Now that we modified the code to fix a race condition, it's clear that
DWC3_EP_END_TRANSFER_PENDING is unnecessary, considering that
DWC3_EP_TRANSFER_STARTED will remain set until End Transfer completes.
Tested-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

There is a rare race condition that may happen during a Disconnect
Interrupt if we have a started request that happens to be
dequeued *after* completion of End Transfer command. If that happens,
that request will be left waiting for completion of an End Transfer
command that will never happen.

If End Transfer command has already completed before, we are safe to
giveback the request straight away.
Tested-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We have a proper flag for testing that we have a valid transfer in
flight, let's use that instead.

This patch is in preparation to fix a rare race condition that happens
upon Disconnect Interrupt.
Tested-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We must wait until End Transfer completes in order to clear
DWC3_EP_TRANSFER_STARTED, otherwise we may confuse the driver.

This patch is in preparation to fix a rare race condition that happens
upon Disconnect Interrupt.
Tested-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Now that we have req->status, we don't need this extra flag
anymore. It's safe to remove it.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Queueing the same request twice can introduce hard-to-debug
problems. At least one function driver - Android's f_mtp.c - is known
to cause this problem.

While that function is out-of-tree, this is a problem that's easy
enough to avoid.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

This patch starts tracking dwc3_request status. A following patch will
build on top of this to prevent a request from being queued twice.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

On v3.10a in dual-role mode, if port is in device mode
and gadget driver isn't loaded, the OTG event interrupts don't
come through.

It seems that if the core is configured to be OTG2.0 only,
then we can't leave the DCFG.DEVSPD at Super-speed (default)
if we expect OTG to work properly. It must be set to High-speed.

Fix this issue by configuring DCFG.DEVSPD to the supported
maximum speed at gadget init. Device tree still needs to provide
correct supported maximum speed for this to work.

This issue wasn't present on v2.40a but is seen on v3.10a.
It doesn't cause any side effects on v2.40a.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NSekhar Nori <nsekhar@ti.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Currently the link_state is uninitialized and the default value is 0(U0)
before the first time we start the udc, and after we start the udc then
 stop the udc, the link_state will be undefined.
We may have the following warnings if we start the udc again with
an undefined link_state:

WARNING: CPU: 0 PID: 327 at drivers/usb/dwc3/gadget.c:294 dwc3_send_gadget_ep_cmd+0x304/0x308
dwc3 100e0000.hidwc3_0: wakeup failed --> -22
[...]
Call Trace:
[<c010f270>] (unwind_backtrace) from [<c010b3d8>] (show_stack+0x10/0x14)
[<c010b3d8>] (show_stack) from [<c034a4dc>] (dump_stack+0x84/0x98)
[<c034a4dc>] (dump_stack) from [<c0118000>] (__warn+0xe8/0x100)
[<c0118000>] (__warn) from [<c0118050>](warn_slowpath_fmt+0x38/0x48)
[<c0118050>] (warn_slowpath_fmt) from [<c0442ec0>](dwc3_send_gadget_ep_cmd+0x304/0x308)
[<c0442ec0>] (dwc3_send_gadget_ep_cmd) from [<c0445e68>](dwc3_ep0_start_trans+0x48/0xf4)
[<c0445e68>] (dwc3_ep0_start_trans) from [<c0446750>](dwc3_ep0_out_start+0x64/0x80)
[<c0446750>] (dwc3_ep0_out_start) from [<c04451c0>](__dwc3_gadget_start+0x1e0/0x278)
[<c04451c0>] (__dwc3_gadget_start) from [<c04452e0>](dwc3_gadget_start+0x88/0x10c)
[<c04452e0>] (dwc3_gadget_start) from [<c045ee54>](udc_bind_to_driver+0x88/0xbc)
[<c045ee54>] (udc_bind_to_driver) from [<c045f29c>](usb_gadget_probe_driver+0xf8/0x140)
[<c045f29c>] (usb_gadget_probe_driver) from [<bf005424>](gadget_dev_desc_UDC_store+0xac/0xc4 [libcomposite])
[<bf005424>] (gadget_dev_desc_UDC_store [libcomposite]) from[<c023d8e0>] (configfs_write_file+0xd4/0x160)
[<c023d8e0>] (configfs_write_file) from [<c01d51e8>] (__vfs_write+0x1c/0x114)
[<c01d51e8>] (__vfs_write) from [<c01d5ff4>] (vfs_write+0xa4/0x168)
[<c01d5ff4>] (vfs_write) from [<c01d6d40>] (SyS_write+0x3c/0x90)
[<c01d6d40>] (SyS_write) from [<c0107400>] (ret_fast_syscall+0x0/0x3c)
Signed-off-by: NZeng Tao <prime.zeng@hisilicon.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

OUT endpoint requests may somtimes have this flag set when
preparing to be submitted to HW indicating that there is an
additional TRB chained to the request for alignment purposes.
If that request is removed before the controller can execute the
transfer (e.g. ep_dequeue/ep_disable), the request will not go
through the dwc3_gadget_ep_cleanup_completed_request() handler
and will not have its needs_extra_trb flag cleared when
dwc3_gadget_giveback() is called.  This same request could be
later requeued for a new transfer that does not require an
extra TRB and if it is successfully completed, the cleanup
and TRB reclamation will incorrectly process the additional TRB
which belongs to the next request, and incorrectly advances the
TRB dequeue pointer, thereby messing up calculation of the next
requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flag is cleared
before it is given back to the function driver.  A good place
to do that is in dwc3_gadget_del_and_unmap_request().

