Hi all,
I am adding Sarah Conway from the Linux Foundation, who will handle PR related matters on behalf of the project. I am also summarizing the discussion on the thread (based on your feedback)
* Non-Linux driver domains [see last e-mail ... this needs to be refined]
* Event channel scalability. Event channels are per-VM resources that allow VMs to communicate with each other. These were previously
limited to either 1024 or 4096 channels per domain. Domain 0 needs several event channels for each guest VM, which limited the total number of VMs available to several hundred. The new event channel implementation allows hundreds of thousands of event channels,
removing this as a limit on the number of VMs which can be started. This benefits cloud operating systems such as MirageOS, ErlangOnXen, OSv, HalVM, ... as well as disaggregated Xen systems in which drivers, services (e.g. Qemu, Tor, ...) and other functionality
that would normally be run in Domain 0 can be run in a separate VM.
* Experimental support for PVH mode for guests. PVH mode combines the best elements of HVM and PV into a mode which allows Xen to take
advantage of many of the hardware virtualization features without needing the overhead of simulating devices of a physical computer. This
will allow for increased efficiency, as well as reduced footprint in Linux and FreeBSD going forward.
More information on PVH: see
https://www.linux.com/news/enterprise/systems-management/658784-the-spectrum-of-paravirtualization-with-xen-part-2 and
* Improved support for SPICE. SPICE is a protocol for virtial desktops which allows a much richer connection than display-only
protocols like VNC. Xen 4.4 adds support for additional SPICE functionality, including vdagent, clipboard sharing, and USB redirection.
* GRUB 2 now supports PV xen images (external). In the past, Xen required a custom implementation of GRUB called pvgrub. The upstream GRUB 2 (see
http://www.gnu.org/software/grub/) project now has a build target which will construct a bootable PV xen image. This ensures 100% GRUB 2
compatibility for pvgrub going forward.
* Indirect descriptors for block PV protocol (Linux). Modern storage devices work much better with larger chunks of data. Indirect
descriptors have allowed the size of each individual request to triple, greatly improving I/O performance when running on fast storage
technologies like SSD and RAID. This support is available in any guest running Linux 3.11 or higher (regardless of Xen version).
* Improved kexec for debug support. kexec functionality is primarily used when a crash happens, to allow a special kernel to come in afterwards and collect information about the cause of the crash, to allow developers to diagnose and fix the root cause.
* Improved XAPI and Mirage OS support in Xen. XAPI and Mirage OS are sub-projects within the Xen Project written in OCaml. Both are also used in XenServer (see
http://xenserver.org/) and rely on the Xen OCaml language bindings to operate well. These language bindings have had a major overhaul, and result in much better compatibility between XAPI, Mirage OS and Linux distros going
forward.
* Experimental support for Guest EFI boot. EFI is the new booting standard that is replacing BIOS. Some operating systems only boot
with EFI; and some features, like SecureBoot, only work with EFI.
* Improved ARM support for Xen. [TODO: clarify whether support has moved from tech preview to experimental or supported]. A number of new features have been implemented:
** 64 bit Xen on ARM now supports booting guests
** Physical disk partitions and LVM volumes can now be used to store guest images using xen-blkback (or is PV drivers better in terms of terminology)
** Significant stability improvements across the board
** ARM/multiboot booting protocol design and implementation in Xen and U-boot
** PSCI support in Xen
** Same DMA in Dom0 even with no hardware IOMMUs (not sure what the implications of this are)
** ARM and ARM64 ABIs in Xen are declared stable and maintained for backwards compatibility
** Significant usability improvements, such as automatic creation of guest device trees and improved handling of host DTBs.
** Adding new hardware platforms to Xen on ARM has been vastly improved, making it easier for Hardware vendors and embedded vendors to port Xen on ARM to their board.
** Xen on ARM now supports the Arndale board, Calxeda ECX-2000 (aka Midway), Applied Micro X-Gene Storm, TI OMAP5 and
Allwinner A20/A30 boards [TODO: check with APM whether we can use this in a press release and whether there is more than Mustang support now].
** ARM server class hardware (Calxeda Midway) has been introduced in the Xen OSSTest automated testing framework.
* Updated to qemu 1.6 and SeaBIOS 1.7.3.1
Does anyone want to add anything?
Best Regards
Lars
On 17/12/2013 09:55, Dario Faggioli wrote:
On lun, 2013-12-16 at 13:23 +0000, Ian Campbell wrote:
There have been a tonne of other changes to Xen on ARM since 4.3 but
perhaps they are mainly under the hood, Stefano any thoughts?
FWIW, I think that, considering...
Some things which spring to mind:
* We now generate guest device trees automatically, instead of
requiring the user to supply one.
...this affects users, and does it positively, as it simplifies their
life, and this...
* It is vastly simpler to add a new platform in 4.4.
...Affects developers/adopters, we should really mention at least these
twos.
Regards,
Dario
_______________________________________________
Publicity mailing list
Publicity@xxxxxxxxxxxxxxxxxxxx
http://lists.xenproject.org/cgi-bin/mailman/listinfo/publicity
