For systems using the Mentor HDRC controllers we get better TX DMA throughput
if we can avoid falling back to PIO to write zero length packets ... so tell
the driver to avoid ZLPs.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

another gcc 4.1 signdness warning:

drivers/usb/gadget/ether.c:2028: warning: comparison of unsigned expression < 0 is always false

length is assigned the value of usb_ep_queue() which returns an int.
Directly after this it is checked for < 0, which can never be true.  Making
length an int makes the error check work again.
Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Acked-by: NDavid Brownell <david-b@pacbell.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Allows compiling g_ether in and fixes a typo with MUSB_HDRC
Signed-off-by: NTony Lindgren <tony@atomide.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The new spinlock debug code turned up a spinlock recursion bug in the
Ethernet gadget driver on a disconnect path; it would show up with any
UDC driver where the cancellation of active requests was synchronous,
rather than e.g. delayed until a controller's completion IRQ.

That recursion is fixed here by creating and using a new spinlock to
protect the relevant lists.
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>

Signed-off-by: NJörn Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: NAdrian Bunk <bunk@stusta.de>

This moves <linux/usb_cdc.h> to <linux/usb/cdc.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>

This removes extraneous whitespace from the Ethernet/RNDIS gadget driver.
It's all space-at-EOL, spaces-before-tabs, or tabs-then-spaces.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Acked-by: NDavid Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Be sure to record the peripheral's ep0 maxpacket size BEFORE using
that to initialize the (high speed) device qualifier; that helps a
lot with USBCV testing.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch allows you to set the iSerialNumber field in the
usb_device_descriptor structure for your USB ethernet gadget.

It also changes the parameters shown through sysfs so they're
no longer declared as __initdata, preventing potential oopses.

That's most useful for the Ethernet addresses, which may in
some cases be random "locally administered" addresses.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This adds __init section annotations to gadget driver bind() routines to
remove calls from .text into .init sections (for endpoint autoconfig).
Likewise it adds __exit section annotations to their unbind() routines.

The specification of the gadget driver register/unregister functions is
updated to explicitly allow use of those sections.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This adds declarations for three USB peripheral controllers:

  - Two high speed USB cores that can be licensed from Mentor Graphics
    to be integrated into silicon:

      * "musbhsfc" is for peripherals only, as found in for example the
        IBM/AMCC 44EP processors.

      * "musbhdrc" is OTG-capable (dual role), and is found in various
        products including OMAP 2430 and the new DaVinci SOCs.

    The "musbh" standing for "Mentor USB Highspeed", the rest standing
    for "Function Controller" or "Dual Role Controller" (OTG-capable).

  - The full speed controller on the FreeScale MPC8272.

Adding these definitions just allows gadget driver code to handle any
controller-specific logic; controller drivers are quite separate.
Signed-off-by: NDavid Brownell <david-b@pacbell.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Resove a minor FIXME:  don't change MTU while RNDIS link is active,
the other end won't expect such things...
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

USB gadget drivers make no use of these, remove the pointless
comments.
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Ensure the the device_driver and usb_gadget_driver
have their .owner fields initialised to associate
the module owner to the driver.
Signed-off-by: NBen Dooks <ben-linux@fluff.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch centralizes the assignment of bcdDevice numbers for different
gadget controllers.  This won't improve the object code at all, but it
does save a lot of repetitive and error-prone source code ... and will
simplify the work of supporting a new controller driver, since most new
gadget drivers will no longer need patches (unless some hardware quirks
limit USB protocol messaging).

Added minor cleanups and identifer hooks for the UDC in the Freescale
iMX series processors.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

I also needed the following on 2.6.13-rc1 without CONFIG_USB_ETH_RNDIS,
symbol fs_status_desc isn't available in that case on PXA255.

This builds both with and without ETH_RNDIS, but I haven't actually
tested either.
Signed-off-by: NIan Campbell <icampbell@arcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NIan Campbell <icampbell@arcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Greg,

This patch fixes the kmalloc() flags argument type in USB
subsystem; hopefully all of its occurences. The patch was
made against patch-2.6.12-git2 from Jun 20.

