Or Gerlitz wrote:
> Mike Christie wrote:
>>> Does open-iscsi support I/OAT now?  If so, from which version?
>> No. It uses some network API that is not converted to support I/OAT. The 
>> intel guys have sent patches to add it, but they were a little behind 
>> upstream so their patches would break. I think they are working on it still. 
> Mike, Here's a pointer for the patches 
> http://www.mail-archive.com/open-iscsi@googlegroups.com/msg00388.html - 
> and now after saving you 5sec of work... a question comes: can you or 
> Supreeth help me get educated what makes the open iscsi usage of the 
> network API not applicable to use receive copy offload by the kernel 
> I/OAT engine -  such that both the network stack and the iscsi code has 

I am not sure what you are asking. Did you look at the patches?

The network api has be aware of the dma functionality. Check out 
tcp_recvmsg's NET_DMA ifdefs or check out the last patch for iscsi that 
has us use the dma helpers that Supreeth exported. The net functions we 
used when Supreeth made the patch did not support ioat like how 
tcp_recvmsg did and the function we use today also does not.

Maybe Supreeth could have thought of a way for iscsi to use some new 
network helper or API in a way that hid if it was using ioat or not, 
like if we used kernel_recvmsg/tcp_recvmsg, but he did not so that is 
why the iscsi code had to be patched as well as the network code.

Does that answer your question or were you asking something else?

> to be patched? for example, does the NFS kernel code can use I/OAT as it 
> is now?
> 




> thanks,
> 
> Or.
> 
> 
> 
> 
> > 


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~----------~----~----~----~------~----~------~--~---

Reply via email to