When clearing HALT on an endpoint, req->complete of in progress
requests must be called with locks off. New request should only be
started if there is not already a pending request on the endpoint.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NJohn Youn <johnyoun@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When a remote wakeup happens during bus_suspend, hcd needs to resume
its root hub.
Acked-by: NJohn Youn <johnyoun@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Fix  using the  bare number  to set the 'wHubCharacteristics' field of the Hub
Descriptor while the values are #define'd in <linux/usb/ch11.h>.
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This reverts commit 0cf884e8.
Even after applying the follow-on patch at
https://patchwork.kernel.org/patch/5325111
there are still problems with device connect on the Altera SOCFPGA
platform at least. One possible fix would be to add a whitelist
to enable suspend/resume on platforms where it does work correctly.
Signed-off-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

If usb bus is reset without a physical disconnection, all endpoints
will remain open. Call s3c_hsotg_disconnect() from reset handler to
report a disconnect to gadget framework. hsotg->connected is checked
in s3c_hsotg_disconnect() before processing disconnect.

In some cases, USBRst is seen before EnumDone and after it as well.
So move setting of hsotg->connected to set-address to avoid reporting
extra disconnection in this case.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When booting with id pin grounded, dwc2 default to host mode.
Thus, force device mode prior initializing gadget part.
Else fifo init will fail since fifo values are not correct
in host mode.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This callback informs the driver about the total amount of current it
is allowed to draw. Share this information with the phy so that
current limits can be set for charging for example.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Gadget must be informed about disconnection when pullup is
removed.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

After all endpoints are disabled, fifo_map should have reached 0.
Its a bug if if didn't, so warn about it and reset it to 0 so that
driver can continue using all the fifos.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When selecting different alt setting, s3c_hsotg_ep_enable can be
called with fifo already allocated. Allocate fifo again only if
required and after deallocating the previous fifo.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Current algorithm picks the first fifo which is equal to or greater
than the required size. This can result in bigger fifos assigned to
endpoints with smaller maxps. Change the algorithm to pick the
smallest fifo which is greater than or equal to the required size.

Moreover, only use signed variables when required.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This flag is set before sending the zlp. So use present tense instead
of the past tense.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Enable phy driver to report vbus session. This allows us to remove
D+ pullup when vbus is not present.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Can't stay in the loop forever. Break it after timeout.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The warning is probably good but it has false positives in both dma and non-dma
cases. So its not very helpful in either.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

According to programming guide, zero length packet should be
programmed on its own and should not be counted in DIEPTSIZ.PktCnt
with other packets.

For ep0, this is the zlp for DATA IN stage (if required) and not for
the STATUS stage.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Manage ep0 state in software to add handling of status OUT stage.
Just toggling hsotg->setup in s3c_hsotg_handle_outdone leaves it in
wrong state in 2-stage control transfers.
Moreover, ensure that for setup-packet s3c_hsotg_handle_outdone is
called either from SetupDone or OutDone but not both. Dwc2 ip v3.00a
generates both SetupDone and OutDone on setup packets.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

kill_all_requests() can flush the fifo. Call it after disabling the
endpoint.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When matching tx fifo to endpoint, consider all fifos instead of
hard limiting to 8
Moreover, print error in case no fifo could be found.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

< 15 check doesn't show debug information for endpoint 15.
It is possible to have less than 15 endpoints so use limit
provided by hardware configuration.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

These members are only occupying space.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

As fifo size can vary between SOCs, add possibility to configure
them from device tree. Fifo sizes used by the legacy driver will
be used If they are not provided by the device tree.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Reviewed-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Current code does not check endpoint 15 interrupt. Use number
of endpoint configured in hardware instead of the hardcoded value.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

GHWCFG1 provides hardware configuration of each endpoint. Use
it to configure the endpoints instead of assuming all even
endpoint are OUT and all odd endpoints are IN.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

* Add an of specific function to parse device node properties.
* Enable dma usage only if device tree property 'g_use_dma' is present.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Do a read-modify-write instead of only setting DMAEn bit.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

HBstLen is GAHBCFG[4:1]. Use GAHBCFG_HBSTLEN_SHIFT to write burst-
length at correct position.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In the error path, s3c_hsotg_phy_disable should be called after a
call to s3c_hsotg_phy_enable is made.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When using DMA, data of the previous setup packet can be read back
from cache because ep0 and ctrl buffers are embedded in struct s3c_hsotg.
Allocate buffers instead of embedding them.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Reviewed-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Only process DOEPINT.XferCompl on data packet as DOEPINTn.SetUp can
occur with or without DOEPINT.XferCompl. When DOEPINT.SetUp occurs
with DOEPINT.XferCompl, only DOEPINT.SetUp needs to be handled.

