This reverts commit 9b0a1de3.

Aaro writes:
	With v3.12-rc4 I can no longer connect to N800 (OMAP2) with USB
	(peripheral, g_ether).

	According to git bisect this is caused by:

	9b0a1de3 is the first bad commit

So revert this patch, as Felipe says:
	It's unfortunate that tusb6010 is so messed up
Reported-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Acked-by: NFelipe Balbi <balbi@ti.com>
Cc: Bin Liu <b-liu@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

I have am335x-evm with one port running in OTG mode. Since commit
fe4cb091 ("usb: musb: gadget: remove hcd initialization") the loaded
gadget does non pop up on the host. All I see is
|usb 4-5: new high-speed USB device number 52 using ehci-pci
|usb 4-5: device descriptor read/64, error -110

Since a later commit 2cc65fea ("usb: musb: add musb_host_setup() and
musb_host_cleanup()) the gadget shows up on the host again but only
in OTG mode (because we have the host init code running). It does not
work in device only mode.
If running in OTG mode and the gadget is removed and added back (rmmod
followed by modprobe of a gadget) then the same error is pops up on the
host side.

This patch ensures that the gadget side also executes musb_start() which
puts the chip in "connect accept" mode. With this change the device
works in OTG & device mode and the gadget can be added & removed
multiple times.
A device (if musb is in OTG mode acting as a host) is only recognized if
it is attached during module load (musb_hdrc module). After the device
unplugged and plugged again the host does not recognize it. We get a
buch of errors if musb running in OTG mode, attached to a host and no
gadget is loaded. Bah.
This is one step forward. Host & device only mode should work. I will
look at OTG later. I looked at this before commit fe4cb091 and OTG wasn't
working there perfectly so I am not sure that it is a regression :)

Cc: <stable@vger.kernel.org> # v3.11
Cc: Daniel Mack <zonque@gmail.com>
Cc: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In gadget mode, musb->is_active should be set only when connected to the
host. musb_g_reset() already takes care of it.
Signed-off-by: NBin Liu <b-liu@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Right after the musb_hdrc driver is loaded, the otg default state
is a_idle, and Mode=Host, which are set by musb_host_setup().

This causes the following kernel message during musb gadget
enumeration.

	CAUTION: musb: Babble Interrupt Occurred

This patch sets the otg default state to b_idle, and its Mode to
Peripheral.

It has been validated on TI AM335x GP EVM USB0 port with g_zero.
Signed-off-by: NBin Liu <b-liu@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

musb_shutdown() removes always USB host and device.
musb_init_controller() adds host and device depending on port_mode. If
port mode is set to HOST then the removal of UDC leads only to:
|(NULL device *): gadget not registered.
and nothing else happens. If port mode is set to DEVICE and we remove
the host then we oops in usb_remove_hcd().
This patch ensures that we only remove host in OTG/host mode and device
only in OTG/device mode to avoid any trouble.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Since commit 511f3c53 (usb: gadget: udc-core: fix a regression during
gadget driver unbinding) usb_gadget_remove_driver will pass NULL for
the driver argument.
Signed-off-by: NMaarten ter Huurne <maarten@treewalker.org>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Since dma_map_single() may fail it is good to actually check the return
code to see if it succeeded.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch makes use of the two function is_cppi_enabled() and
tusb_dma_omap() instead of the ifdef for the proper DMA implementation
setup code. It basically shifts the code right by one indention level
and adds a few line breaks once the chars are crossed.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

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: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This will be done from a more appropriate location and as it doesn't
work anyway, it can safely be removed before the other changes.
Signed-off-by: NDaniel Mack <zonque@gmail.com>
Acked-by: NPeter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

On some platform configurations (like OMAP3+twl4030) it's the platform
code that enables VBUS, not OTG transceiver, so call vbus platform
callback instead, it will then call the transceiver if needed.

This fixes a use case where USB cable is plugged first and gadget
driver is loaded later after that.
Signed-off-by: NGrazvydas Ignotas <notasas@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Some musb glue drivers use gadget_driver pointer to know if any gadget
drivers are loaded at some moment and base further decisions on it,
like to do runtime suspend/resume or not. Right now the pointer is
left alone on stop and OMAP musb glue later does wrong runtime_pm
decisions because of it.

