebuilding kernel.
20051108:
rp(4)'s device files now contain the unit number.
Uses of {cua,tty}R[0-9a-f] should be replaced by {cua,tty}R0[0-9a-f].
20051029:
/etc/rc.d/ppp-user has been renamed to /etc/rc.d/ppp.
Its /etc/rc.conf.d configuration file has been `ppp' from
the beginning, and hence there is no need to touch it.
20051014:
Now most modules get their build-time options from the kernel
configuration file. A few modules still have fixed options
due to their non-conformant implementation, but they will be
corrected eventually. You may need to review the options of
the modules in use, explicitly specify the non-default options
in the kernel configuration file, and rebuild the kernel and
modules afterwards.
20051001:
kern.polling.enable sysctl MIB is now deprecated. Use ifconfig(8)
to turn polling(4) on your interfaces.
20050927:
The old bridge(4) implementation was retired. The new
if_bridge(4) serves as a full functional replacement.
20050722:
The ai_addrlen of a struct addrinfo was changed to a socklen_t
to conform to POSIX-2001. This change broke an ABI
compatibility on 64 bit architecture. You have to recompile
userland programs that use getaddrinfo(3) on 64 bit
architecture.
20050711:
RELENG_6 branched here.
20050629:
The pccard_ifconfig rc.conf variable has been removed and a new
variable, ifconfig_DEFAULT has been introduced. Unlike
pccard_ifconfig, ifconfig_DEFAULT applies to ALL interfaces that
do not have ifconfig_ifn entries rather than just those in
removable_interfaces.
20050616:
Some previous versions of PAM have permitted the use of
non-absolute paths in /etc/pam.conf or /etc/pam.d/* when referring
to third party PAM modules in /usr/local/lib. A change has been
made to require the use of absolute paths in order to avoid
ambiguity and dependence on library path configuration, which may
affect existing configurations.
2005-06-10 16:51:49 +00:00
20050610:
Major changes to network interface API. All drivers must be
recompiled. Drivers not in the base system will need to be
updated to the new APIs.
20050609:
Changes were made to kinfo_proc in sys/user.h. Please recompile
userland, or commands like `fstat', `pkill', `ps', `top' and `w'
will not behave correctly.
The API and ABI for hwpmc(4) have changed with the addition
of sampling support. Please recompile lib/libpmc(3) and
usr.sbin/{pmcstat,pmccontrol}.
20050606:
The OpenBSD dhclient was imported in place of the ISC dhclient
and the network interface configuration scripts were updated
accordingly. If you use DHCP to configure your interfaces, you
must now run devd. Also, DNS updating was lost so you will need
to find a workaround if you use this feature.
The '_dhcp' user was added to support the OpenBSD dhclient. Be
sure to run mergemaster -p (like you are supposed to do every time
anyway).
20050605:
if_bridge was added to the tree. This has changed struct ifnet.
Please recompile userland and all network related modules.
2005-06-03 03:34:21 +00:00
20050603:
The n_net of a struct netent was changed to an uint32_t, and
1st argument of getnetbyaddr() was changed to an uint32_t, to
conform to POSIX-2001. These changes broke an ABI
2005-06-03 10:19:26 +00:00
compatibility on 64 bit architecture. With these changes,
shlib major of libpcap was bumped. You have to recompile
userland programs that use getnetbyaddr(3), getnetbyname(3),
getnetent(3) and/or libpcap on 64 bit architecture.
2005-06-03 03:34:21 +00:00
20050528:
Kernel parsing of extra options on '#!' first lines of shell
scripts has changed. Lines with multiple options likely will
fail after this date. For full details, please see
http://people.freebsd.org/~gad/Updating-20050528.txt
20050503:
The packet filter (pf) code has been updated to OpenBSD 3.7
Please note the changed anchor syntax and the fact that
authpf(8) now needs a mounted fdescfs(5) to function.
2005-04-18 14:33:18 +00:00
20050415:
The NO_MIXED_MODE kernel option has been removed from the i386
amd64 platforms as its use has been superceded by the new local
APIC timer code. Any kernel config files containing this option
should be updated.
20050227:
The on-disk format of LC_CTYPE files was changed to be machine
independent. Please make sure NOT to use NO_CLEAN buildworld
when crossing this point. Crossing this point also requires
recompile or reinstall of all locale depended packages.
20050225:
The ifi_epoch member of struct if_data has been changed to
contain the uptime at which the interface was created or the
statistics zeroed rather then the wall clock time because
wallclock time may go backwards. This should have no impact
unless an snmp implementation is using this value (I know of
none at this point.)
20050224:
The acpi_perf and acpi_throttle drivers are now part of the
acpi(4) main module. They are no longer built separately.
20050223:
The layout of struct image_params has changed. You have to
recompile all compatibility modules (linux, svr4, etc) for use
with the new kernel.
