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: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This prevents gadget drivers from being selected when no controller has
been selected, by adding an additional boolean and depending on it.
It's mostly to help "allmodconfig".
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Florian Echtler  <echtler@fs.tum.de>
Signed-off-by: Andreas Deresch <aderesch@fs.tum.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch makes the dummy_hcd driver use emulated root-hub interrupts
instead of polling.  It's in the spirit of similar changes being made to
the other HCDs.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch adds to the dummy_hcd driver a new routine for keeping track of
all changes in the state of the emulated USB link.  The logic is now kept
in one spot instead of spread around, and it's easier to verify and
update the code.  The behavior of the port features has been corrected in
a few respects as well (for instance, if the POWER feature is clear then
none of the other features can be set).

Also added is support for the (relatively new) _connect() and
_disconnect() calls of the Gadget API.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch makes the dummy_hcd driver create separate platform devices for
the emulated host controller and emulated device controller.  This gives a
more accurate simulation and will permit testing of situations where only
one of the two devices is suspended.

This also changes the name of the host controller platform device to match
the name of the driver.  That way the normal platform bus probe mechanism
will handle binding the driver to the device.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch makes some cosmetic changes to dummy_hcd:

	Minor alterations of comments and whitespace.

	Replace USB_PORT_FEAT_xxx with USB_PORT_STAT_xxx.  This is
	appropriate as the values are stored in a status variable
	and they aren't feature indices.  Also it allows the
	elimination of a bunch of awkward bit shift operations.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Partial OTG support for dummy_hcd, mostly as a framework for further work.
It emulates the new OTG flags in the host and peripheral frameworks, if
that option is configured.  But it's incomplete:

  - Resetting the peripheral needs to clear the OTG state bits;
    a second enumeration won't work correctly.

  - This stops modeling HNP right when roles should switch the first time.
    It should probably disconnect, then set the usb_bus.is_b_host and
    usb_gadget.is_a_peripheral flags; then it'd enumerate almost normally,
    except for the role reversal.  Roles could then switch a second time,
    back to "normal" (with those flags cleared).

  - SRP should be modeled as "resume from port-unpowered", which is
    a state that usbcore doesn't yet use.

HNP can be triggered by enabling the OTG whitelist and configuring a
gadget driver that's not in that list; or by configuring Gadget Zero
to identify itself as the HNP test device.
Sent-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

More omap_udc updates:

  * OMAP 1710 updates
      - new UDC bit for clearing endpoint toggle, affecting CLEAR_HALT
      - new OTG bits affecting wakeup
  * Fix the bug Vladimir noted, that IN-DMA transfer code path kicks in
    for under 1024 bytes (not "up to 1024 bytes")
  * Handle transceiver setup more intelligently
      - use transceiver whenever one's available; this can be handy
        for GPIO based, loopback, or transceiverless configs
      - cleanup correctly after the "unrecognized HMC" case
  * DMA performance tweaks
      - allow burst/pack for memory access
      - use 16 bit DMA access most of the time on TIPB
  * Add workarounds for some DMA errata (not observed "in the wild"):
      - DMA CSAC/CDAC reads returning zero
      - RX/TX DMA config registers bit 12 always reads as zero (TI patch)
  * More "sparse" warnings removed, notably "changing" the SETUP packet
    to return data in USB byteorder (an API change, null effect on OMAP
    except for these warnings).
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg 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: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg 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: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch changes the host controller drivers; they no longer need to
register their root hubs because usbcore will take care of it for them.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch makes usbcore automatically allocate and register the root hub
device for a new host controller when the controller is registered.  This
way the HCDs don't all have to include the same boilerplate code.  As a
pleasant side benefit, the register_root_hub routine can now be made
static and not EXPORTed.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch changes the HCDs that used the old hub_set_power_budget call,
making them use the new hcd->power_budget field instead.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch simplifies the uhci-hcd driver by removing the device pointer
currently stored in the QH and TD structures.  Those pointers weren't
being used for anything other than to increment the device's reference
count, which is unnecessary since the device is used only when an URB
completes, and outstanding URBs take their own reference to the device.
As a useful side effect, this change means that uhci-hcd no longer needs
to have the root-hub device available in the start routine.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch removes the hub_set_power_budget routine, which was used by a
couple of HCDs to indicate that the root hub was running on battery power.
In its place is a new field added to struct usb_hcd, which HCDs can set
before the root hub is registered.  Special-case code in the hub driver
knows to look at this field when configuring a root hub.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

My patch adding PM support for zd1201 didn't check for the device on
resume, which can oops if the device has been removed.