Moreover, ignore DOEPINT.XferCompl when it occurs with
DOEPINT.StupPktRcvd as driver needs to wait for DOEPINT.SetUp to
continue.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When using DMA, keep fifo empty interrupt disabled. Otherwise
core is flooded by interrupts.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Bind peripheral controller to the phy on udc_start. Unbind on
udc_stop.
Tested-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NGregory Herrero <gregory.herrero@intel.com>
Signed-off-by: NMian Yousaf Kaukab <yousaf.kaukab@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch fixes in simpler way the bug described in [1] and [2]. It
looks like DWC2 is the only UDC driver that doesn't force usb requests
to complete in ep_disable() function. This causes described problem,
because we have no guarantee that all requests will be completed before
unbind of usb function.

To fix this problem we force all requests of disabled endpoint to complete.
Also currently running request is not handled. This allowed to simplify
code of kill_all_requests() function, because 'force' parameter is always
set to true, so we don't need it anymore.

In s3c_hsotg_rx_data() we change function used to print message when active
request is NULL from dev_warn() to dev_dbg(), because such situation is
harmless for driver and now it can take place during normal endpoint
disabling.

[1] https://lkml.org/lkml/2014/12/9/283
[2] https://lkml.org/lkml/2014/12/12/360Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Get PHY parameters from devicetree and power off usb PHY during
system suspend.
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NYunzhi Li <lyz@rock-chips.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Fixed 3 typos in comments
Signed-off-by: NMickael Maison <mickael.maison@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This makes us sure that all requests are completed before we unbind
gadget. There are assumptions in gadget API that all requests have to
be completed and leak of complete can break some usb function drivers.

For example unbind of ECM function can cause NULL pointer dereference:

[   26.396595] configfs-gadget gadget: unbind function
'cdc_ethernet'/e79c4c00
[   26.414999] Unable to handle kernel NULL pointer dereference at
virtual address 00000000
(...)
[   26.452223] PC is at ecm_unbind+0x6c/0x9c
[   26.456209] LR is at ecm_unbind+0x68/0x9c
(...)
[   26.603696] [<c033fdb4>] (ecm_unbind) from [<c033661c>]
(purge_configs_funcs+0x94/0xd8)
[   26.611674] [<c033661c>] (purge_configs_funcs) from [<c0336674>]
(configfs_composite_unbind+0x14/0x34)
[   26.620961] [<c0336674>] (configfs_composite_unbind) from
[<c0337124>] (usb_gadget_remove_driver+0x68/0x9c)
[   26.630683] [<c0337124>] (usb_gadget_remove_driver) from [<c03376c8>]
(usb_gadget_unregister_driver+0x64/0x94)
[   26.640664] [<c03376c8>] (usb_gadget_unregister_driver) from
[<c0336be8>] (unregister_gadget+0x20/0x3c)
[   26.650038] [<c0336be8>] (unregister_gadget) from [<c0336c84>]
(gadget_dev_desc_UDC_store+0x80/0xb8)
[   26.659152] [<c0336c84>] (gadget_dev_desc_UDC_store) from
[<c0335120>] (gadget_info_attr_store+0x1c/0x28)
[   26.668703] [<c0335120>] (gadget_info_attr_store) from [<c012135c>]
(configfs_write_file+0xe8/0x148)
[   26.677818] [<c012135c>] (configfs_write_file) from [<c00c8dd4>]
(vfs_write+0xb0/0x1a0)
[   26.685801] [<c00c8dd4>] (vfs_write) from [<c00c91b8>]
(SyS_write+0x44/0x84)
[   26.692834] [<c00c91b8>] (SyS_write) from [<c000e560>]
(ret_fast_syscall+0x0/0x30)
[   26.700381] Code: e30409f8 e34c0069 eb07b88d e59430a8 (e5930000)
[   26.706485] ---[ end trace f62a082b323838a2 ]---

It's because in some cases request is still running on endpoint during
unbind and kill_all_requests() called from s3c_hsotg_udc_stop() function
doesn't cause call of complete() of request. Missing complete() call
causes ecm->notify_req equals NULL in ecm_unbind() function, and this
is reason of this bug.

Similar breaks can be observed in another usb function drivers.

This patch fixes this bug forcing usb request completion in when
s3c_hsotg_ep_disable() is called from s3c_hsotg_udc_stop().
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Since we have assigned clk=NULL, which is a valid clk, we should not
be returning when a clock node is not provide. Instead, we should return
only when we cannot enable the clock.
Signed-off-by: NDinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Building with bcm2835_defconfig, which has CONFIG_PM_SLEEP=n causes the
following build warning:

drivers/usb/dwc2/platform.c:227:12: warning: 'dwc2_suspend' defined but not used [-Wunused-function]
drivers/usb/dwc2/platform.c:237:12: warning: 'dwc2_resume' defined but not used [-Wunused-function]

Annotate these functions with '__maybe_unused' to prevent the warnings.
Reported-by: NOlof's autobuilder <build@lixom.net>
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Suspend/resume code assumed that the gadget was always started and
enabled to connect to usb bus. This means that the actual state of the
gadget (started/stopped or connected/disconnected) was not correctly
preserved on suspend/resume cycle. This patch fixes this issue.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch adds mutex, which protects initialization and
deinitialization procedures against suspend/resume methods. This mutex
will be needed by the updated suspend/resume calls, which tracks gadget
state.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Acked-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>