GitHub/MotorolaMobilityLLC/kernel-slsi.git
10 years agoUSB: usb_wwan: use interface-data accessors
Johan Hovold [Mon, 26 May 2014 17:23:26 +0000 (19:23 +0200)]
USB: usb_wwan: use interface-data accessors

Use usb_get_serial_data() rather than accessing the private pointer
directly.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: make resume error messages uniform
Johan Hovold [Mon, 26 May 2014 17:23:25 +0000 (19:23 +0200)]
USB: usb_wwan: make resume error messages uniform

Make resume error messages uniform.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: kill interrupt urb explicitly at suspend
Johan Hovold [Mon, 26 May 2014 17:23:24 +0000 (19:23 +0200)]
USB: usb_wwan: kill interrupt urb explicitly at suspend

As the port interrupt URB is submitted by the subdriver at open, we
should also kill it explicitly at suspend (even though this will be
taken care of by USB serial core otherwise).

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: remove redundant urb kill from port remove
Johan Hovold [Mon, 26 May 2014 17:23:23 +0000 (19:23 +0200)]
USB: usb_wwan: remove redundant urb kill from port remove

Remove redundant usb_kill_urb from port remove, which is called
post-shutdown (close).

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: remove unimplemented set_termios
Johan Hovold [Mon, 26 May 2014 17:23:22 +0000 (19:23 +0200)]
USB: usb_wwan: remove unimplemented set_termios

The driver does not implement set_termios so the operation can be left
unset (tty will do the tty_termios_copy_hw for us).

Note that the send_setup call is bogus as it really only sets DTR/RTS
to their current values.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: remove redundant modem-control request
Johan Hovold [Mon, 26 May 2014 17:23:21 +0000 (19:23 +0200)]
USB: usb_wwan: remove redundant modem-control request

The tty-port implementation has already made sure that DTR/RTS have been
raised by calling dtr_rts so remove the redundant call from open.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix remote wakeup
Johan Hovold [Mon, 26 May 2014 17:23:20 +0000 (19:23 +0200)]
USB: usb_wwan: fix remote wakeup

Make sure that needs_remote_wake up is always set when there are open
ports.

Currently close() would unconditionally set needs_remote_wakeup to 0
even though there might still be open ports. This could lead to blocked
input and possibly dropped data on devices that do not support remote
wakeup (and which must therefore not be runtime suspended while open).

Add an open_ports counter (protected by the susp_lock) and only clear
needs_remote_wakeup when the last port is closed.

Note that there are currently no multi-port drivers using the usb_wwan
implementation.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix discarded writes on resume errors
Johan Hovold [Mon, 26 May 2014 17:23:19 +0000 (19:23 +0200)]
USB: usb_wwan: fix discarded writes on resume errors

There's no reason not to try sending off any further delayed write urbs,
should one urb-submission fail.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix potential blocked I/O after resume
Johan Hovold [Mon, 26 May 2014 17:23:18 +0000 (19:23 +0200)]
USB: usb_wwan: fix potential blocked I/O after resume

Keep trying to submit urbs rather than bail out on first read-urb
submission error, which would also prevent I/O for any further ports
from being resumed.

Instead keep an error count, for all types of failed submissions, and
let USB core know that something went wrong.

Also make sure to always clear the suspended flag. Currently a failed
read-urb submission would prevent cached writes as well as any
subsequent writes from being submitted until next suspend-resume cycle,
something which may not even necessarily happen.

Note that USB core currently only logs an error if an interface resume
failed.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix potential NULL-deref at resume
Johan Hovold [Mon, 26 May 2014 17:23:17 +0000 (19:23 +0200)]
USB: usb_wwan: fix potential NULL-deref at resume

The interrupt urb was submitted unconditionally at resume, something
which could lead to a NULL-pointer dereference in the urb completion
handler as resume may be called after the port and port data is gone.

Fix this by making sure the interrupt urb is only submitted and active
when the port is open.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Cc: <stable@vger.kernel.org> # v2.6.32: 032129cb03df
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix urb leak at shutdown
Johan Hovold [Mon, 26 May 2014 17:23:16 +0000 (19:23 +0200)]
USB: usb_wwan: fix urb leak at shutdown

The delayed-write queue was never emptied at shutdown (close), something
which could lead to leaked urbs if the port is closed before being
runtime resumed due to a write.

When this happens the output buffer would not drain on close
(closing_wait timeout), and after consecutive opens, writes could be
corrupted with previously buffered data, transfered with reduced
throughput or completely blocked.

Note that unbusy_queued_urb() was simply moved out of CONFIG_PM.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix write and suspend race
Johan Hovold [Mon, 26 May 2014 17:23:15 +0000 (19:23 +0200)]
USB: usb_wwan: fix write and suspend race

Fix race between write() and suspend() which could lead to writes being
dropped (or I/O while suspended) if the device is runtime suspended
while a write request is being processed.

Specifically, suspend() releases the susp_lock after determining the
device is idle but before setting the suspended flag, thus leaving a
window where a concurrent write() can submit an urb.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix race between write and resume
xiao jin [Mon, 26 May 2014 17:23:14 +0000 (19:23 +0200)]
USB: usb_wwan: fix race between write and resume

We find a race between write and resume. usb_wwan_resume run play_delayed()
and spin_unlock, but intfdata->suspended still is not set to zero.
At this time usb_wwan_write is called and anchor the urb to delay
list. Then resume keep running but the delayed urb have no chance
to be commit until next resume. If the time of next resume is far
away, tty will be blocked in tty_wait_until_sent during time. The
race also can lead to writes being reordered.

