d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Remove entries that:

Browse Source 
- exist in a release [1]
 - do not have a corresponding revision class

[1] This is done to make ensure for 11.0-RELEASE, we do not end
up with entries from older releases, like happened with 10.0-R.

Approved by:	re (implicit)
Sponsored by:	The FreeBSD Foundation
This commit is contained in:
Glen Barber 2014-01-20 21:50:00 +00:00
parent 54f9beb117
commit 683e1de39d
Notes: svn2git 2020-12-20 02:59:44 +00:00 
svn path=/head/; revision=260928
1 changed files with 11 additions and 241 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

252
release/doc/en_US.ISO8859-1/relnotes/article.xml
View File

@ -104,132 +104,28 @@
<sect2 xml:id="kernel">
<title>Kernel Changes</title>
<para revision="248508">The use of unmapped VMIO buffers
eliminates the need to perform TLB shootdown for mapping on
buffer creation and reuse, greatly reducing the amount of IPIs
for shootdown on big-SMP machines and eliminating up to 25-30%
of the system time on i/o intensive workloads.</para>
<para arch="amd64" revision="254466">The maximum amount of memory
the &os; kernel can address has been increased from 1TB to
4TB.</para>
<para role="merged">The &man.ddb.4; kernel debugger now has an
output capture facility. Input and output from &man.ddb.4; can
now be captured to a memory buffer for later inspection using
&man.sysctl.8; or a textdump. The new
<command>capture</command> command controls this feature.</para>
<para role="merged">The &man.ddb.4; debugger now supports a simple
scripting facility, which supports a set of named scripts
consisting of a set of &man.ddb.4; commands. These commands can
be managed from within &man.ddb.4; or with the use of the new
&man.ddb.8; utility. More details can be found in the
&man.ddb.4; manual page.</para>
<para arch="amd64,i386" revision="240135">Support was added for
the new Intel on-CPU Bull Mountain random number generator,
found on IvyBridge and supposedly later CPUs, accessible with
RDRAND instruction.</para>
<para>&nbsp;</para>
<sect3 xml:id="kernel-virtualization">
<title>Virtualization support</title>
<para arch="amd64" revision="245652">The BSD Hypervisor,
&man.bhyve.8; is included with &os;. &man.bhyve.8; requires
Intel CPUs with VT-x and Extended Page Table (EPT) support.
These features are on all Nehalem models and beyond (e.g.
Nehalem and newer), but not on the lower-end Atom CPUs.</para>
<para revision="227652">&man.virtio.4; support has been added.
&man.virtio.4; is the name for the paravirtualization
interface developed for the Linux KVM, but since adopted to
other virtual machine hypervisors (with the notable exception
of Xen). This work brings in a BSD-licensed clean-room
implementation of the virtio kernel drivers for disk IO
(&man.virtio_blk.4; and &man.virtio_scsi.4;), network IO
(&man.vtnet.4;), memory ballooning (&man.virtio_balloon.4;),
and PCI. Tested with on Qemu/KVM, VirtualBox, and
&man.bhyve.4;.</para>
<para arch="amd64,i386" revision="255524">Paravirtualized
drivers which support Microsoft Hyper-V have been imported and
made part of the amd64 GENERIC kernel. For i386, these
drivers are not part of GENERIC, so the following lines must
be added to <filename>/boot/loader.conf</filename> to load
these drivers:</para>
<programlisting>hv_ata_pci_disengage_load="YES"
hv_netsvc_load="YES"
hv_utils_load="YES"
hv_vmbus_load="YES"</programlisting>
<para>Alternatively, the Hyper-V drivers can be added to the
i386 kernel by adding <literal>device hyperv</literal> to the
kernel config, and then recompiling the kernel. Please refer
to: <link xlink:href="http://wiki.freebsd.org/HyperV">FreeBSD
and Microsoft Windows Server Hyper-V support</link> for full
instructions on how to set up Hyper-V support under
FreeBSD.</para>
<para revision="254738">The &man.vmx.4; driver has been added.
&man.vmx.4; is a VMware VMXNET3 ethernet driver ported from
OpenBSD.</para>
<para revision="255744" arch="amd64,i386">Xen PVHVM
virtualization is now part of the GENERIC kernel.</para>
<para>&nbsp;</para>
</sect3>
<sect3 xml:id="kernel-arm">
<title>ARM support</title>
