2019-01-08 15:53:44 +01:00
|
|
|
From: Olaf Hering <olaf@aepfle.de>
|
|
|
|
|
Date: Tue, 8 Jan 2019 14:20:08 +0100
|
|
|
|
|
Subject: xen: ignore live parameter from xen-save-devices-state
|
|
|
|
|
|
2019-09-12 17:54:03 +02:00
|
|
|
References: bsc#1079730, bsc#1101982, bsc#1063993
|
|
|
|
|
|
2019-01-08 15:53:44 +01:00
|
|
|
The final step of xl migrate|save for an HVM domU is saving the state of
|
|
|
|
|
qemu. This also involves releasing all block devices. While releasing
|
|
|
|
|
backends ought to be a separate step, such functionality is not
|
|
|
|
|
implemented.
|
|
|
|
|
|
|
|
|
|
Unfortunately, releasing the block devices depends on the optional
|
|
|
|
|
'live' option. This breaks offline migration with 'virsh migrate domU
|
|
|
|
|
dom0' because the sending side does not release the disks, as a result
|
|
|
|
|
the receiving side can not properly claim write access to the disks.
|
|
|
|
|
|
|
|
|
|
As a minimal fix, remove the dependency on the 'live' option. Upstream
|
|
|
|
|
may fix this in a different way, like removing the newly added 'live'
|
|
|
|
|
parameter entirely.
|
|
|
|
|
|
|
|
|
|
Fixes: 5d6c599fe1 ("migration, xen: Fix block image lock issue on live migration")
|
|
|
|
|
|
|
|
|
|
Signed-off-by: Olaf Hering <olaf@aepfle.de>
|
2019-09-12 17:54:03 +02:00
|
|
|
Signed-off-by: Bruce Rogers <brogers@suse.com>
|
2019-01-08 15:53:44 +01:00
|
|
|
---
|
|
|
|
|
migration/savevm.c | 2 +-
|
|
|
|
|
1 file changed, 1 insertion(+), 1 deletion(-)
|
|
|
|
|
|
|
|
|
|
diff --git a/migration/savevm.c b/migration/savevm.c
|
2021-12-17 11:07:39 +01:00
|
|
|
index d59e976d50e7c81c20bbf5b930d1..7f7cedddb36dcc536fc242349882 100644
|
2019-01-08 15:53:44 +01:00
|
|
|
--- a/migration/savevm.c
|
|
|
|
|
+++ b/migration/savevm.c
|
2021-12-17 11:07:39 +01:00
|
|
|
@@ -2942,7 +2942,7 @@ void qmp_xen_save_devices_state(const char *filename, bool has_live, bool live,
|
2019-01-08 15:53:44 +01:00
|
|
|
* So call bdrv_inactivate_all (release locks) here to let the other
|
Accepting request 854151 from home:bfrogers:branches:Virtualization
- Update to v5.2.0: See http://wiki.qemu.org/ChangeLog/5.2
Take note that ongoing feature deprecation is tracked at both
http://wiki.qemu-project.org/Features/LegacyRemoval and in
the deprecated.html file installed with the qemu package
Some noteworthy changes:
* Dropped system emulators: qemu-system-lm32, qemu-system-unicore32
* Dropped linux user emulator: qemu-ppc64abi32
* Added linux user emulator: qemu-extensaeb
* Unicore32 and lm32 guest support dropped
* New sub-packages (most due to ongoing modularization of QEMU):
qemu-audio-spice, qemu-hw-chardev-spice, qemu-hw-display-virtio-vga,
qemu-hw-display-virtio-gpu, qemu-hw-display-virtio-gpu-pci,
qemu-ui-spice-core, qemu-ui-opengl, qemu-ivshmem-tools
* x86: A new KVM feature which improves the handling of asynchronous page
faults is available with -cpu ...,kvm-async-pf-int (requires Linux 5.8)
* s390: More instructions emulated under TCG
* PowerPC: nvdimm= machine option now functions correctly; misc improvements
* ARM: new boards: mps2-an386 (Cortex-M4 based) and mps2-an500
(Cortex-M7 based), raspi3ap (the Pi 3 model A+), raspi0 (the Pi Zero)
and raspi1ap (the Pi A+)
* RISC-V: OpenSBI v0.8 included by default; Generic OpenSBI platform used
when no -bios argument is supplied; Support for NUMA sockets on Virt
and Spike Machines; Support for migrating machines; misc improvements
* Misc NVMe improvements
* The 'vhost-user-blk' export type has been added, allowing
qemu-storage-daemon to act as a vhost-user-blk device backend
* The SMBIOS OEM strings can now come from a file
* 9pfs - misc performance related improvements
* virtiofs - misc improvements
* migration: The default migration bandwidth has been increased to 1Gbps
(users are still encouraged to tune it to their own hardware); The new
'calc-dirty-rate' and 'query-dirty-rate' QMP commands can help determine
the likelihood of precopy migration success; TLS+multifd now supported
