Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

I don't much like the trick with multiple inclusions of solos-attrlist.c
but don't really see a saner way to do it without repeating the list.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

There are still a _lot_ of attributes, but for at least the basic ones
we want to be able to get/set them from the kernel. Especially the ones
we want to inform the ATM core about (link state, speed).
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

We no longer try to load firmware while the ATM is up and running.
However, this means that we _do_ make init_module() wait for it, and it
takes a long time for now (since we're using ultra-conservative code in
the FPGA for that too).

The inner loop which uses swahb32p() was by Simon Farnsworth.

Simon has patches which migrate us to request_firmware_nowait(), for
which we'll actually need to take down the ATM devices, do the upgrade,
then reregister them.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

This is just a straight pull in of changes, syncing us up to 0.07 from
openadsl.sf.net
Signed-off-by: NNathan Williams <nathan@traverse.com.au>
Signed-off-by: NSimon Farnsworth <simon@farnz.org.uk>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Print a message if pskb_expand_head fails.

Make atmdebug writable by root, so that you can turn printing of data sent to
and received from the card on and off at runtime - useful for tracking
corruption.
Signed-off-by: NSimon Farnsworth <simon@farnz.org.uk>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

The length field shouldn't ever include the size of the header itself.
This fixes the problem that some people were seeing with 1500-byte
packets.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

This adds basic support for the 'Solos' PCI ADSL2+ cards being developed
by Traverse Technologies and Xrio Ltd:
http://www.traverse.com.au/productview.php?product_id=116Signed-off-by: NNathan Williams <nathan@traverse.com.au>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>