I've found some new datasheets which describe some additionally
supported standard baud rates and I've verified them with my HX
(rev. 3A) device. But adding support for individual (chip type
specific) baud rates would add a good amount of extra code (especially
when support for further chips will be added to the driver one day),
which makes no sense as long as we are not using the direct baud rate
encoding method for newer chips.
So for now, just drop a comment about these additionally supported baud
rates.

The second comment is about the baud rate differences between the two
encoding methods. In theory, we could optimize the code a bit by
comparing the resulting baud rates of both methods and selecting the
one which is closer to the requested baud rate. But that seems to be a
bit overkill, because the differences are very small and the device
likely uses the same baud rate generator for both methods so that the
resulting baud rate would be the same.
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: pl2303: also use the divisor based baud rate encoding method for baud rates < 115200 with HX chips

Now that the divisor based baud rate encoding method has been fixed and
extended, it can also be used for baud rates < 115200 baud with HX
chips.
This makes it possible to adjust the baud rate almost continuously
instead of just beeing able to select between 16 fixed standard values.

Tested with a PL2303HX 04463A (week 46, 2004, rev 3A).
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Reinhard Max has done some tests with a PL2303HX (rev A) and a logic
analyzer and it seems, that although the PL2303HX is specified for baud
rates from 75 to 6M baud, the full divisor range can be used with the
divisor based baud rate encoding method. This corresponds to baud rates
from 46 to 24M baud.
Baud rates down to 46 baud (max. divisor) have been confirmed to work
even under heavy/permanent load, so remove the lower limit.
Baud rates up to 24M baud should really be tested carefully in "real
life" scenarios before removing the upper limit completely.
Anyway, the Windows driver allows maximum baud rates of 110% of the
specified limit, so for now, increase the upper limit to this value.
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NReinhard Max <max@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Commit 0c967e7e "USB: serial: pl2303 works at 500kbps" added 500000
baud to the list of supported standard baud rates.
But the reason why the driver works with this baud rate is, that since
commit 8d48fdf6 "USB: PL2303: correctly handle baudrates above 115200"
a second (divisor based) baud rate encoding method is used for values
above 115200 baud, which is not limited to a fixed set of standard baud
rates.

Remove the 500000 baud value from the list of standard baud rates
again, because this list is only used with the direct baud rate
encoding method and 500000 baud is not supported with this method.
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: pl2303: do not round to the next nearest standard baud rate for the divisor based baud rate encoding method

In opposition to the direct baud rate encoding method, the divisor
based method is not limited to a fixed set of standard baud rates.
Hence, there is no need to round to the next nearest standard value.
Reported-by: NMastro Gippo <gipmad@gmail.com>
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NReinhard Max <max@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Based on the formula in the code description, Reinhard Max and me have
investigated the devices behavior / functional principle of the divisor
based baud rate encoding method.

It turned out, that (although beeing a good starting point) the current
code has some flaws. It doesn't work correctly for a wide range of baud
rates and the divisor resolution can be improved. It also doesn't
report the actually set baud rate.

This patch fixes and improves the code for the divisor based baud rate
encoding method a lot. It can now be used for the whole range of baud
rates from 46 baud to 24M baud with a very good divisor resolution and
userspace can read back the resulting baud rate.

It also documents the formula used for encoding and the hardware
behavior (including special cases).

The basic algorithm, rounding and several code comments/explanations
are provided by Reinhard Max.
I've added some minor fixes, the handling of the special cases and
further code/algorithm descriptions.
Signed-off-by: NFrank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: NReinhard Max <max@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In the disconnect routine for the hwa_hc interface, it calls
uwb_pal_unregister to unregister itself from the UWB subsystem.  This
function attempts to clean up the link to the host controller directory in
the device's UWB radio control interface directory.  If the disconnect
routine for the radio control interface has already run, the uwb directory
will be gone so the call to sysfs_remove_link generates a warning.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Prevent the USB core from suspending the HWA root hub since bus_suspend
and bus_resume are not yet supported.  Otherwise the PM system will chew
up CPU time constantly attempting to suspend and resume the root hub but
never succeeding.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes a kernel panic that can occur when unplugging the HWA
dongle while a downstream device is in the process of disconnecting.
This involved 2 changes.  First, call usb_lock_device_for_reset before
usb_reset_device to synchronize the HWA's post_rest and disconnect
routines.  Second, set the hwarc->neep_urb and hwarc->rd_buffer to NULL
when they are freed in the error path in the post_reset routine.  This
prevents a double free when the disconnect routine is called and attempts
to free those resources again.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Both could want to submit the same URB. Some checks of the flag
intended to prevent that were missing.
Signed-off-by: NOliver Neukum <oneukum@suse.de>
CC: stable@vger.kernel.org
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

