firefly-linux-kernel-4.4.55.git
10 years agousb: dwc3: omap: add dwc3_omap_extcon_register function
George Cherian [Wed, 16 Jul 2014 13:07:08 +0000 (18:37 +0530)]
usb: dwc3: omap: add dwc3_omap_extcon_register function

Move the extcon related code to its own function.
Improve code readability, decrease the dwc3_probe() size.

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: omap: add dwc3_omap_set_utmi_mode() function
George Cherian [Wed, 16 Jul 2014 13:07:07 +0000 (18:37 +0530)]
usb: dwc3: omap: add dwc3_omap_set_utmi_mode() function

Move find and set the utmi mode to its own separate function.
Improve code readability, decrease the dwc3_probe() size.

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: omap: add dwc3_omap_map_offset() function
George Cherian [Wed, 16 Jul 2014 13:07:06 +0000 (18:37 +0530)]
usb: dwc3: omap: add dwc3_omap_map_offset() function

Move map offset to its own separate function.
Improve code readability, decrease the dwc3_probe() size.

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: omap: remove x_major calculation from revision register
George Cherian [Wed, 16 Jul 2014 13:07:05 +0000 (18:37 +0530)]
usb: dwc3: omap: remove x_major calculation from revision register

Remove the x_major calculation logic from the wrapper revision register
to differentiate between OMAP5 and AM437x. This was done to find the
register offsets of wrapper register. Now that We do it using dt
compatible, remove the whole logic.

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: NCM: Stop RX TCP Bursts getting dropped.
Jim Baxter [Mon, 7 Jul 2014 17:33:19 +0000 (18:33 +0100)]
usb: gadget: NCM: Stop RX TCP Bursts getting dropped.

This fixes a problem with dropped packets over 16k CDC-NCM
when the connection is being heavily used.

The issue was that the extracted frames cloned from the
received frame were consuming more memory than necessary
resulting in the truesize being ~32KB instead of ~2KB, this
meant there was a high chance of reaching the sk_rcvbuf
limit.

Signed-off-by: Jim Baxter <jim_baxter@mentor.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: NCM: Add transmit multi-frame.
Jim Baxter [Mon, 7 Jul 2014 17:33:18 +0000 (18:33 +0100)]
usb: gadget: NCM: Add transmit multi-frame.

This adds multi-frame support to the NCM NTB's for
the gadget driver. This allows multiple network
packets to be put inside a single USB NTB with a
maximum size of 16kB.

It has a time out of 300ms to ensure that smaller
number of packets still maintain a normal latency.

Also the .fp_index and .next_fp_index have been
changed to .ndp_index and .next_ndp_index to
match the latest CDC-NCM specification and
help with maintenance.

Results transmitting from gadget to host.

Before the change:

TCP_STREAM Throughput (10^6bits/sec): 22.72
UDP_STREAM Throughput (10^6bits/sec): 25.94

Latency:
netperf -H 192.168.1.101 -v2 -l 50 -t TCP_RR -- -r 16384,16384
Trans.   RoundTrip  Throughput
Rate     Latency    10^6bits/s
per sec  usec/Tran  Outbound

100.83   9918.116   13.215

After the change:

TCP_STREAM Throughput (10^6bits/sec): 124.26
UDP_STREAM Throughput (10^6bits/sec): 227.48

Latency:
netperf -H 192.168.1.101 -v2 -l 50 -t TCP_RR -- -r 16384,16384
Trans.   RoundTrip  Throughput
Rate     Latency    10^6bits/s
per sec  usec/Tran  Outbound

156.80   6377.730   20.552

Signed-off-by: Jim Baxter <jim_baxter@mentor.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: NCM: RX function support multiple NDPs
Jim Baxter [Mon, 7 Jul 2014 17:33:17 +0000 (18:33 +0100)]
usb: gadget: NCM: RX function support multiple NDPs

The NDP was ignoring the wNextNdpIndex in the NDP which
means that NTBs containing multiple NDPs would have missed
frames.

Signed-off-by: Jim Baxter <jim_baxter@mentor.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fix eem_wrap cloned skb logic
Nathan Sullivan [Mon, 7 Jul 2014 14:50:14 +0000 (09:50 -0500)]
usb: gadget: fix eem_wrap cloned skb logic

Even if the skb is cloned, we still need a ZLP or USB will stall.

Signed-off-by: Nathan Sullivan <nathan.sullivan@ni.com>
Acked-by: Brad Mouring <brad.mouring@ni.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: gadget: remove unnecessary 'start_new' variable
Jingoo Han [Fri, 4 Jul 2014 06:00:51 +0000 (15:00 +0900)]
usb: dwc3: gadget: remove unnecessary 'start_new' variable

Remove 'start_new' variable from dwc3_endpoint_transfer_complete(),
since this variable has not been used.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: am335x: Use SIMPLE_DEV_PM_OPS macro
Jingoo Han [Tue, 8 Jul 2014 11:51:34 +0000 (20:51 +0900)]
usb: phy: am335x: Use SIMPLE_DEV_PM_OPS macro

Use SIMPLE_DEV_PM_OPS macro and remove DEV_PM_OPS macro, in order
to make the code simpler.

Acked-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: Convert drivers to use module_usb_composite_driver()
Tobias Klauser [Wed, 9 Jul 2014 16:09:56 +0000 (18:09 +0200)]
usb: gadget: Convert drivers to use module_usb_composite_driver()

Use the module_usb_composite_driver() macro where applicable to
eliminate the module_init/module_exit boilerplate in USB gadget composite
drivers.

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: Add helper macro for usb_composite_driver boilerplate
Tobias Klauser [Wed, 9 Jul 2014 16:09:55 +0000 (18:09 +0200)]
usb: gadget: Add helper macro for usb_composite_driver boilerplate

Introduce the module_usb_composite_driver macro as a convenience macro
for USB gadget composite driver modules, similar to module_usb_driver.
It is intended to be used by drivers which init/exit section does
nothing but calling usb_composite_probe/usb_composite_unrregister. By
using this macro it is possible to eliminate a few lines of boilerplate
code per USB gadget composite driver.