This patch put play_Delayed and intfdata->suspended together in the
spinlock, it's to avoid the write race during resume.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Signed-off-by: xiao jin <jin.xiao@intel.com>
Signed-off-by: Zhang, Qi1 <qi1.zhang@intel.com>
Reviewed-by: David Cohen <david.a.cohen@linux.intel.com>
Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: usb_wwan: fix urb leak in write error path
xiao jin [Mon, 26 May 2014 17:23:13 +0000 (19:23 +0200)]
USB: usb_wwan: fix urb leak in write error path

When enable usb serial for modem data, sometimes the tty is blocked
in tty_wait_until_sent because portdata->out_busy always is set and
have no chance to be cleared.

We find a bug in write error path. usb_wwan_write set portdata->out_busy
firstly, then try autopm async with error. No out urb submit and no
usb_wwan_outdat_callback to this write, portdata->out_busy can't be
cleared.

This patch clear portdata->out_busy if usb_wwan_write try autopm async
with error.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Signed-off-by: xiao jin <jin.xiao@intel.com>
Signed-off-by: Zhang, Qi1 <qi1.zhang@intel.com>
Reviewed-by: David Cohen <david.a.cohen@linux.intel.com>
Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: option: add missing usb_mark_last_busy
Johan Hovold [Mon, 26 May 2014 17:23:12 +0000 (19:23 +0200)]
USB: option: add missing usb_mark_last_busy

We should call usb_mark_last_busy in all input paths, including the
interrupt completion handler.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: option: fix line-control pipe direction
Johan Hovold [Mon, 26 May 2014 17:23:11 +0000 (19:23 +0200)]
USB: option: fix line-control pipe direction

The option line-control request has been using the wrong pipe direction,
while relying on USB core to fix it up.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: option: fix runtime PM handling
Johan Hovold [Mon, 26 May 2014 17:23:10 +0000 (19:23 +0200)]
USB: option: fix runtime PM handling

Fix potential I/O while runtime suspended due to missing PM operations
in send_setup.

Fixes: 383cedc3bb43 ("USB: serial: full autosuspend support for the
option driver")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: do not resume I/O on closing ports
Johan Hovold [Mon, 26 May 2014 17:23:09 +0000 (19:23 +0200)]
USB: sierra: do not resume I/O on closing ports

Use tty-port initialised flag rather than private flag to determine when
port is closing down.

Since the tty-port flag is set prior to dropping DTR/RTS (when HUPCL is
set) this avoid submitting the read urbs when resuming the interface in
dtr_rts() only to immediately kill them again in shutdown().

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: minimise no-suspend window during close
Johan Hovold [Mon, 26 May 2014 17:23:08 +0000 (19:23 +0200)]
USB: sierra: minimise no-suspend window during close

Move usb_autopm_get_interface_no_resume to the end of close(). This
makes the window during which suspend is prevented before the final put
in USB serial core slightly smaller.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: refactor delayed-urb submission
Johan Hovold [Mon, 26 May 2014 17:23:07 +0000 (19:23 +0200)]
USB: sierra: refactor delayed-urb submission

Refactor and clean up delayed-urb submission at resume.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: clean up suspend
Johan Hovold [Mon, 26 May 2014 17:23:06 +0000 (19:23 +0200)]
USB: sierra: clean up suspend

Clean up suspend() somewhat and make sure to always set the suspended
flag (although it's only used for runtime PM) in order to match
resume().

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: use interface-data accessors
Johan Hovold [Mon, 26 May 2014 17:23:05 +0000 (19:23 +0200)]
USB: sierra: use interface-data accessors

Use usb_get_serial_data() rather than accessing the private pointer
directly.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: remove redundant modem-control requests
Johan Hovold [Mon, 26 May 2014 17:23:04 +0000 (19:23 +0200)]
USB: sierra: remove redundant modem-control requests

The tty-port implementation has already made sure that DTR/RTS have been
raised and lowered by calling dtr_rts so remove the redundant calls from
open and close.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: do not resume I/O on closed ports
Johan Hovold [Mon, 26 May 2014 17:23:03 +0000 (19:23 +0200)]
USB: sierra: do not resume I/O on closed ports

Do not resume any I/O, including the delayed write queue, on closed
ports.

Note that this currently has no functional impact due to the
usb_autopm_get_interface() in close(), but that call is about to be
removed by a follow-up patch.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: remove disconnected test from close
Johan Hovold [Mon, 26 May 2014 17:23:02 +0000 (19:23 +0200)]
USB: sierra: remove disconnected test from close

Remove no longer needed disconnected test from close, which is never
called post disconnect (and drivers must handle failed I/O during
disconnect anyway).

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: remove unimplemented set_termios
Johan Hovold [Mon, 26 May 2014 17:23:01 +0000 (19:23 +0200)]
USB: sierra: remove unimplemented set_termios

The driver does not implement set_termios so the operation can be left
unset (tty will do the tty_termios_copy_hw for us).

Note that the send_setup call is bogus as it really only sets DTR/RTS
to their current values.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: remove unused variable
Johan Hovold [Mon, 26 May 2014 17:23:00 +0000 (19:23 +0200)]
USB: sierra: remove unused variable

Remove unused variable from sierra_release_urb.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: remove bogus endpoint test
Johan Hovold [Mon, 26 May 2014 17:22:59 +0000 (19:22 +0200)]
USB: sierra: remove bogus endpoint test

Remove bogus endpoint-address test which is never true.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix line-control pipe direction
Johan Hovold [Mon, 26 May 2014 17:22:58 +0000 (19:22 +0200)]
USB: sierra: fix line-control pipe direction

The sierra line-control request has been using the wrong pipe direction,
while relying on USB core to fix it up.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix resume error reporting
Johan Hovold [Mon, 26 May 2014 17:22:57 +0000 (19:22 +0200)]
USB: sierra: fix resume error reporting

Add error message to resume error path and make sure to also return an
error when failing to submit a cached write.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix urbs not being killed on shutdown
Johan Hovold [Mon, 26 May 2014 17:22:56 +0000 (19:22 +0200)]
USB: sierra: fix urbs not being killed on shutdown

Make sure to stop all I/O, including any active write urbs, at shutdown.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix characters being dropped at close
Johan Hovold [Mon, 26 May 2014 17:22:55 +0000 (19:22 +0200)]
USB: sierra: fix characters being dropped at close

Fix characters potentially being dropped at close due to missing
chars_in_buffer implementation.

