On Thu, Aug 06, 2020 at 06:46:06PM +0300, Nir Soffer wrote:
Given the poor results, I suspect that that handles created using same connection share a lock. This also makes sense if connection abstract a blocking socket.
With multiple nbd connections we got 60% improvement. I expect to see similar results if we use multiple connection+handle pairs.
Can you try to move the connection into the vddk_handle struct?
It's a good point, and would also explain why multi-conn was faster than this. I'll give that a go later. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 100 libraries supported. http://fedoraproject.org/wiki/MinGW