openwrt/staging/blogic.git
11 years agousb: dwc3: core: clarify usb-phy array binding
Kumar Gala [Fri, 9 Aug 2013 15:40:32 +0000 (10:40 -0500)]
usb: dwc3: core: clarify usb-phy array binding

The binding spec wasn't clear that the order of the phandles in the
usb-phy array has meaning.  Clarify this point in the binding that
it should be <USB2-HS-PHY, USB3-SS-PHY>.

Signed-off-by: Kumar Gala <galak@codeaurora.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: uvc: Fix error handling in uvc_queue_buffer()
Laurent Pinchart [Mon, 29 Apr 2013 20:18:01 +0000 (22:18 +0200)]
usb: gadget: uvc: Fix error handling in uvc_queue_buffer()

The conversion to videobuf2 failed to check the return value of
vb2_qbuf(). Fix it.

Cc: stable@vger.kernel.org
Reported-by: Michael Grzeschik <mgr@pengutronix.de>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-By: Michael Grzeschik <mgr@pengutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb dma: add cppi41 dma driver
Sebastian Andrzej Siewior [Thu, 20 Jun 2013 10:13:04 +0000 (12:13 +0200)]
usb: musb dma: add cppi41 dma driver

This driver is currently used by musb' cppi41 couter part. I may merge
both dma engine user of musb at some point but not just yet.

The driver seems to work in RX/TX mode in host mode, tested on mass
storage. I increaed the size of the TX / RX transfers and waited for the
core code to cancel a transfers and it seems to recover.

v2..3:
- use mall transfers on RX side and check data toggle.
- use rndis mode on tx side so we haveon interrupt for 4096 transfers.
- remove custom "transferred" hack and use dmaengine_tx_status() to
  compute the total amount of data that has been transferred.
- cancel transfers and reclaim descriptors

v1..v2:
- RX path added
- dma mode 0 & 1 is working
- device tree nodes re-created.

Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Dan Williams <djbw@fb.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dsps: use proper child nodes
Sebastian Andrzej Siewior [Fri, 5 Jul 2013 12:51:33 +0000 (14:51 +0200)]
usb: musb: dsps: use proper child nodes

This moves the two instances from the big node into two child nodes. The
glue layer ontop does almost nothing.

There is one devices containing the control module for USB (2) phy,
(2) usb and later the dma engine. The usb device is the "glue device"
which contains the musb device as a child. This is what we do ever since.

The new file musb_am335x is just here to prob the new bus and populate
child devices.

There are a lot of changes to the dsps file as a result of the changes:

- musb_core_offset
  This is gone. The device tree provides memory ressources information
  for the device there is no need to "fix" things

- instances
  This is gone as well. If we have two instances then we have have two
  child enabled nodes in the device tree. For instance the SoC in beagle
  bone has two USB instances but only one has been wired up so there is
  no need to load and init the second instance since it won't be used.

- dsps_glue is now per glue device
  In the past there was one of this structs but with an array of two and
  each instance accessed its variable depending on the platform device
  id.

- no unneeded copy of structs
  I do not know why struct dsps_musb_wrapper is copied but it is not
  necessary. The same goes for musb_hdrc_platform_data which allocated
  on demand and then again by platform_device_add_data(). One copy is
  enough.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dsps: remove the hardcoded phy pieces
Sebastian Andrzej Siewior [Fri, 5 Jul 2013 10:33:33 +0000 (12:33 +0200)]
usb: musb: dsps: remove the hardcoded phy pieces

dsps uses a nop driver which is added in dsps itself and does the PHY
on/off calls within dsps. Since those calls are now moved the nop driver
itself, we can now request the phy proper phy and remove those calls.
Currently only the first musb interface is used so we only add one phy
node for now.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Add AM335x PHY driver
Sebastian Andrzej Siewior [Tue, 30 Jul 2013 19:43:45 +0000 (21:43 +0200)]
usb: phy: Add AM335x PHY driver

This driver is a redo of my earlier attempt. It uses parts of the
generic PHY driver and uses the new control driver for the register
the phy needs to power on/off the phy. It also enables easy access for
the wakeup register which is not yet implemented.
The difference between the omap attempt is:
- no static holding variable
- one global visible function which exports a struct with callbacks to
  access the "control" registers.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: phy-generic: export init functions
Sebastian Andrzej Siewior [Tue, 30 Jul 2013 15:20:06 +0000 (17:20 +0200)]
usb: phy: phy-generic: export init functions

This patch exports the mostly generic functions so they can be used from
other phy driver instead of duplicating the code.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoMerge branch 'nop-phy-rename' into next
Felipe Balbi [Fri, 9 Aug 2013 14:31:23 +0000 (17:31 +0300)]
Merge branch 'nop-phy-rename' into next

Signed-off-by: Felipe Balbi <balbi@ti.com>
Conflicts:
drivers/usb/phy/phy-generic.c

11 years agousb: phy: rename nop_usb_xceiv => usb_phy_gen_xceiv
Sebastian Andrzej Siewior [Fri, 26 Jul 2013 10:16:42 +0000 (12:16 +0200)]
usb: phy: rename nop_usb_xceiv => usb_phy_gen_xceiv

The "nop" driver isn't a do-nothing-stub but supports a couple functions
like clock on/off or is able to use a voltage regulator. This patch
simply renames the driver to "generic" since it is easy possible to
extend it by a simple function istead of writing a complete driver.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: mv-usb: remove incorrect __exit_p annotation
Jingoo Han [Mon, 5 Aug 2013 05:17:53 +0000 (14:17 +0900)]
usb: phy: mv-usb: remove incorrect __exit_p annotation

When platform_driver_probe() is not used, bind/unbind via sysfs
is enabled.  Thus, __exit_p annotation should be removed from
remove().

Also, mv_otg_remove() is staticized, because this function is
used only in this file. Fix the following sparse warning:

drivers/usb/phy/phy-mv-usb.c:656:5: warning: symbol 'mv_otg_remove' was not declared. Should it be static?