Based on f3a6a4b6 ("USB: Add helper macro for usb_driver boilerplate")
which introduced the according macro for USB drivers.

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: f_fs: OS descriptors support
Andrzej Pietrasiewicz [Wed, 9 Jul 2014 10:20:08 +0000 (12:20 +0200)]
usb: gadget: f_fs: OS descriptors support

Add support for OS descriptors. The new format of descriptors is used,
because the "flags" field is required for extensions. os_count gives
the number of OSDesc[] elements.
The format of descriptors is given in include/uapi/linux/usb/functionfs.h.

For extended properties descriptor the usb_ext_prop_desc structure covers
only a part of a descriptor, because the wPropertyNameLength is unknown
up front.

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: u_os_desc: helper functions for accessing ext prop buffer
Andrzej Pietrasiewicz [Wed, 9 Jul 2014 10:20:07 +0000 (12:20 +0200)]
usb: gadget: u_os_desc: helper functions for accessing ext prop buffer

Provide helper functions to get pointers to particular locations within
a buffer holding an extended properties descriptor.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: f_fs: rename descriptor parsing functions
Andrzej Pietrasiewicz [Wed, 9 Jul 2014 10:20:06 +0000 (12:20 +0200)]
usb: gadget: f_fs: rename descriptor parsing functions

ffs_do_desc() handles one descriptor, while ffs_do_descs() handles a number
of descriptors. The tho names are so similar that it causes confusion.
Rename to reflect their purpose better.

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: tegra: Do not include asm/mach-types.h
Thierry Reding [Mon, 7 Jul 2014 12:06:35 +0000 (14:06 +0200)]
usb: phy: tegra: Do not include asm/mach-types.h

It is no longer needed and keeping it will break 64-bit ARM builds.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: register nop transceiver driver for jz4740
Apelete Seketeli [Sun, 6 Jul 2014 17:58:03 +0000 (19:58 +0200)]
usb: musb: register nop transceiver driver for jz4740

Following the name change of the NOP transceiver driver in commit
4525bee (usb: phy: rename usb_nop_xceiv to usb_phy_generic), the
transceiver driver was no longer operable under its old name.

Register the transceiver driver before calling usb_get_phy() to make
sure we are noticed by an error message if it is not available.

Signed-off-by: Apelete Seketeli <apelete@seketeli.net>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Fix typo on #ifdef
Ricardo Ribalda Delgado [Fri, 4 Jul 2014 09:27:03 +0000 (11:27 +0200)]
usb: gadget: net2280: Fix typo on #ifdef

Commit e56e69cc0ff4 ("usb: gadget: net2280: Use pr_* function")
includes a editing mistake on one of the #ifdef.

This patch fixes it.

Reported-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: pxa25x_udc: use correct header for gpio devm_ functions
Arnd Bergmann [Wed, 9 Jul 2014 09:24:44 +0000 (11:24 +0200)]
usb: gadget: pxa25x_udc: use correct header for gpio devm_ functions

commit c63d2225e7be ("usb: gadget: pxa25x_udc: use devm_ functions")
introduced the use of devm_gpio_request in this driver, but did not
correctly include the header file declaring this function, which
causes a build failure.

This changes pxa25x_udc to include linux/gpio.h instead of asm/gpio.h
to fix this.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Himangi Saraogi <himangi774@gmail.com>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Felipe Balbi <balbi@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: renesas_usbhs: gadget: fix re-enabling pipe without re-connecting
Yoshihiro Shimoda [Wed, 9 Jul 2014 11:30:13 +0000 (20:30 +0900)]
usb: renesas_usbhs: gadget: fix re-enabling pipe without re-connecting

This patch fixes an issue that the renesas_usbhs driver in gadget mode
cannot work correctly even if I disabled DMAC of the driver when I used
the g_zero driver and the testusb tool.

When a usb cable is re-connected, the renesas_usbhs driver calls the
usbhsp_flags_init() (via usbhs_hotplug() --> usbhs_mod_call(start) -->
usbhsg_try_start() --> usbhs_pipe_init()). However, the driver doesn't
call the usbhsp_flags_init() when usbhsg_ep_disable() is called.
So, if a gadget driver calls usb_ep_enable() and usb_ep_disable() again
and again, the renesas_usbhs driver will output the following log:

  renesas_usbhs renesas_usbhs: can't get pipe (BULK)
  renesas_usbhs renesas_usbhs: wrong recip request

Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: renesas_usbhs: fix usbhs_pipe_malloc() to re-enable a pipe.
Yoshihiro Shimoda [Wed, 9 Jul 2014 11:30:03 +0000 (20:30 +0900)]
usb: renesas_usbhs: fix usbhs_pipe_malloc() to re-enable a pipe.

This patch fixes an issue that the driver cannot push a new data when
a pipe is re-enabled after the pipe is queued.

Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: s3c2410: Move to clk_prepare_enable/clk_disable_unprepare
Vasily Khoruzhick [Mon, 30 Jun 2014 19:13:29 +0000 (22:13 +0300)]
usb: gadget: s3c2410: Move to clk_prepare_enable/clk_disable_unprepare

Use clk_prepare_enable/clk_disable_unprepare to make the driver
work properly with common clock framework.

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: cppi41: fire hrtimer according to programmed channel length
Daniel Mack [Thu, 19 Jun 2014 22:20:44 +0000 (00:20 +0200)]
usb: musb: cppi41: fire hrtimer according to programmed channel length

The musb/cppi41 code installs a hrtimer to work around DMA completion
interrupts that have fired too early on AM335x hardware. This timer
is currently programmed to first fire 140 microseconds after the DMA
completion callback. According to the commit which introduced it
(a655f481d83, "usb: musb: musb_cppi41: handle pre-mature TX complete
interrupt"), that value is is considered a 'rule of thumb' that worked
well with the test case described in the commit log.

Test show, however, that for USB audio devices and much smaller packet
sizes, the timer has to fire earlier in order to correctly handle the audio
stream. The original test case had output transfer sizes of 1514 bytes, and
a delay of 140 microseconds. For audio devices with 24 bytes channel size, 3
microseconds seem to work well.