Note that currently the write urbs are not even killed at close (will be
fixed separately), but this could still lead to dropped data since we
have lowered DTR/RTS.

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix remote wakeup
Johan Hovold [Mon, 26 May 2014 17:22:54 +0000 (19:22 +0200)]
USB: sierra: fix remote wakeup

Make sure that needs_remote_wake up is always set when there are open
ports.

Currently close() would unconditionally set needs_remote_wakeup to 0
even though there might still be open ports. This could lead to blocked
input and possibly dropped data on devices that do not support remote
wakeup (and which must therefore not be runtime suspended while open).

Add an open_ports counter (protected by the susp_lock) and only clear
needs_remote_wakeup when the last port is closed.

Fixes: e6929a9020ac ("USB: support for autosuspend in sierra while
online")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix urb and memory leak on disconnect
Johan Hovold [Mon, 26 May 2014 17:22:53 +0000 (19:22 +0200)]
USB: sierra: fix urb and memory leak on disconnect

The delayed-write queue was never emptied on disconnect, something which
would lead to leaked urbs and transfer buffers if the device is
disconnected before being runtime resumed due to a write.

Fixes: e6929a9020ac ("USB: support for autosuspend in sierra while
online")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix urb and memory leak in resume error path
Johan Hovold [Mon, 26 May 2014 17:22:52 +0000 (19:22 +0200)]
USB: sierra: fix urb and memory leak in resume error path

Neither the transfer buffer or the urb itself were released in the
resume error path for delayed writes. Also on errors, the remainder of
the queue was not even processed, which leads to further urb and buffer
leaks.

The same error path also failed to balance the outstanding-urb counter,
something which results in degraded throughput or completely blocked
writes.

Fix this by releasing urb and buffer and balancing counters on errors,
and by always processing the whole queue even when submission of one urb
fails.

Fixes: e6929a9020ac ("USB: support for autosuspend in sierra while
online")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix use after free at suspend/resume
Johan Hovold [Mon, 26 May 2014 17:22:51 +0000 (19:22 +0200)]
USB: sierra: fix use after free at suspend/resume

Fix use after free or NULL-pointer dereference during suspend and
resume.

The port data may never have been allocated (port probe failed)
or may already have been released by port_remove (e.g. driver is
unloaded) when suspend and resume are called.

Fixes: e6929a9020ac ("USB: support for autosuspend in sierra while
online")

Cc: <stable@vger.kernel.org> # v2.6.32
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: sierra: fix AA deadlock in open error path
Johan Hovold [Mon, 26 May 2014 17:22:50 +0000 (19:22 +0200)]
USB: sierra: fix AA deadlock in open error path

Fix AA deadlock in open error path that would call close() and try to
grab the already held disc_mutex.

Fixes: b9a44bc19f48 ("sierra: driver urb handling improvements")

Cc: <stable@vger.kernel.org> # v2.6.31
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: iowarrior: Convert local dbg macro to dev_dbg
Joe Perches [Tue, 20 May 2014 20:09:33 +0000 (13:09 -0700)]
USB: iowarrior: Convert local dbg macro to dev_dbg

Use a more standard logging style.

Add terminating newlines to formats.
Remove __func__ as that can be added via dynamic debug.
Remove now unnecessary debug module parameter.
Remove the dbg macro too.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoUSB: appledisplay: Convert /n to \n
Joe Perches [Fri, 25 Apr 2014 01:51:02 +0000 (18:51 -0700)]
USB: appledisplay: Convert /n to \n

Use a newline character appropriately.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: qcserial: remove interface number matching
Bjørn Mork [Sun, 27 Apr 2014 14:47:46 +0000 (16:47 +0200)]
usb: qcserial: remove interface number matching

Matching on interface numbers was not such a good idea
for multi-function serial devices after all. It is much
better do create well defined device layouts, allowing
a single match entry per device.

Remove this now unused code.

Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: qcserial: define and use Sierra Wireless layout
Bjørn Mork [Sun, 27 Apr 2014 14:47:45 +0000 (16:47 +0200)]
usb: qcserial: define and use Sierra Wireless layout

All the "non Gobi" Qualcomm based devices handled by this
driver share a common standard Sierra Wireless specific
layout. Adding code specifically for this layout allow
us to reduce the number of match entries per device from
three to one.

This change will result in a penalty wrt stable backports,
but simplifies new Sierra device addtitions in the long
term.

Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: qcserial: refactor device layout selection
Bjørn Mork [Sun, 27 Apr 2014 14:47:44 +0000 (16:47 +0200)]
usb: qcserial: refactor device layout selection

Preparing for more supported standard device layouts. Keeping
the matching macros unchanged to avoid breaking stable
backporting of new device additions.

Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: qcserial: fix multiline comment coding style
Bjørn Mork [Sun, 27 Apr 2014 14:47:43 +0000 (16:47 +0200)]
usb: qcserial: fix multiline comment coding style

Use a consistent style for all multiline comments.

Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agouwb: add error messages when reservation establish fails
Thomas Pugliese [Fri, 25 Apr 2014 15:30:33 +0000 (10:30 -0500)]
uwb: add error messages when reservation establish fails

Add better error messages during the channel change/reservation establish
process.