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: mv-u3d: Staticize mv_u3d_phy_shutdown()
Jingoo Han [Mon, 5 Aug 2013 05:17:11 +0000 (14:17 +0900)]
usb: phy: mv-u3d: Staticize mv_u3d_phy_shutdown()

mv_u3d_phy_shutdown() is used only in this file.
Fix the following sparse warning:

drivers/usb/phy/phy-mv-u3d-usb.c:85:6: warning: symbol 'mv_u3d_phy_shutdown' was not declared. Should it be static?

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_uac1: Staticize local functions
Jingoo Han [Mon, 5 Aug 2013 03:15:27 +0000 (12:15 +0900)]
usb: gadget: f_uac1: Staticize local functions

control_selector_init() is used only in this file.
audio_bind_config() is used only in audio.c file to which
f_uac1.c is included. Thus, these functions are staticized
to fix the following warnings.

drivers/usb/gadget/f_uac1.c:698:12: warning: symbol 'control_selector_init' was not declared. Should it be static?
drivers/usb/gadget/f_uac1.c:722:12: warning: symbol 'audio_bind_config' was not declared. Should it be static?

Acked-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: u_uac1: add __user annotation
Jingoo Han [Mon, 5 Aug 2013 03:14:17 +0000 (12:14 +0900)]
usb: gadget: u_uac1: add __user annotation

Added __user annotation to fix the following sparse warning.

drivers/usb/gadget/u_uac1.c:194:52: warning: incorrect type in argument 2 (different address spaces)
drivers/usb/gadget/u_uac1.c:194:52:    expected void const [noderef] <asn:1>*buf
drivers/usb/gadget/u_uac1.c:194:52:    got void *buf

Acked-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: rndis: Staticize rndis_init()/rndis_exit()
Jingoo Han [Mon, 5 Aug 2013 03:12:42 +0000 (12:12 +0900)]
usb: gadget: rndis: Staticize rndis_init()/rndis_exit()

rndis_init() and rndis_exit() are used only in this file.
Fix the following sparse warnings:

drivers/usb/gadget/rndis.c:1145:5: warning: symbol 'rndis_init' was not declared. Should it be static?
drivers/usb/gadget/rndis.c:1179:6: warning: symbol 'rndis_exit' was not declared. Should it be static?

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: use NULL instead of 0
Jingoo Han [Mon, 5 Aug 2013 03:11:05 +0000 (12:11 +0900)]
usb: gadget: f_mass_storage: use NULL instead of 0

The local variables such as 'filename', 'vendor_name', and
'product_name' are pointers; thus, use NULL instead of 0 to fix
the following sparse warnings

drivers/usb/gadget/f_mass_storage.c:3046:27: warning: Using plain integer as NULL pointer
drivers/usb/gadget/f_mass_storage.c:3050:28: warning: Using plain integer as NULL pointer
drivers/usb/gadget/f_mass_storage.c:3051:29: warning: Using plain integer as NULL pointer

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: amd5536udc: unconditionally use GFP_ATOMIC in udc_queue()
Alexey Khoroshilov [Thu, 1 Aug 2013 19:50:47 +0000 (23:50 +0400)]
usb: gadget: amd5536udc: unconditionally use GFP_ATOMIC in udc_queue()

As far as prep_dma() is called with spinlock held,
we have to pass GFP_ATOMIC regardless of gfp argument.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: at91_udc: add usb_clk for transition to common clk framework
Boris BREZILLON [Fri, 2 Aug 2013 08:37:34 +0000 (10:37 +0200)]
usb: gadget: at91_udc: add usb_clk for transition to common clk framework

The AT91 PMC (Power Management Controller) provides an USB clock used by
USB Full Speed host (ohci) and USB Full Speed device (udc).
The usb drivers (ohci and udc) must configure this clock to 48Mhz.
This configuration was formely done in mach-at91/clock.c, but this
implementation will be removed when moving to common clk framework.

This patch adds support for usb clock retrieval and configuration, and is
backward compatible with the current at91 clk implementation (if usb clk
is not found, it does not configure/enable it).

Signed-off-by: Boris BREZILLON <b.brezillon@overkiz.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: at91_udc: add missing clk_put on fclk and iclk
Boris BREZILLON [Fri, 2 Aug 2013 08:37:27 +0000 (10:37 +0200)]
usb: gadget: at91_udc: add missing clk_put on fclk and iclk

This patch adds missing clk_put on fclk and iclk in case the probe function
fails after these clocks have been retrieved.

Signed-off-by: Boris BREZILLON <b.brezillon@overkiz.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: fusb300_udc: Staticize fusb300_rdcxf()
Jingoo Han [Fri, 2 Aug 2013 08:17:05 +0000 (17:17 +0900)]
usb: gadget: fusb300_udc: Staticize fusb300_rdcxf()

fusb300_rdcxf() used only in this file.
Fix the following sparse warning:

drivers/usb/gadget/fusb300_udc.c:560:6: warning: symbol 'fusb300_rdcxf' was not declared. Should it be static?

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: goku_udc: use NULL instead of 0
Jingoo Han [Fri, 2 Aug 2013 08:16:33 +0000 (17:16 +0900)]
usb: gadget: goku_udc: use NULL instead of 0

'req' is a pointer; thus, use NULL instead of 0
to fix the following sparse warning:

drivers/usb/gadget/goku_udc.c:775:13: warning: Using plain integer as NULL pointer

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: pci: add CONFIG_PM_SLEEP to suspend/resume functions
Jingoo Han [Fri, 2 Aug 2013 08:13:17 +0000 (17:13 +0900)]
usb: dwc3: pci: add CONFIG_PM_SLEEP to suspend/resume functions

Add CONFIG_PM_SLEEP to suspend/resume functions to fix the following
build warning when CONFIG_PM_SLEEP is not selected. This is because
sleep PM callbacks defined by SET_SYSTEM_SLEEP_PM_OPS are only used
when the CONFIG_PM_SLEEP is enabled. Unnecessary CONFIG_PM ifdefs
are removed.

drivers/usb/dwc3/dwc3-pci.c:215:12: warning: 'dwc3_pci_suspend' defined but not used [-Wunused-function]
drivers/usb/dwc3/dwc3-pci.c:224:12: warning: 'dwc3_pci_resume' defined but not used [-Wunused-function]

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: use dev_get_platdata()
Felipe Balbi [Wed, 31 Jul 2013 06:21:25 +0000 (09:21 +0300)]
usb: dwc3: use dev_get_platdata()

Use the wrapper function for retrieving the
platform_data instead of accessing dev->platform_data
directly.