Clear the gadget_driver pointer on remove, it's invalid after stop
anyway.
Signed-off-by: NGrazvydas Ignotas <notasas@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

all other functions under drivers/usb/ start
with usb_, let's do the same thing.

This patch is in preparation for moving otg_state_string
to usb-common.c and deleting otg.c completely.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

udc-core provides a better way to handle release
methods, let's use it.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

udc-core now sets dma-related and parent fields
for us, we don't need to do it ourselves.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Fix the following compile warning:

drivers/usb/musb/musb_gadget.c: In function ‘rxstate’:
drivers/usb/musb/musb_gadget.c:714:22: warning: comparison of distinct pointer types lacks a cast [enabled by default]
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Those comments haven't been updated for a long
time, so much that they don't make sense anymore.

Best to remove them.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

We don't need the ->register_my_device flag
anymore because all UDC drivers have been
properly converted.

Let's remove every history of it.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

By simply setting a flag, we can delete a little
boilerplate code.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

musb does not use DMA buffer for ep0 but it uses the same giveback
function *musb_g_giveback* for all endpoints (*musb_g_ep0_giveback* calls
*musb_g_giveback*). So for ep0 case request.dma will be '0'
and will result in kernel OOPS if tried to *unmap_dma_buffer* for requests in
ep0. Fixed it by doing *unmap_dma_buffer* only for valid DMA addr and
checking that musb_ep->dma is valid when unmapping.
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The TXMAXP register is not set correctly for full speed bulk case
when the can_bulk_split() is used. Without this PIO transfers will
not take place correctly

The "mult" factor needs to be updated correctly for the
can_bulk_split() case

The AUTOSET bit in the TXCSR is not being set if the "mult"
factor is greater than 0 for the High Bandwidth ISO case.
But the "mult" factor is also greater than 0 in case of Full speed
bulk transfers with the packet splitting in TXMAXP register

Without the AUTOSET the DMA transfers will not progress in mode1

[ balbi@ti.com : add braces to both branches ]
Signed-off-by: Nsupriya karanth <supriya.karanth@stericsson.com>
Signed-off-by: NPraveena NADAHALLY <praveen.nadahally@stericsson.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

CONFIG_HOTPLUG is going away as an option so __devinit is no longer
needed.
Signed-off-by: NBill Pemberton <wfp5p@virginia.edu>
Cc: Peter Korsgaard <jacmet@sunsite.dk>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Cc: Li Yang <leoli@freescale.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Cc: Geoff Levand <geoff@infradead.org>
Cc: Wan ZongShun <mcuos.com@gmail.com>
Cc: Olav Kongas <ok@artecdesign.ee>
Cc: Lennert Buytenhek <kernel@wantstofly.org>
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This is part of the workaround for AM35x advisory Advisory 1.1.20.
The advisory says that the IPSS bridge can't handle 8 & 16 bit read
access. An 16bit read access to MUSB_INTRTXE results in an 32bit read
access which also reads INTRRX and therefore may lose interrupts.
This patch uses a shadow register of MUSB_INTRTXE so we only perform
write access to it.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This is part of the workaround for AM35x advisory Advisory 1.1.20.
The advisory says that the IPSS bridge can't handle 8 & 16 bit read
access. An 16bit read access to MUSB_INTRRXE results in an 32bit read
access which also reads INTRUSB and therefore may lose interrupts.
This patch uses a shadow register of MUSB_INTRRXE so we only perform
write access to it.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This is part of the workaround for AM35x advisory Advisory 1.1.20.
The advisory says that the IPSS bridge can't handle 8 & 16 bit read
access. An 8bit read access to FADDR results in an 32bit read
access which also reads INTRTX and therefore may lose interrupts.
This patch removes any reads to FADDR as they are not really required.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This reverts commit 4f3e8d26.

it turns out that current implementation of Mode 1
DMA added a few regressions to some users, so we
need to revert this patch and let author work on
a better version of Mode 1 DMA support.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Remove unnecessary if condition. No change in logic.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

we are compiling the driver always with full OTG
capabilities, so that board_mode trick becomes
useless.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Do not rely on any hints from gadget drivers and use DMA mode 1
whenever we expect data of at least the endpoint's packet size and
have not yet received a short packet.

The last packet if short is always transferred using DMA mode 0.