Cleanup of flags for kmalloc() in USB subsystem.
Signed-off-by: NOlav Kongas <ok@artecdesign.ee>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This fixes a build error on pxa25x processes with pxa2xx_udc and

        CONFIG_USB_ETH=m
        # CONFIG_USB_ETH_RNDIS is not set

The error is because on that CPU there's no status transfer support
except with RNDIS.  Workaround, enable the RNDIS support too.
Signed-off-by: NIan Campbell <icampbell@arcom.com>
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This fixes a compile glitch with CONFIG_USB_ETH_RNDIS disabled, and
replaces some inline #ifdeffery (and other code) with inline functions
which can evaluate to constants.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This updates most of the gadget framework to expect SETUP packets use
USB byteorder (matching the annotation in <linux/usb_ch9.h> and usage
in the host side stack):

  - definition in <linux/usb_gadget.h>
  - gadget drivers:  Ethernet/RNDIS, serial/ACM, file_storage, gadgetfs.
  - dummy_hcd

It also includes some other similar changes as suggested by "sparse",
which was used to detect byteorder bugs.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Some cleanup for the the Ethernet part of the Ethernet/RNDIS gadget driver:

  - Remove remnants of ancient endpoint init logic; this is simpler, clearer

  - Save a smidgeon of space in the object file

  - Get rid of some #ifdeffery, mostly by using some newish inlines

  - Reset more driver state as part of USB reset

  - Remove a needless wrapper around an RNDIS call

  - Improve and comment the status interrupt handling:
      * RNDIS sometimes needs to queue these transfers (rarely in normal
        cases, but reproducibly while Windows was deadlocking its USB stack)
      * Mark requests as busy/not

  - Enable the SET_NETDEV_DEV() call; sysfs seems to behave sanely now

This is a net shrink of source code.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Some bugfixes and lots of cleanup (net code shrink):

  - On reset, force the RNDIS state machine its initial state

  - Hook up the RNDIS (outgoing) filters to the CDC mechanism

  - Lots of cleanup:
     * Eliminate duplicate copy of OID table;
     * Unify handlying of the OID "query" response data pointer;
     * Reduce code duplication for calculating query response lengths;
     * Remove some checks for "can't happen" errors;
     * Get rid of debugging #ifdefs by making the debug flag an integer level

Most of the patch, by volume, relates to those query response cleanups.
It incidentally shaves off a few hundred bytes of object code.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Here are some spelling corrections for drivers/usb.

cancelation -> cancellation
succesful -> successful
cancelation -> cancellation
decriptor -> descriptor
Initalize -> Initialize
wierd -> weird
Protocoll -> Protocol
occured -> occurred
successfull -> successful
Procesing -> Processing
devide -> divide
Isochronuous -> Isochronous
noticable -> noticeable
Basicly -> Basically
transfering -> transferring
intialize -> initialize
Incomming -> Incoming
additionnal -> additional
asume -> assume
Unfortunatly -> Unfortunately
retreive -> retrieve
tranceiver -> transceiver
Compatiblity -> Compatibility
Incorprated -> Incorporated
existance -> existence
Ununsual -> Unusual
Signed-off-by: NSteven Cole <elenstev@mesatop.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Updates to the Ethernet/RNDIS gadget driver (mostly for RNDIS):

  - Fix brown-paper bag goof with RNDIS packet TX ... the wrong length
    field got set, so Windows would ignore data packets it received.

  - More consistent handling of CDC output filters (but not yet hooking
    things up so RNDIS uses the mechanism).

  - Zerocopy RX for RNDIS packets too (saving CPU cycles).

  - Use the pre-allocated interrupt/status request and buffer, rather
    than allocating and freeing one of each every few seconds (which
    could fail).

  - Some more "sparse" tweaks, making both dual-speed and single-speed
    configurations happier.

  - RNDIS speeds are reported in units of 100bps, not bps.

Plus two minor cleanups (whitespace, messaging).
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!