While at that also make change 'node' initialization
to use the dev pointer.

Inspired-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: renesas: use dev_get_platdata()
Jingoo Han [Tue, 30 Jul 2013 08:06:20 +0000 (17:06 +0900)]
usb: renesas: use dev_get_platdata()

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: use dev_get_platdata()
Jingoo Han [Tue, 30 Jul 2013 08:03:12 +0000 (17:03 +0900)]
usb: musb: use dev_get_platdata()

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: use dev_get_platdata()
Jingoo Han [Tue, 30 Jul 2013 08:02:13 +0000 (17:02 +0900)]
usb: phy: use dev_get_platdata()

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: use dev_get_platdata()
Jingoo Han [Tue, 30 Jul 2013 08:00:51 +0000 (17:00 +0900)]
usb: gadget: use dev_get_platdata()

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: tegra: Fix typo in tegra20-usb-phy documentation
Mikko Perttunen [Tue, 30 Jul 2013 06:47:57 +0000 (09:47 +0300)]
usb: tegra: Fix typo in tegra20-usb-phy documentation

The device tree binding documentation for tegra20-usb-phy used to claim
that the vbus-supply property is required for host phys and optional
for otg phys. This should be the other way around.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: of: fix build breakage caused by recent patches
Felipe Balbi [Tue, 30 Jul 2013 06:09:15 +0000 (09:09 +0300)]
usb: of: fix build breakage caused by recent patches

commit 052a11d (usb: phy: make PHY driver selection
possible by controller drivers) changed the rules
on how drivers/usb/phy/of.c would be compiled and
failed to update include/linux/usb/of.h accordingly.

Because of that, we can fall into situations where
of_usb_get_phy_mode() is redefined. In order to fix
the error, we update the IS_ENABLED() check in
include/linux/usb/of.h to reflect the condition
where of.c is built.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: modify IO memory resource after deferred probe completes
Ivan T. Ivanov [Mon, 29 Jul 2013 07:27:02 +0000 (10:27 +0300)]
usb: dwc3: core: modify IO memory resource after deferred probe completes

When deferred probe happens driver will try to ioremap multiple times
and will fail. Memory resource.start variable is a global variable,
modifications in this field will be accumulated on every probe.
Fix this by moving the above operations after driver hold all
required PHY's.

Signed-off-by: Ivan T. Ivanov <iivanov@mm-sol.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Use switch instead of if-else
Tuomas Tynkkynen [Thu, 25 Jul 2013 19:24:09 +0000 (22:24 +0300)]
usb: phy: tegra: Use switch instead of if-else

Use switch() instead of if-else when checking for the PHY type.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dsps: remove EOI access
Sebastian Andrzej Siewior [Fri, 26 Jul 2013 09:11:38 +0000 (11:11 +0200)]
usb: musb: dsps: remove EOI access

The EOI register is not present in the AM335x memory space according to
the TRM and thus removed.
Should any platform using the EOI register get merged then it may be
used again if the register address is not zero.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dsps: rename ti81xx_driver_data to am33xx_driver_data
Sebastian Andrzej Siewior [Fri, 26 Jul 2013 09:11:37 +0000 (11:11 +0200)]
usb: musb: dsps: rename ti81xx_driver_data to am33xx_driver_data

The ti81xx platform is not fully supported right now. This patch renames
the date structure to a am33xx prefix which is actually used.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Use DT helpers for dr_mode
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:09 +0000 (21:38 +0300)]
usb: phy: tegra: Use DT helpers for dr_mode

Use the new of_usb_get_dr_mode helper function for parsing dr_mode
from the device tree. Also replace the usage of the custom
tegra_usb_phy_mode enum with the standard enum.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Use DT helpers for phy_type
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:08 +0000 (21:38 +0300)]
usb: phy: tegra: Use DT helpers for phy_type

Use the new of_usb_get_phy_mode helper function for parsing phy_type
from the device tree.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Remove custom PHY locating APIs
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:07 +0000 (21:38 +0300)]
usb: phy: tegra: Remove custom PHY locating APIs

The Tegra EHCI driver is no longer using these custom functions, so they
can be removed.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: host: tegra: Locate a PHY via standard API
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:06 +0000 (21:38 +0300)]
usb: host: tegra: Locate a PHY via standard API

Use devm_get_phy_by_phandle to get a PHY device instead of the custom
Tegra functions.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Register as an USB PHY.
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:05 +0000 (21:38 +0300)]
usb: phy: tegra: Register as an USB PHY.

Register the Tegra PHY device instances with the PHY subsystem so that
the Tegra EHCI driver can locate a PHY via the standard APIs.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoARM: tegra: Remove USB platform data
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:04 +0000 (21:38 +0300)]
ARM: tegra: Remove USB platform data

USB-related platform data is not used anymore in the Tegra USB drivers,
so remove all of it.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: tegra: host: Remove references to plat data
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:03 +0000 (21:38 +0300)]
usb: tegra: host: Remove references to plat data

Platform data is not used in tegra-ehci anymore, so remove all
references to it.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: host: tegra: Remove leftover code
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:02 +0000 (21:38 +0300)]
usb: host: tegra: Remove leftover code

ehci-tegra calls devm_usb_get_phy, which will never succeed since the Tegra
PHY does not register itself with the PHY subsystem. It is also completely
redundant since the code has already located a PHY via an internal API.