Signed-off-by: Thomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agouwb: fix channel change failure
Thomas Pugliese [Fri, 25 Apr 2014 15:30:32 +0000 (10:30 -0500)]
uwb: fix channel change failure

Make the transition to the UWB_RSV_STATE_NONE state synchronous so that
there is not a race between uwb_rsv_terminate and uwb_rsv_establish.
uwb_rsv_terminate would set the rsv->state to UWB_RSV_STATE_NONE but did
not release the stream resource until a 320ms timeout had expired.  If a
user called uwb_rsv_establish during that time, it could fail to
establish the reservation because no stream resources were available.
This patch removes the timer from the uwb_rsv_terminate process since it
is not needed when transitioning to UWB_RSV_STATE_NONE.

Signed-off-by: Thomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ohci-exynos: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:10 +0000 (17:30 +0530)]
usb: host: ohci-exynos: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ehci-tegra: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:09 +0000 (17:30 +0530)]
usb: host: ehci-tegra: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ehci-spear: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:08 +0000 (17:30 +0530)]
usb: host: ehci-spear: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ehci-mv: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:07 +0000 (17:30 +0530)]
usb: host: ehci-mv: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ehci-msm: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:06 +0000 (17:30 +0530)]
usb: host: ehci-msm: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: host: ehci-exynos: Use devm_ioremap_resource instead of devm_ioremap
Vivek Gautam [Sat, 10 May 2014 12:00:05 +0000 (17:30 +0530)]
usb: host: ehci-exynos: Use devm_ioremap_resource instead of devm_ioremap

Using devm_ioremap_resource() API should actually be preferred over
devm_ioremap(), since the former request the mem region first and then
gives back the ioremap'ed memory pointer.
devm_ioremap_resource() calls request_mem_region(), therby preventing
other drivers to make any overlapping call to the same region.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: ehci-exynos: Change to use phy provided by the generic phy framework
Kamil Debski [Mon, 5 May 2014 05:02:28 +0000 (10:32 +0530)]
usb: ehci-exynos: Change to use phy provided by the generic phy framework

Add the phy provider, supplied by new Exynos-usb2phy using
Generic phy framework.
Keeping the support for older USB phy intact right now, in order
to prevent any functionality break in absence of relevant
device tree side change for ehci-exynos.
Once we move to new phy in the device nodes for ehci, we can
remove the support for older phys.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
[gautam.vivek@samsung.com: Addressed review comments from mailing list]
[gautam.vivek@samsung.com: Kept the code for old usb-phy, and just
added support for new exynos5-usb2phy in generic phy framework]
[gautam.vivek@samsung.com: Edited the commit message]
Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Reviewed-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: ohci-exynos: Add facility to use phy provided by the generic phy framework
Vivek Gautam [Mon, 5 May 2014 05:02:57 +0000 (10:32 +0530)]
usb: ohci-exynos: Add facility to use phy provided by the generic phy framework

Add support to consume phy provided by Generic phy framework.
Keeping the support for older usb-phy intact right now, in order
to prevent any functionality break in absence of relevant
device tree side change for ohci-exynos.
Once we move to new phy in the device nodes for ohci, we can
remove the support for older phys.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Cc: Jingoo Han <jg1.han@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Reviewed-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: ehci-exynos: Use struct device instead of platform_device
Vivek Gautam [Mon, 5 May 2014 05:04:25 +0000 (10:34 +0530)]
usb: ehci-exynos: Use struct device instead of platform_device

Change to use struct device instead of struct platform_device
for some static functions.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: ohci-exynos: Use struct device instead of platform_device
Vivek Gautam [Mon, 5 May 2014 05:03:42 +0000 (10:33 +0530)]
usb: ohci-exynos: Use struct device instead of platform_device

Change to use struct device instead of struct platform_device
for some static functions.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: phy: add run-time dependencies to R-Car driver
Jean Delvare [Fri, 23 May 2014 08:36:22 +0000 (10:36 +0200)]
usb: phy: add run-time dependencies to R-Car driver

The Renesas R-Car USB PHY driver only supports the R8A7778 and
R8A7779, it isn't useful on other systems unless build testing.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Felipe Balbi <balbi@ti.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: phy: msm: fix bug in probe()
Dan Carpenter [Mon, 19 May 2014 20:35:19 +0000 (23:35 +0300)]
usb: phy: msm: fix bug in probe()

My previous patch introduced a bug which prevented this driver from
loading.  devm_ioremap_resource() has a call to
devm_request_mem_region() which will fail because the address space is
shared between this PHY driver and CI device controller driver.

Fixes: 10f0577aa5cb ('usb: phy: msm: change devm_ioremap() to devm_ioremap_resource()')
Reported-by:"Ivan T. Ivanov" <iivanov@mm-sol.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: udc: update gadget states according to ch9
Peter Chen [Sun, 4 May 2014 01:24:44 +0000 (09:24 +0800)]
usb: chipidea: udc: update gadget states according to ch9

Update device states according to ch9 in USB 2.0 specification

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoDoc: usb: chipidea: need to build both kernel Image and modules
Peter Chen [Sun, 4 May 2014 01:24:43 +0000 (09:24 +0800)]
Doc: usb: chipidea: need to build both kernel Image and modules

When tried to enable OTG FSM, we need to rebuild both kernel Image
and modules, since there are some codes at gadget modules which are
controlled by related configurations.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: msm: Initialize offset of the capability registers
Ivan T. Ivanov [Sun, 4 May 2014 01:24:42 +0000 (09:24 +0800)]
usb: chipidea: msm: Initialize offset of the capability registers

Since commit 62bb84e (usb: gadget: ci13xxx: convert to platform device)
start address of the capability registers is not passed correctly to
udc_probe(). Fix this.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Ivan T. Ivanov <iivanov@mm-sol.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: msm: Add device tree support
Ivan T. Ivanov [Sun, 4 May 2014 01:24:41 +0000 (09:24 +0800)]
usb: chipidea: msm: Add device tree support

