On Fri, Mar 18, 2011 at 03:16:22PM -0700, Dan Williams wrote:

From: Dave Jiang <dave.jiang@intel.com>

The efivars module already scans all available variables, normalizes the
variable names, and stores them in a list.  Rather than duplicate this
to efi runtime services interface let drivers query variable data by
GUID.

This is needed by the isci driver which relies on an efi variable to
store critical platform parameters like gloablly unique sas addresses
and phy configuration parameters.  This is similar to the
pci_map_biosrom() enabling that allows the isci driver to retrieve the
same data in the non-efi case.

For the built-in case efivars is moved to subsys_initcall.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Dan Williams <redacted>
---
Not sure who looks after efivars.c, but get_maintainer.pl and git shortlog
fingered Greg as a likely target.

Yes, that is me.

We are currently targeting a late merge of the isci driver through the
staging tree into 2.6.39.  This is a pre-requisite to be able to use the
driver on an efi enabled platform.

You are?  That's news to me, why didn't you ask the staging tree
maintainer about this?

I needed all patches in linux-next _before_ the merge window opened to
be able to accept it.  So, sorry, it will have to wait until .40.
Please send me the isci driver and I will be glad to queue it up for
staging for that kernel.

thanks,

greg k-h