wa_urb_enqueue_run locks and unlocks its list lock as it traverses the
list of queued transfers.  This was done to prevent deadlocking due to
acquiring locks in reverse order in different places.  The problem is that
releasing the lock during the list traversal could allow the dequeue
routine to corrupt the list while it is being iterated over.  This patch
moves all list entries to a temp list while holding the list lock, then
traverses the temp list with no lock held.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The hub_status_data function on the wireless USB root hub controller
(wusbhc_rh_status_data) always returns a positive value even if no ports
have changed.  This patch updates wusbhc_rh_status_data to only return a
positive value if the root hub status needs to be queried.  The current
implementation can also leave the upper bits of the port bitmap
uninitialized if wusbhc->ports_max is not one less than an even multiple
of 8.  This patch fixes that as well by initializing the buffer to 0.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The USB Embedded High-speed Host Electrical Test (EHSET) defines the
SINGLE_STEP_SET_FEATURE test as follows:

1) The host enumerates the test device with VID:0x1A0A, PID:0x0108
2) The host sends the SETUP stage of a GetDescriptor(Device)
3) The device ACKs the request
4) The host issues SOFs for 15 seconds allowing the test operator to
   raise the scope trigger just above the SOF voltage level
5) The host sends the IN packet
6) The device sends data in response, triggering the scope
7) The host sends an ACK in response to the data

This patch adds additional handling to the EHCI hub driver and allows
the EHSET driver to initiate this test mode by issuing a a SetFeature
request to the root hub with a Test Selector value of 0x06. From there
it mimics ehci_urb_enqueue() but separately submits QTDs for the
SETUP and DATA/STATUS stages in order to insert a delay in between.
Signed-off-by: NManu Gautam <mgautam@codeaurora.org>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
[jackp@codeaurora.org: imported from commit c2084930 on codeaurora.org;
 minor cleanup and updated author email]
Signed-off-by: NJack Pham <jackp@codeaurora.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If someone provided meaningful error codes from reset() we should tell the
user what they were.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Correct use of devnum in supports_autosuspend documentation, the sysfs path
contains busnum-port.port.port not busnum-devnum (which is the usb bus device
address).
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The config descriptors as read from /proc/bus/usb/BBB/DDD are in *bus* endian
format. Correct proc_usb_info.txt to correctly reflect that.
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

While reading the config parsing code I noticed this check is missing, without
this check config->desc.wTotalLength can end up with a value larger then the
dev->rawdescriptors length for the config, and when userspace then tries to
get the rawdescriptors bad things may happen.
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Refactor so that register writes for configuration are only performed if
the device has a regmap provided and also register as a platform driver.
This allows the driver to be used to manage GPIO based control of the
device.
Signed-off-by: NMark Brown <broonie@linaro.org>
Cc: devicetree@vger.kernel.org
Reviewed-by: NDongjin Kim <tobetter@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The binding document says that all properties are required but in fact
almost all are optional (and should be) - update the document to reflect
this.
Signed-off-by: NMark Brown <broonie@linaro.org>
Cc: devicetree@vger.kernel.org
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There are no software visible differences that I am aware of but in case
any are discovered allow the DTS to specify exactly which device is
present.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Since there is no runtime interface for changing modes this is probably
the most sensible default.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for supporting operation without an I2C control interface
factor out the I2C-specific parts of the probe routine from those that
don't do any register I/O.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This will give access to the diagnostic infrastructure regmap has but
the main point is to support future refactoring.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If the connect signal is pulled high then the device will start up meaning
that if we just pull it high on probe then the device will start running
prior to the configuration being written out. Fix this by pulling the GPIO
low when we reset and only pulling it high when configuration is finished.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The /RESET GPIO is not manipulated from atomic context so support GPIOs
that can't be written from atomic context by using _cansleep().
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When CONFIG_USB_SUPPORT is not selected we get things like:

scripts/kconfig/mconf Kconfig
warning: (MIPS_SEAD3 && PMC_MSP && CPU_CAVIUM_OCTEON) selects USB_EHCI_BIG_ENDIAN_MMIO which has unmet direct dependencies (USB_SUPPORT && USB)

It is much cleaner to make the various system Kconfigs select
USB_EHCI_BIG_ENDIAN_MMIO rather than move the system config
information into USB's Kconfig, but the warnings are annoying.