Hence, let's assume that the time it takes to clear the bit correlates with
the number of bytes transferred. The referenced commit log mentions such a
suspicion as well. Let the timer fire in cppi41_channel->total_len/10
microseconds to correctly handle both cases.

Also, shorten the interval in which the timer fires again in case of
a non-empty early_tx list.

With these changes in place, both FS and HS audio devices appear to work
well on AM335x hardware.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Reported-by: Sebastian Reimers <sebastian.reimers@googlemail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agoRevert "usb: musb: musb_cppi41: Handle ISOCH differently and not use the hrtimer."
Daniel Mack [Thu, 19 Jun 2014 22:20:43 +0000 (00:20 +0200)]
Revert "usb: musb: musb_cppi41: Handle ISOCH differently and not use the hrtimer."

This reverts commit 1af54b7a4.

The commit tried to address cases in which isochronous transfers are 'not
reliable', most probably in the tests conducted, polling for the
MUSB_TXCSR_TXPKTRDY bit in MUSB_TXCSR is done too late.

Hence, it installs a work struct which basically busy-polls for the bit in a
rather agressive way by rescheduling the work if the FIFO is not empty. With
USB audio devices, tests have shown that it takes approximately 100
iterations of the asynchronous worker until the FIFO signals completion,
which leads to 100% CPU loads when streaming audio.

The issue the patch tried to address can be handled differently, which is
what the next patch does.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Reported-by: Sebastian Reimers <sebastian.reimers@googlemail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: fix wrong indentation in musb_host.c
Daniel Mack [Mon, 26 May 2014 12:52:40 +0000 (14:52 +0200)]
usb: musb: fix wrong indentation in musb_host.c

Just a cosmetic cleanup with no functional change.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb/cppi41: call musb_ep_select() before accessing an endpoint's CSR
Daniel Mack [Mon, 26 May 2014 12:52:39 +0000 (14:52 +0200)]
usb: musb/cppi41: call musb_ep_select() before accessing an endpoint's CSR

Before accessing any of an endpoint's CSR registers, make sure the
correct endpoint is selected. Otherwise, data read from or written to
the registers is likely to affect the wrong endpoint as long as the
connected device has more than one endpoint.

This, of course, leads to all sorts of strange effects such as stream
starvation and driver internal state machine confusion due to spurious
interrupts.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: introduce dma_channel.rx_packet_done
Daniel Mack [Mon, 26 May 2014 12:52:38 +0000 (14:52 +0200)]
usb: musb: introduce dma_channel.rx_packet_done

The musb/cppi41 glue layer is capable of handling transactions that span
over more than one USB packet by reloading the DMA descriptors
partially. An urb is considered completed when either its transfer
buffer has been filled entirely (actual_length ==
transfer_buffer_length) or if a packet in the stream has less bytes than
the endpoint's wMaxPacketSize.

Once one of the above conditions is met, musb_dma_completion() is called
from cppi41_trans_done(). However, the final decision whether or not to
return the urb to its owner is made by the core and its determination of
the variable 'done' in musb_host_rx(). This code has currently no way of
knowing what the size of the last packet was, and whether or not to
give back the urb due to a short read.

Fix this by introducing a new boolean flag in 'struct dma_channel', and
set it from musb_cppi41.c. If set, it will make the core do what the
DMA layer decided and complete the urb.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: fix bit mask for CSR in musb_h_tx_flush_fifo()
Daniel Mack [Mon, 26 May 2014 12:52:37 +0000 (14:52 +0200)]
usb: musb: fix bit mask for CSR in musb_h_tx_flush_fifo()

The datasheet says that MUSB_TXCSR_FLUSHFIFO is only valid when
MUSB_TXCSR_TXPKTRDY is set as well.

With this patch applied, the warning in this function does no longer
kick in when an USB soundcard is unplugged while the stream is active.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: use is_host_active() to distinguish between host and gadget mode
Daniel Mack [Mon, 26 May 2014 12:52:36 +0000 (14:52 +0200)]
usb: musb: use is_host_active() to distinguish between host and gadget mode

On AM33xx platforms, unplugging a device in the middle of an active
transfer leads to a drop of MUSB_DEVCTL_HM in MUSB_DEVCTL before the
system is informed about a disconnect. This consequently makes the musb
core call the gadget code to handle the interrupt request, which then
crashes the kernel because the relevant pointers haven't been set up
for gadget mode.

To fix this, use is_host_active() rather than (devctl & MUSB_DEVCTL_HM)
in musb_interrupt() and musb_dma_completion() to detect whether the
controller is in host or peripheral mode. This information is provided
by the driver logic and does not rely on register contents.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: remove unnecessary (void) prefix at function calls
Daniel Mack [Mon, 26 May 2014 12:52:35 +0000 (14:52 +0200)]
usb: musb: remove unnecessary (void) prefix at function calls

Just a little cleanup that removes unnecessary casts.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: FunctionFS: Return -ENOENT instead of -ENODEV when device not found.
Krzysztof Opasiak [Wed, 21 May 2014 12:05:35 +0000 (14:05 +0200)]
usb: gadget: FunctionFS: Return -ENOENT instead of -ENODEV when device not found.

Syscall mount returns -ENODEV error if requested FS type
has not been found. Returning the same error from FFS mount
callback makes value returned to userspace misleading.

Other file systems returns -ENOENT if requested device
has not been found. Adjust FFS to this convention to make
error codes meaningfull.

Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: omap_udc: should not call gadget driver's .unbind
Peter Chen [Wed, 21 May 2014 01:04:21 +0000 (09:04 +0800)]
usb: gadget: omap_udc: should not call gadget driver's .unbind

It has already been covered by udc core

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2272: do not need to judge gadget driver's .unbind
Peter Chen [Wed, 21 May 2014 01:04:20 +0000 (09:04 +0800)]
usb: gadget: net2272: do not need to judge gadget driver's .unbind

It has already been covered by udc core, besides, we do not
need unbind at .udc_start

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: m66592-udc: should not call gadget driver's .unbind
Peter Chen [Wed, 21 May 2014 01:04:19 +0000 (09:04 +0800)]
usb: gadget: m66592-udc: should not call gadget driver's .unbind

It has already been covered by udc core

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fusb300_udc: should not call gadget driver's .unbind
Peter Chen [Wed, 21 May 2014 01:04:18 +0000 (09:04 +0800)]
usb: gadget: fusb300_udc: should not call gadget driver's .unbind

It has already been covered by udc core

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fsl_udc_core: should not call gadget driver's .unbind
Peter Chen [Wed, 21 May 2014 01:04:17 +0000 (09:04 +0800)]
usb: gadget: fsl_udc_core: should not call gadget driver's .unbind

It has already been covered by udc core

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: msm: Make phy_reset clk and reset line optional.
Srinivas Kandagatla [Mon, 30 Jun 2014 17:29:49 +0000 (18:29 +0100)]
usb: phy: msm: Make phy_reset clk and reset line optional.