<para revision="239922">Raspberry PI support has been added.
Refer to these <link
xlink:href="http://kernelnomicon.org/?p=164">setup
instructions</link> and <link
xlink:href="http://www.raspberrypi.org/quick-start-guide">quick
start guide</link>.</para>
<para revision="253396">The default ABI on ARM is now the ARM
EABI. This brings a number of improvements and allows future
support for VFP and Thumb-2.</para>
<para revision="239268">ARM support has been greatly improved,
including support for ARMv6 and ARMv7, SMP and thread-local
storage (TLS). Additionally support for some newer SoC like
the MV78x60 and OMAP4 was added. See <link
xlink:href="http://lists.freebsd.org/pipermail/freebsd-arm/2012-August/003757.html">this
announcement</link> for further details.</para>
<para revision="254918">Superpages support on ARM has been
added. Superpages support provides improved performance and
scalability by allowing TLB translations to dynamically cover
large physical memory regions. All ARMv6 and ARMv7-based
platforms can take advantage of this feature. See <link
xlink:href="https://wiki.freebsd.org/ARMSuperpages">this
page</link> for further details.</para>
<para>&nbsp;</para>
</sect3>
<!--
<sect3 xml:id="boot">
<title>Boot Loader Changes</title>
<para arch="amd64,i386" role="merged">The BTX kernel used by the
boot loader has been changed to invoke BIOS routines from real
mode. This change makes it possible to boot &os; from USB
devices.</para>
<para arch="amd64,i386" role="merged">A new gptboot boot loader
has been added to support booting from a GPT labeled disk.
A new <command>boot</command> command has been added to
&man.gpt.8;, which makes a GPT disk bootable by writing the
required bits of the boot loader, creating a new boot
partition if required.</para>
<para>&nbsp;</para>
</sect3>
-->
<sect3 xml:id="proc">
<title>Hardware Support</title>
@ -237,168 +133,42 @@ hv_vmbus_load="YES"</programlisting>
<sect4 xml:id="mm">
<title>Multimedia Support</title>
<para revision="240609">Support for version 2.0 of the USB
Audio reference design has been added. New devices should
support higher bandwidth, increased sampling frequency and
wider dynamic range.</para>
<para>&nbsp;</para>
</sect4>
<sect4 xml:id="net-if">
<title>Network Interface Support</title>
<para>The firmware for the &man.mxge.4; driver has been
1.4.55.</para>
<para>&nbsp;</para>
<para>The &man.re.4; driver has been overhauled to fix a
number of issues. This driver now has Wake On LAN (WOL)
support.</para>
<para>The &man.vr.4; driver has been overhauled to fix a
number of outstanding issues. It also now works on all
architectures.</para>
<para arch="amd64,i386" role="merged">The &man.wpi.4; driver
has been updated to include a number of stability
fixes.</para>
<para revision="248925">The &man.cxgbe.4; driver has been
updated to support 40G/10G Ethernet NICs based on Chelsio's
Terminator 5 (T5) ASIC.</para>
<para revision="256694">The iw_cxgbe driver has been added.
This is an experimental iWARP/RDMA driver (kernel verbs
only) for Chelsio's T4 and T5 based cards.</para>
<para revision="255932">The Open Fabrics Enterprise
Distribution (OFED) and OFED Infiniband core has been
updated to the same version as supplied by Linux version
3.7</para>
<para revision="255932">The Mellanox Infiniband driver has
been updated to firmware version 2.30.3200 for ConnectX3
NICs. Support has been added for ConnectX3 VPI NICs, where
each port can be used as Infiniband 56 GB/s or Ethernet 40
GB/s. Support has been added for dynamically loading kernel
modules for Infiniband core (ibcore) and IP over Infiniband
(ipoib).</para>
<para revision="227614">&man.netmap.4; has been added.
&man.netmap.4; is a framework for high-performance
direct-to-hardware packet IO, offering low latency and high
PPS rates to userland applications while bypassing any
kernel-side packet processing. With &man.netmap.4; it is
trivially possible to fully saturate a 10 Gbps network
interface with minimal packet sizes. For more information,
see: <link
xlink:href="http://info.iet.unipi.it/~luigi/netmap/">Netmap
Project</link>.</para>
</sect4>
</sect3>
<sect3 xml:id="net-proto">
<title>Network Protocols</title>
