Kais Belgaied Wrote: > Cecilia Hu wrote: >>> Cecilia Hu wrote: >>>> >>>> The architecture of the Intel 10GbE PCI Express Adapter differs >>>> significantly from the Intel 82597EX based PCIX Adapter, which is >>>> supported by ixgb. An important new feature of this adapter is >>>> I/OAT >>>> (I/O Acceleration Technology) from Intel which will be helpful for >>>> performance improvement. So we introduce a new driver to support >>>> them. >>> >>> let's see. I/O AT is a collection of new capabilities that may >>> involve the NIC, the chipset and/or >>> the CPU. >>> Could you say more about which of these capabilities that this case >>> will be using/supporting? >> >> This driver can use the capabilities of NIC, chipset and CPU, but it will >> support the I/OAT features only after framework is ready. > > so I take it there is another framework in Solaris that exposes the I/O > AT features, right? > This case depends on that framework in order to deliver its support for > key performance functionality. > This case imports interfaces from that framework. Is that a fair > understanding?
It's probably that my words are misleaded. This case only requests to provide a driver which uses existing interfaces and framework, and such request is exactly same as any other new NIC drivers. >>> Also looking at Intel's docs (and previous presentations), there is a >>> hint to a need for an "optimized >>> TCP/IP stack" in order to benefit from I/O AT. (see for instance >>> http://download.intel.com/technology/comms/perfnet/download/98856.pdf). >>> Is this case introducing changes for the OpenSolaris TCP/IP stack to >>> be able to use I/OAT ? Any new >>> interfaces needed to negotiate such capabs? >> >> This case only covers the implementation of this driver. It will not >> introduce changes for the OpenSolaris TCP/IP stack either. > > Given the fact that the case included support for I/O AT features to > help performance, and > given the observation that other Operating systems needed to change > their TCP/IP stack is addition > to the driver in order to benefit from I/O AT, can this case be > considered complete without > the TCP/IP stack changes? can it deliver any performance benefit at all > without those changes? I/OAT feature support is excluded in this case, and will be supported later when framework is ready. -- Cecilia