This patch makes the phy reset clk and reset line optional as this clk
is not available on boards like IFC6410 with APQ8064.

Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: pxa27x_udc: prepare and unprepare the clock
Robert Jarzmik [Sun, 29 Jun 2014 14:01:10 +0000 (16:01 +0200)]
usb: gadget: pxa27x_udc: prepare and unprepare the clock

Add clock prepare and unprepare as required by clock framework.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agotools: ffs-aio-example: add license information
Robert Baldyga [Wed, 25 Jun 2014 06:26:49 +0000 (08:26 +0200)]
tools: ffs-aio-example: add license information

Add missing information about license. Some people will probably want to
reuse this code in their projects released under variety of licenses. For this
reason this example is under Public Domain license to avoid GPL limitations.

Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agotools: ffs-aio-example: convert to new descriptor format
Robert Baldyga [Wed, 25 Jun 2014 06:26:48 +0000 (08:26 +0200)]
tools: ffs-aio-example: convert to new descriptor format

Since commit [ac8dde11: “Add flags to descriptors block”] functionfs
supports a new descriptor format, so we update example application
to make it using recomended version of descriptors.

Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agotools: ffs-aio-example: fix header values endianess
Robert Baldyga [Wed, 25 Jun 2014 06:26:47 +0000 (08:26 +0200)]
tools: ffs-aio-example: fix header values endianess

We wrap numeric values of fs_count and hs_count fields in htole32,
because they should be in little-endian format.

Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Use quirks instead of pci id
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:12 +0000 (18:30 +0200)]
usb: gadget: net2280: Use quirks instead of pci id

Use of quirks improve readability and will be easier to add new devices
to this driver.

Suggested-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Use pr_* function
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:11 +0000 (18:30 +0200)]
usb: gadget: net2280: Use pr_* function

Driver was using custom functions WARNING, ERROR, DEBUG, instead of
 pr_err, pr_dgb...

New ep_* macros have been created that use standard pr_* functions.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Code Cleanup
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:10 +0000 (18:30 +0200)]
usb: gadget: net2280: Code Cleanup

- Move logical continuations to end of line
- Improve spacing

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Pass checkpacth.pl test
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:09 +0000 (18:30 +0200)]
usb: gadget: net2280: Pass checkpacth.pl test

Fix Code Style using checkpatch.pl criteria

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Refactor queues_show
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:08 +0000 (18:30 +0200)]
usb: gadget: net2280: Refactor queues_show

Replace a long and ugly expresion with an already available function.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Use module_pci_driver macro
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:07 +0000 (18:30 +0200)]
usb: gadget: net2280: Use module_pci_driver macro

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Use true/false instead of 1/0
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:06 +0000 (18:30 +0200)]
usb: gadget: net2280: Use true/false instead of 1/0

For bool variables

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Use BIT() macro
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:05 +0000 (18:30 +0200)]
usb: gadget: net2280: Use BIT() macro

Improves readability of the code

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Dont use magic numbers
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:04 +0000 (18:30 +0200)]
usb: gadget: net2280: Dont use magic numbers

Instead of using magic numbers use #defines

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: net2280: Add support for PLX USB338X
Ricardo Ribalda Delgado [Tue, 20 May 2014 16:30:03 +0000 (18:30 +0200)]
usb: gadget: net2280: Add support for PLX USB338X

This patch adds support for the PLX USB3380 and USB3382.

This driver is based on the driver from the manufacturer.

Since USB338X is register compatible with NET2280, I thought that it
would be better to include this hardware into net2280 driver.

Manufacturer's driver only supported the USB33X, did not follow the
Kernel Style and contain some trivial errors. This patch has tried to
address this issues.

This patch has only been tested on USB338x hardware, but the merge has
been done trying to not affect the behaviour of NET2280.

Tested-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:07 +0000 (15:19 +0800)]
usb: gadget: r8a66597-udc: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fusb300_udc: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:06 +0000 (15:19 +0800)]
usb: gadget: fusb300_udc: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: m66592-udc: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:05 +0000 (15:19 +0800)]
usb: gadget: m66592-udc: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: lpc32xx: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:04 +0000 (15:19 +0800)]
usb: gadget: lpc32xx: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fsl_udc_core: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:03 +0000 (15:19 +0800)]
usb: gadget: fsl_udc_core: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: atmel_usba_udc: delete __init marker for probe
Peter Chen [Sun, 18 May 2014 07:19:02 +0000 (15:19 +0800)]
usb: gadget: atmel_usba_udc: delete __init marker for probe

The probe function may be probed deferal and called after .init
section has freed.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: add support for USB 2.0-only core configuration
Paul Zimmerman [Fri, 23 May 2014 18:39:24 +0000 (11:39 -0700)]
usb: dwc3: add support for USB 2.0-only core configuration

Newer DWC3 controllers can be built for USB 2.0-only mode, where
most of the USB 3.0 circuitry is left out. To support this mode,
the driver must limit the speed programmed into the DCFG register
to Hi-Speed or lower.

Reads and writes to the PIPECTL register are left as-is, since
they should be no-ops in USB 2.0-only mode. Calls to phy_init()
etc. for the USB3 phy are also left as-is, since the no-op USB3
phy should be used for USB 2.0-only mode controllers.