Call otg_set_host unconditionally to simplify the code since it should
be safe to do so.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Remove unnecessary 'dev' field
Tuomas Tynkkynen [Thu, 25 Jul 2013 18:38:01 +0000 (21:38 +0300)]
usb: phy: tegra: Remove unnecessary 'dev' field

struct usb_phy already has a field for the device pointer, so this
unnecessary field can be removed.

Signed-off-by: Tuomas Tynkkynen <ttynkkynen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoarm: tegra: Remove obsolete nvidia,vbus-gpio properties
Mikko Perttunen [Wed, 17 Jul 2013 07:37:51 +0000 (10:37 +0300)]
arm: tegra: Remove obsolete nvidia,vbus-gpio properties

USB VBUS regulators are now specified with the vbus-supply property
instead of nvidia,vbus-gpio, so remove the obsolete properties.

The equivalent vbus-supply properties were already added in patch
"ARM: tegra: update device trees for USB binding rework".

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: tegra: Add vbus-supply property for host mode PHYs
Mikko Perttunen [Wed, 17 Jul 2013 07:37:50 +0000 (10:37 +0300)]
usb: tegra: Add vbus-supply property for host mode PHYs

Document vbus-supply as an optional property for host mode phy-tegra-usb PHYs.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: tegra: Use regulators instead of GPIOs for USB PHY VBUS
Mikko Perttunen [Wed, 17 Jul 2013 07:37:49 +0000 (10:37 +0300)]
usb: tegra: Use regulators instead of GPIOs for USB PHY VBUS

The tegra ehci driver has enabled USB vbus regulators directly using
GPIOs and the device tree attribute nvidia,vbus-gpio. This is ugly
and causes error messages on boot when both the regulator driver
and the ehci driver want access to the same GPIO.

After this patch, usb vbus regulators for tegra usb phy devices are specified
with the device tree attribute vbus-supply = <&x> where x is a regulator defined
in the device tree. The old nvidia,vbus-gpio property is no longer supported.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra: Read UTMIP parameters from device tree
Mikko Perttunen [Wed, 17 Jul 2013 06:31:01 +0000 (09:31 +0300)]
usb: phy: tegra: Read UTMIP parameters from device tree

UTMIP parameters used to be hardcoded into tables in the
PHY driver. This patch reads them from the device tree instead
in accordance with the phy-tegra-usb DT documentation.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoarm: dts: tegra20: Rename USB UTMI parameters according to new definitions
Mikko Perttunen [Wed, 17 Jul 2013 06:31:00 +0000 (09:31 +0300)]
arm: dts: tegra20: Rename USB UTMI parameters according to new definitions

This patch changes the Tegra20 USB PHY nodes to use the UTMI configuration
parameter names as specified in the device tree binding documentation
after patch "ARM: tegra: finalize USB EHCI and PHY bindings".

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: do not change dev's dma_mask
Sebastian Andrzej Siewior [Mon, 22 Jul 2013 18:10:04 +0000 (20:10 +0200)]
usb: musb: do not change dev's dma_mask

Commit 8d2421e ("usb: musb: kill global and static for multi instance")
removed the global dma_mask copy and replaced by parent's DMA mask. The
problem here is that if the parent does not have a dma_mask then
musb_remove() goes kaboom.
Instead trying to fix this I was thinking we do we need to erase
dma_mask in the first place?

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: remove ti81xx pieces from musb
Sebastian Andrzej Siewior [Mon, 22 Jul 2013 18:09:55 +0000 (20:09 +0200)]
usb: musb: remove ti81xx pieces from musb

ti81xx does not have a baseport mainline i.e. it should not boot. The
amount of rework that is required makes it easier to simply remove that
platform (i.e. that possible platform device) and add it later once it
comes back with DT support.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dsps: init / shutdown the phy
Sebastian Andrzej Siewior [Mon, 22 Jul 2013 18:09:52 +0000 (20:09 +0200)]
usb: musb: dsps: init / shutdown the phy

If the init / shutdown function of the phy moves out of dsps into the
phy driver, then dsps needs to call the callbacks of the phy driver to
ensure that this happens.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: don't check resource with devm_ioremap_resource
Wolfram Sang [Tue, 23 Jul 2013 18:01:52 +0000 (20:01 +0200)]
usb: phy: don't check resource with devm_ioremap_resource

devm_ioremap_resource does sanity checks on the given resource. No need to
duplicate this in the driver.

Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: atmel_usba: prepare clk before calling enable
Boris BREZILLON [Thu, 18 Jul 2013 07:43:08 +0000 (09:43 +0200)]
usb: gadget: atmel_usba: prepare clk before calling enable

Replace clk_enable/disable with clk_prepare_enable/disable_unprepare to
avoid common clk framework warnings.

Signed-off-by: Boris BREZILLON <b.brezillon@overkiz.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: renesas_usbhs: tidyup original usbhsx_for_each_xxx macro
Kuninori Morimoto [Fri, 12 Jul 2013 05:32:31 +0000 (22:32 -0700)]
usb: renesas_usbhs: tidyup original usbhsx_for_each_xxx macro

Current usbhsx_for_each_xxx macro will read out-of-array's
memory after last loop operation.
It was not good C language operation, and the binary which was
compiled by (at least) gcc 4.8.1 is broken
This patch tidyup these issues

Reported-by: Yusuke Goda <yusuke.goda.sx@renesas.com>
Reviewed-by: Takashi Yoshii <takashi.yoshii.zj@renesas.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: omap-usb3: Improve DPLL parameter lookup code
Roger Quadros [Tue, 9 Jul 2013 11:38:31 +0000 (14:38 +0300)]
usb: phy: omap-usb3: Improve DPLL parameter lookup code

Use a mapping table (dpll_map) to match the possible system clock rates
to the appropriate DPLL parameters.

