1999-01-14 05:59:52 +00:00
|
|
|
Updating Information for FreeBSD current users
|
|
|
|
|
2009-09-03 17:04:42 +00:00
|
|
|
This file is maintained and copyrighted by M. Warner Losh <imp@freebsd.org>.
|
|
|
|
See end of file for further details. For commonly done items, please see the
|
|
|
|
COMMON ITEMS: section later in the file. These instructions assume that you
|
|
|
|
basically know what you are doing. If not, then please consult the FreeBSD
|
|
|
|
handbook.
|
2001-06-29 06:00:44 +00:00
|
|
|
|
2004-07-29 15:06:19 +00:00
|
|
|
Items affecting the ports and packages system can be found in
|
2009-09-03 17:04:42 +00:00
|
|
|
/usr/ports/UPDATING. Please read that file before running portupgrade.
|
2004-07-29 15:06:19 +00:00
|
|
|
|
2009-08-22 23:44:37 +00:00
|
|
|
NOTE TO PEOPLE WHO THINK THAT FreeBSD 9.x IS SLOW:
|
2009-09-03 17:04:42 +00:00
|
|
|
FreeBSD 9.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
|
2008-01-22 15:05:22 +00:00
|
|
|
ln -s aj /etc/malloc.conf.)
|
2008-12-19 23:10:55 +00:00
|
|
|
|
2009-09-05 08:08:14 +00:00
|
|
|
20090825:
|
|
|
|
The old tunable hw.bus.devctl_disable has been superseded by
|
|
|
|
hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
|
|
|
|
replaced by hw.bus.devctl_queue=0. The default for this new tunable
|
|
|
|
is 1000.
|
|
|
|
|
2009-08-13 17:09:45 +00:00
|
|
|
20090813:
|
2009-09-03 17:04:42 +00:00
|
|
|
Remove the option STOP_NMI. The default action is now to use NMI only
|
|
|
|
for KDB via the newly introduced function stop_cpus_hard() and
|
|
|
|
maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
|
|
|
|
|
|
|
|
20090803:
|
2009-09-03 17:13:54 +00:00
|
|
|
stable/8 branch created in subversion.
|
2009-08-13 17:09:45 +00:00
|
|
|
|
2009-07-19 17:25:24 +00:00
|
|
|
20090719:
|
2009-09-03 17:04:42 +00:00
|
|
|
Bump the shared library version numbers for all libraries that do not
|
|
|
|
use symbol versioning as part of the 8.0-RELEASE cycle. Bump
|
|
|
|
__FreeBSD_version to 800105.
|
2009-07-19 17:25:24 +00:00
|
|
|
|
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator
(DPCPU), as suggested by Peter Wemm, and implement a new per-virtual
network stack memory allocator. Modify vnet to use the allocator
instead of monolithic global container structures (vinet, ...). This
change solves many binary compatibility problems associated with
VIMAGE, and restores ELF symbols for virtualized global variables.
Each virtualized global variable exists as a "reference copy", and also
once per virtual network stack. Virtualized global variables are
tagged at compile-time, placing the in a special linker set, which is
loaded into a contiguous region of kernel memory. Virtualized global
variables in the base kernel are linked as normal, but those in modules
are copied and relocated to a reserved portion of the kernel's vnet
region with the help of a the kernel linker.
Virtualized global variables exist in per-vnet memory set up when the
network stack instance is created, and are initialized statically from
the reference copy. Run-time access occurs via an accessor macro, which
converts from the current vnet and requested symbol to a per-vnet
address. When "options VIMAGE" is not compiled into the kernel, normal
global ELF symbols will be used instead and indirection is avoided.
This change restores static initialization for network stack global
variables, restores support for non-global symbols and types, eliminates
the need for many subsystem constructors, eliminates large per-subsystem
structures that caused many binary compatibility issues both for
monitoring applications (netstat) and kernel modules, removes the
per-function INIT_VNET_*() macros throughout the stack, eliminates the
need for vnet_symmap ksym(2) munging, and eliminates duplicate
definitions of virtualized globals under VIMAGE_GLOBALS.
Bump __FreeBSD_version and update UPDATING.
Portions submitted by: bz
Reviewed by: bz, zec
Discussed with: gnn, jamie, jeff, jhb, julian, sam
Suggested by: peter
Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
|
|
|
20090714:
|
2009-09-03 17:04:42 +00:00
|
|
|
Due to changes in the implementation of virtual network stack support,
|
|
|
|
all network-related kernel modules must be recompiled. As this change
|
|
|
|
breaks the ABI, bump __FreeBSD_version to 800104.
|
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator
(DPCPU), as suggested by Peter Wemm, and implement a new per-virtual
network stack memory allocator. Modify vnet to use the allocator
instead of monolithic global container structures (vinet, ...). This
change solves many binary compatibility problems associated with
VIMAGE, and restores ELF symbols for virtualized global variables.
Each virtualized global variable exists as a "reference copy", and also
once per virtual network stack. Virtualized global variables are
tagged at compile-time, placing the in a special linker set, which is
loaded into a contiguous region of kernel memory. Virtualized global
variables in the base kernel are linked as normal, but those in modules
are copied and relocated to a reserved portion of the kernel's vnet
region with the help of a the kernel linker.
Virtualized global variables exist in per-vnet memory set up when the
network stack instance is created, and are initialized statically from
the reference copy. Run-time access occurs via an accessor macro, which
converts from the current vnet and requested symbol to a per-vnet
address. When "options VIMAGE" is not compiled into the kernel, normal
global ELF symbols will be used instead and indirection is avoided.