Signed-off-by: Paul Zimmerman <paulz@synopsys.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: dsps: Call usb_phy(_shutdown/_init) during musb_platform_reset()
George Cherian [Mon, 26 May 2014 09:20:10 +0000 (14:50 +0530)]
usb: musb: dsps: Call usb_phy(_shutdown/_init) during musb_platform_reset()

For DSPS platform usb_phy_vbus(_off/_on) are NOPs.
So during musb_platform_reset() call usb_phy(_shutdown/_init)

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: Use kmemdup instead of kmalloc + memcpy
Benoit Taine [Mon, 26 May 2014 15:21:20 +0000 (17:21 +0200)]
usb: gadget: Use kmemdup instead of kmalloc + memcpy

This issue was reported by coccicheck using the semantic patch
at scripts/coccinelle/api/memdup.cocci

Signed-off-by: Benoit Taine <benoit.taine@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: musb_host.c: Cleaning up uninitialized variables
Rickard Strandqvist [Sun, 1 Jun 2014 13:48:12 +0000 (15:48 +0200)]
usb: musb: musb_host.c: Cleaning up uninitialized variables

There is a risk that the variable will be used without being initialized.

This was largely found by using a static code analysis program called cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: backfin: Introduce the use of the managed version of kzalloc
Himangi Saraogi [Sun, 1 Jun 2014 20:43:21 +0000 (02:13 +0530)]
usb: musb: backfin: Introduce the use of the managed version of kzalloc

This patch moves data allocated using kzalloc to managed data allocated
using devm_kzalloc and cleans now unnecessary kfrees in probe and remove
functions. Also, a label is done away with and err2 and err3 renamed.

The following Coccinelle semantic patch was used for making the change:

@platform@
identifier p, probefn, removefn;
@@
struct platform_driver p = {
  .probe = probefn,
  .remove = removefn,
};

@prb@
identifier platform.probefn, pdev;
expression e, e1, e2;
@@
probefn(struct platform_device *pdev, ...) {
  <+...
- e = kzalloc(e1, e2)
+ e = devm_kzalloc(&pdev->dev, e1, e2)
  ...
?-kfree(e);
  ...+>
}

@rem depends on prb@
identifier platform.removefn;
expression e;
@@
removefn(...) {
  <...
- kfree(e);
  ...>
}

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: tusb6010: Introduce the use of the managed version of kzalloc
Himangi Saraogi [Mon, 2 Jun 2014 15:45:05 +0000 (21:15 +0530)]
usb: musb: tusb6010: Introduce the use of the managed version of kzalloc

This patch moves data allocated using kzalloc to managed data allocated
using devm_kzalloc and cleans now unnecessary kfrees in probe and remove
functions. Also, the unnecesary labels are removed and linux/device.h is
added to make sure the devm_*() routine declarations are unambiguously
available.

The following Coccinelle semantic patch was used for making the change:

@platform@
identifier p, probefn, removefn;
@@
struct platform_driver p = {
  .probe = probefn,
  .remove = removefn,
};

@prb@
identifier platform.probefn, pdev;
expression e, e1, e2;
@@
probefn(struct platform_device *pdev, ...) {
  <+...
- e = kzalloc(e1, e2)
+ e = devm_kzalloc(&pdev->dev, e1, e2)
  ...
?-kfree(e);
  ...+>
}

@rem depends on prb@
identifier platform.removefn;
expression e;
@@
removefn(...) {
  <...
- kfree(e);
  ...>
}

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: davinci: use devm_ functions.
Himangi Saraogi [Mon, 2 Jun 2014 16:26:01 +0000 (21:56 +0530)]
usb: musb: davinci: use devm_ functions.

This patch moves data allocated using kzalloc to managed data allocated
using devm_kzalloc and cleans now unnecessary kfrees in probe and remove
functions. Also, a label is done away with and clk_get is replaced by it
corresponding devm version and the clk_puts are done away with. The
labels are renamed to make them ordered.

The following Coccinelle semantic patch was used for making the change:

@platform@
identifier p, probefn, removefn;
@@
struct platform_driver p = {
  .probe = probefn,
  .remove = removefn,
};

@prb@
identifier platform.probefn, pdev;
expression e, e1, e2;
@@
probefn(struct platform_device *pdev, ...) {
  <+...
- e = kzalloc(e1, e2)
+ e = devm_kzalloc(&pdev->dev, e1, e2)
  ...
?-kfree(e);
  ...+>
}

@rem depends on prb@
identifier platform.removefn;
expression e;
@@
removefn(...) {
  <...
- kfree(e);
  ...>
}

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: phy-gpio-vbus-usb: use devm_ functions
Himangi Saraogi [Tue, 3 Jun 2014 21:19:53 +0000 (02:49 +0530)]
usb: phy: phy-gpio-vbus-usb: use devm_ functions

The various devm_ functions allocate memory that is released when a
driver detaches.  This patch uses devm_kzalloc, devm_request_irq,
devm_gpio_request, devm_regulator_get etc. for data that is
allocated in the probe function of a platform device and is only
freed in the remove function. The corresponding free functions are
removed and the labels are done away with.

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: dwc3: Keeping 'resource' related code together
Vivek Gautam [Wed, 4 Jun 2014 09:04:52 +0000 (14:34 +0530)]
usb: dwc3: Keeping 'resource' related code together

Putting together the code related to getting the 'IORESOURCE_MEM'
and assigning the same to dwc->xhci_resources, for increasing
the readability.

Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: fsl_qe_udc: Introduce use of managed version of kzalloc
Himangi Saraogi [Sat, 14 Jun 2014 15:18:27 +0000 (20:48 +0530)]
usb: gadget: fsl_qe_udc: Introduce use of managed version of kzalloc

This patch moves data allocated using kzalloc to managed data allocated
using devm_kzalloc and cleans now unnecessary kfrees in probe and remove
functions. Also, the unnecesary labels are removed and some labels are
renamed to preserve ordering.

