On 06/14/2011 10:18 AM, Michal Privoznik wrote:
For virtio disks and interfaces, qemu allows users to enable or disable ioeventfd feature. This means, qemu can execute domain code, while another thread waits for I/O event. Basically, in some cases it is win, in some loss. This feature is available via 'asyncio' attribute in disk and interface <driver> element. It accepts 'on' and 'off'. Leaving this attribute out defaults to hypervisor decision. --- this is rework as suggested: https://www.redhat.com/archives/libvir-list/2011-May/msg01269.html
docs/formatdomain.html.in | 34 ++++++++++++- docs/schemas/domain.rng | 14 +++++ src/conf/domain_conf.c | 49 ++++++++++++++++++- src/conf/domain_conf.h | 11 ++++ src/libvirt_private.syms | 2 + src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 23 +++++++++ tests/qemuhelptest.c | 3 +- .../qemuxml2argv-disk-asyncio.args | 11 ++++ .../qemuxml2argvdata/qemuxml2argv-disk-asyncio.xml | 51 ++++++++++++++++++++ tests/qemuxml2argvtest.c | 3 + 12 files changed, 201 insertions(+), 4 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-asyncio.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-asyncio.xml
ACK.
+ The optional <code>asyncio</code> attribute allows users to + set <a href='https://patchwork.kernel.org/patch/43390/'> + domain I/O asynchronous handling</a> for disk device. + The default is left to the discretion of the hypervisor. + Accepted values are "on" and "off". Enabling this allows + qemu to execute VM while a separate thread handles I/O. + Typically guests experiencing high system CPU utilization + during I/O will benefit from this. On the other hand, + on overloaded host it could increase guest I/O latency. + <span class="since">Since 0.9.3 (QEMU and KVM only)</span> + <b>In general you should leave this option alone, unless you + are very certain you know what you are doing.</b>
And nice description.
+static int +qemuBuildAsyncIoStr(virBufferPtr buf, + enum virDomainAsyncIo use, + virBitmapPtr qemuCaps) +{ + if (qemuCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_IOEVENTFD)) { + switch (use) { + case VIR_DOMAIN_ASYNC_IO_ON: + case VIR_DOMAIN_ASYNC_IO_OFF: + virBufferAsprintf(buf, ",ioeventfd=%s", + virDomainAsyncIoTypeToString(use)); + break; + default: + /* In other cases (_DEFAULT, _LAST) we don't + * want to add anything */ + break; + } + }
I would have avoided the switch statement, and gone with the simpler: if (use && qemuCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_IOEVENTFD)) virBufferAsprintf(buf, ",ioeventfd=%s", virDomainAsyncIoTypeToString(use)); which automatically filters out VIR_DOMAIN_ASYNC_IO_DEFAULT, and nothing in the code should ever be setting the value of 'use' to VIR_DOMAIN_ASYNC_IO_LAST or greater. But your approach is correct as-is, so it's up to you if you want to simplify. -- Eric Blake eblake@redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org