This change restores static initialization for network stack global
variables, restores support for non-global symbols and types, eliminates
the need for many subsystem constructors, eliminates large per-subsystem
structures that caused many binary compatibility issues both for
monitoring applications (netstat) and kernel modules, removes the
per-function INIT_VNET_*() macros throughout the stack, eliminates the
need for vnet_symmap ksym(2) munging, and eliminates duplicate
definitions of virtualized globals under VIMAGE_GLOBALS.
Bump __FreeBSD_version and update UPDATING.
Portions submitted by: bz
Reviewed by: bz, zec
Discussed with: gnn, jamie, jeff, jhb, julian, sam
Suggested by: peter
Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
|
|
|
|
2009-07-13 11:51:02 +00:00
|
|
|
20090713:
|
2009-09-03 17:04:42 +00:00
|
|
|
The TOE interface to the TCP syncache has been modified to remove
|
|
|
|
struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
|
|
|
|
The cxgb driver is the only TOE consumer affected by this change, and
|
|
|
|
needs to be recompiled along with the kernel. As this change breaks
|
|
|
|
the ABI, bump __FreeBSD_version to 800103.
|
2009-07-13 11:51:02 +00:00
|
|
|
|
2009-09-03 17:04:42 +00:00
|
|
|
20090712:
|
2009-07-12 09:14:28 +00:00
|
|
|
Padding has been added to struct tcpcb, sackhint and tcpstat in
|
|
|
|
<netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
|
|
|
|
maintainig the ABI. However, this change breaks the ABI, so bump
|
|
|
|
__FreeBSD_version to 800102. User space tools that rely on the size of
|
|
|
|
any of these structs (e.g. sockstat) need to be recompiled.
|
|
|
|
|
2009-07-01 07:35:57 +00:00
|
|
|
20090630:
|
2009-09-03 17:04:42 +00:00
|
|
|
The NFS_LEGACYRPC option has been removed along with the old kernel
|
|
|
|
RPC implementation that this option selected. Kernel configurations
|
|
|
|
may need to be adjusted.
|
2009-07-01 07:35:57 +00:00
|
|
|
|
2009-06-29 19:46:29 +00:00
|
|
|
20090629:
|
2009-09-03 17:04:42 +00:00
|
|
|
The network interface device nodes at /dev/net/<interface> have been
|
|
|
|
removed. All ioctl operations can be performed the normal way using
|
|
|
|
routing sockets. The kqueue functionality can generally be replaced
|
|
|
|
with routing sockets.
|
2009-06-29 19:46:29 +00:00
|
|
|
|
2009-06-28 08:59:46 +00:00
|
|
|
20090628:
|
2009-09-03 17:04:42 +00:00
|
|
|
The documentation from the FreeBSD Documentation Project (Handbook,
|
|
|
|
FAQ, etc.) is now installed via packages by sysinstall(8) and under
|
|
|
|
the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
|
2009-06-28 08:59:46 +00:00
|
|
|
|
2009-06-26 17:50:52 +00:00
|
|
|
20090624:
|
2009-09-03 17:04:42 +00:00
|
|
|
The ABI of various structures related to the SYSV IPC API have been
|
|
|
|
changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
|
|
|
|
options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
|
|
|
|
800100.
|
2009-06-26 17:50:52 +00:00
|
|
|
|
2009-06-22 17:48:16 +00:00
|
|
|
20090622:
|
2009-09-03 17:04:42 +00:00
|
|
|
Layout of struct vnet has changed as routing related variables were
|
|
|
|
moved to their own Vimage module. Modules need to be recompiled. Bump
|
|
|
|
__FreeBSD_version to 800099.
|
2009-06-22 17:48:16 +00:00
|
|
|
|
2009-06-19 17:10:35 +00:00
|
|
|
20090619:
|
2009-09-03 17:04:42 +00:00
|
|
|
NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
|
|
|
|
respectively. As long as no more than 16 groups per process are used,
|
|
|
|
no changes should be visible. When more than 16 groups are used, old
|
|
|
|
binaries may fail if they call getgroups() or getgrouplist() with
|
|
|
|
statically sized storage. Recompiling will work around this, but
|
|
|
|
applications should be modified to use dynamically allocated storage
|
|
|
|
for group arrays as POSIX.1-2008 does not cap an implementation's
|
|
|
|
number of supported groups at NGROUPS_MAX+1 as previous versions did.
|
|
|
|
|
|
|
|
NFS and portalfs mounts may also be affected as the list of groups is
|
|
|
|
truncated to 16. Users of NFS who use more than 16 groups, should
|
|
|
|
take care that negative group permissions are not used on the exported
|
|
|
|
file systems as they will not be reliable unless a GSSAPI based
|
|
|
|
authentication method is used.
|
|
|
|
|
|
|
|
20090616:
|
|
|
|
The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
|
|
|
|
option compiles in the support for adaptive spinning for lockmgrs
|
|
|
|
which want to enable it. The lockinit() function now accepts the flag
|
|
|
|
LK_ADAPTIVE in order to make the lock object subject to adaptive
|
|
|
|
spinning when both held in write and read mode.
|
Introduce support for adaptive spinning in lockmgr.
Actually, as it did receive few tuning, the support is disabled by
default, but it can opt-in with the option ADAPTIVE_LOCKMGRS.