The following Coccinelle semantic patch was used for making the change:

@platform@
identifier p, probefn, removefn;
@@
struct platform_driver p = {
  .probe = probefn,
  .remove = removefn,
};

@prb@
identifier platform.probefn, pdev;
expression e, e1, e2;
@@
probefn(struct platform_device *pdev, ...) {
  <+...
- e = kzalloc(e1, e2)
+ e = devm_kzalloc(&pdev->dev, e1, e2)
  ...
?-kfree(e);
  ...+>
}

@rem depends on prb@
identifier platform.removefn;
expression e;
@@
removefn(...) {
  <...
- kfree(e);
  ...>
}

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: remove now unused clean_up and clean_up3 label.
Ben Dooks [Tue, 17 Jun 2014 15:15:01 +0000 (16:15 +0100)]
usb: gadget: r8a66597-udc: remove now unused clean_up and clean_up3 label.

With the devm additions, the clean_up and clean_up3 are now
not needed or used. Change clean_up3 and make everything use
clean_up2 and just remove clean_up.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: use devm_request_irq() to get device irq
Ben Dooks [Tue, 17 Jun 2014 15:15:00 +0000 (16:15 +0100)]
usb: gadget: r8a66597-udc: use devm_request_irq() to get device irq

Use the devm_request_irq() call to get the interrupt for the
device and have it automatically free on exit.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: use devm_clk_get() to get clock
Ben Dooks [Tue, 17 Jun 2014 15:14:59 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: use devm_clk_get() to get clock

Change to using the devm_clk_get() to get the clock and
have it automatically freed on exit.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: cleanup error path
Ben Dooks [Tue, 17 Jun 2014 15:14:58 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: cleanup error path

With the updates for devm, the cleanup path no longer needs to
check for NULL device state, so remove it and return directly
if the irq resource missing

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: handle sudmac registers with devm_ioremap_resource()
Ben Dooks [Tue, 17 Jun 2014 15:14:57 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: handle sudmac registers with devm_ioremap_resource()

Change the sudmac register handling in the devm_ioremap_resource
to use the devm variant.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: use devm_kzalloc() to allocate driver state
Ben Dooks [Tue, 17 Jun 2014 15:14:56 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: use devm_kzalloc() to allocate driver state

Update driver to use devm_kzalloc() to make tracking of resources
easier. Also remove the exit point via cleanup as there's no
cleanup necessary from this point now.

As a note, also removes the error print as the allocation calls
produce errors if they do not return memory.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: keep dev as reference to &pdev->dev
Ben Dooks [Tue, 17 Jun 2014 15:14:55 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: keep dev as reference to &pdev->dev

Remove usages of &pdev->dev in the driver probe function
with just dev to make the references to it easier to
write. Convert all the current users of it to use it.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: r8a66597-udc: use devm_ioremap_resource() for registers
Ben Dooks [Tue, 17 Jun 2014 15:14:54 +0000 (16:14 +0100)]
usb: gadget: r8a66597-udc: use devm_ioremap_resource() for registers

trivial patch removing boilerplate clode.

Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: lpc32xx_udc: Make of_device_id array const
Jingoo Han [Wed, 18 Jun 2014 04:41:39 +0000 (13:41 +0900)]
usb: gadget: lpc32xx_udc: Make of_device_id array const

Make of_device_id array const, because all OF functions handle
it as const.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: gr_udc: Make of_device_id array const
Jingoo Han [Wed, 18 Jun 2014 04:40:31 +0000 (13:40 +0900)]
usb: gadget: gr_udc: Make of_device_id array const

Make of_device_id array const, because all OF functions handle
it as const.

Acked-by: Andreas Larsson <andreas@gaisler.com>
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: tegra: Make of_device_id array const
Jingoo Han [Wed, 18 Jun 2014 04:43:50 +0000 (13:43 +0900)]
usb: phy: tegra: Make of_device_id array const

Make of_device_id array const, because all OF functions handle
it as const.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: phy: msm: Make of_device_id array const
Jingoo Han [Wed, 18 Jun 2014 04:42:44 +0000 (13:42 +0900)]
usb: phy: msm: Make of_device_id array const

Make of_device_id array const, because all OF functions handle
it as const.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: ux500: use devm_ functions
Himangi Saraogi [Fri, 20 Jun 2014 20:40:25 +0000 (02:10 +0530)]
usb: musb: ux500: use devm_ functions

This patch introduces the use of managed interfaces for clk_get and
kzalloc and removes the corresponding free function calls in the probe
and remove functions.

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: gadget: pxa25x_udc: use devm_ functions
Himangi Saraogi [Sun, 22 Jun 2014 07:17:51 +0000 (12:47 +0530)]
usb: gadget: pxa25x_udc: use devm_ functions

This patch introduces the use of devm_request_irq, devm_gpio_request,
devm_clk_get etc. instead of the corresponding unmanaged interfaces. The
calls to the functions like free_irq to free the allocated resources are
removed as they are no longer required. Some labels in the probe function
are also done away with and the name of the label err_gpio_pullup is
changed to make it less specific to the context.

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agousb: musb: dsps: coding style cleanup
Lothar Waßmann [Mon, 30 Jun 2014 11:17:26 +0000 (13:17 +0200)]
usb: musb: dsps: coding style cleanup

There is no reason for the register accessor functions not to adhere
to the CodingStyle rules.

Signed-off-by: Lothar Waßmann <LW@KARO-electronics.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
10 years agoLinux 3.16-rc3
Linus Torvalds [Sun, 29 Jun 2014 21:11:36 +0000 (14:11 -0700)]
Linux 3.16-rc3

10 years agoMerge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Linus Torvalds [Sun, 29 Jun 2014 20:40:08 +0000 (13:40 -0700)]
Merge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM fixes from Russell King:
 "Another round of ARM fixes.  The largest change here is the L2 changes
  to work around problems for the Armada 37x/380 devices, where most of
  the size comes down to comments rather than code.

  The other significant fix here is for the ptrace code, to ensure that
  rewritten syscalls work as intended.  This was pointed out by Kees
  Cook, but Will Deacon reworked the patch to be more elegant.