20050223:
The p4tcc driver has been merged into cpufreq(4). This makes
"options CPU_ENABLE_TCC" obsolete. Please load cpufreq.ko or
compile in "device cpufreq" to restore this functionality.
20050220:
The responsibility of recomputing the file system summary of
a SoftUpdates-enabled dirty volume has been transferred to the
background fsck. A rebuild of fsck(8) utility is recommended
if you have updated the kernel.
To get the old behavior (recompute file system summary at mount
time), you can set vfs.ffs.compute_summary_at_mount=1 before
mounting the new volume.
20050206:
The cpufreq import is complete. As part of this, the sysctls for
acpi(4) throttling have been removed. The power_profile script
has been updated, so you can use performance/economy_cpu_freq in
rc.conf(5) to set AC on/offline cpu frequencies.
20050206:
NG_VERSION has been increased. Recompiling kernel (or ng_socket.ko)
requires recompiling libnetgraph and userland netgraph utilities.
20050114:
Support for abbreviated forms of a number of ipfw options is
now deprecated. Warnings are printed to stderr indicating the
correct full form when a match occurs. Some abbreviations may
be supported at a later date based on user feedback. To be
considered for support, abbreviations must be in use prior to
this commit and unlikely to be confused with current key words.
20041221:
By a popular demand, a lot of NOFOO options were renamed
to NO_FOO (see bsd.compat.mk for a full list). The old
spellings are still supported, but will cause annoying
warnings on stderr. Make sure you upgrade properly (see
the COMMON ITEMS: section later in this file).
20041219:
Auto-loading of ancillary wlan modules such as wlan_wep has
been temporarily disabled; you need to statically configure
the modules you need into your kernel or explicitly load them
prior to use. Specifically, if you intend to use WEP encryption
with an 802.11 device load/configure wlan_wep; if you want to
use WPA with the ath driver load/configure wlan_tkip, wlan_ccmp,
and wlan_xauth as required.
20041213:
The behaviour of ppp(8) has changed slightly. If lqr is enabled
(``enable lqr''), older versions would revert to LCP ECHO mode on
negotiation failure. Now, ``enable echo'' is required for this
behaviour. The ppp version number has been bumped to 3.4.2 to
reflect the change.
2004-12-11 23:21:31 +00:00
20041201:
The wlan support has been updated to split the crypto support
into separate modules. For static WEP you must configure the
wlan_wep module in your system or build and install the module
in place where it can be loaded (the kernel will auto-load
the module when a wep key is configured).
20041201:
The ath driver has been updated to split the tx rate control
algorithm into a separate module. You need to include either
ath_rate_onoe or ath_rate_amrr when configuring the kernel.
20041116:
Support for systems with an 80386 CPU has been removed. Please
use FreeBSD 5.x or earlier on systems with an 80386.
20041110:
We have had a hack which would mount the root filesystem
R/W if the device were named 'md*'. As part of the vnode
work I'm doing I have had to remove this hack. People
building systems which use preloaded MD root filesystems
may need to insert a "/sbin/mount -u -o rw /dev/md0 /" in
their /etc/rc scripts.
2005-02-07 09:15:52 +00:00
20041104:
FreeBSD 5.3 shipped here.
20041102:
The size of struct tcpcb has changed again due to the removal
of RFC1644 T/TCP. You have to recompile userland programs that
read kmem for tcp sockets directly (netstat, sockstat, etc.)
20041022:
The size of struct tcpcb has changed. You have to recompile
userland programs that read kmem for tcp sockets directly
(netstat, sockstat, etc.)
20041016:
RELENG_5 branched here. For older entries, please see updating
in the RELENG_5 branch.
COMMON ITEMS:
General Notes
-------------
Avoid using make -j when upgrading. From time to time in the
past there have been problems using -j with buildworld and/or
installworld. This is especially true when upgrading between
"distant" versions (eg one that cross a major release boundary
or several minor releases, or when several months have passed
on the -current branch).
Sometimes, obscure build problems are the result of environment
poisoning. This can happen because the make utility reads its
environment when searching for values for global variables.
To run your build attempts in an "environmental clean room",
prefix all make commands with 'env -i '. See the env(1) manual
page for more details.
When upgrading from one major version to another it is generally
best to upgrade to the latest code in the currently installed branch
first, then do an upgrade to the new branch. This is the best-tested
upgrade path, and has the highest probability of being successful.
Please try this approach before reporting problems with a major
version upgrade.