Due to the nature of lockmgrs, adaptive spinning needs to be
selectively enabled for any interested lockmgr.
The support is bi-directional, or, in other ways, it will work in both
cases if the lock is held in read or write way. In particular, the
read path is passible of further tunning using the sysctls
debug.lockmgr.retries and debug.lockmgr.loops . Ideally, such sysctls
should be axed or compiled out before release.
Addictionally note that adaptive spinning doesn't cope well with
LK_SLEEPFAIL. The reason is that many (and probabilly all) consumers
of LK_SLEEPFAIL are mainly interested in knowing if the interlock was
dropped or not in order to reacquire it and re-test initial conditions.
This directly interacts with adaptive spinning because lockmgr needs
to drop the interlock while spinning in order to avoid a deadlock
(further details in the comments inside the patch).
Final note: finding someone willing to help on tuning this with
relevant workloads would be either very important and appreciated.
Tested by: jeff, pho
Requested by: many
2009-06-17 01:55:42 +00:00
|
|
|
|
2009-06-13 23:44:56 +00:00
|
|
|
20090613:
|
2009-09-03 17:04:42 +00:00
|
|
|
The layout of the structure returned by IEEE80211_IOC_STA_INFO has
|
|
|
|
changed. User applications that use this ioctl need to be rebuilt.
|
2009-06-13 23:44:56 +00:00
|
|
|
|
Introduce a mechanism for detecting calls from outbound path of the
network stack when reentering the inbound path from netgraph, and
force queueing of mbufs at the outbound netgraph node.
The mechanism relies on two components. First, in netgraph nodes
where outbound path of the network stack calls into netgraph, the
current thread has to be appropriately marked using the new
NG_OUTBOUND_THREAD_REF() macro before proceeding to call further
into the netgraph topology, and unmarked using the
NG_OUTBOUND_THREAD_UNREF() macro before returning to the caller.
Second, netgraph nodes which can potentially reenter the network
stack in the inbound path have to mark their inbound hooks using
NG_HOOK_SET_TO_INBOUND() macro. The netgraph framework will then
detect when there is a danger of a call graph looping back from
outbound to inbound path via netgraph, and defer handing off the
mbufs to the "inbound" node to a worker thread with a clean stack.
In this first pass only the most obvious netgraph nodes have been
updated to ensure no outbound to inbound calls can occur. Nodes
such as ng_ipfw, ng_gif etc. should be further examined whether a
potential for outbound to inbound call looping exists.
This commit changes the layout of struct thread, but due to
__FreeBSD_version number shortage a version bump has been omitted
at this time, nevertheless kernel and modules have to be rebuilt.
Reviewed by: julian, rwatson, bz
Approved by: julian (mentor)
2009-06-11 16:50:49 +00:00
|
|
|
20090611:
|
2009-09-03 17:04:42 +00:00
|
|
|
The layout of struct thread has changed. Kernel and modules need to
|
|
|
|
be rebuilt.
|
Introduce a mechanism for detecting calls from outbound path of the
network stack when reentering the inbound path from netgraph, and
force queueing of mbufs at the outbound netgraph node.
The mechanism relies on two components. First, in netgraph nodes
where outbound path of the network stack calls into netgraph, the
current thread has to be appropriately marked using the new
NG_OUTBOUND_THREAD_REF() macro before proceeding to call further
into the netgraph topology, and unmarked using the
NG_OUTBOUND_THREAD_UNREF() macro before returning to the caller.
Second, netgraph nodes which can potentially reenter the network
stack in the inbound path have to mark their inbound hooks using
NG_HOOK_SET_TO_INBOUND() macro. The netgraph framework will then
detect when there is a danger of a call graph looping back from
outbound to inbound path via netgraph, and defer handing off the
mbufs to the "inbound" node to a worker thread with a clean stack.
In this first pass only the most obvious netgraph nodes have been
updated to ensure no outbound to inbound calls can occur. Nodes
such as ng_ipfw, ng_gif etc. should be further examined whether a
potential for outbound to inbound call looping exists.
This commit changes the layout of struct thread, but due to
__FreeBSD_version number shortage a version bump has been omitted
at this time, nevertheless kernel and modules have to be rebuilt.
Reviewed by: julian, rwatson, bz
Approved by: julian (mentor)
2009-06-11 16:50:49 +00:00
|
|
|
|
Introduce an infrastructure for dismantling vnet instances.
Vnet modules and protocol domains may now register destructor
functions to clean up and release per-module state. The destructor
mechanisms can be triggered by invoking "vimage -d", or a future
equivalent command which will be provided via the new jail framework.
While this patch introduces numerous placeholder destructor functions,
many of those are currently incomplete, thus leaking memory or (even
worse) failing to stop all running timers. Many of such issues are
already known and will be incrementaly fixed over the next weeks in
smaller incremental commits.
Apart from introducing new fields in structs ifnet, domain, protosw
and vnet_net, which requires the kernel and modules to be rebuilt, this
change should have no impact on nooptions VIMAGE builds, since vnet
destructors can only be called in VIMAGE kernels. Moreover,
destructor functions should be in general compiled in only in
options VIMAGE builds, except for kernel modules which can be safely
kldunloaded at run time.
Bump __FreeBSD_version to 800097.