for higher bandwidth encrypted migration; misc minor features added
* Misc minor block features added
* Misc doc improvements
* qemu-microvm subpackage change: the bios-microvm.bin is now SeaBIOS based,
and the qboot based on is now qboot.rom
* elf2dmp is no longer part of qemu-tools (it was never intended to be
a packaged binary)
* Some subpackages which were 'Requires' are now 'Recommends', allowing for
a smaller qemu packaging footprint if needed
* Patches dropped (included in release tarball, unless otherwise noted):
docs-fix-trace-docs-build-with-sphinx-3..patch (fixed differently)
hw-hyperv-vmbus-Fix-32bit-compilation.patch
linux-user-properly-test-for-infinite-ti.patch
Switch-order-of-libraries-for-mpath-supp.patch (fixed differently)
Conditionalize-ui-bitmap-installation-be.patch (fixed differently)
hw-usb-hcd-xhci-Fix-GCC-9-build-warning.patch (no longer using gcc9)
hw-usb-dev-mtp-Fix-GCC-9-build-warning.patch (no longer using gcc9)
roms-Makefile-enable-cross-compile-for-b.patch (fixed with different patch)
libvhost-user-handle-endianness-as-manda.patch
virtio-add-vhost-user-fs-ccw-device.patch
Fix-s-directive-argument-is-null-error.patch
build-Workaround-compilation-error-with-.patch
build-Be-explicit-about-fcommon-compiler.patch
intel-Avoid-spurious-compiler-warning-on.patch
golan-Add-explicit-type-casts-for-nodnic.patch
Do-not-apply-WORKAROUND_CFLAGS-for-host-.patch
ensure-headers-included-are-compatible-w.patch
Enable-cross-compile-prefix-for-C-compil.patch (fixed differently)
hw-net-net_tx_pkt-fix-assertion-failure-.patch
hw-net-xgmac-Fix-buffer-overflow-in-xgma.patch
s390x-protvirt-allow-to-IPL-secure-guest.patch
usb-fix-setup_len-init-CVE-2020-14364.patch
* Patches added:
meson-install-ivshmem-client-and-ivshmem.patch
Revert-roms-efirom-tests-uefi-test-tools.patch
Makefile-Don-t-check-pc-bios-as-pre-requ.patch
roms-Makefile-add-cross-file-to-qboot-me.patch
qboot-add-cross.ini-file-to-handle-aarch.patch
usb-Help-compiler-out-to-avoid-a-warning.patch
- In spec file, where reasonable, switch BuildRequires: XXX-devel
to be pkgconfig(XXX') instead
- No longer disable link time optimization for qemu for x86. It looks like
either the build service, qemu code changes and/or the switch to meson
have resolved issues previously seen there. We still see problems for
other architectures however.
- For the record, the following issues reported for SUSE SLE15-SP2
are either fixed in this current package, or are otherwise no longer
an issue: bsc#1172384 bsc#1174386 bsc#1174641 bsc#1174863 bsc#1175370
bsc#1175441 bsc#1176494 CVE-2020-13361 CVE-2020-14364 CVE-2020-15863
CVE-2020-16092 CVE-2020-24352
and the following feature requests are satisfied by this package:
jsc#SLE-13689 jsc#SEL-13780 jsc#SLE-13840
- To be more accurate, and to align with other qemu packaging
practices, rename the qemu-s390 package to qemu-s390x. The old
name (in the rpm namespace) is provided with a "Provides"
directive, and an "Obsoletes" done against that name for prior
qemu versions, as is standard practice (boo#1177764 jsc#SLE-17060)
- Take this opportunity to remove some ancient Split-Provides
mechanisms which can't conceivably be needed any more:
qemu-block-curl provided: qemu:%_libdir/%name/block-curl.so
qemu-guest-agent provided: qemu:%_bindir/qemu-ga
qemu-tools provided: qemu:%_libexecdir/qemu-bridge-helper
- Disable linux-user 'ls' test on 32 bit arm. It's failing with
"Allocating guest commpage: Cannot allocate memory" error, which
we should hunt down, but for now we don't want it to prevent the
package from being built
OBS-URL: https://build.opensuse.org/request/show/854151
OBS-URL: https://build.opensuse.org/package/show/Virtualization/qemu?expand=0&rev=597
2020-12-08 23:01:20 +01:00
|
|
|
* side of the migration take control of the images.
|
2019-01-08 15:53:44 +01:00
|
|
|
*/
|
|
|
|
|
- if (live && !saved_vm_running) {
|
|
|
|
|
+ if (!saved_vm_running) {
|
|
|
|
|
ret = bdrv_inactivate_all();
|
|
|
|
|
if (ret) {
|
|
|
|
|
error_setg(errp, "%s: bdrv_inactivate_all() failed (%d)",
|
