Signed-off-by: NCliff Cai <cliff.cai@analog.com>
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Normally, the musb uses ep1 as the bidirectional bulk endpoint.  This won't
work on the Blackfin musb as all endpoints (except ep0) are unidirectional.
Further, ep1-ep4 have a small 128 byte FIFO which makes them undesirable
for bulk endpoints (which need more like a 512 byte FIFO).  This leaves us
with ep5-ep7 which have 1024 byte FIFOs and can be configured as either
in/out and bulk/interrupt/iso on the fly.
Signed-off-by: NBryan Wu <cooloney@kernel.org>
Signed-off-by: NCliff Cai <cliff.cai@analog.com>
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The base versions handle constant folding now.
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Since num is unsigned, it would seem better to use simple_strtoul that
simple_strtol.

A simplified version of the semantic patch that makes this change is as
follows: (http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@r2@
long e;
position p;
@@

e = simple_strtol@p(...)

@@
position p != r2.p;
type T;
T e;
@@

e = 
- simple_strtol@p
+ simple_strtoul
  (...)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Acked-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Define three new descriptor manipulation utilities, for use when
setting up functions that may have multiple instances:

	usb_copy_descriptors() to copy a vector of descriptors
	usb_free_descriptors() to free the copy
	usb_find_endpoint() to find a copied version

These will be used as follows.  Functions will continue to have static
tables of descriptors they update, now used as __initdata templates.

When a function creates a new instance, it patches those tables with
relevant interface and string IDs, plus endpoint assignments.  Then it
copies those morphed descriptors, associates the copies with the new
function instance, and records the endpoint descriptors to use when
activating the endpoints.  When initialization is done, only the copies
remain in memory.  The copies are freed on driver removal.

This ensures that each instance has descriptors which hold the right
instance-specific data.  Two instances in the same configuration will
obviously never share the same interface IDs or use the same endpoints.
Instances in different configurations won't do so either, which means
this is slightly less memory-efficient in some cases.

This also includes a bugfix to the epautoconf code that shows up with
this usage model.  It must replace the previous endpoint number when
updating the template descriptors, not just mask in a few more bits.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Restore some section annotations:  they were switched to "__devinit"
while they should have been "__init", because of bogus warnings.  The
warnings are now fixed, so the runtime footprint of various drivers
can now shrink a bit.  On ARMv5, it's about 600 bytes except for the
Ethernet gadget, where it can save a bit more.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Move <linux/usb_gadget.h> to <linux/usb/gadget.h>, reducing
some of the clutter in the main include directory.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This removes complaints about the gadget stack which are generated by
the currrent "sparse":  it doesn't like the fact that zero is the null
pointer.  (Last I checked, C guarantees that's correct ...)
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

It turns out that le16_to_cpup() and le32_to_cpup() aren't always safe
to call with pointers into packed structures, since those are inlined
functions and GCC may lose the "packed" attribute.  So those references
can become unaligned kernel accesses, which are evil on some hardware.

This patch updates uses of those routines in the gadget stack.  The
references into packed structures can just use leXX_to_cpu(*x), which
in most cases is more natural.  Some other uses in RNDIS, mostly in
debug code, were wrong in the first place; those use get_unaligned().
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This moves <linux/usb_ch9.h> to <linux/usb/ch9.h> to reduce some of the
clutter of usb header files.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Recent section changes broke gadget builds on some platforms.  This patch
is the best fix that's available until better section markings exist:

 - There's a lot of cleanup code that gets used in both init and exit paths;
   stop marking it as "__exit".

   (Best fix for this would be an "__init_or_exit" section marking, putting
   the cleanup in __init when __exit sections get discarded else in __exit.)

 - Stop marking the use-once probe routines as "__init" since references
   to those routines are not allowed from driver structures.  They're now
   marked "__devinit", which in practice is a net lose.

   (Best fix for this is likely to separate such use-once probe routines
   from the driver structure ... but in general, all busses that aren't
   hotpluggable will be forced to waste memory for all probe-only code.)

In general these broken section rules waste an average of two to four kBytes
per driver of code bloat ... because none of the relevant code can ever be
reused after module initialization.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!