Reviewed by: bz, julian
Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
|
|
|
20090608:
|
2009-09-03 17:04:42 +00:00
|
|
|
The layout of structs ifnet, domain, protosw and vnet_net has changed.
|
|
|
|
Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
|
Introduce an infrastructure for dismantling vnet instances.
Vnet modules and protocol domains may now register destructor
functions to clean up and release per-module state. The destructor
mechanisms can be triggered by invoking "vimage -d", or a future
equivalent command which will be provided via the new jail framework.
While this patch introduces numerous placeholder destructor functions,
many of those are currently incomplete, thus leaking memory or (even
worse) failing to stop all running timers. Many of such issues are
already known and will be incrementaly fixed over the next weeks in
smaller incremental commits.
Apart from introducing new fields in structs ifnet, domain, protosw
and vnet_net, which requires the kernel and modules to be rebuilt, this
change should have no impact on nooptions VIMAGE builds, since vnet
destructors can only be called in VIMAGE kernels. Moreover,
destructor functions should be in general compiled in only in
options VIMAGE builds, except for kernel modules which can be safely
kldunloaded at run time.
Bump __FreeBSD_version to 800097.
Reviewed by: bz, julian
Approved by: rwatson, kib (re), julian (mentor)
2009-06-08 17:15:40 +00:00
|
|
|
|
2009-06-02 13:44:36 +00:00
|
|
|
20090602:
|
|
|
|
window(1) has been removed from the base system. It can now be
|
|
|
|
installed from ports. The port is called misc/window.
|
|
|
|
|
2009-06-01 15:49:42 +00:00
|
|
|
20090601:
|
2009-09-03 17:04:42 +00:00
|
|
|
The way we are storing and accessing `routing table' entries has
|
|
|
|
changed. Programs reading the FIB, like netstat, need to be
|
|
|
|
re-compiled.
|
2009-06-01 15:49:42 +00:00
|
|
|
|
2009-06-01 16:00:36 +00:00
|
|
|
20090601:
|
|
|
|
A new netisr implementation has been added for FreeBSD 8. Network
|
|
|
|
file system modules, such as igmp, ipdivert, and others, should be
|
|
|
|
rebuilt.
|
|
|
|
Bump __FreeBSD_version to 800096.
|
|
|
|
|
2009-05-30 23:52:23 +00:00
|
|
|
20090530:
|
2009-09-03 17:04:42 +00:00
|
|
|
Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
|
|
|
|
more valid.
|
2009-05-30 23:52:23 +00:00
|
|
|
|
2009-05-30 14:01:01 +00:00
|
|
|
20090530:
|
|
|
|
Add VOP_ACCESSX(9). File system modules need to be rebuilt.
|
|
|
|
Bump __FreeBSD_version to 800094.
|
|
|
|
|
2009-05-29 18:50:27 +00:00
|
|
|
20090529:
|
2009-09-03 17:04:42 +00:00
|
|
|
Add mnt_xflag field to 'struct mount'. File system modules need to be
|
|
|
|
rebuilt.
|
2009-05-29 18:50:27 +00:00
|
|
|
Bump __FreeBSD_version to 800093.
|
|
|
|
|
2009-05-29 01:49:27 +00:00
|
|
|
20090528:
|
|
|
|
The compiling option ADAPTIVE_SX has been retired while it has been
|
|
|
|
introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
|
|
|
|
The KPI for sx_init_flags() changes as accepting flags:
|
2009-09-03 17:04:42 +00:00
|
|
|
SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
|
|
|
|
been introduced in order to handle the reversed logic.
|
2009-05-29 01:49:27 +00:00
|
|
|
Bump __FreeBSD_version to 800092.
|
|
|
|
|
2009-05-27 14:11:23 +00:00
|
|
|
20090527:
|
|
|
|
Add support for hierarchical jails. Remove global securelevel.
|
|
|
|
Bump __FreeBSD_version to 800091.
|
|
|
|
|
2009-05-23 21:43:44 +00:00
|
|
|
20090523:
|
|
|
|
The layout of struct vnet_net has changed, therefore modules
|
|
|
|
need to be rebuilt.
|
|
|
|
Bump __FreeBSD_version to 800090.
|
|
|
|
|
2009-05-23 06:31:50 +00:00
|
|
|
20090523:
|
2009-09-03 17:04:42 +00:00
|
|
|
The newly imported zic(8) produces a new format in the output. Please
|
|
|
|
run tzsetup(8) to install the newly created data to /etc/localtime.
|
2009-05-23 06:31:50 +00:00
|
|
|
|
2009-05-21 01:48:42 +00:00
|
|
|
20090520:
|
|
|
|
The sysctl tree for the usb stack has renamed from hw.usb2.* to
|
|
|
|
hw.usb.* and is now consistent again with previous releases.
|
|
|
|
|
2009-05-20 20:05:56 +00:00
|
|
|
20090520:
|
2009-09-03 17:04:42 +00:00
|
|
|
802.11 monitor mode support was revised and driver api's were changed.
|
|
|
|
Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
|
|
|
|
of DLT_IEEE802_11. No user-visible data structures were changed but
|
|
|
|
applications that use DLT_IEEE802_11 may require changes.
|
2009-05-20 20:05:56 +00:00
|
|
|
Bump __FreeBSD_version to 800088.
|
|
|
|
|
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.
|
|
|
|
|
2009-04-24 09:58:50 +00:00
|
|
|
20090422:
|
2009-04-22 15:54:27 +00:00
|
|
|
Implement low-level Bluetooth HCI API.