<para revision="228571">&man.carp.4; has been rewritten to make
addresses more sane from the viewpoint of routing daemons such
as quagga/zebra. It also brings support for a single redundant
address on the subnet (carpdev), switching state with
&man.ifconfig.8;, better locking and using modern kernel
interfaces to allocate multicast memberships. Configuration
of the CARP protocol via &man.ifconfig.8; has changed, as well
as format of CARP events submitted to &man.devd.8; has
changed. See &man.carp.4; for more information. The arpbalance
feature of &man.carp.4; is currently not supported
anymore.</para>
<para>&nbsp;</para>
<para revision="240233">The &man.pf.4; firewall now supports
fine-grain locking and better utilization on multi-cpu
machines resulting in significant improvements in
performance.</para>
<para revision="250700">Support for up to 65536 routing tables
has been introduced.</para>
<para revision="248552">Support for setting/matching
differentiated services codepoints (DSCP) in IP header has
been added to &man.ipfw.8;.</para>
</sect3>
<sect3 xml:id="disks">
<title>Disks and Storage</title>
<para role="merged">The &man.aac.4; driver now supports volumes
larger than 2TB in size.</para>
<para>&nbsp;</para>
<para>The &man.ata.4; driver now supports a spindown command for
disks; after a configurable amount of time, if no requests
have been received for a disk, the disk will be spun down
until the next request. The &man.atacontrol.8; utility now
supports a <command>spindown</command> command to configure
this feature.</para>
<para role="merged">The &man.hptrr.4; driver has been updated to
version 1.2 from Highpoint.</para>
<para revision="240616">&man.nvme.4; has been added and provides
NVM Express support. NVM Express is an optimized register
interface, command set and feature set of PCI Express
(PCIe)-based Solid-State Drives (SSDs). For more information,
see <link
xlink:href="http://http://www.nvmexpress.org/">nvmexpress.org</link>.</para>
</sect3>
<sect3 xml:id="fs">
<title>File Systems</title>
<para revision="255570">A new kernel-based iSCSI target and
initiator has been added.</para>
<para revision="243246">UFS filesystems can now be enlarged with
&man.growfs.8; while mounted read-write. This is especially
useful for virtual machines, allowing the addition of more
harddrive space without interruption of service.</para>
<para revision="241519">A state of the art FUSE implementation
is now part of the base system. It allows the use of nearly
all fusefs file systems.</para>
<para>&nbsp;</para>
<sect4 xml:id="fs-zfs">
<title>ZFS</title>
<para revision="256361">&man.bsdinstall.8; now supports
installing ZFS on the root file system. It includes
a single configuration menu that allows you to select all of
the required details, including which drives to use, what
ZFS RAID level to use (taking into consideration the
selected number of drives), GPT or MBR, GELI encryption,
forcing 4K sectors, pool name, etc.</para>
<para>&nbsp;</para>
<para revision="240868">TRIM support has been added for
ZFS.</para>
<para revision="246586">Support for the high performance LZ4
compression algorithm has been added to ZFS. LZ4 is usually
faster and can achieve a higher compression ratio than LZJB,
the default compression algorithm.</para>
<para revision="252140">Support for L2ARC compression has been
added to ZFS.</para>
<para revision="243524">The zio nop-write improvement from
Illumos was imported into &os;. To reduce I/O, nop-write
skips overwriting data if the checksum (cryptographically
secure) of new data matches the checksum of existing data.
It also saves space if snapshots are in use. This
improvement only works only on datasets with enabled
compression, disabled deduplication and sha256
checksums.</para>
<para>ZFS will now compare the checksums of incoming writes to
the checksum of the existing on-disk data and avoid issuing
any write I/O for data that has not changed. This will
reduce I/O as well as space usage because if the old block
is referenced by a snapshot, both copies of the block are
kept even though both contain the same data.</para>
</sect4>
</sect3>
</sect2>