This patch fixes it.
Signed-off-by: Colin Leroy <colin@colino.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch enables power management (suspend, resume) support for zd1201.
It fixes problems after wakeup for me, but these problems did not appear
everytime without this patch.  it's a bit empirical, based on what the
usbnet does, so maybe not correct...  Maybe someone can give it a look
before it's applied.
Signed-off-by: Colin Leroy <colin@colino.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

[PATCH] USB: add ability for usb-serial drivers to determine if their write urb is currently being used.

This removes a lot of racy and buggy code by trying to check the status of the urb.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Adds a reboot notifier to OHCI, mostly to benefit kexec; plus
minor #include tweaks.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Updates to "usbtest" driver:

  * Improve some diagnostics.  One path that never generated diagnostics
    before should now generate two ... unless you hit a GCC bug that
    all my compilers seem to have, go figure.

  * Add suspend/resume support, so this behaves when the Linux host
    being used for testing suspends.

  * Don't test the "zero byte ep0 read" case unless real-world relevance
    for the testing is is irrelevant.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch fixes lost LF when ACM device is used with getty/login/bash,
in case of a modem which takes calls.
Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Oliver Neukum <oliver@neukum.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This is a revised version of an earlier patch to add support to usbcore
for driving root hubs by interrupts rather than polling.

There's a temporary flag added to struct usb_hcd, marking devices whose
drivers are aware of the new mechanism.  By default that flag doesn't get
set so drivers will continue to see the same polling behavior as before.
This way we can convert the HCDs one by one to use interrupt-based event
reporting, and the temporary flag can be removed when they're all done.

Also included is a small change to the hcd_disable_endpoint routine.
Although endpoints normally shouldn't be disabled while a controller is
suspended, it's legal to do so when the controller's driver is being
rmmod'ed.

Lastly the patch adds a new callback, .hub_irq_enable, for use by HCDs
where the root hub's port-change interrupts are level-triggered rather
than edge-triggered.  The callback is invoked each time khubd has finished
processing a root hub, to let the HCD know that the interrupt can safely
be re-enabled.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

After all the discussion you might not be interested in this still, but
nevertheless here it is.  This patch adds a shutdown method to the
uhci-hcd driver.  Its prerequisite is the patch you wrote adding shutdown
support for PCI.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch improves the strategy uhci-hcd uses for performing controller
resets and checking whether they are needed.

	The HCRESET command doesn't affect the Suspend, Resume,
	or Reset bits in the port status & control registers, so
	the driver must clear them by itself.  This means the
	code to figure out how many ports there are has to be moved
	to an earlier spot in the driver.

	The R/WC bits in the USBLEGSUP register can be set by the
	hardware even in the absence of BIOS meddling with legacy
	support features.  Hence it's not a good idea to check them
	while trying to determine whether the BIOS has altered the
	controller's state.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch, which has as478b as a prerequisite, enables the uhci-hcd
driver to take advantage of root-hub IRQs rather than polling during the
time it is suspended.  (Unfortunately the hardware doesn't support
port-change interrupts while the controller is running.)  It also turns
off the driver's private timer while the controller is suspended, as it
isn't needed then.  The combined elimination of polling interrupts and
timer interrupts ought to be enough to allow some systems to save a
noticeable amount of power while they are otherwise idle.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch tidies up a few loose ends left by the preceding patches.
It indicates the controller supports remote wakeup whenever the PM
capability is present -- which shouldn't cause any harm if the
assumption turns out to be wrong.  It refuses to suspend the
controller if the root hub is still active, and it refuses to resume
the root hub if the controller is suspended.  It adds checks for a
dead controller in several spots, and it adds memory barriers as
needed to insure that I/O operations are completed before moving on.