Allows controller to be specified via device tree.
Pass PHY phandle specified in DT to core driver.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Ivan T. Ivanov <iivanov@mm-sol.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: msm: Add device tree binding information
Ivan T. Ivanov [Sun, 4 May 2014 01:24:40 +0000 (09:24 +0800)]
usb: chipidea: msm: Add device tree binding information

Document device tree binding information as required by
the Qualcomm USB controller.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Ivan T. Ivanov <iivanov@mm-sol.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agochipidea: usbmisc_imx: Allow USB OTG to work on mx51
Fabio Estevam [Sun, 4 May 2014 01:24:39 +0000 (09:24 +0800)]
chipidea: usbmisc_imx: Allow USB OTG to work on mx51

The field PLLDIVVALUE of register PHY_CTRL_1 selects the reference clock source
for the PHY:
00 = sysclock uses 19.2 MHz
01 = sysclock uses 24 MHz
10 = sysclock uses 26 MHz
11 = sysclock uses 27 MHz

The reset value for this field is 10 according to the reference manual, and
even though this reset value works for mx53, it does not work for mx51.

So instead of relying on the reset value for the PLLDIVVALUE field, explicitly
set it to 01 so that a 24MHz clock can be selected for the PHY and allowing both
mx51 and mx53 to have USB OTG port functional.

Succesfully tested 'g_ether' on a imx51-babbage and on a imx53-qsb boards.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: using one inline function to cover queue work operations
Peter Chen [Fri, 23 May 2014 00:12:49 +0000 (08:12 +0800)]
usb: chipidea: using one inline function to cover queue work operations

The otg queue work include operations: one is disable interrupt,
another one is call kernel queue work API. Many codes do this
operation, using one inline function to instead of them.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: udc: delete useless code
Peter Chen [Fri, 23 May 2014 00:12:48 +0000 (08:12 +0800)]
usb: chipidea: udc: delete useless code

Delete useless code

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: chipidea: update TODO list
Peter Chen [Fri, 23 May 2014 00:12:47 +0000 (08:12 +0800)]
usb: chipidea: update TODO list

Update TODO list

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoMerge tag 'usb-for-v3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi...
Greg Kroah-Hartman [Fri, 23 May 2014 02:28:21 +0000 (11:28 +0900)]
Merge tag 'usb-for-v3.16' of git://git./linux/kernel/git/balbi/usb into usb-next

Felipe writes:

usb: patches for v3.16 merge window

Not a lot here during this merge window. Mostly we just have
the usual miscellaneous patches (removal of unnecessary prints,
proper dependencies being added to Kconfig, build warning fixes,
new device ID, etc.

Other than those, the only important new features are the
new support for OS Strings which should help Linux Gadget
Drivers behave better under MS Windows. Also Babble Recovery
implementation for MUSB on AM335x. Lastly, we also have
ARCH_QCOM PHY support though phy-msm.

Signed-of-by: Felipe Balbi <balbi@ti.com>
Conflicts:
drivers/usb/phy/phy-mv-u3d-usb.c

10 years agoxhci: rework command timeout and cancellation,
Mathias Nyman [Thu, 8 May 2014 16:26:03 +0000 (19:26 +0300)]
xhci: rework command timeout and cancellation,

Use one timer to control command timeout.

start/kick the timer every time a command is completed and a
new command is waiting, or a new command is added to a empty list.

If the timer runs out, then tag the current command as "aborted", and
start the xhci command abortion process.

Previously each function that submitted a command had its own timer.
If that command timed out, a new command structure for the
command was created and it was put on a cancel_cmd_list list,
then a pci write to abort the command ring was issued.

when the ring was aborted, it checked if the current command
was the one to be canceled, later when the ring was stopped the
driver got ownership of the TRBs in the command ring,
compared then to the TRBs in the cancel_cmd_list,
and turned them into No-ops.

Now, instead, at timeout we tag the status of the command in the
command queue to be aborted, and start the ring abortion.
Ring abortion stops the command ring and gives control of the
commands to us.
All the aborted commands are now turned into No-ops.

If the ring is already stopped when the command times outs its not possible
to start the ring abortion, in this case the command is turnd to No-op
right away.

All these changes allows us to remove the entire cancel_cmd_list code.

The functions waiting for a command to finish no longer have their own timeouts.
They will wait either until the command completes normally,
or until the whole command abortion is done.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: Use completion and status in global command queue
Mathias Nyman [Thu, 8 May 2014 16:26:02 +0000 (19:26 +0300)]
xhci: Use completion and status in global command queue

Remove the per-device command list and handle_cmd_in_cmd_wait_list()
and use the completion and status variables found in the
command structure in the global command list.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: Add a global command queue
Mathias Nyman [Thu, 8 May 2014 16:26:01 +0000 (19:26 +0300)]
xhci: Add a global command queue

Create a list to store command structures, add a structure to it every time
a command is submitted, and remove it from the list once we get a
command completion event matching the command.

Callers that wait for completion will free their command structures themselves.
The other command structures are freed in the command completion event handler.

Also add a check that prevents queuing commands if host is dying

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: Use command structures when queuing commands on the command ring
Mathias Nyman [Thu, 8 May 2014 16:26:00 +0000 (19:26 +0300)]
xhci: Use command structures when queuing commands on the command ring

To create a global command queue we require that each command put on the
command ring is submitted with a command structure.

Functions that queue commands and wait for completion need to allocate a command
before submitting it, and free it once completed. The following command queuing
functions need to be modified.

xhci_configure_endpoint()
xhci_address_device()
xhci_queue_slot_control()
xhci_queue_stop_endpoint()
xhci_queue_new_dequeue_state()
xhci_queue_reset_ep()
xhci_configure_endpoint()

xhci_configure_endpoint() could already be called with a command structure,
and only xhci_check_maxpacket and xhci_check_bandwidth did not do so. These
are changed and a command structure is now required. This change also simplifies
the configure endpoint command completion handling and the "goto bandwidth_change"
handling code can be removed.

