For 'make modules_install', install any firmware required by
the modules which are being installed.

Also add a 'make firmware_install' target which doesn't depend on the
configuration, but installs _all_ available in-kernel-tree firmware into
$(INSTALL_FW_PATH), which defaults to /lib/firmware. This is intended
for distributors to make arch-independent (and config-independent)
packages containing firmware.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

This will control whether we build firmware into the kernel image for
_every_ driver which we convert to request_firmware(), to avoid a
proliferation of 'CONFIG_XXX_FIRMWARE' options for each one.

Default to 'y' for now, which is the wrong thing to do but people seem
to be insisting on it and refusing to even review patches until it's
done. And it does preserve the existing behaviour for built-in drivers.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

This allows arbitrary firmware files to be included in the static kernel
where the firmware loader can find them without requiring userspace to
be alive.

(Updated and CONFIG_EXTRA_FIRMWARE_DIR added with lots of help from
Johannes Berg).
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>

Some drivers have their own hacks to bypass the kernel's firmware loader
and build their firmware into the kernel; this renders those unnecessary.

Other drivers don't use the firmware loader at all, because they always
want the firmware to be available. This allows them to start using the
firmware loader.

A third set of drivers already use the firmware loader, but can't be
used without help from userspace, which sometimes requires an initrd.
This allows them to work in a static kernel.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

In preparation for supporting firmware files linked into the static
kernel, make fw->data const to ensure that users aren't modifying it (so
that we can pass a pointer to the original in-kernel copy, rather than
having to copy it).
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Fix a const pointer usage warning in the Digigram miXart soundcard driver.  A
const pointer is being passed to copy_from_user() to load the firmware into.
This is okay in this case because the function has allocated the firmware
struct itself, but the const qualifier is part of the firmware struct - so the
patch casts the const away.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Fix a const pointer usage warning in the Digigram pcxhr compatible soundcard
driver.  A const pointer is being passed to copy_from_user() to load the
firmware into.  This is okay in this case because the function has allocated
the firmware struct itself, but the const qualifier is part of the firmware
struct - so the patch casts the const away.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Fix a const pointer to non-const pointer assignment error in the Conexant
cx23418 MPEG encoder driver.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Fix an assignment of a const pointer to a non-const pointer in moxa_load_fw().
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Fix a const pointer usage warning in the Digigram VX soundcard driver.  A
const pointer is being passed to copy_from_user() to load the firmware into.
This is okay in this case because the function has allocated the firmware
struct itself, but the const qualifier will be part of the firmware 
struct - so the patch casts the const away.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Standardise both in-kernel and loaded firmware to be stored as
little-endian instead of host-endian.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

The maestro3 driver is byte-swapping its firmware to be host-endian in
advance, when it doesn't seem to be necessary -- we could just use
le16_to_cpu() as we load it.

Doing that means that we need to switch the in-tree firmware to be
little-endian too.

Take the least intrusive way of doing this, which is to switch the
existing snd_m3_convert_from_le() function to convert _to_ little-endian
instead, and use it on the in-tree firmware instead of the loaded
firmware. It's a bit suboptimal but doesn't matter much right now
because we're about to remove the special cases for the in-tree version
anyway.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

...which means allocating our own copy when we want to modify it.

(stupid thinko fixed by mkrufky)
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Signed-off-by: NMichael Krufky <mkrufky@linuxtv.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

... which means allocating our own buffer for reading it back.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Acked-by: NHans Verkuil <hverkuil@xs4all.nl>
Acked-by: NTyler Trafford <ttrafford@gmail.com>
Acked-by: NMike Isely <isely@pobox.com>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/cxgb3: Fix regression caused by class_device -> device conversion