Eliminate the warning by moving the definition of
USB_EHCI_BIG_ENDIAN_MMIO outside of all the Kconfig if statements.
While we are at it move USB_OHCI_BIG_ENDIAN_DESC,
USB_OHCI_BIG_ENDIAN_MMIO, USB_OHCI_LITTLE_ENDIAN and
USB_EHCI_BIG_ENDIAN_DESC too, as they could very well suffer similar
problems for other systems.

Get rid of the redundant "default n" in USB_OHCI_BIG_ENDIAN_DESC and
USB_OHCI_BIG_ENDIAN_MMIO
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Saves us a bit of code.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The intn and connect GPIO properties are swapped in the code which will
cause failures at runtime if these are connected, fix the code.

There are currently no in-tree users of this device to check or update.
Signed-off-by: NMark Brown <broonie@linaro.org>
Acked-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Instead of having to create a new driver for a "simple" usb to serial
device, mush them all into one file, with a macro, so as to make it easy
to add new ones.

Cc: "René Bürgel" <rene.buergel@sohard.de>
Acked-by: NWei Shuai <cpuwolf@gmail.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Acked-by: NFrans Klaver <frans.klaver@xsens.com>
Cc: "Wesley W. Terpstra" <w.terpstra@gsi.de>
Cc: Johan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch enables 'can_dma_sg' flag for ax88179_178a device
if the attached host controller supports building packet from
discontinuous buffers(DMA SG is possible), so TSO can be enabled
and skb fragment buffers can be passed to usb stack via urb->sg
directly.

With the patch, system CPU utilization decreased ~50% and throughput
increased by ~10% when doing iperf client test on one ARM A15 dual
core board.

Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Grant Grundler <grundler@google.com>
Cc: Oliver Neukum <oneukum@suse.de>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Freddy Xin <freddy@asix.com.tw>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Acked-by: NEric Dumazet <edumazet@gmail.com>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch introduces support of DMA SG if the USB host controller
which usbnet device is attached to is capable of building packet from
discontinuous buffers.

The patch supports passing the skb fragment buffers to usb stack directly
via urb->sg.

Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Grant Grundler <grundler@google.com>
Cc: Freddy Xin <freddy@asix.com.tw>
Cc: Alan Stern <stern@rowland.harvard.edu>
Acked-by: NOliver Neukum <oneukum@suse.de>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch marks all xHCI controllers as no_sg_constraint
since xHCI supports building packet from discontinuous buffers.

Cc: Alan Stern <stern@rowland.harvard.edu>
Acked-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Some host controllers(such as xHCI) can support building
packet from discontinuous buffers, so introduce one flag
and helper for this kind of host controllers, then the
feature can help some applications(such as usbnet) by
supporting arbitrary length of sg buffers.
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

All 4 transfer types can work well on EHCI HCD after switching to run
URB giveback in tasklet context, so mark all HCD drivers to support
it.

Also we don't need to release ehci->lock during URB giveback any more.

>From below test results on 3 machines(2 ARM and one x86), time
consumed by EHCI interrupt handler droped much without performance
loss.