This patch fixes USB throughput issues in mass storage mode for
host to device transfers.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This is to reduce the overhead of dma programming for zero byte
transmit as same can be done using pio mode.
Signed-off-by: NAjay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In rxstate(), improper types are given to 'fifo_count' and 'len' variables, and
these variables are not used in accordance to their names (up to the certain
point), i.e. 'len' to hold the size of a packet in the RX FIFO, and 'fifo_count'
to hold a difference between 'request->length' and 'request->actual'...
Interchange the variables up to the point where their use starts to make sense
again.
Signed-off-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

usb_ep_ops.disable must clear external copy of the endpoint descriptor,
otherwise musb crashes after loading/unloading several gadget modules
in a row:

Unable to handle kernel paging request at virtual address bf013730
pgd = c0004000
[bf013730] *pgd=8f26d811, *pte=00000000, *ppte=00000000
Internal error: Oops: 7 [#1]
Modules linked in: g_cdc [last unloaded: g_file_storage]
CPU: 0    Not tainted  (3.2.17 #647)
PC is at musb_gadget_enable+0x4c/0x24c
LR is at _raw_spin_lock_irqsave+0x4c/0x58
[<c027c030>] (musb_gadget_enable+0x4c/0x24c) from [<bf01b760>] (gether_connect+0x3c/0x19c [g_cdc])
[<bf01b760>] (gether_connect+0x3c/0x19c [g_cdc]) from [<bf01ba1c>] (ecm_set_alt+0x15c/0x180 [g_cdc])
[<bf01ba1c>] (ecm_set_alt+0x15c/0x180 [g_cdc]) from [<bf01ecd4>] (composite_setup+0x85c/0xac4 [g_cdc])
[<bf01ecd4>] (composite_setup+0x85c/0xac4 [g_cdc]) from [<c027b744>] (musb_g_ep0_irq+0x844/0x924)
[<c027b744>] (musb_g_ep0_irq+0x844/0x924) from [<c027a97c>] (musb_interrupt+0x79c/0x864)
[<c027a97c>] (musb_interrupt+0x79c/0x864) from [<c027aaa8>] (generic_interrupt+0x64/0x7c)
[<c027aaa8>] (generic_interrupt+0x64/0x7c) from [<c00797cc>] (handle_irq_event_percpu+0x28/0x178)
...

Cc: stable@vger.kernel.org # v3.1+
Signed-off-by: NGrazvydas Ignotas <notasas@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

User can trigger disabling of gadget at run time while the
transfers are going on.
Eg: 1: rmmod of musb driver while transfers are going on

Eg: 2: On android doing:
 echo 0       > /sys/class/android_usb/android0/enable
While a big file transfer is going on via PTP/MTP.

In such a case, musb_gadget_disable() calls nuke()
but the dma interrupt may still happen for an endpoint since hw
would raise the interrupt in anycase.

This can result in a NULL pointer access crash:

[  314.030426] PC is at txstate+0x74/0x20c
[  314.034759] LR is at musb_g_tx+0x140/0x204
[  314.039489] pc : [<c03506f4>]    lr : [<c0350bcc>]    psr: 20000193
[  314.039520] sp : c783bc68  ip : 00000002  fp : c783bc9c
[  314.052429] r10: 00000018  r9 : 00000000  r8 : 00000200
[  314.058258] r7 : 00000000  r6 : fc0ab130  r5 : c781a410  r4 : c6caf640
[  314.065643] r3 : 00000000  r2 : 00000000  r1 : 00000000  r0 : c781a000
[  315.083251] Backtrace:
[  315.086242] [<c0350680>] (txstate+0x0/0x20c) from [<c0350bcc>] (musb_g_tx+0x140/0x204)
[  315.095123] [<c0350a8c>] (musb_g_tx+0x0/0x204) from [<c034eb00>] (musb_dma_completion+0x40/0x54)
[  315.104980] [<c034eac0>] (musb_dma_completion+0x0/0x54) from [<c0351e6c>] (dma_controller_irq+0x118/0x184)
[  315.115661] [<c0351d54>] (dma_controller_irq+0x0/0x184) from [<c00d86b8>] (handle_irq_event_percpu+0x54/0x188)

So put protection in code to handle possiblity of getting an interrupt for an
endpoint that might have been already nuked.
Reported-by: NTodd Poynor <toddpoynor@google.com>
Signed-off-by: NVikram Pandita <vikram.pandita@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

There's really no point in having hcd->irq as a
signed integer when we consider the fact that
IRQ 0 means NO_IRQ. In order to avoid confusion,
make hcd->irq unsigned and fix users who were
passing -1 as the IRQ number to usb_add_hcd.
Tested-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This changes the otg functions so that they receive struct
otg instead of struct usb_phy as parameter and
converts all users of these functions to pass the otg member
of their usb_phy.