|
|
|
|
Bump __FreeBSD_version to 800083.
|
|
|
|
|
2009-04-24 09:58:50 +00:00
|
|
|
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.
|
|
|
|
|
2009-04-15 22:09:42 +00:00
|
|
|
20090415:
|
2009-04-24 15:38:13 +00:00
|
|
|
Anticipate overflowing inp_flags - add inp_flags2.
|
2009-04-15 22:09:42 +00:00
|
|
|
This changes most offsets in inpcb, so checking v4 connection
|
2009-04-24 15:38:13 +00:00
|
|
|
state will require a world rebuild.
|
2009-04-15 22:09:42 +00:00
|
|
|
Bump __FreeBSD_version to 800080.
|
2009-04-24 15:38:13 +00:00
|
|
|
|
2009-04-15 20:34:19 +00:00
|
|
|
20090415:
|
|
|
|
Add an llentry to struct route and struct route_in6. Modules
|
2009-04-24 15:38:13 +00:00
|
|
|
embedding a struct route will need to be recompiled.
|
2009-04-15 20:34:19 +00:00
|
|
|
Bump __FreeBSD_version to 800079.
|
|
|
|
|
2009-04-14 23:05:36 +00:00
|
|
|
20090414:
|
|
|
|
The size of rt_metrics_lite and by extension rtentry has changed.
|
|
|
|
Networking administration apps will need to be recompiled.
|
2009-04-24 15:38:13 +00:00
|
|
|
The route command now supports show as an alias for get, weighting
|
2009-04-14 23:05:36 +00:00
|
|
|
of routes, sticky and nostick flags to alter the behavior of stateful
|
|
|
|
load balancing.
|
|
|
|
Bump __FreeBSD_version to 800078.
|
2009-04-24 11:34:59 +00:00
|
|
|
|
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.
|
|
|
|
|
2009-04-08 18:30:42 +00:00
|
|
|
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.
|
|
|
|
|
2009-03-20 21:51:27 +00:00
|
|
|
20090320:
|
|
|
|
GEOM_PART has become the default partition slicer for storage devices,
|
|
|
|
replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
|
2009-03-20 23:13:32 +00:00
|
|
|
introduces some changes:
|
2009-04-24 15:38:13 +00:00
|
|
|
|
2009-03-20 23:13:32 +00:00
|
|
|
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.
|
2009-04-24 15:38:13 +00:00
|
|
|
|
2009-03-20 23:13:32 +00:00
|
|
|
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.
|
2009-03-20 21:51:27 +00:00
|
|
|
|
2009-03-19 20:33:26 +00:00
|
|
|
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).
|
|
|
|
|
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'.
|
|
|
|
|
2009-03-15 16:12:50 +00:00
|
|
|
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.
|
|
|
|
|
2009-03-13 16:40:56 +00:00
|
|
|
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.
|
|
|
|
|
2009-03-13 16:30:33 +00:00
|
|
|
20090313:
|
|
|
|
The k8temp(4) driver has been renamed to amdtemp(4) since
|
|
|
|
support for K10 and K11 CPU families was added.
|
|
|
|
|
2009-03-09 17:53:05 +00:00
|
|
|
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.
|
|
|
|
|
2009-03-09 17:05:31 +00:00
|
|
|
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
|
2009-03-09 19:56:37 +00:00
|
|
|
rebuild all ports which uses libusb. More specific directions are given
|
2009-03-09 22:42:01 +00:00
|
|
|
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.
|
2009-03-09 17:05:31 +00:00
|
|
|
|
2009-03-02 18:53:30 +00:00
|
|
|
20090302:
|
2009-03-05 12:04:42 +00:00
|
|
|
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
|
2009-03-02 18:53:30 +00:00
|
|
|
of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
|
2009-05-23 09:24:07 +00:00
|
|
|
wrong for large segments. Note that limits must be explicitly
|
2009-03-02 18:53:30 +00:00
|
|
|
raised to allow such segments to be created.
|
|
|
|
|
2009-03-01 12:44:33 +00:00
|
|
|
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.
|
|
|
|
|
2009-02-23 18:56:49 +00:00
|
|
|
20090223:
|
|
|
|
The new USB2 stack has now been permanently moved in and all kernel and
|
2009-02-23 19:30:00 +00:00
|
|
|
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
|
2009-02-23 18:56:49 +00:00
|
|
|
with the letter 'o', the old usb modules have been removed.
|
2009-03-01 18:57:59 +00:00
|
|
|
Updating entry 20090216 for xorg and 20090215 for libmap may still
|
|
|
|
apply.
|
2009-02-23 18:56:49 +00:00
|
|
|
|
2009-02-17 11:55:50 +00:00
|
|
|
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.
|
|
|
|
|
2009-02-16 18:59:18 +00:00
|
|
|
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.
|
2009-02-16 18:59:18 +00:00
|
|
|
|
2009-02-15 22:33:44 +00:00
|
|
|
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.
|
2009-02-15 22:33:44 +00:00
|
|
|
|
2009-03-01 18:57:59 +00:00
|
|
|
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
|
|
|
|
|
2009-02-06 15:03:14 +00:00
|
|
|
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.
|
|
|
|
|
2009-02-01 21:11:08 +00:00
|
|
|
20090201:
|
|
|
|
INET6 statistics (struct ip6stat) was updated.
|
|
|
|
netstat(1) needs to be recompiled.
|
|
|
|
|
2009-01-19 17:00:42 +00:00
|
|
|
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.
