Planning/5.2: Difference between revisions
(19 intermediate revisions by 5 users not shown) | |||
Line 21: | Line 21: | ||
|- | |- | ||
| 2020-12-01 | | 2020-12-01 | ||
| Release; or tag rc4 if needed | | <strike>Release; or tag rc4 if needed</strike> Tag rc4 | ||
|- | |- | ||
| 2020-12-08 | | 2020-12-08 | ||
| Release if we needed an rc4 | | <strike>Release if we needed an rc4</strike> Release | ||
|} | |} | ||
Line 31: | Line 31: | ||
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. | 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 | === Fixed in rc1 === | ||
* | * s390 is failing --without-default-devices build (https://travis-ci.com/gitlab/bonzini/qemu/jobs/426142849) | ||
** fixed: | ** fixed in -rc1 via https://git.qemu.org/?p=qemu.git;a=commit;h=77280d33bc9cfdbfb5b5d462259d644f5aefe9b3 | ||
* SD command ACMD41 misbehaves (https://bugs.launchpad.net/qemu/+bug/1809291) | |||
** fixed: 89c6700fe7e ("hw/sd/sdcard: Zero out function selection fields before being populated") | |||
=== Fixed in rc2 === | === Fixed in rc2 === | ||
Line 48: | Line 50: | ||
** fixed: 575094b786e ("hw/sd: Fix 2 GiB card CSD register values") | ** fixed: 575094b786e ("hw/sd: Fix 2 GiB card CSD register values") | ||
=== Fixed in | === Fixed in rc3 === | ||
* Regression on POWER9 when using VSMT breaks guests (https://bugs.launchpad.net/qemu/+bug/1900241) | |||
** fixed: 6d24795ee7e3 ("Revert series "spapr/xive: Allocate vCPU IPIs from the vCPU contexts"") | |||
* net: Use correct default-path macro for downscript | |||
** fixed: 9925990d01a ("net: Use correct default-path macro for downscript") | |||
* "make check" breakage because of recently added "local" tests for 9pfs | |||
** fix: https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg05794.html | |||
* qcow2: Regression that potentially causes corruption on write_zeroes with MAY_UNMAP | |||
** fix: https://lists.gnu.org/archive/html/qemu-block/2020-11/msg00674.html | |||
* set_pci_host_devaddr() fails parsing bus > 0x1f | |||
** fix: https://www.mail-archive.com/qemu-devel@nongnu.org/msg761406.html | |||
* pci broken for s390x-on-x86 tcg guests, bisected to 28dc86a07299 ("s390x/pci: use a PCI Group structure") (https://lore.kernel.org/qemu-devel/20201117124349.2e1e2fe3.cohuck@redhat.com/) | |||
** fix: https://lore.kernel.org/qemu-devel/20201118104202.1301363-1-cohuck@redhat.com/ | |||
===Fixed in rc4 === | |||
* | * nsis: Fix build for 64 bit installer | ||
* | * drop setuptools dependency requirement | ||
* | * meson: use dependency() to find libjpeg | ||
* | * "[PATCH] qxl: fix segfault" regression fix | ||
* slirp: update to fix CVE-2020-29129 CVE-2020-29130 | |||
* ide: atapi: assert that the buffer pointer is in range | |||
=== | === Will not be fixed in 5.2 === | ||
* microblaze s3adsp1800 stuck while booting (https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg02183.html) - not a recent regression, also happens with v5.0 already | |||
* error condition in hash_tbl_search_fd_by_ifid (not a regression) | |||
** fix: https://www.mail-archive.com/qemu-devel@nongnu.org/msg755553.html | |||
* pica61 -device isa-vga aborts (RAMBlock "vga.vram" already registered) - not a recent regression, thus this should likely not block the release | * pica61 -device isa-vga aborts (RAMBlock "vga.vram" already registered) - not a recent regression, thus this should likely not block the release | ||
* | * qga: Correct loop count in qmp_guest_get_vcpus (not a regression) | ||
* | ** proposed fix: https://www.mail-archive.com/qemu-devel@nongnu.org/msg761395.html | ||
* | * HMP netdev_add doesn't let you reuse the ID of a deleted netdev, QMP netdev_add fails to reject duplicate ID (not a regression for 5.2 as error was already released in 5.0) | ||
** proposed fix: https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg06188.html | |||
* ide/atapi: check io_buffer_index in ide_atapi_cmd_reply_end | |||