Introduce a function "omap_usb3_get_dpll_params()" that will
return the matching DPLL parameters for the given clock rate.

Also, bail out on phy init if DPLL locking fails.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: ehci-omap: select NOP_USB_XCEIV PHY driver
Roger Quadros [Fri, 14 Jun 2013 09:51:42 +0000 (12:51 +0300)]
usb: ehci-omap: select NOP_USB_XCEIV PHY driver

ehci-omap needs NOP_USB_XCEIV PHY driver to function
properly, so select it. As the USB PHY drivers no longer
depend on USB_PHY, it is safe to select the PHY drivers.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Tested-by: Adrien Vergé <adrienverge@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: make PHY driver selection possible by controller drivers
Roger Quadros [Fri, 14 Jun 2013 09:51:41 +0000 (12:51 +0300)]
usb: phy: make PHY driver selection possible by controller drivers

Convert PHY Drivers from menuconfig to menu so that the PHY drivers
can be explicitely selected by the controller drivers.

USB_PHY is no longer a user visible option. It is upto to the PHY
drivers to select it if needed. This patch does so for the existing
PHY drivers that use the USB_PHY library.

Doing so moves the USB_PHY and PHY driver selection problem from the
end user to the PHY and controller driver developer.

e.g.

Earlier, a controller driver (e.g. EHCI_OMAP) that needs to select
a PHY driver (e.g. NOP_PHY) couldn't do so because the PHY driver
depended on USB_PHY. Making the controller driver depend on USB_PHY
has a negative effect i.e. it becomes invisible to the user till
USB_PHY is enabled. Most end users will not familiar with this.

With this patch, the end user just needs to select the controller driver
needed for his/her platform without worrying about which PHY driver to
select.

Also update USB_EHCI_MSM, USB_LPC32XX and USB_OMAP to not depend
on USB_PHY any more. They can safely select the necessary PHY drivers.

[ balbi@ti.com : refreshed on top of my next branch. Changed bool
followed by default n into def_bool n ]

CC: Pavankumar Kondeti <pkondeti@codeaurora.org>
Acked-by: Roland Stigge <stigge@antcom.de>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: ep0: don't change to configured state too early
Felipe Balbi [Mon, 22 Jul 2013 09:41:47 +0000 (12:41 +0300)]
usb: dwc3: ep0: don't change to configured state too early

before changing to configured state, we need
to wait until gadget driver has had a chance
to process the request.

In case of USB_GADGET_DELAYED_STATUS, that means
we need to defer usb_gadget_set_state() until
the upcoming usb_ep_queue().

Reported-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: ep0: only change to ADDRESS if set_config() succeeds
Felipe Balbi [Mon, 22 Jul 2013 09:31:31 +0000 (12:31 +0300)]
usb: dwc3: ep0: only change to ADDRESS if set_config() succeeds

In case we're switching back to USB_STATE_ADDRESS
from USB_STATE_CONFIGURED (if host sends
a set configuration command for configuration
zero), we should only switch if the request
is successfully processed by the gadget driver.

Reported-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: udc-core: move sysfs_notify() to a workqueue
Felipe Balbi [Wed, 17 Jul 2013 08:09:49 +0000 (11:09 +0300)]
usb: gadget: udc-core: move sysfs_notify() to a workqueue

usb_gadget_set_state() will call sysfs_notify()
which might sleep. Some users might want to call
usb_gadget_set_state() from the very IRQ handler
which actually changes the gadget state.

Instead of having every UDC driver add their own
workqueue for such a simple notification, we're
adding it generically to our struct usb_gadget,
so the details are hidden from all UDC drivers.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: fix typo in comment of dwc3_ep
Huang Rui [Wed, 12 Jun 2013 15:43:11 +0000 (23:43 +0800)]
usb: dwc3: fix typo in comment of dwc3_ep

Change intervall into interval.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: introduce dwc3_process_event_buf
Felipe Balbi [Wed, 12 Jun 2013 18:25:08 +0000 (21:25 +0300)]
usb: dwc3: gadget: introduce dwc3_process_event_buf

in order to make our IRQ handler thread easier
to read, we re-factor the inner loop to a separate
function.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: rename dwc3_process_event_buf
Felipe Balbi [Wed, 12 Jun 2013 18:16:11 +0000 (21:16 +0300)]
usb: dwc3: gadget: rename dwc3_process_event_buf

that function isn't really processing any event
buffer, rather just checking whether it contains
events to be processed.

Due to that reason, we're reaning it to dwc3_check_event_buf()
instead.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: get rid of IRQF_ONESHOT
Felipe Balbi [Wed, 12 Jun 2013 18:11:14 +0000 (21:11 +0300)]
usb: dwc3: gadget: get rid of IRQF_ONESHOT

We can make use of bit 31 of the GEVNTSIZ(n)
registers to mask/unmask interrupts from that
particular interrupter.

With that feature, we can easily drop IRQF_ONESHOT
from our driver which makes it possible to
properly change IRQ priorities when using RT
patchset *and* it allows us to make use of the
scheduler to choose the proper time to handle
this IRQ thread.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: introduce and use macros for Event Size register
Felipe Balbi [Wed, 12 Jun 2013 18:09:26 +0000 (21:09 +0300)]
usb: dwc3: core: introduce and use macros for Event Size register

That register has more than just the event buffer
size; we can also mask and unmask that particular
interrupter on bit 31 of that register.

In this patch we introduce the necessary macros
and make sure to use the new macros while also
making sure we mask interrupts during driver
removal.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: don't enable LPM early
Felipe Balbi [Mon, 15 Jul 2013 08:31:27 +0000 (11:31 +0300)]
usb: dwc3: gadget: don't enable LPM early

LPM is enabled in Connection Done interrupt,
there's no need to enable it early.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: don't redefine DWC3_DCFG_LPM_CAP
Felipe Balbi [Mon, 15 Jul 2013 08:30:45 +0000 (11:30 +0300)]
usb: dwc3: core: don't redefine DWC3_DCFG_LPM_CAP

the macro DWC3_DCFG_LPM_CAP was defined twice.

