The following series implements Intel I/O Acceleration Technology (I/
OAT)
support for open-iscsi. This series works on the 2.6.24.x kernels. We
are
working on modifying the patches to work with the new tcp recv path in
open-iscsi from the 2.6.25-rcx kernels. I recommend using the latest
stable
2.6.24 kernel to apply the patches. We're putting these out now to
solicit
feedback from the community, and will submit new patches against the
upstream
open-iscsi kernel. Please feel free to send any comments/suggestions
to me or
the list. We also continue to work on performance tuning these
patches. Each
patch in this series has additional descriptive notes. Please apply
the
patches in the order 01-tcp_dma, 02-networking, 03-iscsibase, and 04-
iscsioffload.
I have just applied these patches using stg on the stable version of
kernel
2.6.24.4 from git.kernel.org successfully. Please let me know if for
some reason
these don't apply cleanly for you.

The following additional features may need to be enabled in .config by
the
user if not active by default by running "make menuconfig."

1. Under "Device Drivers" enable DMA Engine Support, and Intel I/OAT.
2. Under "Device Drivers" enable SCSI low level drivers support and
the iSCSI
initiator over TCP/IP.

To enable the features in the patches, please ensure that the
"ioatdma" module
is loaded using "modprobe ioatdma." We noticed a marginal increase in
throughput using Intel 1Gb ethernet cards (e1000) and a significant
increase
in throughput using the Intel 10Gb cards.

NOTE: THIS SERIES WILL ONLY COMPILE ON THE 2.6.24.X KERNELS AND NOT
THE LATEST
(2.6.25-RCx)UPSTREAM KERNELS.


--
Supreeth Venkataraman

--~--~---------~--~----~------------~-------~--~----~
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 [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~----------~----~----~----~------~----~------~--~---

Reply via email to