In some cases the command queuing function is called in interrupt context.
These commands needs to be allocated atomically, and they can't wait for
completion. These commands will in this patch be freed directly after queuing,
but freeing will be moved to the command completion event handler in a later
patch once we get the global command queue up.(Just so that we won't leak
memory in the middle of the patch set)

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: Report max device limit when Enable Slot command fails.
Sarah Sharp [Thu, 8 May 2014 16:25:59 +0000 (19:25 +0300)]
xhci: Report max device limit when Enable Slot command fails.

xHCI host controllers may only support a limited number of device slot
IDs, which is usually far less than the theoretical maximum number of
devices (255) that the USB specifications advertise.  This is
frustrating to consumers that expect to be able to plug in a large
number of devices.

Add a print statement when the Enable Slot command fails to show how
many devices the host supports.  We can't change hardware manufacturer's
design decisions, but hopefully we can save customers a little bit of
time trying to debug why their host mysteriously fails when too many
devices are plugged in.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Reported-by: Amund Hov <Amund.Hov@silabs.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: Use pci_enable_msix_exact() instead of pci_enable_msix()
Alexander Gordeev [Thu, 8 May 2014 16:25:58 +0000 (19:25 +0300)]
xhci: Use pci_enable_msix_exact() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range()  or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-usb@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: xhci: Use IS_ENABLED() macro
Fabio Estevam [Thu, 8 May 2014 16:25:57 +0000 (19:25 +0300)]
usb: xhci: Use IS_ENABLED() macro

Using the IS_ENABLED() macro can make the code shorter and easier to read.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: catch attempts to submit urbs with a vmalloc'd transfer buffer
Dan Williams [Thu, 8 May 2014 16:25:56 +0000 (19:25 +0300)]
usb: catch attempts to submit urbs with a vmalloc'd transfer buffer

Save someone else the debug cycles of figuring out why a driver's
transfer request is failing or causing undefined system behavior.
Buffers submitted for dma must come from GFP allocated / DMA-able
memory.

Return -EAGAIN matching the return value for dma_mapping_error() cases.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoxhci: fix wrong port number reported when setting USB2.0 hardware LPM.
Lin Wang [Thu, 8 May 2014 16:25:54 +0000 (19:25 +0300)]
xhci: fix wrong port number reported when setting USB2.0 hardware LPM.

This patch fix wrong port number reported when trying to enable/disable
USB2.0 hardware LPM.

Signed-off-by: Lin Wang <lin.x.wang@intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoMerge tag 'for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/kishon/linux...
Greg Kroah-Hartman [Tue, 20 May 2014 00:56:05 +0000 (09:56 +0900)]
Merge tag 'for-3.16' of git://git./linux/kernel/git/kishon/linux-phy into usb-next

Kishon writes:

for 3.16 merge window

Bunch of fixes and a new driver for Exynos5 USB 3.0 PHY.

10 years agoMerge 3.15-rc5 into usb-next
Greg Kroah-Hartman [Tue, 20 May 2014 00:49:41 +0000 (09:49 +0900)]
Merge 3.15-rc5 into usb-next

We need these USB fixes in here as well.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agousb: musb: tusb6010: Use musb->tusb_revision instead of tusb_get_revision call.
Matwey V. Kornilov [Fri, 16 May 2014 14:20:52 +0000 (18:20 +0400)]
usb: musb: tusb6010: Use musb->tusb_revision instead of tusb_get_revision call.

The value of the revision is stored in musb->tusb_revision,
so don't re-read it every time.
Exporting tusb_get_revision is not needed anymore,
so the dependency loop between tusb6010 and tusb6010_omap is resolved.

Signed-off-by: Matwey V. Kornilov <matwey@sai.msu.ru>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: tusb6010: Add tusb_revision to struct musb to store the revision.
Matwey V. Kornilov [Fri, 16 May 2014 14:19:54 +0000 (18:19 +0400)]
usb: musb: tusb6010: Add tusb_revision to struct musb to store the revision.

Add field to store tusb6010 revision value. Read the revision at
the startup and store to the variable.

Signed-off-by: Matwey V. Kornilov <matwey@sai.msu.ru>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Fix NULL pointer dereference
Ricardo Ribalda Delgado [Thu, 15 May 2014 12:28:45 +0000 (14:28 +0200)]
usb: gadget: net2280: Fix NULL pointer dereference

When DEBUG is enabled driver->driver.name is accessed, but driver
can be NULL

[  174.411689] BUG: unable to handle kernel NULL pointer dereference at 0000000000000040
[  174.429043] RIP: 0010:[<ffffffffa0131ca3>]  [<ffffffffa0131ca3>] net2280_stop+0xa3/0x100 [net2280]
[  174.457910] Call Trace:
[  174.459503]  [<ffffffffa00dd92a>] usb_gadget_remove_driver+0x5a/0xb0 [udc_core]
[  174.462693]  [<ffffffffa00ddd84>] usb_del_gadget_udc+0xb4/0x110 [udc_core]
[  174.464316]  [<ffffffffa012e2bf>] net2280_remove+0x2f/0x1c0 [net2280]

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: uvc: Set the vb2 queue timestamp flags
Laurent Pinchart [Sun, 23 Mar 2014 15:19:50 +0000 (16:19 +0100)]
usb: gadget: uvc: Set the vb2 queue timestamp flags

The vb2 queue timestamp_flags field must be set by drivers, as enforced
by a WARN_ON in vb2_queue_init. The UVC gadget driver failed to do so.
This resulted in the following warning.

[    2.104371] g_webcam gadget: uvc_function_bind
[    2.105567] ------------[ cut here ]------------
[    2.105567] ------------[ cut here ]------------
[    2.106779] WARNING: CPU: 0 PID: 1 at drivers/media/v4l2-core/videobuf2-core.c:2207 vb2_queue_init+0xa3/0x113()

Fix it.