This patch just removes one of the definitions,
no functional changes.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: add a debugging print when initializing endpoints
Felipe Balbi [Fri, 12 Jul 2013 16:11:57 +0000 (19:11 +0300)]
usb: dwc3: gadget: add a debugging print when initializing endpoints

that way we get debugging information when
enabling verbose debug of the driver.

It will be no-op otherwise.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: move direction setting up
Felipe Balbi [Fri, 12 Jul 2013 16:10:59 +0000 (19:10 +0300)]
usb: dwc3: gadget: move direction setting up

no functional changes, just grouping dep
initialization.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: move debugging print around
Felipe Balbi [Fri, 12 Jul 2013 16:09:39 +0000 (19:09 +0300)]
usb: dwc3: gadget: move debugging print around

by moving that dev_vdbg() to the internal
__dwc3_gadget_ep_enable() we get the print
even when enable ep0, which calls the internal
function directly.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: switch over to devm_ioremap_resource()
Felipe Balbi [Fri, 12 Jul 2013 12:34:12 +0000 (15:34 +0300)]
usb: dwc3: core: switch over to devm_ioremap_resource()

use the new devm_ioremap_resource() on core.c

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: omap: switch over to devm_ioremap_resource()
Felipe Balbi [Fri, 12 Jul 2013 12:33:35 +0000 (15:33 +0300)]
usb: dwc3: omap: switch over to devm_ioremap_resource()

use the new devm_ioremap_resource on dwc3-omap.c

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: drop dwc3 manual phy control
Felipe Balbi [Fri, 12 Jul 2013 10:19:52 +0000 (13:19 +0300)]
usb: dwc3: gadget: drop dwc3 manual phy control

Recent versions of the core, can suspend and resume
the PHYs automatically, so we don't need to fiddle
with dwc3's Global PHY registers at all.

On versions prior to 1.94a this patch will mean
that we will never ask dwc3 to suspend the PHY.

This is an acceptable behavior or such old versions
of the core, specially considering that the only
chip known to have a version prior to 1.94a was
OMAP5 ES1.0 and that's not supported in mainline
kernel, because it was just a test spin of OMAP5.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: adapt to use dr_mode device tree helper
Ruchika Kharwar [Sat, 6 Jul 2013 12:52:49 +0000 (07:52 -0500)]
usb: dwc3: adapt to use dr_mode device tree helper

This patch adapts the dwc3 to use the device tree helper
"of_usb_get_dr_mode" for the mode of operation of the dwc3 instance
being probed.

[ balbi@ti.com : make of_usb_get_dr_mode() conditional on
dev->of_node and let pdata pass dr_mode too ]

Reviewed-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Ruchika Kharwar <ruchika@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: phy-mxs-usb: Check the return value from stmp_reset_block()
Fabio Estevam [Wed, 3 Jul 2013 19:34:13 +0000 (16:34 -0300)]
usb: phy: phy-mxs-usb: Check the return value from stmp_reset_block()

stmp_reset_block() may fail, so let's check its return value and propagate it
in the case of error.

Acked-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: get rid of unused proc_dir_entry
Greg Kroah-Hartman [Tue, 2 Jul 2013 04:34:58 +0000 (21:34 -0700)]
usb: musb: get rid of unused proc_dir_entry

The musb driver no longer uses procfs, so get rid of the proc_dir_entry
variable in struct musb.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: Allow driver instantiation using device tree
Tomasz Figa [Tue, 25 Jun 2013 15:38:23 +0000 (17:38 +0200)]
usb: gadget: s3c-hsotg: Allow driver instantiation using device tree

This patch adds OF match table to the driver to allow instantiating it
using device tree.

Signed-off-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: switch to snps,dwc3
Felipe Balbi [Tue, 2 Jul 2013 18:20:24 +0000 (21:20 +0300)]
usb: dwc3: core: switch to snps,dwc3

all other drivers using Synopsys IPs with DT
have a compatible of snps,$driver, in order
to add consistency, we are switching over to
snps,dwc3 but keeping synopsys,dwc3 in the core
driver to maintain backwards compatibility.

New DTS bindings should NOT use synopsys,dwc3.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: make maximum-speed a per-instance attribute
Felipe Balbi [Sun, 30 Jun 2013 11:29:51 +0000 (14:29 +0300)]
usb: dwc3: make maximum-speed a per-instance attribute

in order to allow different instances of the
core work in different maximum speeds, we will
move the maximum_speed module_parameter to
both DeviceTree (making use the new maximum-speed
DT property) and platform_data.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: let non-DT platforms pass tx-fifo-resize flag;
Felipe Balbi [Sun, 30 Jun 2013 11:19:33 +0000 (14:19 +0300)]
usb: dwc3: let non-DT platforms pass tx-fifo-resize flag;

in case we're not in a DT boot, we should
still be able to tell the driver how to behave.

In order to be able to pass flags to the driver,
we introduce platform_data structure which the
core driver should use.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: common: introduce of_usb_get_maximum_speed()
Felipe Balbi [Sun, 30 Jun 2013 10:56:45 +0000 (13:56 +0300)]
usb: common: introduce of_usb_get_maximum_speed()

this helper will be used for controllers which
want to work at a lower speed even though they
support higher USB transfer rates.

One such case is Texas Instruments' AM437x
SoC where it uses a USB3 controller without
a USB3 PHY, rendering the controller USB2-only.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: protect against NULL phy pointers
Felipe Balbi [Sun, 30 Jun 2013 12:27:26 +0000 (15:27 +0300)]
usb: phy: protect against NULL phy pointers

In order to decrease the amount of work done
by PHY users, allow NULL phy pointers to be
passed.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: switch to GPL v2 only
Felipe Balbi [Sun, 30 Jun 2013 11:15:11 +0000 (14:15 +0300)]
usb: dwc3: switch to GPL v2 only

This is a Linux-only driver which makes use
of GPL-only symbols. It makes no sense to
maintain Dual BSD/GPL licensing for this driver.