** https://www.mail-archive.com/qemu-devel@nongnu.org/msg760877.html -- patch is not the correct fix; in rc4 we put in an assert so that the worst-case is just "QEMU asserts" rather than reading off the end of the buffer | |||
=== Not fixed yet === | |||
== Targeted Features == | == Targeted Features == | ||
See the [[ChangeLog/5.2]] for full details. | See the [[ChangeLog/5.2]] for full details. |
Latest revision as of 10:12, 3 December 2020
Release Schedule
2020-08-12 | Beginning of development phase |
2020-10-27 | 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. |
2020-11-03 | Hard feature freeze. Tag rc0 |
2020-11-10 | Tag rc1 |
2020-11-17 | Tag rc2 |
2020-11-24 | Tag rc3 |
2020-12-01 | |
2020-12-08 |
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 rc1
- s390 is failing --without-default-devices build (https://travis-ci.com/gitlab/bonzini/qemu/jobs/426142849)
- SD command ACMD41 misbehaves (https://bugs.launchpad.net/qemu/+bug/1809291)
- fixed: 89c6700fe7e ("hw/sd/sdcard: Zero out function selection fields before being populated")
Fixed in rc2
- ghes migration broken (looped vmstate https://lists.gnu.org/archive/html/qemu-devel/2020-10/msg05187.html )
- graphic_hw_update() may crash with -nodefaults fixed
- Xen support no longer built on arm64 fixed
- Crash with "qemu-system-ppc64 -M powernv -device macio-newworld" or "qemu-system-ppc64 -M mac99 -device macio-oldworld" (fix is available here: https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg02630.html)
- fixed: 6bfa035ec31 ("macio: set user_creatable to false in macio_class_init")
- migration incompatibility of nec-usb-xhci ( https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg03128.html )
- fixed: 172bc8520db ("xhci: fix nec-usb-xhci properties")
- 2GB SD card announced as 1GB (https://www.mail-archive.com/qemu-devel@nongnu.org/msg753834.html)
- fixed: 575094b786e ("hw/sd: Fix 2 GiB card CSD register values")
Fixed in rc3
- Regression on POWER9 when using VSMT breaks guests (https://bugs.launchpad.net/qemu/+bug/1900241)
- fixed: 6d24795ee7e3 ("Revert series "spapr/xive: Allocate vCPU IPIs from the vCPU contexts"")
- net: Use correct default-path macro for downscript
- fixed: 9925990d01a ("net: Use correct default-path macro for downscript")
- "make check" breakage because of recently added "local" tests for 9pfs
- qcow2: Regression that potentially causes corruption on write_zeroes with MAY_UNMAP
- set_pci_host_devaddr() fails parsing bus > 0x1f
- pci broken for s390x-on-x86 tcg guests, bisected to 28dc86a07299 ("s390x/pci: use a PCI Group structure") (https://lore.kernel.org/qemu-devel/20201117124349.2e1e2fe3.cohuck@redhat.com/)
Fixed in rc4
- nsis: Fix build for 64 bit installer
- drop setuptools dependency requirement
- meson: use dependency() to find libjpeg
- "[PATCH] qxl: fix segfault" regression fix
- slirp: update to fix CVE-2020-29129 CVE-2020-29130
- ide: atapi: assert that the buffer pointer is in range
Will not be fixed in 5.2
- microblaze s3adsp1800 stuck while booting (https://lists.gnu.org/archive/html/qemu-devel/2020-11/msg02183.html) - not a recent regression, also happens with v5.0 already
- error condition in hash_tbl_search_fd_by_ifid (not a regression)
- pica61 -device isa-vga aborts (RAMBlock "vga.vram" already registered) - not a recent regression, thus this should likely not block the release
- qga: Correct loop count in qmp_guest_get_vcpus (not a regression)
- HMP netdev_add doesn't let you reuse the ID of a deleted netdev, QMP netdev_add fails to reject duplicate ID (not a regression for 5.2 as error was already released in 5.0)
- ide/atapi: check io_buffer_index in ide_atapi_cmd_reply_end
- https://www.mail-archive.com/qemu-devel@nongnu.org/msg760877.html -- patch is not the correct fix; in rc4 we put in an assert so that the worst-case is just "QEMU asserts" rather than reading off the end of the buffer
Not fixed yet
Targeted Features
See the ChangeLog/5.2 for full details.