To build a kernel
-----------------
If you are updating from a prior version of FreeBSD (even one just
a few days old), you should follow this procedure. It is the most
failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
make kernel-toolchain
make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
To test a kernel once
---------------------
If you just want to boot a kernel once (because you are not sure
if it works, or if you want to boot a known bad kernel to provide
debugging information) run
make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
nextboot -k testkernel
To just build a kernel when you know that it won't mess you up
--------------------------------------------------------------
This assumes you are already running a 5.X system. Replace
${arch} with the architecture of your machine (e.g. "i386",
"alpha", "amd64", "ia64", "pc98", "sparc64", etc).
cd src/sys/${arch}/conf
config KERNEL_NAME_HERE
cd ../compile/KERNEL_NAME_HERE
make depend
make
2004-06-22 10:33:58 +00:00
make install
If this fails, go to the "To build a kernel" section.
To rebuild everything and install it on the current system.
-----------------------------------------------------------
# Note: sometimes if you are running current you gotta do more than
# is listed here if you are upgrading from a really old current.
<make sure you have good level 0 dumps>
make buildworld
make kernel KERNCONF=YOUR_KERNEL_HERE
[1]
<reboot in single user> [3]
mergemaster -p [5]
make installworld
make delete-old
mergemaster [4]
<reboot>
To cross-install current onto a separate partition
--------------------------------------------------
# In this approach we use a separate partition to hold
# current's root, 'usr', and 'var' directories. A partition
# holding "/", "/usr" and "/var" should be about 2GB in
# size.
<make sure you have good level 0 dumps>
<boot into -stable>
make buildworld
make buildkernel KERNCONF=YOUR_KERNEL_HERE
<maybe newfs current's root partition>
<mount current's root partition on directory ${CURRENT_ROOT}>
make installworld DESTDIR=${CURRENT_ROOT}
make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
<edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
<reboot into current>
<do a "native" rebuild/install as described in the previous section>
<maybe install compatibility libraries from ports/misc/compat*>
<reboot>
To upgrade in-place from 5.x-stable to current
----------------------------------------------
<make sure you have good level 0 dumps>
make buildworld [9]
make kernel KERNCONF=YOUR_KERNEL_HERE [8]
[1]
<reboot in single user> [3]
mergemaster -p [5]
make installworld
make delete-old
mergemaster -i [4]
<reboot>
2000-08-28 03:54:51 +00:00
Make sure that you've read the UPDATING file to understand the
tweaks to various things you need. At this point in the life
cycle of current, things change often and you are on your own
to cope. The defaults can also change, so please read ALL of
the UPDATING entries.
Also, if you are tracking -current, you must be subscribed to
freebsd-current@freebsd.org. Make sure that before you update
your sources that you have read and understood all the recent
messages there. If in doubt, please track -stable which has
much fewer pitfalls.
[1] If you have third party modules, such as vmware, you
should disable them at this point so they don't crash your
system on reboot.
[3] From the bootblocks, boot -s, and then do
fsck -p
mount -u /
mount -a
cd src
adjkerntz -i # if CMOS is wall time
2002-04-25 02:14:48 +00:00
Also, when doing a major release upgrade, it is required that
you boot into single user mode to do the installworld.
[4] Note: This step is non-optional. Failure to do this step
can result in a significant reduction in the functionality of the
system. Attempting to do it by hand is not recommended and those
that pursue this avenue should read this file carefully, as well
as the archives of freebsd-current and freebsd-hackers mailing lists
for potential gotchas.
[5] Usually this step is a noop. However, from time to time
you may need to do this if you get unknown user in the following
step. It never hurts to do it all the time. You may need to
install a new mergemaster (cd src/usr.sbin/mergemaster && make
install) after the buildworld before this step if you last updated
from current before 20020224 or from -stable before 20020408.
[8] In order to have a kernel that can run the 4.x binaries
needed to do an installworld, you must include the COMPAT_FREEBSD4
option in your kernel. Failure to do so may leave you with a system
that is hard to boot to recover. A similar kernel option COMPAT_FREEBSD5
is required to run the 5.x binaries on more recent kernels.
Make sure that you merge any new devices from GENERIC since the
last time you updated your kernel config file.
[9] When checking out sources, you must include the -P flag to have
cvs prune empty directories.
If CPUTYPE is defined in your /etc/make.conf, make sure to use the
"?=" instead of the "=" assignment operator, so that buildworld can
override the CPUTYPE if it needs to.
MAKEOBJDIRPREFIX must be defined in an environment variable, and
not on the command line, or in /etc/make.conf. buildworld will
warn if it is improperly defined.
FORMAT:
This file contains a list, in reverse chronological order, of major
breakages in tracking -current. Not all things will be listed here,
and it only starts on October 16, 2004. Updating files can found in
previous releases if your system is older than this.
Copyright information:
Copyright 1998-2005 M. Warner Losh. All Rights Reserved.
Redistribution, publication, translation and use, with or without
modification, in full or in part, in any form or format of this
document are permitted without further permission from the author.
THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
If you find this document useful, and you want to, you may buy the
author a beer.
Contact Warner Losh if you have any questions about your use of
this document.
1999-08-28 01:35:59 +00:00
$FreeBSD$