|
|
|
|
|
2009-01-15 06:44:22 +00:00
|
|
|
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.
|
|
|
|
|
2008-12-25 10:05:00 +00:00
|
|
|
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.
|
|
|
|
|
2008-12-19 23:10:55 +00:00
|
|
|
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).
|
2008-12-19 23:10:55 +00:00
|
|
|
|
2008-12-17 00:11:56 +00:00
|
|
|
20081216:
|
2009-02-17 10:50:18 +00:00
|
|
|
The afdata and ifnet locks have been changed from mutexes to
|
2008-12-17 00:11:56 +00:00
|
|
|
rwlocks, network modules will need to be re-compiled.
|
2006-11-15 20:02:20 +00:00
|
|
|
|
This main goals of this project are:
1. separating L2 tables (ARP, NDP) from the L3 routing tables
2. removing as much locking dependencies among these layers as
possible to allow for some parallelism in the search operations
3. simplify the logic in the routing code,
The most notable end result is the obsolescent of the route
cloning (RTF_CLONING) concept, which translated into code reduction
in both IPv4 ARP and IPv6 NDP related modules, and size reduction in
struct rtentry{}. The change in design obsoletes the semantics of
RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland
applications such as "arp" and "ndp" have been modified to reflect
those changes. The output from "netstat -r" shows only the routing
entries.
Quite a few developers have contributed to this project in the
past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and
Andre Oppermann. And most recently:
- Kip Macy revised the locking code completely, thus completing
the last piece of the puzzle, Kip has also been conducting
active functional testing
- Sam Leffler has helped me improving/refactoring the code, and
provided valuable reviews
- Julian Elischer setup the perforce tree for me and has helped
me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
|
|
|
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
|
This main goals of this project are:
1. separating L2 tables (ARP, NDP) from the L3 routing tables
2. removing as much locking dependencies among these layers as
possible to allow for some parallelism in the search operations
3. simplify the logic in the routing code,
The most notable end result is the obsolescent of the route
cloning (RTF_CLONING) concept, which translated into code reduction
in both IPv4 ARP and IPv6 NDP related modules, and size reduction in
struct rtentry{}. The change in design obsoletes the semantics of
RTF_CLONING, RTF_WASCLONE and RTF_LLINFO routing flags. The userland
applications such as "arp" and "ndp" have been modified to reflect
those changes. The output from "netstat -r" shows only the routing
entries.
Quite a few developers have contributed to this project in the
past: Glebius Smirnoff, Luigi Rizzo, Alessandro Cerri, and
Andre Oppermann. And most recently:
- Kip Macy revised the locking code completely, thus completing
the last piece of the puzzle, Kip has also been conducting
active functional testing
- Sam Leffler has helped me improving/refactoring the code, and
provided valuable reviews
- Julian Elischer setup the perforce tree for me and has helped
me maintaining that branch before the svn conversion
2008-12-15 06:10:57 +00:00
|
|
|
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.
|
|
|
|
|
2008-12-01 16:53:01 +00:00
|
|
|
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
|
2008-12-01 16:53:01 +00:00
|
|
|
|
|
|
|
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.
|
|
|
|
|
2008-11-22 05:55:56 +00:00
|
|
|
20081121:
|
2008-12-08 17:12:40 +00:00
|
|
|
__FreeBSD_version 800054 adds memory barriers to
|
|
|
|
<machine/atomic.h>, new interfaces to ifnet to facilitate
|
2008-11-22 05:55:56 +00:00
|
|
|
multiple hardware transmit queues for cards that support
|
|
|
|
them, and a lock-less ring-buffer implementation to
|
2008-12-08 17:12:40 +00:00
|
|
|
enable drivers to more efficiently manage queueing of
|
2008-11-22 05:55:56 +00:00
|
|
|
packets.
|
|
|
|
|
2008-11-18 21:41:09 +00:00
|
|
|
20081117:
|
|
|
|
A new version of ZFS (version 13) has been merged to -HEAD.
|
2008-11-19 00:25:15 +00:00
|
|
|
This version has zpool attribute "listsnapshots" off by
|
|
|
|
default, which means "zfs list" does not show snapshots,
|
2008-11-18 21:41:09 +00:00
|
|
|
and is the same as Solaris behavior.
|
|
|
|
|
2008-10-28 14:14:57 +00:00
|
|
|
20081028:
|
|
|
|
dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
|
|
|
|
|
2008-10-10 06:37:51 +00:00
|
|
|
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"
|
|
|
|
|
2008-10-10 04:23:40 +00:00
|
|
|
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:
|
2008-09-14 19:25:57 +00:00
|
|
|
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:
|
2008-09-14 19:25:57 +00:00
|
|
|
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.
|
|
|
|
|
2008-09-01 23:50:56 +00:00
|
|
|
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.
|
|
|
|
|
2008-10-08 01:31:00 +00:00
|
|
|
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.
|
|
|
|
|
2008-07-13 07:20:14 +00:00
|
|
|
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
|
|
|
|
2008-07-13 07:20:14 +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.
|
|
|
|
|
2008-08-18 10:38:16 +00:00
|
|
|
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.
|
|
|
|
|
2008-06-03 18:09:10 +00:00
|
|
|
20080603:
|
2008-06-03 17:50:13 +00:00
|
|
|
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
|
2008-06-03 17:50:13 +00:00
|
|
|
if it fixes the problem contact emulation mailing list.