Considering that the amount of work to use this
driver in any different operating system would likely
be as large as developing the driver from scratch and
considering that we depend on GPL-only symbols, we
will switch over to a GPL v2-only license.

Cc: Anton Tikhomirov <av.tikhomirov@samsung.com>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: clean up redundant parameter comment
Huang Rui [Wed, 26 Jun 2013 17:08:10 +0000 (01:08 +0800)]
usb: dwc3: clean up redundant parameter comment

@list is not as a parameter of dwc3_event_buffer, so remove it in
comments.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: gadget: don't request IRQs in atomic
Felipe Balbi [Thu, 27 Jun 2013 07:00:18 +0000 (10:00 +0300)]
usb: dwc3: gadget: don't request IRQs in atomic

We cannot request an IRQ with spinlocks held
as that would trigger a sleeping inside
spinlock warning.

Cc: <stable@vger.kernel.org> # v3.10
Reported-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: remove imx_udc
Felipe Balbi [Thu, 27 Jun 2013 07:20:44 +0000 (10:20 +0300)]
usb: gadget: remove imx_udc

That driver hasn't been really maintained for
a long time. It doesn't compile in any way, it
includes non-existent headers, has no users,
and is just plain broken.

The person who used to work with that driver
has publicly stated that he has no plans to
touch that driver again and is ok with removal[1].

Due to these factors, imx_udc is now removed from
the tree, if someone really believe it needs to
be kept, please fix the bugs in that driver.

[1] http://marc.info/?l=linux-usb&m=136197620417636&w=2

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: make glue layers selectable
Felipe Balbi [Wed, 26 Jun 2013 12:14:06 +0000 (15:14 +0300)]
usb: dwc3: make glue layers selectable

Glue layers are starting to have separate
requirements. For example, OMAP's glue layer
is starting to use extcon framework which
no one else needs.

In order to make it clear the proper dependencies,
we are now allowing glue layers to be selectable
so that each glue layer can list their own dependencies
without messing with the core IP driver.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: unmap reqs in musb_gadget_queue()'s error case
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:15 +0000 (17:38 +0200)]
usb: musb: unmap reqs in musb_gadget_queue()'s error case

If the descriptor is missing the reqeust is never unmapped. This patch
changes this and renames the cleanup label to unlock since there is no
cleanup done. The cleanup would revert the allocation of ressource (i.e.
this dma mapping) but it does not, it simply unlocks and returns.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: core: call dma_controller_destroy() in the err path
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:14 +0000 (17:38 +0200)]
usb: musb: core: call dma_controller_destroy() in the err path

The cleanup in the error is missing the dma controller. The structure is
allocated at runtime and ux500 allocates even a little more than just
this struct. So cleanup!

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: remove a few is_dma_capable() in init/exit code
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:13 +0000 (17:38 +0200)]
usb: musb: remove a few is_dma_capable() in init/exit code

This patch removes is_dma_capable() and an ifdef in the init/exit path
around init/de-init of the dma_controller. Since we have the empty stubs
in the PIO code we can call it without gcc trouble. Earlier we had an
ifdef and the is_dma_capable() macro where gcc ignored the if (0) path
even that the function was not around :)

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: provide empty dma_controller_create() in PIO mode
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:12 +0000 (17:38 +0200)]
usb: musb: provide empty dma_controller_create() in PIO mode

Add a dma_controller_create() returning NULL so a few ifdefs can
dropped.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: dma: merge ->start/stop into create/destroy
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:11 +0000 (17:38 +0200)]
usb: musb: dma: merge ->start/stop into create/destroy

The core code creates a controller and immediately after that it calls
the ->start() callback. This one might drop an error but nobody cares.
The same thing happens in the destroy corner: First ->stop() called
followed by destroy callback. So why not merge those two into the same
function since there is no difference.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: musbhsdma: drop the controller check in dma_controller_destroy()
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:10 +0000 (17:38 +0200)]
usb: musb: musbhsdma: drop the controller check in dma_controller_destroy()

This check is hardly required and alas is wrong. 'c' might be NULL but
the chances are low that 'controller' after the container_of() becomes
NULL.

Since no other DMA implementation is doing that and musb-core does not
call it with a NULL pointer it can dropped.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: replace ifndef with ifdef for CONFIG_MUSB_PIO_ONLY
Sebastian Andrzej Siewior [Wed, 19 Jun 2013 15:38:09 +0000 (17:38 +0200)]
usb: musb: replace ifndef with ifdef for CONFIG_MUSB_PIO_ONLY

The ifdef reads somehow better than an ifndef

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoLinux 3.11-rc3
Linus Torvalds [Mon, 29 Jul 2013 03:53:33 +0000 (20:53 -0700)]
Linux 3.11-rc3

11 years agoMerge tag 'pinctrl-for-v3.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 29 Jul 2013 01:19:27 +0000 (18:19 -0700)]
Merge tag 'pinctrl-for-v3.11-2' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pin control fixes from Linus Walleij:
 - Driver fixes for AM33xx, SIRF and PFC pin controllers
 - Fix a compile warning from the pinctrl single-register driver
 - Fix a little nasty memory leak

* tag 'pinctrl-for-v3.11-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: fix a memleak when freeing maps
  pinctrl: pinctrl-single: fix compile warning when no CONFIG_PM
  pinctrl: sh-pfc: fix SDHI0 VccQ regulator on sh73a0 with DT
  arm/dts: sirf: fix the pingroup name mismatch between drivers and dts
  pinctrl: sirf: add usp0_uart_nostreamctrl pin group for usp-uart without flowctrl
  pinctrl: sirf: fix the pin number and mux bit for usp0
  pinctrl: am33xx dt binding: correct include path

11 years agoMerge tag 'trace-fixes-3.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 29 Jul 2013 01:10:39 +0000 (18:10 -0700)]
Merge tag 'trace-fixes-3.11-rc2' of git://git./linux/kernel/git/rostedt/linux-trace

Pull tracing fixes from Steven Rostedt:
 "Oleg is working on fixing a very tight race between opening a event
  file and deleting that event at the same time (both must be done as
  root).