  The remainder are fairly trivial changes"

* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  ARM: 8087/1: ptrace: reload syscall number after secure_computing() check
  ARM: 8086/1: Set memblock limit for nommu
  ARM: 8085/1: sa1100: collie: add top boot mtd partition
  ARM: 8084/1: sa1100: collie: revert back to cfi_probe
  ARM: 8080/1: mcpm.h: remove unused variable declaration
  ARM: 8076/1: mm: add support for HW coherent systems in PL310 cache

10 years agoMAINTAINERS: exceptions for Documentation maintainer
Randy Dunlap [Sat, 28 Jun 2014 01:28:56 +0000 (18:28 -0700)]
MAINTAINERS: exceptions for Documentation maintainer

Note that I don't maintain Documentation/ABI/,
Documentation/devicetree/, or the language translation files.

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoDocumentation: add section about git to email-clients.txt
Dan Carpenter [Sat, 28 Jun 2014 01:28:46 +0000 (18:28 -0700)]
Documentation: add section about git to email-clients.txt

These days most people use git to send patches so I have added a section
about that.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoARM: 8087/1: ptrace: reload syscall number after secure_computing() check
Will Deacon [Fri, 27 Jun 2014 16:01:47 +0000 (17:01 +0100)]
ARM: 8087/1: ptrace: reload syscall number after secure_computing() check

On the syscall tracing path, we call out to secure_computing() to allow
seccomp to check the syscall number being attempted. As part of this, a
SIGTRAP may be sent to the tracer and the syscall could be re-written by
a subsequent SET_SYSCALL ptrace request. Unfortunately, this new syscall
is ignored by the current code unless TIF_SYSCALL_TRACE is also set on
the current thread.

This patch slightly reworks the enter path of the syscall tracing code
so that we always reload the syscall number from
current_thread_info()->syscall after the potential ptrace traps.

Acked-by: Kees Cook <keescook@chromium.org>
Tested-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8086/1: Set memblock limit for nommu
Laura Abbott [Fri, 27 Jun 2014 09:17:27 +0000 (10:17 +0100)]
ARM: 8086/1: Set memblock limit for nommu

Commit 1c2f87c (ARM: 8025/1: Get rid of meminfo) changed find_limits
to use memblock_get_current_limit for calculating the max_low pfn.
nommu targets never actually set a limit on memblock though which
means memblock_get_current_limit will just return the default
value. Set the memblock_limit to be the end of DDR to make sure
bounds are calculated correctly.

Signed-off-by: Laura Abbott <lauraa@codeaurora.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8085/1: sa1100: collie: add top boot mtd partition
Andrea Adami [Wed, 25 Jun 2014 21:32:26 +0000 (22:32 +0100)]
ARM: 8085/1: sa1100: collie: add top boot mtd partition

The CFI mapping is now perfect so we can expose the top block, read only.
There isn't much to read, though, just the sharpsl_params values.

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8084/1: sa1100: collie: revert back to cfi_probe
Andrea Adami [Wed, 25 Jun 2014 21:31:15 +0000 (22:31 +0100)]
ARM: 8084/1: sa1100: collie: revert back to cfi_probe

Reverts commit d26b17edafc45187c30cae134a5e5429d58ad676
ARM: sa1100: collie.c: fall back to jedec_probe flash detection

Unfortunately the detection was challenged on the defective unit used for tests:
one of the NOR chips did not respond to the CFI query.
Moreover that bad device needed extra delays on erase-suspend/resume cycles.

Tested personally on 3 different units and with feedback of two other users.

Signed-off-by: Andrea Adami <andrea.adami@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8080/1: mcpm.h: remove unused variable declaration
Nicolas Pitre [Thu, 19 Jun 2014 21:57:01 +0000 (22:57 +0100)]
ARM: 8080/1: mcpm.h: remove unused variable declaration

The sync_phys variable has been replaced by link time computation in
mcpm_head.S before the code was submitted upstream.

Signed-off-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8076/1: mm: add support for HW coherent systems in PL310 cache
Thomas Petazzoni [Fri, 13 Jun 2014 09:58:38 +0000 (10:58 +0100)]
ARM: 8076/1: mm: add support for HW coherent systems in PL310 cache

When a PL310 cache is used on a system that provides hardware
coherency, the outer cache sync operation is useless, and can be
skipped. Moreover, on some systems, it is harmful as it causes
deadlocks between the Marvell coherency mechanism, the Marvell PCIe
controller and the Cortex-A9.

To avoid this, this commit introduces a new Device Tree property
'arm,io-coherent' for the L2 cache controller node, valid only for the
PL310 cache. It identifies the usage of the PL310 cache in an I/O
coherent configuration. Internally, it makes the driver disable the
outer cache sync operation.

Note that technically speaking, a fully coherent system wouldn't
require any of the other .outer_cache operations. However, in
practice, when booting secondary CPUs, these are not yet coherent, and
therefore a set of cache maintenance operations are necessary at this
point. This explains why we keep the other .outer_cache operations and
only ->sync is disabled.

While in theory any write to a PL310 register could cause the
deadlock, in practice, disabling ->sync is sufficient to workaround
the deadlock, since the other cache maintenance operations are only
used in very specific situations.

Contrary to previous versions of this patch, this new version does not
simply NULL-ify the ->sync member, because the l2c_init_data
structures are now 'const' and therefore cannot be modified, which is
a good thing. Therefore, this patch introduces a separate
l2c_init_data instance, called of_l2c310_coherent_data.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoMerge tag 'spi-v3.16-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Sat, 28 Jun 2014 18:32:32 +0000 (11:32 -0700)]
Merge tag 'spi-v3.16-rc2' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A few driver specific fixes, the biggest one being a fix for the newly
  added Qualcomm SPI controller driver to make it not use its internal
  chip select due to hardware bugs, replacing it with GPIOs"

* tag 'spi-v3.16-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: qup: Remove chip select function
  spi: qup: Fix order of spi_register_master
  spi: sh-sci: fix use-after-free in sh_sci_spi_remove()
  spi/pxa2xx: fix incorrect SW mode chipselect setting for BayTrail LPSS SPI

10 years agoMerge tag 'regulator-v3.16-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 28 Jun 2014 18:31:58 +0000 (11:31 -0700)]
Merge tag 'regulator-v3.16-rc2' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "Several driver specific fixes here, the palmas fixes being especially
  important for a range of boards - the recent updates to support new
  devices have introduced several regressions"