|
|
|
|
|
2008-05-26 10:40:09 +00:00
|
|
|
20080525:
|
|
|
|
ISDN4BSD (I4B) was removed from the src tree. You may need to
|
|
|
|
update a your kernel configuration and remove relevant entries.
|
|
|
|
|
2008-05-09 23:14:01 +00:00
|
|
|
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).
|
2008-05-09 23:14:01 +00:00
|
|
|
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.
|
2008-04-27 04:07:36 +00:00
|
|
|
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.
|
|
|
|
|
2008-04-08 17:55:26 +00:00
|
|
|
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.
|
|
|
|
|
2008-03-12 09:48:42 +00:00
|
|
|
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
|
|
|
|
|
2008-03-01 22:54:42 +00:00
|
|
|
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.
|
|
|
|
|
2008-02-29 22:08:49 +00:00
|
|
|
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.
|
|
|
|
|
2008-02-20 07:50:13 +00:00
|
|
|
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.
|
|
|
|
|
2008-02-11 23:23:21 +00:00
|
|
|
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.
|
|
|
|
|
2008-01-23 22:21:36 +00:00
|
|
|
20080123:
|
2008-01-21 22:04:37 +00:00
|
|
|
To upgrade to -current after this date, you must be running
|
2008-01-23 22:21:36 +00:00
|
|
|
FreeBSD not older than 6.0-RELEASE. Upgrading to -current
|
2008-01-21 22:04:37 +00:00
|
|
|
from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
|
|
|
|
|
2007-11-28 13:04:11 +00:00
|
|
|
20071128:
|
|
|
|
The ADAPTIVE_GIANT kernel option has been retired because its
|
|
|
|
functionality is the default now.
|
|
|
|
|
2007-11-18 18:11:16 +00:00
|
|
|
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"
|
|
|
|
|
2007-10-24 20:51:44 +00:00
|
|
|
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.
|
|
|
|
|
2007-10-21 04:27:07 +00:00
|
|
|
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.
|
|
|
|
|
2007-10-11 04:28:08 +00:00
|
|
|
20071010:
|
|
|
|
RELENG_7 branched.
|
|
|
|
|
2000-01-30 23:15:21 +00:00
|
|
|
COMMON ITEMS:
|
|
|
|
|
2000-09-04 03:18:01 +00:00
|
|
|
General Notes
|
|
|
|
-------------
|
2009-09-03 17:04:42 +00:00
|
|
|
Avoid using make -j when upgrading. While generally safe, there are
|
|
|
|
sometimes problems using -j to upgrade. If your upgrade fails with
|
|
|
|
-j, please try again wtihout -j. 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).
|
2000-09-04 03:18:01 +00:00
|
|
|
|
2002-01-26 21:33:07 +00:00
|
|
|
Sometimes, obscure build problems are the result of environment
|
|
|
|
poisoning. This can happen because the make utility reads its
|
2009-09-03 17:04:42 +00:00
|
|
|
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.
|
2005-12-07 20:49:42 +00:00
|
|
|
|
2000-01-30 23:15:21 +00:00
|
|
|
To build a kernel
|
|
|
|
-----------------
|
2000-08-10 05:14:31 +00:00
|
|
|
If you are updating from a prior version of FreeBSD (even one just
|
2005-09-09 15:59:17 +00:00
|
|
|
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
|
2003-10-03 18:29:06 +00:00
|
|
|
make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
|
|
|
|
make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
|
2000-08-10 05:14:31 +00:00
|
|
|
|
2005-05-25 21:03:13 +00:00
|
|
|
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
|
|
|
|
|
2000-08-10 05:14:31 +00:00
|
|
|
To just build a kernel when you know that it won't mess you up
|
|
|
|
--------------------------------------------------------------
|
2009-09-03 17:04:42 +00:00
|
|
|
This assumes you are already running a CURRENT system. Replace
|
2004-04-11 03:30:09 +00:00
|
|
|
${arch} with the architecture of your machine (e.g. "i386",
|
2009-09-03 17:04:42 +00:00
|
|
|
"arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
|
2004-04-11 03:30:09 +00:00
|
|
|
|
|
|
|
cd src/sys/${arch}/conf
|
2002-01-06 20:18:13 +00:00
|
|
|
config KERNEL_NAME_HERE
|
2004-04-11 03:30:09 +00:00
|
|
|
cd ../compile/KERNEL_NAME_HERE
|
2000-08-10 05:14:31 +00:00
|
|
|
make depend
|
|
|
|
make
|
2004-06-22 10:33:58 +00:00
|
|
|
make install
|
2000-08-10 05:14:31 +00:00
|
|
|
|
|
|
|
If this fails, go to the "To build a kernel" section.
|
|
|
|
|
|
|
|
To rebuild everything and install it on the current system.
