CHANGELOG.md 6.36 KB
Newer Older
Paul Durrant's avatar
Paul Durrant committed
1 2 3 4 5 6
# Changelog

Notable changes to Xen will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)

7 8
## [unstable UNRELEASED](https://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=staging) - TBD

9 10 11 12
### Changed
 - On x86 "vga=current" can now be used together with GrUB2's gfxpayload setting. Note that
   this requires use of "multiboot2" (and "module2") as the GrUB commands loading Xen.

13
### Removed / support downgraded
14
 - dropped support for the (x86-only) "vesa-mtrr" and "vesa-remap" command line options
15

16
## [4.16.0](https://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=staging) - 2021-12-02
17

18 19 20 21
### Removed
 - XENSTORED_ROOTDIR environment variable from configuartion files and
   initscripts, due to being unused.

22 23 24 25 26 27
### Changed
 - Quarantining of passed-through PCI devices no longer defaults to directing I/O to a scratch
   page, matching original post-XSA-302 behavior (albeit the change was also backported, first
   appearing in 4.12.2 and 4.11.4). Prior (4.13...4.15-like) behavior can be arranged for
   either by enabling the IOMMU_QUARANTINE_SCRATCH_PAGE setting at build (configuration) time
   or by passing "iommu=quarantine=scratch-page" on the hypervisor command line.
28 29
 - pv-grub stubdoms will no longer be built per default. In order to be able to use pv-grub
   configure needs to be called with "--enable-pv-grub" as parameter.
30 31 32
 - qemu-traditional based device models (both, qemu-traditional and ioemu-stubdom) will
   no longer be built per default. In order to be able to use those, configure needs to
   be called with "--enable-qemu-traditional" as parameter.
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
 - Fixes for credit2 scheduler stability in corner case conditions.
 - Ongoing improvements in the hypervisor build system.
 - vtpmmgr miscellaneous fixes in preparation for TPM 2.0 support.
 - 32bit PV guests only supported in shim mode.
 - Improved PVH dom0 debug key handling.
 - Fix booting on some Intel systems without a PIT (i8254).
 - Cleanup of the xenstore library interface.
 - Fix truncation of return value from xencall2 by introducing a new helper
   that returns a long instead.
 - Fix system register accesses on Arm to use the proper 32/64bit access size.
 - Various fixes for Arm OP-TEE mediator.
 - Switch to domheap for Xen page tables.

### Added
 - 32bit Arm builds to the gitlab-ci automated tests.
 - x86 full system tests to the gitlab-ci automated tests.
 - Arm limited vPMU support for guests.
 - Static physical memory allocation for dom0less on arm64.
 - dom0less EFI support on arm64.
 - GICD_ICPENDR register handling in vGIC emulation to support Zephyr OS.
 - CPU feature leveling on arm64 platform with heterogeneous cores.
 - Report unpopulated memory regions safe to use for external mappings, Arm and
   device tree only.
 - Support of generic DT IOMMU bindings for Arm SMMU v2.
 - Limit grant table version on a per-domain basis.
58

59
## [4.15.0](https://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=RELEASE-4.15.0) - 2021-04-08
60

61
### Added / support upgraded
62 63 64
 - ARM IOREQ servers (device emulation etc.) (Tech Preview)
 - Renesas IPMMU-VMSA (Supported, not security supported; was Tech Preview)
 - ARM SMMUv3 (Tech Preview)
65
 - Switched MSR accesses to deny by default policy.
66
 - Intel Processor Trace support (Tech Preview)
67
 - Named PCI devices for xl/libxl
68
 - Improved documentation for xl PCI configuration format
69
 - Support for zstd-compressed dom0 (x86) and domU kernels
70 71 72
 - EFI: Enable booting unified hypervisor/kernel/initrd/DT images
 - Reduce ACPI verbosity by default
 - Add ucode=allow-same option to test late microcode loading path
73
 - Library improvements from NetBSD ports upstreamed
74 75 76
 - CI loop: Add Alpine Linux, Ubuntu Focal targets; drop CentOS 6
 - CI loop: Add qemu-based dom0 / domU test for ARM
 - CI loop: Add dom0less aarch64 smoke test
77
 - x86: Allow domains to use AVX-VNNI instructions
78
 - Factored out HVM-specific shadow code, improving code clarity and reducing the size of PV-only hypervisor builds
79
 - Added XEN_SCRIPT_DIR configuration option to specify location for Xen scripts, rather than hard-coding /etc/xen/scripts
80
 - xennet: Documented a way for the backend (or toolstack) to specify MTU to the frontend
81
 - xenstore can now be live-updated on a running system. (Tech preview)
82
 - Some additional affordances in various xl subcommands.
83
 - Added workarounds for the following ARM errata: Cortex A53 #843419, Cortex A55 #1530923, Cortex A72 #853709, Cortex A73 #858921, Cortex A76 #1286807, Neoverse-N1 #1165522
84
 - On detecting a host crash, some debug key handlers can automatically triggered to aid in debugging
George Dunlap's avatar
George Dunlap committed
85
 - Increase the maximum number of guests which can share a single IRQ from 7 to 16, and make this configurable with irq-max-guests
86

87
### Removed / support downgraded
88 89 90 91 92

 - qemu-xen-traditional as host process device model, now "No security
   support, not recommended".  (Use as stub domain device model is still
   supported - see SUPPORT.md.)

93
## [4.14.0](https://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=RELEASE-4.14.0) - 2020-07-23
Paul Durrant's avatar
Paul Durrant committed
94 95 96

### Added
 - This file and MAINTAINERS entry.
97 98 99 100 101 102
 - Use x2APIC mode whenever available, regardless of interrupt remapping
   support.
 - Performance improvements to guest assisted TLB flushes, either when using
   the Xen hypercall interface or the viridian one.
 - Assorted pvshim performance and scalability improvements plus some bug
   fixes.
103 104
 - Hypervisor framework to ease porting Xen to run on hypervisors.
 - Initial support to run on Hyper-V.
105
 - Initial hypervisor file system (hypfs) support.
106
 - libxl support for running qemu-xen device model in a linux stubdomain.
107 108
 - New 'domid_policy', allowing domain-ids to be randomly chosen.
 - Option to preserve domain-id across migrate or save+restore.
109
 - Support in kdd for initial KD protocol handshake for Win 7, 8 and 10 (64 bit).
110 111 112 113 114 115 116 117 118
 - Tech preview support for Control-flow Execution Technology, with Xen using
   Supervisor Shadow Stacks for its own protection.

### Changed
 - The CPUID data seen by a guest on boot is now moved in the migration
   stream.  A guest migrating between non-identical hardware will now no
   longer observe details such as Family/Model/Stepping, Cache, etc changing.
   An administrator still needs to take care to ensure the features visible to
   the guest at boot are compatible with anywhere it might migrate.
Paul Durrant's avatar
Paul Durrant committed
119 120 121 122

## [4.13.0](https://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=RELEASE-4.13.0) - 2019-12-17

> Pointer to release from which CHANGELOG tracking starts