* tag 'regulator-v3.16-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: tps65218: Correct the the config register for LDO1
  regulator: tps65218: Add the missing of_node assignment in probe
  regulator: palmas: fix typo in enable_reg calculation
  regulator: bcm590xx: fix vbus name
  regulator: palmas: Fix SMPS enable/disable/is_enabled

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
Linus Torvalds [Sat, 28 Jun 2014 16:43:58 +0000 (09:43 -0700)]
Merge git://git./linux/kernel/git/nab/target-pending

Pull SCSI target fixes from Nicholas Bellinger:
 "Mostly minor fixes this time around.  The highlights include:

   - iscsi-target CHAP authentication fixes to enforce explicit key
     values (Tejas Vaykole + rahul.rane)
   - fix a long-standing OOPs in target-core when a alua configfs
     attribute is accessed after port symlink has been removed.
     (Sebastian Herbszt)
   - fix a v3.10.y iscsi-target regression causing the login reject
     status class/detail to be ignored (Christoph Vu-Brugier)
   - fix a v3.10.y iscsi-target regression to avoid rejecting an
     existing ITT during Data-Out when data-direction is wrong (Santosh
     Kulkarni + Arshad Hussain)
   - fix a iscsi-target related shutdown deadlock on UP kernels (Mikulas
     Patocka)
   - fix a v3.16-rc1 build issue with vhost-scsi + !CONFIG_NET (MST)"

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  iscsi-target: fix iscsit_del_np deadlock on unload
  iovec: move memcpy_from/toiovecend to lib/iovec.c
  iscsi-target: Avoid rejecting incorrect ITT for Data-Out
  tcm_loop: Fix memory leak in tcm_loop_submission_work error path
  iscsi-target: Explicily clear login response PDU in exception path
  target: Fix left-over se_lun->lun_sep pointer OOPs
  iscsi-target; Enforce 1024 byte maximum for CHAP_C key value
  iscsi-target: Convert chap_server_compute_md5 to use kstrtoul

10 years agoMerge remote-tracking branches 'spi/fix/pxa2xx', 'spi/fix/qup' and 'spi/fix/sh-sci...
Mark Brown [Sat, 28 Jun 2014 13:01:23 +0000 (14:01 +0100)]
Merge remote-tracking branches 'spi/fix/pxa2xx', 'spi/fix/qup' and 'spi/fix/sh-sci' into spi-linus

10 years agoMerge remote-tracking branches 'regulator/fix/bcm590xx', 'regulator/fix/palmas' and...
Mark Brown [Sat, 28 Jun 2014 13:01:04 +0000 (14:01 +0100)]
Merge remote-tracking branches 'regulator/fix/bcm590xx', 'regulator/fix/palmas' and 'regulator/fix/tps65218' into regulator-linus

10 years agoiscsi-target: fix iscsit_del_np deadlock on unload
Mikulas Patocka [Mon, 23 Jun 2014 17:42:37 +0000 (13:42 -0400)]
iscsi-target: fix iscsit_del_np deadlock on unload

On uniprocessor preemptible kernel, target core deadlocks on unload. The
following events happen:
* iscsit_del_np is called
* it calls send_sig(SIGINT, np->np_thread, 1);
* the scheduler switches to the np_thread
* the np_thread is woken up, it sees that kthread_should_stop() returns
  false, so it doesn't terminate
* the np_thread clears signals with flush_signals(current); and goes back
  to sleep in iscsit_accept_np
* the scheduler switches back to iscsit_del_np
* iscsit_del_np calls kthread_stop(np->np_thread);
* the np_thread is waiting in iscsit_accept_np and it doesn't respond to
  kthread_stop

The deadlock could be resolved if the administrator sends SIGINT signal to
the np_thread with killall -INT iscsi_np

The reproducible deadlock was introduced in commit
db6077fd0b7dd41dc6ff18329cec979379071f87, but the thread-stopping code was
racy even before.

This patch fixes the problem. Using kthread_should_stop to stop the
np_thread is unreliable, so we test np_thread_state instead. If
np_thread_state equals ISCSI_NP_THREAD_SHUTDOWN, the thread exits.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
10 years agoMerge tag 'iommu-fixes-v3.16-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 28 Jun 2014 02:00:45 +0000 (19:00 -0700)]
Merge tag 'iommu-fixes-v3.16-rc1' of git://git./linux/kernel/git/joro/iommu

Pull IOMMU fixes from Joerg Roedel:

 - fix VT-d regression with handling multiple RMRR entries per device

 - fix a small race that was left in the mmu_notifier handling in the
   AMD IOMMUv2 driver

* tag 'iommu-fixes-v3.16-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu/amd: Fix small race between invalidate_range_end/start
  iommu/vt-d: fix bug in handling multiple RMRRs for the same PCI device

10 years agoMerge branch 'x86/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Linus Torvalds [Sat, 28 Jun 2014 01:43:03 +0000 (18:43 -0700)]
Merge branch 'x86/urgent' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Peter Anvin:
 "A pile of fixes related to the VDSO, EFI and 32-bit badsys handling.

  It turns out that removing the section headers from the VDSO breaks
  gdb, so this puts back most of them.  A very simple typo broke
  rt_sigreturn on some versions of glibc, with obviously disastrous
  results.  The rest is pretty much fixes for the corresponding fallout.

  The EFI fixes fixes an arithmetic overflow on 32-bit systems and
  quiets some build warnings.

  Finally, when invoking an invalid system call number on x86-32, we
  bypass a bunch of handling, which can make the audit code oops"

* 'x86/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi-pstore: Fix an overflow on 32-bit builds
  x86/vdso: Error out in vdso2c if DT_RELA is present
  x86/vdso: Move DISABLE_BRANCH_PROFILING into the vdso makefile
  x86_32, signal: Fix vdso rt_sigreturn
  x86_32, entry: Do syscall exit work on badsys (CVE-2014-4508)
  x86/vdso: Create .build-id links for unstripped vdso files
  x86/vdso: Remove some redundant in-memory section headers
  x86/vdso: Improve the fake section headers
  x86/vdso2c: Use better macros for ELF bitness
  x86/vdso: Discard the __bug_table section
  efi: Fix compiler warnings (unused, const, type)