On Wed, Jan 18, 2017 at 03:26:20PM +0800, Eli Qiao wrote:
This patch expose cache information to host's capabilites xml.
<cache>
<bank id='0' type='l3' size='56320' unit='KiB'
cpus='0-21,44-65' min='2816' scope='L3DATA'/>
<bank id='1' type='l3' size='56320' unit='KiB'
cpus='22-43,66-87' min='2816' scope='L3DATA'/>
<bank id='2' type='l3' size='56320' unit='KiB'
cpus='0-21,44-65' min='2816' scope='L3CODE'/>
<bank id='3' type='l3' size='56320' unit='KiB'
cpus='22-43,66-87' min='2816' scope='L3CODE'/>
</cache>
There are some nits difference from RFC on mailing list.
https://www.redhat.com/archives/libvir-list/2017-January/msg00644.html
I think this is somewhat misleading - it is makes it look like we have 56320 KiB of L3
cache for DATA and 56320 KiB of L3 cache for CODE, ie 112640 KiB total per socket. IIUC,
what we actually have is just 56320 KiB of L3 cache, which can be split between DATA or
CODE. I think the nested syntax shows that better
<cache>
<bank id='0' type='l3' size='56320' unit='KiB'
cpus='0-21,44-65'>
<control min='2816' scope='L3DATA'/>
<control min='2816' scope='L3CODE'/>
</bank>
<bank id='1' type='l3' size='56320' unit='KiB'
cpus='22-43,66-87'>
<control min='2816' scope='L3DATA'/>
<control min='2816' scope='L3CODE'/>
</bank>
</cache>
Regards,
Daniel
--
|:
http://berrange.com -o-
http://www.flickr.com/photos/dberrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org :|
|:
http://entangle-photo.org -o-
http://search.cpan.org/~danberr/ :|