1 test description
1.1 mass storage performance test:
- run below command 10 times and compute the average performance

    dd if=/dev/sdN iflag=direct of=/dev/null bs=200M count=1

- two usb mass storage device:
A: sandisk extreme USB 3.0 16G(used in test case 1 & case 2)
B: kingston DataTraveler G2 4GB(only used in test case 2)

1.2 uvc function test:
- run one simple capture program in the below link

   http://kernel.ubuntu.com/~ming/up/capture.c

- capture format 640*480 and results in High Bandwidth mode on the
uvc device: Z-Star 0x0ac8/0x3450

- on T410(x86) laptop, also use guvcview to watch video capture/playback

1.3 about test2 and test4
- both two devices involved are tested concurrently by above test items

1.4 how to compute irq time(the time consumed by ehci_irq)
- use trace points of irq:irq_handler_entry and irq:irq_handler_exit

1.5 kernel
3.10.0-rc3-next-20130528

1.6 test machines
Pandaboard A1: ARM CortexA9 dural core
Arndale board: ARM CortexA15 dural core
T410: i5 CPU 2.67GHz quad core

2 test result
2.1 test case1: single mass storage device performance test
--------------------------------------------------------------------
		upstream 		| patched
		perf(MB/s)+irq time(us)	| perf(MB/s)+irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  25.280(avg:145,max:772)	| 25.540(avg:14, max:75)
Arndale board:  29.700(avg:33, max:129)	| 29.700(avg:10,  max:50)
T410: 		34.430(avg:17, max:154*)| 34.660(avg:12, max:155)
---------------------------------------------------------------------

2.2 test case2: two mass storage devices' performance test
--------------------------------------------------------------------
		upstream 			| patched
		perf(MB/s)+irq time(us)		| perf(MB/s)+irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  15.840/15.580(avg:158,max:1216)	| 16.500/16.160(avg:15,max:139)
Arndale board:  17.370/16.220(avg:33 max:234)	| 17.480/16.200(avg:11, max:91)
T410: 		21.180/19.820(avg:18 max:160)	| 21.220/19.880(avg:11, max:149)
---------------------------------------------------------------------

2.3 test case3: one uvc streaming test
- uvc device works well(on x86, luvcview can be used too and has
same result with uvc capture)
--------------------------------------------------------------------
		upstream 		| patched
		irq time(us)		| irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  (avg:445, max:873)	| (avg:33, max:44)
Arndale board:  (avg:316, max:630)	| (avg:20, max:27)
T410: 		(avg:39,  max:107)	| (avg:10, max:65)
---------------------------------------------------------------------

2.4 test case4: one uvc streaming plus one mass storage device test
--------------------------------------------------------------------
		upstream 		| patched
		perf(MB/s)+irq time(us)	| perf(MB/s)+irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  20.340(avg:259,max:1704)| 20.390(avg:24, max:101)
Arndale board:  23.460(avg:124,max:726)	| 23.370(avg:15, max:52)
T410: 		28.520(avg:27, max:169)	| 28.630(avg:13, max:160)
---------------------------------------------------------------------

2.5 test case5: read single mass storage device with small transfer
- run below command 10 times and compute the average speed

 dd if=/dev/sdN iflag=direct of=/dev/null bs=4K count=4000

1), test device A:
--------------------------------------------------------------------
		upstream 		| patched
		perf(MB/s)+irq time(us)	| perf(MB/s)+irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  6.5(avg:21, max:64)	| 6.5(avg:10, max:24)
Arndale board:  8.13(avg:12, max:23)	| 8.06(avg:7,  max:17)
T410: 		6.66(avg:13, max:131)   | 6.84(avg:11, max:149)
---------------------------------------------------------------------

2), test device B:
--------------------------------------------------------------------
		upstream 		| patched
		perf(MB/s)+irq time(us)	| perf(MB/s)+irq time(us)
--------------------------------------------------------------------
Pandaboard A1:  5.5(avg:21,max:43)	| 5.49(avg:10, max:24)
Arndale board:  5.9(avg:12, max:22)	| 5.9(avg:7, max:17)
T410: 		5.48(avg:13, max:155)	| 5.48(avg:7, max:140)
---------------------------------------------------------------------

* On T410, sometimes read ehci status register in ehci_irq takes more
than 100us, and the problem has been reported on the link:

	http://marc.info/?t=137065867300001&r=1&w=2Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

ehci-hcd currently unlinks an interrupt QH when it becomes empty, that
is, after its last URB completes.  This works well because in almost
all cases, the completion handler for an interrupt URB resubmits the
URB; therefore the QH doesn't become empty and doesn't get unlinked.

When we start using tasklets for URB completion, this scheme won't work
as well.  The resubmission won't occur until the tasklet runs, which
will be some time after the completion is queued with the tasklet.
During that delay, the QH will be empty and so will be unlinked
unnecessarily.