|
|
|
|
-----------------------------------------------------------
|
2002-10-26 06:25:11 +00:00
|
|
|
# Note: sometimes if you are running current you gotta do more than
|
|
|
|
# is listed here if you are upgrading from a really old current.
|
|
|
|
|
2002-10-26 22:55:43 +00:00
|
|
|
<make sure you have good level 0 dumps>
|
2002-10-26 06:25:11 +00:00
|
|
|
make buildworld
|
2004-08-25 19:39:13 +00:00
|
|
|
make kernel KERNCONF=YOUR_KERNEL_HERE
|
2002-10-26 06:25:11 +00:00
|
|
|
[1]
|
|
|
|
<reboot in single user> [3]
|
|
|
|
mergemaster -p [5]
|
|
|
|
make installworld
|
2005-07-23 14:23:30 +00:00
|
|
|
make delete-old
|
2002-10-26 06:25:11 +00:00
|
|
|
mergemaster [4]
|
|
|
|
<reboot>
|
|
|
|
|
2000-02-06 04:07:11 +00:00
|
|
|
|
2003-02-07 08:26:23 +00:00
|
|
|
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
|
2005-02-23 20:37:11 +00:00
|
|
|
make buildkernel KERNCONF=YOUR_KERNEL_HERE
|
2003-02-07 08:26:23 +00:00
|
|
|
<maybe newfs current's root partition>
|
|
|
|
<mount current's root partition on directory ${CURRENT_ROOT}>
|
|
|
|
make installworld DESTDIR=${CURRENT_ROOT}
|
2006-04-07 11:36:25 +00:00
|
|
|
make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
|
2005-02-23 20:37:11 +00:00
|
|
|
make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
|
2003-02-07 08:26:23 +00:00
|
|
|
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>
|
2008-10-08 01:31:00 +00:00
|
|
|
<maybe install compatibility libraries from ports/misc/compat*>
|
2008-12-08 17:12:40 +00:00
|
|
|
<reboot>
|
2003-02-07 08:26:23 +00:00
|
|
|
|
|
|
|
|
2009-09-03 17:04:42 +00:00
|
|
|
To upgrade in-place from 8.x-stable to current
|
2003-02-07 08:26:23 +00:00
|
|
|
----------------------------------------------
|
2002-10-26 22:55:43 +00:00
|
|
|
<make sure you have good level 0 dumps>
|
2002-10-28 21:33:10 +00:00
|
|
|
make buildworld [9]
|
2004-09-04 21:03:10 +00:00
|
|
|
make kernel KERNCONF=YOUR_KERNEL_HERE [8]
|
2002-10-26 06:21:00 +00:00
|
|
|
[1]
|
|
|
|
<reboot in single user> [3]
|
|
|
|
mergemaster -p [5]
|
2000-06-14 15:42:50 +00:00
|
|
|
make installworld
|
2005-07-23 14:23:30 +00:00
|
|
|
make delete-old
|
2004-06-26 21:58:56 +00:00
|
|
|
mergemaster -i [4]
|
2000-06-14 15:42:50 +00:00
|
|
|
<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.
|
2000-06-14 15:42:50 +00:00
|
|
|
|
2000-08-06 22:16:34 +00:00
|
|
|
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.
|
|
|
|
|
2000-08-10 05:03:49 +00:00
|
|
|
[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.
|
|
|
|
|
2001-06-04 16:17:06 +00:00
|
|
|
[3] From the bootblocks, boot -s, and then do
|
|
|
|
fsck -p
|
|
|
|
mount -u /
|
|
|
|
mount -a
|
2004-08-25 19:39:13 +00:00
|
|
|
cd src
|
2002-01-06 20:18:13 +00:00
|
|
|
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.
|
2001-06-04 16:17:06 +00:00
|
|
|
|
2001-08-11 02:03:51 +00:00
|
|
|
[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.
|
|
|
|
|
2002-04-19 04:22:35 +00:00
|
|
|
[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.
|
|
|
|
|
2009-09-03 17:04:42 +00:00
|
|
|
[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. And so on
|
|
|
|
for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
|
2002-10-27 04:48:31 +00:00
|
|
|
|
2004-09-04 21:03:10 +00:00
|
|
|
Make sure that you merge any new devices from GENERIC since the
|
|
|
|
last time you updated your kernel config file.
|
|
|
|
|
2002-10-28 21:33:10 +00:00
|
|
|
[9] When checking out sources, you must include the -P flag to have
|
2004-09-04 21:03:10 +00:00
|
|
|
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.
|
2000-01-30 23:15:21 +00:00
|
|
|
FORMAT:
|
|
|
|
|
2000-09-04 16:59:32 +00:00
|
|
|
This file contains a list, in reverse chronological order, of major
|
1999-02-14 05:18:35 +00:00
|
|
|
breakages in tracking -current. Not all things will be listed here,
|
2005-03-03 08:43:33 +00:00
|
|
|
and it only starts on October 16, 2004. Updating files can found in
|
2000-03-23 05:57:10 +00:00
|
|
|
previous releases if your system is older than this.
|
1999-02-14 05:18:35 +00:00
|
|
|
|
2001-06-29 06:00:44 +00:00
|
|
|
Copyright information:
|
|
|
|
|
2009-09-03 17:04:42 +00:00
|
|
|
Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
|
2001-06-29 06:00:44 +00:00
|
|
|
|
2001-09-23 06:36:41 +00:00
|
|
|
Redistribution, publication, translation and use, with or without
|
|
|
|
modification, in full or in part, in any form or format of this
|
2002-04-11 05:45:17 +00:00
|
|
|
document are permitted without further permission from the author.
|
2001-06-29 06:00:44 +00:00
|
|
|
|
|
|
|
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.
|
|
|
|
|
2001-09-23 06:44:07 +00:00
|
|
|
Contact Warner Losh if you have any questions about your use of
|
2001-09-23 06:36:41 +00:00
|
|
|
this document.
|
|
|
|
|
1999-08-28 01:35:59 +00:00
|
|
|
$FreeBSD$
|