Actually I'm not certain the last part is being done correctly.  With
code like this:

	outw(..., ...);
	mb();
	udelay(5);

do we know for certain that the outw() will complete _before_ the
delay begins?  If not, how should this be written?
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch implements (finally!) separate suspend and resume routines
for the root hub and the controller in the UHCI driver.  It also
changes the sequence used to reset the controller during initial
probing, so as to preserve the existing state during a Resume-From-Disk.
(This new sequence is what should be used in the PCI Quirks code for
early USB handoffs, incidentally.)  Lastly it adds a notion of the
controller being "inaccessible" while in a PCI low-power state, when
normal I/O operations shouldn't be allowed.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch starts making some serious changes to the UHCI driver.
There's a set of private states for the root hub, and the internal
routines for suspending and resuming work completely differently, with
transitions based on the new states.  Now the driver distinguishes
between a privately auto-stopped state and a publicly suspended state,
and it will properly suspend controllers with broken resume-detect
interrupts instead of resetting them.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch makes a few small improvements in the UHCI driver.  Some
code is moved between different source files and a more useful pointer
is passed to a callback routine.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch moves a few subroutines around in the uhci-hcd source file.
Nothing else is changed.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch turns a user mode driver error into a hard error, and updates
the relevant diagnostic slightly to help troubleshooting.  gphoto was
known to have this problem, hopefully it is now fixed (they have had
plenty of warning...)

This had been left as a soft error to give various user mode drivers a
change to be properly fixed, with the statement that starting in about
2.6.10 it would be changed.  It had been mostly safe as a soft error ...
but that can not be guaranteed.  Now that a year has passed, it's time to
really insist that the user mode drivers finally fix their relevant bugs.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch fixes an oops triggered at rmmod of isp116x-hcd
after the probe() has failed.

Also, it extends the error message printed, if the driver
cannot detect "Chip's Clock Ready" after a software reset.
As Ian Campbell recently reported, this happens if the
chip's H_WAKEUP pin is not pulled low during software reset.
Several people have already had this issue, hence the update
to the error message.

Also, extend the error message about the failed clock
detection after the software reset.
Signed-off-by: Olav Kongas <ok@artecdesign.ee>

This patch provides an "isp116x-hcd" driver for Philips'
ISP1160/ISP1161 USB host controllers.

The driver:
 - is relatively small, meant for use on embedded platforms.
 - runs usbtests 1-14 without problems for days.
 - has been in use by 6-7 different people on ARM and PPC platforms,
   running a range of devices including USB hubs.
 - supports suspend/resume of both the platform device and the root hub;
   supports remote wakeup of the root hub (but NOT the platform device)
   by USB devices.
 - does NOT support ISO transfers (nobody has asked for them).
 - is PIO-only.
Signed-off-by: Olav Kongas <ok@artecdesign.ee>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Various USB patches, mostly for portability:

  - Fifo mode 1 didn't work previously (oopsed), so now it's fixed and
    (why not) defines even more endpoints for composite devices.

  - OMAP 1710 doesn't have an internal transceiver.

  - Small PM update:  if the USB link is suspended, don't disconnect on
    entry to deep sleep.

  - Be more correct about handling zero length control reads.  OMAP
    seems to mis-handle that protocol peculiarity though; best avoided.

  - Platform device resources (for UDC and OTG controllers) now use
    physical addresses, so /proc/iomem is more consistent.

  - Minor cleanups, notably (by volume) for "sparse" NULL warnings.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>

This patch changes the g_file_storage driver to make the "stall" module
parameter generally available; currently it is available only if the
testing version of the module has been configured.  It also fixes a typo
in a comment -- thanks, Pat!
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This patch simplifies the g_file_storage driver by consolidating a bunch
of min() calculations at a single spot.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

From: <pageexec@freemail.hu>

$subject was fixed in 2.4 already, 2.6 needs it as well.

The impact of the bugs is a kernel stack overflow and privilege escalation
from CAP_NET_ADMIN via the IP_VS_SO_SET_STARTDAEMON/IP_VS_SO_GET_DAEMON
ioctls.  People running with 'root=all caps' (i.e., most users) are not
really affected (there's nothing to escalate), but SELinux and similar
users should take it seriously if they grant CAP_NET_ADMIN to other users.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>