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: uvc: Set the V4L2 buffer field to V4L2_FIELD_NONE
Laurent Pinchart [Sun, 23 Mar 2014 15:25:09 +0000 (16:25 +0100)]
usb: gadget: uvc: Set the V4L2 buffer field to V4L2_FIELD_NONE

The UVC gadget driver doesn't support interlaced video but left the
buffer field uninitialized. Set it to V4L2_FIELD_NONE.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: uvc: Switch to monotonic clock for buffer timestamps
Laurent Pinchart [Sun, 23 Mar 2014 15:25:09 +0000 (16:25 +0100)]
usb: gadget: uvc: Switch to monotonic clock for buffer timestamps

The wall time clock isn't useful for applications as it can jump around
due to time adjustement. Switch to the monotonic clock.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: Add SMSC USB334x PHY ID
Liviu Dudau [Wed, 14 May 2014 20:32:39 +0000 (21:32 +0100)]
usb: phy: Add SMSC USB334x PHY ID

adding new device id for SMSC USB334x devices.

Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Ryan Harkin <ryan.harkin@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: f_uac2: don't queue new requests when shutting down
Andrzej Pietrasiewicz [Thu, 15 May 2014 11:43:50 +0000 (13:43 +0200)]
usb: gadget: f_uac2: don't queue new requests when shutting down

In some circumstances when g_audio is being unloaded there happens
an endless loop in udc driver. It has happend on a board with
s3c-hsotg. If there are requests in endpoint's queue, they are completed
in a loop. But completing them might cause appending new requests
to the queue. This patch causes agdev_iso_complete() to return immediately
if request's status is -ESHUTDOWN. If it does not return immediately,
then although the current request is removed from the queue, a new one
is appended to the queue, so the above mentioned loop cannot end.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: convert to pcim_enable_device()
Andy Shevchenko [Thu, 15 May 2014 12:53:33 +0000 (15:53 +0300)]
usb: dwc3: convert to pcim_enable_device()

This fixes a bug when dwc3_pci_register_phys() fails and leaves device enabled.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: no need to initialize ret variable
Andy Shevchenko [Thu, 15 May 2014 12:53:32 +0000 (15:53 +0300)]
usb: dwc3: no need to initialize ret variable

First usage of ret variable will re-write initial value. Thus, there is no need
to initialize it.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: configfs: OS Extended Properties descriptors support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:28 +0000 (14:06 +0200)]
usb: gadget: configfs: OS Extended Properties descriptors support

Add handling of OS Extended Properties descriptors from configfs interface.
One kind of "OS Descriptors" are "Extended Properties" descriptors, which
need to be specified per interface or per group of interfaces described
by an IAD. This patch adds support for creating subdirectories
in interface.<n> directory located in the function's directory.
Names of subdirectories created become names of properties.
Each property contains two attributes: "type" and "data".
The type can be a numeric value 1..7 while data is a blob interpreted
depending on the type specified.
The types are:
1 - unicode string
2 - unicode string with environment variables
3 - binary
4 - little-endian 32-bit
5 - big-endian 32-bit
6 - unicode string with a symbolic link
7 - multiple unicode strings

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: f_rndis: OS Descriptors configfs support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:27 +0000 (14:06 +0200)]
usb: gadget: f_rndis: OS Descriptors configfs support

Added handling of OS Descriptors support for f_rndis.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: configfs: OS Extended Compatibility descriptors support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:26 +0000 (14:06 +0200)]
usb: gadget: configfs: OS Extended Compatibility descriptors support

Add handling of OS Extended Compatibility descriptors from configfs interface.
Hosts which expect the "OS Descriptors" ask only for configurations @ index 0,
but linux-based USB devices can provide more than one configuration.
This patch adds marking one of gadget's configurations the configuration
to be reported at index 0, regardless of the actual sequence of usb_add_config
invocations used for adding the configurations. The configuration is selected
by creating a symbolic link pointing to it from the "os_desc" directory
located at the top of a gadget's directory hierarchy.

One kind of "OS Descriptors" are "Extended Compatibility Descriptors",
which need to be specified per interface. This patch adds interface.<n>
directory in function's configfs directory to represent each interface
defined by the function. Each interface's directory contains two attributes:
"compatible_id" and "sub_compatible_id", which represent 8-byte
strings to be reported to the host as the "Compatible ID" and "Sub Compatible
ID".

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: configfs: OS String support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:25 +0000 (14:06 +0200)]
usb: gadget: configfs: OS String support

Add handling of OS String extension from the configfs interface.
A directory "os_desc" is added at the top level of a gadget's
directories hierarchy. In the "os_desc" directory there are
three attributes: "use", "b_vendor_code" and "qw_sign".
If "use" contains "0" the OS string is not reported to the host.
"b_vendor_code" contains a one-byte value which is used
for custom per-device and per-interface requests.
"qw_sign" contains an identifier to be reported as the "OS String"
proper.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: f_rndis: OS descriptors support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:24 +0000 (14:06 +0200)]
usb: gadget: f_rndis: OS descriptors support

In order for usb functions to expose OS descriptors they
need to be made aware of OS descriptors. This involves
extending the "options" structure and setting up
appropriate associations.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: OS Feature Descriptors support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:23 +0000 (14:06 +0200)]
usb: gadget: OS Feature Descriptors support

There is a custom (non-USB IF) extension to the USB standard:

http://msdn.microsoft.com/library/windows/hardware/gg463182

They grant permission to use the specification - there is
"Microsoft OS Descriptor Specification License Agreement"
under the link mentioned above, and its Section 2 "Grant
of License", letter (b) reads:

"Patent license. Microsoft hereby grants to You a nonexclusive,
royalty-free, nontransferable, worldwide license under Microsoft’s
patents embodied solely within the Specification and that are owned
or licensable by Microsoft to make, use, import, offer to sell,
sell and distribute directly or indirectly to Your Licensees Your
Implementation. You may sublicense this patent license to Your
Licensees under the same terms and conditions."