  I also found a bug while testing Oleg's patches which has to do with a
  race with kprobes using the function tracer.

  There's also a deadlock fix that was introduced with the previous
  fixes"

* tag 'trace-fixes-3.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  tracing: Remove locking trace_types_lock from tracing_reset_all_online_cpus()
  ftrace: Add check for NULL regs if ops has SAVE_REGS set
  tracing: Kill trace_cpu struct/members
  tracing: Change tracing_fops/snapshot_fops to rely on tracing_get_cpu()
  tracing: Change tracing_entries_fops to rely on tracing_get_cpu()
  tracing: Change tracing_stats_fops to rely on tracing_get_cpu()
  tracing: Change tracing_buffers_fops to rely on tracing_get_cpu()
  tracing: Change tracing_pipe_fops() to rely on tracing_get_cpu()
  tracing: Introduce trace_create_cpu_file() and tracing_get_cpu()

11 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Mon, 29 Jul 2013 01:09:34 +0000 (18:09 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is five bug fixes, two of which fix long standing problems
  causing crashes (sd and mvsas).  The remaining three are hung (isci
  race) or lost (qla2xxx, isci) devices"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] isci: fix breakage caused by >16byte CDB patch
  [SCSI] mvsas: Fix kernel panic on tile due to unaligned data access
  [SCSI] sd: fix crash when UA received on DIF enabled device
  [SCSI] qla2xxx: Properly set the tagging for commands.
  [SCSI] isci: Fix a race condition in the SSP task management path

11 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Fri, 26 Jul 2013 21:40:10 +0000 (14:40 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "This is just a regular fixes pull apart from the qxl one, it has
  radeon and intel bits in it,

  The intel fixes are for a regression with the RC6 fix and a 3.10 hdmi
  regression, whereas radeon is more DPM fixes, a few lockup fixes and
  some rn50/r100 DAC fixes"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/radeon/dpm: fix r600_enable_sclk_control()
  drm/radeon/dpm: implement force performance levels for rv6xx
  drm/radeon/dpm: fix displaygap programming on rv6xx
  drm/radeon/dpm: fix a typo in the rv6xx mclk setup
  drm/i915: initialize gt_lock early with other spin locks
  drm/i915: fix hdmi portclock limits
  drm/radeon: fix combios tables on older cards
  drm/radeon: improve dac adjust heuristics for legacy pdac
  drm/radeon: Another card with wrong primary dac adj
  drm/radeon: fix endian issues with DP handling (v3)
  drm/radeon/vm: only align the pt base to 32k
  drm/radeon: wait for 3D idle before using CP DMA

11 years agoMerge branch 'qxl-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Fri, 26 Jul 2013 21:38:50 +0000 (14:38 -0700)]
Merge branch 'qxl-fixes' of git://people.freedesktop.org/~airlied/linux

Pull qxl drm fixes from Dave Airlie:
 "Okay as I warned, the qxl driver was running a bit free and loose with
  its ttm object reservations and the new lockdep enabled reservation
  tracking shone a bright light into it, it also with the new
  reservations mutexes hits a possible deadlock during boot.

  The first patch is a real fix to render the console correctly as the
  driver used to just drop irq renderering as too hard, this also fixes
  a sleeping while atomic warning.

  The other two patches are the big ugly ones that redo how the driver
  allocates objects and reserves them and makes things all work
  properly, I've tested this in a VM, and compared to the current code
  which hits a lockdep warning and the sleep while atomic warning before
  failing.

  So sorry this is coming in late, I should have tested qxl before
  merging the mutex code, but I'd rather just fix qxl with this than
  revert the reservations code at this point"

* 'qxl-fixes' of git://people.freedesktop.org/~airlied/linux:
  qxl: convert qxl driver to proper use for reservations
  qxl: allow creation of pre-pinned objects and use for releases.
  drm/qxl: add delayed fb operations

11 years agoMerge tag 'pm+acpi-3.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 26 Jul 2013 21:37:18 +0000 (14:37 -0700)]
Merge tag 'pm+acpi-3.11-rc3' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management fixes from Rafael Wysocki:
 "These are just two fixes, a revert of the would-be backlight fix that
  didn't work and an intel_pstate fix for two problems related to
  maximum P-state selection.

  Specifics:

   - Revert of the ACPI video commit that I hoped would help fix
     backlight problems related to Windows 8 compatibility on some
     systems.  Unfortunately, it turned out to cause problems to happen
     too.

   - Fix for two problems in intel_pstate, a possible failure to respond
     to a load change on a quiet system and a possible failure to select
     the highest available P-state on some systems.  From Dirk
     Brandewie"

* tag 'pm+acpi-3.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  Revert "ACPI / video / i915: No ACPI backlight if firmware expects Windows 8"
  cpufreq / intel_pstate: Change to scale off of max P-state

11 years agoMerge tag 'arm64-stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas...
Linus Torvalds [Fri, 26 Jul 2013 19:43:59 +0000 (12:43 -0700)]
Merge tag 'arm64-stable' of git://git./linux/kernel/git/cmarinas/linux-aarch64

Pull arm64 fixes from Catalin Marinas:
 - Stack size increased to 16K (similar to other 64-bit architectures)
 - Additional cache flushing for secondary CPUs boot mode

* tag 'arm64-stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64:
  arm64: Change kernel stack size to 16K
  arm64: Fix definition of arm_pm_restart to match the declaration
  arm64: virt: ensure visibility of __boot_cpu_mode