On 05/06/14 15:27, Ján Tomko wrote:
Not yet merged in upstream QEMU: https://lists.gnu.org/archive/html/qemu-devel/2014-04/msg05024.html
Add support for invariant TSC timer running at constant rate in all ACPI P-, C- and T-states.
It can be enabled by specifying: <clock> <timer name='invtsc' present='yes'/> </clock> in the domain XML.
Migration and saving the domain does not work with this timer.
The support for this timer is indicated by bit 8 of EDX after calling CPUID with 0x80000007. It does not show up in /proc/cpuinfo [1] and since we're calling qemu without 'enforce', it doesn't error out if the host doesn't support this.
Alternatively, we could expose it in libvirt as a cpu flag: <cpu mode='custom' match='exact'> <model fallback='forbid'>qemu64</model> <feature policy='require' name='invtsc'/> </cpu> or maybe add +invtsc to qemu args when the 'nonstop_tsc' flag is requested?
[1]: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/arch/x8... --- docs/formatdomain.html.in | 9 ++++++-- docs/schemas/domaincommon.rng | 1 + src/conf/domain_conf.c | 6 +++-- src/conf/domain_conf.h | 1 + src/qemu/qemu_command.c | 6 +++++ src/qemu/qemu_migration.c | 14 ++++++++++++ .../qemuxml2argv-clock-timer-inv-tsc.args | 5 +++++ .../qemuxml2argv-clock-timer-inv-tsc.xml | 26 ++++++++++++++++++++++ tests/qemuxml2argvtest.c | 1 + tests/qemuxml2xmltest.c | 1 + 10 files changed, 66 insertions(+), 4 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-clock-timer-inv-tsc.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-clock-timer-inv-tsc.xml
...
diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 4249ed5..5154826 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -918,6 +918,7 @@ <choice> <value>kvmclock</value> <value>hypervclock</value> + <value>invtsc</value>
I'd prefer to change the name of the feature to "invarianttsc" in libvirt's representation, but that's just bikeshedding.
</choice> </attribute> </group>
...
diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index a9f7fea..c1ffc0f 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -1513,6 +1513,20 @@ qemuMigrationIsAllowed(virQEMUDriverPtr driver, virDomainObjPtr vm, return false; }
+ for (i = 0; i < def->clock.ntimers; i++) { + virDomainTimerDefPtr timer = def->clock.timers[i]; + + if (timer->present != 1) + continue; + + if (timer->name == VIR_DOMAIN_TIMER_NAME_INVTSC) { + virReportError(VIR_ERR_OPERATION_INVALID, + _("domain has '%s' timer"), + virDomainTimerNameTypeToString(timer->name)); + return false; + } + } + return true; }
It's a shame that this doesn't work across migration in a way HyperV has designed it. ACK once the qemu functionality will be released Peter