Includes fixes to IMX code from Sascha Hauer.

[ balbi@ti.com : fixed a compile warning on ehci-mv.c ]
Signed-off-by: NHeikki Krogerus <heikki.krogerus@linux.intel.com>
Acked-by: NSascha Hauer <s.hauer@pengutronix.de>
Acked-by: NIgor Grinberg <grinberg@compulab.co.il>
Acked-by: NPavankumar Kondeti <pkondeti@codeaurora.org>
Acked-by: NLi Yang <leoli@freescale.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Reviewed-by: NMarek Vasut <marek.vasut@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Use the new usb_phy_* functions with transceiver
operations instead of the old otg functions.

Includes fixes from Sascha Hauer.
Signed-off-by: NHeikki Krogerus <heikki.krogerus@linux.intel.com>
Acked-by: NSascha Hauer <s.hauer@pengutronix.de>
Acked-by: NPavankumar Kondeti <pkondeti@codeaurora.org>
Acked-by: NLi Yang <leoli@freescale.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Reviewed-by: NMarek Vasut <marek.vasut@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

musb INDEX register is getting modified/corrupted during temporary
un-locking in a SMP system. Set this register with proper value
after re-acquiring the lock

Scenario:
---------
CPU1 is handling a data transfer completion interrupt received for
the CLASS1 EP
CPU2 is handling a CLASS2 thread which is queuing data to musb for
transfer

Below is the error sequence:

         CPU1                   |             CPU2
--------------------------------------------------------------------
Data transfer completion inter- |
rupt recieved.                  |
                                |
musb INDEX reg set to CLASS1 EP |
                                |
musb LOCK is acquired.          |
                                |
                                | CLASS2 thread queues data.
                                |
                                | CLASS2 thread tries to acquire musb
                                | LOCK but lock is already taken by
                                | CLASS1, so CLASS2 thread is
                                | spinning.
                                |
From Interrupt Context musb     |
giveback function is called     |
                                |
The giveback function releases  | CLASS2 thread now acquires LOCK
LOCK                            |
                                |
ClASS1 Request's completion cal-| ClASS2 schedules the data transfer and
lback is called                 | sets the MUSB INDEX to Class2 EP number
                                |
Interrupt handler for CLASS1 EP |
tries to acquire LOCK and is    |
spinning                        |
                                |
Interrupt for Class1 EP acquires| Class2 completes the scheduling etc and
the MUSB LOCK                   | releases the musb LOCK
                                |
Interrupt for Class1 EP schedul-|
es the next data transfer       |
but musb INDEX register is still|
set to CLASS2 EP                |

Since the MUSB INDEX register is set to a different endpoint, we
read and modify the wrong registers. Hence data transfer will not
happen properly. This results in unpredictable behavior

So, the MUSB INDEX register is set to proper value again when
interrupt re-acquires the lock

Cc: stable@vger.kernel.org
Signed-off-by: NSupriya Karanth <supriya.karanth@stericsson.com>
Signed-off-by: NPraveena Nadahally <praveen.nadahally@stericsson.com>
Reviewed-by: Nsrinidhi kasagar <srinidhi.kasagar@stericsson.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Use struct usb_otg members with OTG specific functions instead
of usb_phy members.

[ balbi@ti.com: added a missing change on musb_gadget.c to avoid
	a compile error on a later patch ]
Signed-off-by: NHeikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: NMarek Vasut <marek.vasut@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

There's really no point in doing all that
initcall trickery when we can safely let
udev handle module probing for us.

Remove all of that trickery, by moving everybody
to module_init() and making proper use of
platform_device_register() rather than
platform_device_probe().
Tested-by: NRajashekhara, Sudhakar <sudhakar.raj@ti.com>
Tested-by: NTasslehoff Kjappfot <tasskjapp@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Remove a few unnecessary headers from a
few files.
Signed-off-by: NFelipe Balbi <balbi@ti.com>