To prevent this problem, this patch adds a 5-ms time delay before empty
interrupt QHs are unlinked.  Most often, during that time the interrupt
URB will be resubmitted and thus we can avoid unlinking the QH.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The patch does the below improvement:

- think QH_STATE_COMPLETING as unlinking state since all URBs on the
endpoint should be in unlinking or unlinked when doing endpoint_disable()

- add "WARN_ON(!list_empty(&qh->qtd_list));" if qh->qh_state is
QH_STATE_LINKED because there shouldn't be any active transfer in qh

- when qh->qh_state is QH_STATE_LINKED, the QH(async or periodic)
should be in its corresponding list, so the search through the async
list isn't necessary.

- unlink periodic QH to speed up unlinking if the QH is in linked
state

Basically, only the last one is related with this patchset because
the assumption of "periodic qh self-unlinks on empty" isn't true
any more when we introduce unlink-wait for periodic qh.
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There is no good reason to run complete() in hard interrupt
disabled context.

After switch to run complete() in tasklet, we will enable local IRQs
when calling complete() since we can do it at that time.

Even though we still disable IRQs now when calling complete()
in tasklet, the URB documentation is updated to claim complete()
will be run in tasklet context and local IRQs will be enabled, so
that USB drivers can know the change and avoid one deadlock caused
by: assume IRQs disabled in complete() and call spin_lock() to
hold lock which might be acquired in interrupt context.

Current spin_lock() usages in drivers' complete() will be cleaned
up at the same time, and once the cleanup is finished, local IRQs
will be enabled when calling complete() in tasklet.

Also fix description about type of usb_complete_t, and remove the
advice of running completion handler in tasklet for decreasing
system latency.

Cc: Oliver Neukum <oliver@neukum.org>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch implements the mechanism of giveback of URB in
tasklet context, so that hardware interrupt handling time for
usb host controller can be saved much, and HCD interrupt handling
can be simplified.

Motivations:

1), on some arch(such as ARM), DMA mapping/unmapping is a bit
time-consuming, for example: when accessing usb mass storage
via EHCI on pandaboard, the common length of transfer buffer is 120KB,
the time consumed on DMA unmapping may reach hundreds of microseconds;
even on A15 based box, the time is still about scores of microseconds

2), on some arch, reading DMA coherent memoery is very time-consuming,
the most common example is usb video class driver[1]

3), driver's complete() callback may do much things which is driver
specific, so the time is consumed unnecessarily in hardware irq context.

4), running driver's complete() callback in hardware irq context causes
that host controller driver has to release its lock in interrupt handler,
so reacquiring the lock after return may busy wait a while and increase
interrupt handling time. More seriously, releasing the HCD lock makes
HCD becoming quite complicated to deal with introduced races.

So the patch proposes to run giveback of URB in tasklet context, then
time consumed in HCD irq handling doesn't depend on drivers' complete and
DMA mapping/unmapping any more, also we can simplify HCD since the HCD
lock isn't needed to be released during irq handling.

The patch should be reasonable and doable:

1), for drivers, they don't care if the complete() is called in hard irq
context or softirq context

2), the biggest change is the situation in which usb_submit_urb() is called
in complete() callback, so the introduced tasklet schedule delay might be a
con, but it shouldn't be a big deal:

	- control/bulk asynchronous transfer isn't sensitive to schedule
	  delay

	- the patch schedules giveback of periodic URBs using
	  tasklet_hi_schedule, so the introduced delay should be very
	  small

	- for ISOC transfer, generally, drivers submit several URBs
	  concurrently to avoid interrupt delay, so it is OK with the
	  little schedule delay.

	- for interrupt transfer, generally, drivers only submit one URB
	  at the same time, but interrupt transfer is often used in event
	  report, polling, ... situations, and a little delay should be OK.

Considered that HCDs may optimize on submitting URB in complete(), the
patch may cause the optimization not working, so introduces one flag to mark
if the HCD supports to run giveback URB in tasklet context. When all HCDs
are ready, the flag can be removed.

[1], http://marc.info/?t=136438111600010&r=1&w=2

Cc: Oliver Neukum <oliver@neukum.org>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>