Planning/3.0: Difference between revisions

From QEMU
No edit summary
 
(22 intermediate revisions by 5 users not shown)
Line 21: Line 21:
|-
|-
| 2018-08-07
| 2018-08-07
| Release, or if necessary, tag rc4
| Tag rc4
|-
|-
| 2018-08-14
| 2018-08-14
| Release, if we needed rc4
| Release
|-
| 2019-04-08
| Freeze for v3.0.1
|-
| 2019-04-11
| Tag v3.0.1
|}
|}


Line 47: Line 53:
* On s390x, maxram is broken for large guests - fix posted: 2018-07-30: "s390x/sclp: fix maxram calculation"
* On s390x, maxram is broken for large guests - fix posted: 2018-07-30: "s390x/sclp: fix maxram calculation"


=== Not yet fixed in any rc ===
=== Fixed in rc4 ===


* QEMU (frontend) crashes upon warm reboot with vga=775 on Linux cmdline: https://bugs.launchpad.net/qemu/+bug/1784900
* assert when using vIOMMU in vhost (vhost: check region type before casting)
* assert when using vIOMMU in vhost (vhost: check region type before casting)
* memory hotplug regression when using two numa nodes
* corrupt SRAT when using numa node without memory (not a regression)
* hw/misc/macio: Fix device introspection problems in macio devices
* sam460ex: Fix PCI interrupts with multiple devices (not a regression)
* sam460ex: Fix PCI interrupts with multiple devices (not a regression)
* hw/misc/macio: Fix device introspection problems in macio devices
* bad TCG codegen for dup_vec: "[PATCH] tcg/optimize: Do not skip default processing of dup_vec"
* ICMP handling on OSX hosts (not a regression)
* Arm migration of v8M CPUs is broken (not quite a regression)
* corrupt SRAT when using numa node without memory (not a regression)  
* Arm migration of KVM GICv3 is broken
* memory hotplug regression when using two numa nodes
* xtensa introspection bug: https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg03965.html
* heap overflow in slirp: https://bugs.launchpad.net/qemu/+bug/1785670
* QEMU (frontend) crashes upon warm reboot with vga=775 on Linux cmdline: https://bugs.launchpad.net/qemu/+bug/1784900
 
=== Not yet fixed in any rc ===


=== Issues that will not be fixed ===
=== Issues that will not be fixed ===
* monitor: fix oob command leak (patch on ML, regression, but only for monitor OOB mode which is still experimental and disabled by default)
* RFC: fix megasas leak (patch on ML, may not be a regression)
* qemu NBD server kills connection on NBD_CMD_BLOCK_STATUS when serving non-sector-aligned file (not a regression)
* ICMP handling on OSX hosts (not a regression)
* qemu_kill() gives poor log results when tests fail with core dump (not a regression; debug aid)
* monitor cleanup dead-locks when connected to a Spice port: "[PATCH] monitor: avoid potential dead-lock when cleaning up" (not a regression)
* A couple of minor fixes for structured load/stores and rounding cases for [[Features/ARM/SVE|SVE]] will be in 3.0.1


== Targeted Features ==
== Targeted Features ==


See the [[ChangeLog/3.0]] for full details.
See the [[ChangeLog/3.0]] for full details.

Latest revision as of 20:58, 1 April 2019

Release Schedule

2018-04-24 Beginning of development phase
2018-07-03 Soft feature freeze. Only bug fixes after this point. All feature changes must be already in a sub maintainer tree and all pull requests from submaintainers must have been sent to the list by this date.
2018-07-10 Hard feature freeze. Tag rc0
2018-07-17 Tag rc1
2018-07-24 Tag rc2
2018-07-31 Tag rc3
2018-08-07 Tag rc4
2018-08-14 Release
2019-04-08 Freeze for v3.0.1
2019-04-11 Tag v3.0.1

Known issues

Please list all patch series or bugs which need to be fixed for the release here. That way I can be sure I don't miss any when I'm rolling release candidates.

Fixed in rc0

Fixed in rc1

  • Crash with: qemu-system-aarch64 -M integratorcp,accel=qtest -nographic and then at the HMP monitor: device_add bcm2837 followed by device_add corgi-ssp

Fixed in rc2

Fixed in rc3

  • Valgrind complains about uninitialized value here: valgrind -q aarch64-softmmu/qemu-system-aarch64 -M mps2-an505,accel=qtest - fix posted: 2018-07-24 "hw/misc/tz-mpc: Zero the LUT on initialization, not just reset"
  • migration crash with -device hda-duplex - fix posted: 2018-07-24: audio/hda: Fix migration
  • On s390x, maxram is broken for large guests - fix posted: 2018-07-30: "s390x/sclp: fix maxram calculation"

Fixed in rc4

  • assert when using vIOMMU in vhost (vhost: check region type before casting)
  • memory hotplug regression when using two numa nodes
  • corrupt SRAT when using numa node without memory (not a regression)
  • hw/misc/macio: Fix device introspection problems in macio devices
  • sam460ex: Fix PCI interrupts with multiple devices (not a regression)
  • bad TCG codegen for dup_vec: "[PATCH] tcg/optimize: Do not skip default processing of dup_vec"
  • Arm migration of v8M CPUs is broken (not quite a regression)
  • Arm migration of KVM GICv3 is broken
  • xtensa introspection bug: https://lists.gnu.org/archive/html/qemu-devel/2018-07/msg03965.html
  • heap overflow in slirp: https://bugs.launchpad.net/qemu/+bug/1785670
  • QEMU (frontend) crashes upon warm reboot with vga=775 on Linux cmdline: https://bugs.launchpad.net/qemu/+bug/1784900

Not yet fixed in any rc

Issues that will not be fixed

  • monitor: fix oob command leak (patch on ML, regression, but only for monitor OOB mode which is still experimental and disabled by default)
  • RFC: fix megasas leak (patch on ML, may not be a regression)
  • qemu NBD server kills connection on NBD_CMD_BLOCK_STATUS when serving non-sector-aligned file (not a regression)
  • ICMP handling on OSX hosts (not a regression)
  • qemu_kill() gives poor log results when tests fail with core dump (not a regression; debug aid)
  • monitor cleanup dead-locks when connected to a Spice port: "[PATCH] monitor: avoid potential dead-lock when cleaning up" (not a regression)
  • A couple of minor fixes for structured load/stores and rounding cases for SVE will be in 3.0.1

Targeted Features

See the ChangeLog/3.0 for full details.