The said extension is maintained by Microsoft for Microsoft.

Yet it is fairly common for various devices to use it, and a
popular proprietary operating system expects devices to provide
"OS descriptors", so Linux-based USB gadgets whishing to be able
to talk to a variety of operating systems should be able to provide
the "OS descriptors".

This patch adds optional support for gadgets whishing to expose
the so called "OS Feature Descriptors", that is "Extended Compatibility ID"
and "Extended Properties".

Hosts which do request "OS descriptors" from gadgets do so during
the enumeration phase and before the configuration is set with
SET_CONFIGURATION. What is more, those hosts never ask for configurations
at indices other than 0. Therefore, gadgets whishing to provide
"OS descriptors" must designate one configuration to be used with
this kind of hosts - this is what os_desc_config is added for in
struct usb_composite_dev. There is an additional advantage to it:
if a gadget provides "OS descriptors" and designates one configuration
to be used with such non-USB-compliant hosts it can invoke
"usb_add_config" in any order because the designated configuration
will be reported to be at index 0 anyway.

This patch also adds handling vendor-specific requests addressed
at device or interface and related to handling "OS descriptors".

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: OS String support
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:22 +0000 (14:06 +0200)]
usb: gadget: OS String support

There is a custom (non-USB IF) extension to the USB standard:

http://msdn.microsoft.com/library/windows/hardware/gg463182

They grant permission to use the specification - there is
"Microsoft OS Descriptor Specification License Agreement"
under the link mentioned above, and its Section 2 "Grant
of License", letter (b) reads:

"Patent license. Microsoft hereby grants to You a nonexclusive,
royalty-free, nontransferable, worldwide license under Microsoft’s
patents embodied solely within the Specification and that are owned
or licensable by Microsoft to make, use, import, offer to sell,
sell and distribute directly or indirectly to Your Licensees Your
Implementation. You may sublicense this patent license to Your
Licensees under the same terms and conditions."

The said extension is maintained by Microsoft for Microsoft.

Yet it is fairly common for various devices to use it, and a
popular proprietary operating system expects devices to provide
"OS descriptors", so Linux-based USB gadgets whishing to be able
to talk to a variety of operating systems should be able to provide
the "OS descriptors".

This patch adds optional support for gadgets whishing to expose
the so called "OS String" under index 0xEE of language 0.
The contents of the string is generated based on the qw_sign
array and b_vendor_code.

Interested gadgets need to set the cdev->use_os_string flag,
fill cdev->qw_sign with appropriate values and fill cdev->b_vendor_code
with a value of their choice.

This patch does not however implement responding to any vendor-specific
USB requests.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: FunctionFS: share VLA macros with all usb gadget files
Andrzej Pietrasiewicz [Thu, 8 May 2014 12:06:21 +0000 (14:06 +0200)]
usb: gadget: FunctionFS: share VLA macros with all usb gadget files

Variable Length Array macros allow portable (compilable with both gcc
and clang) way of allocating a number of structures using a single
memory chunk. They can be useful for files other than f_fs.c,
so move them to a header file.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: msm: change devm_ioremap() to devm_ioremap_resource()
Dan Carpenter [Wed, 14 May 2014 13:54:47 +0000 (16:54 +0300)]
usb: phy: msm: change devm_ioremap() to devm_ioremap_resource()

There are several issues here:
1) platform_get_resource() can return NULL and that wasn't handled.
2) We should request the memory before we remap it, and
   devm_ioremap_resource() does that.
3) devm_ioremap() returns a NULL but we were checking for IS_ERR().

Fixes: 6b99c68ec1f9 ('usb: phy: msm: Migrate to Managed Device Resource allocation')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: msm: reset controller is mandatory now
Arnd Bergmann [Tue, 13 May 2014 19:46:52 +0000 (21:46 +0200)]
usb: phy: msm: reset controller is mandatory now

Commit a27345434134 "usb: phy: msm: Use reset framework for LINK
and PHY resets" introduced a mandatory call to reset_control_get
into the msm usb phy driver, which means we have to add a Kconfig
dependency on the API to avoid this build error:

phy/phy-msm-usb.c: In function 'msm_otg_read_dt':
phy/phy-msm-usb.c:1461:2: error: implicit declaration of function 'devm_reset_control_get' [-Werror=implicit-function-declaration]
  motg->link_rst = devm_reset_control_get(&pdev->dev, "link");
  ^

Since the usb-ehci-msm driver currently selects the OTG driver,
we could still get a broken dependency here. To solve that,
this patch also removes the 'select', which turns out to be
unnecessary.

Reviewed-by: Ivan T. Ivanov <iivanov@mm-sol.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: omap2plus bus glue needs USB host support
Arnd Bergmann [Thu, 8 May 2014 13:52:16 +0000 (15:52 +0200)]
usb: musb: omap2plus bus glue needs USB host support

The musb/omap2430.c bus glue driver calls usb_hcd_poll_rh_status,
which is only available if CONFIG_USB is also set, i.e. we
are building USB host mode and not just endpoint mode.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: tusb-dma can't be built-in if tusb is not
Arnd Bergmann [Thu, 8 May 2014 13:52:15 +0000 (15:52 +0200)]
usb: musb: tusb-dma can't be built-in if tusb is not

A configuration with CONFIG_USB_MUSB_HDRC=y, CONFIG_USB_TUSB_OMAP_DMA=y
and CONFIG_USB_MUSB_TUSB6010=m causes a link failure because of the
dependency on the tusb_get_revision symbol:

(.text+0x154ce8): undefined reference to `tusb_get_revision'

This patch ensures that either MUSB_HDRC and MUSB_TUSB6010 are
both modules or both built-in, which are the valid configurations.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Felipe Balbi <balbi@ti.com>