Fixes: c6267a51 ("usb: dwc3: gadget: align transfers to wMaxPacketSize")
Cc: stable@vger.kernel.org
Signed-off-by: NJack Pham <jackp@codeaurora.org>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We see dwc3 endpoint stopped by unwanted irq during
suspend resume test, which is caused dwc3 ep can't be started
with error "No Resource".

Here, add synchronize_irq before suspend to sync the
pending IRQ handlers complete.
Signed-off-by: NBo He <bo.he@intel.com>
Signed-off-by: NYu Wang <yu.y.wang@intel.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

In stream mode, when fast-forwarding TRBs, the stream number
is not cleared causing the new stream to not get assigned. So
we don't want controller to carry on transfers when short packet
is received. So disable the CSP for stream capable endpoint.

This is based on the 3.30a Programming guide, where table 3-1
device descriptor structure field definitions says for CSP bit
If this bit is 0, the controller generates an XferComplete event
and remove the stream. So if we keep CSP as 1 then switching between
streams would not happen as in stream mode, when fast-forwarding
TRBs, the stream number is not cleared causing the new stream to not get
assigned.
Signed-off-by: NTejas Joglekar <joglekar@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Implement the new frame_number API to report the isochronous interval
frame number. This patch checks and reports the interval in which the
isoc transfer was transmitted or received via the Isoc-First TRB SOF
number field.
Signed-off-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Availability of TRB's is calculated using dwc3_calc_trbs_left(), which
determines total available TRB's based on the HWO bit set in a TRB.

In the present code, __dwc3_prepare_one_trb() is called with a TRB which
needs to be prepared for transfer. This __dwc3_prepare_one_trb() calls
dwc3_calc_trbs_left() to determine total available TRBs and set IOC bit
if the total available TRBs are zero. Since the present working TRB (which
is passed as an argument to __dwc3_prepare_one_trb() )  doesn't yet have
the HWO bit set before calling dwc3_calc_trbs_left(), there are chances
that dwc3_calc_trbs_left() wrongly calculates this present working TRB
as free(since the HWO bit is not yet set) and returns the total available
TRBs as greater than zero (including the present working TRB). This could
be a problem.

This patch corrects the above mentioned problem in __dwc3_prepare_one_trb()
by increementing the dep->trb_enqueue at the last (after preparing the TRB)
instead of increementing at the start and setting the IOC bit only if the
total available TRBs returned by dwc3_calc_trbs_left() is 1 . Since we are
increementing the dep->trb_enqueue at the last, the present working TRB is
also considered as available by dwc3_calc_trbs_left() and non zero value is
returned . So, according to the modified logic, when the total available
TRBs is equal to 1 that means the total available TRBs in the pool are 0.
Signed-off-by: NAnurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
Reviewed-by: NThinh Nguyen <thinhn@synopsys.com>
Tested-by: NTejas Joglekar <tejas.joglekar@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

The stream capable endpoints require stream id to be given
when issuing START TRANSFER. While issuing no-op trb the
stream id is not yet known, so don't issue no-op trb's on
stream capable endpoints.
Signed-off-by: NAnurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

For stream capable endpoints, stream id related information
needs to be updated into DEPCMD while issuing START TRANSFER.
This patch does the same.
Signed-off-by: NAnurag Kumar Vulisha <anurag.kumar.vulisha@xilinx.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Gadget driver may take an unbounded amount of time to queue requests
after XferNotReady. This is important for isochronous endpoints which
need to be started for a specific (micro-)frame.

If we fail to start a transfer for isochronous endpoint, let's try
queueing to a future interval and see if that helps. We will stop trying
if we fail a start transfer for 5 intervals in the future.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Running out of requests on isochronous endpoints is part of normal
operation. We don't really need to know about it every time it
happens.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Sometimes, errors happen when kicking transfers from
__dwc3_gadget_start_isoc(). In those cases, we need to pass along the
error so gadget driver can make informed decisions.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Now that we have a list of cancelled requests, we can skip over TRBs
when END_TRANSFER command completes.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Whenever we have a request in flight, we can move it to the cancelled
list and later simply iterate over that list and skip over any TRBs we
find.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

This list will host cancelled requests who still have TRBs being
processed.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Extract the logic for skipping over TRBs to its own function. This
makes the code slightly more readable and makes it easier to move this
call to its final resting place as a following patch.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Now that we track how many TRBs a request uses, it's easier to skip
over them in case of a call to usb_ep_dequeue(). Let's do so and
simplify the code a bit.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

This will help us remove the wait_event() from our ->dequeue().
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Both flags are used for the same purpose in dwc3: appending an extra
TRB at the end to deal with controller requirements. By combining both
flags into one, we make it clear that the situation is the same and
that they should be treated equally.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

In DWC_usb31 version 1.70a-ea06 and prior, for highspeed and fullspeed
isochronous IN, BIT[15:14] of the 16-bit microframe number reported by
the XferNotReady event are invalid. The driver uses this number to
schedule the isochronous transfer and passes it to the START TRANSFER
command. Because this number is invalid, the command may fail. If
BIT[15:14] matches the internal 16-bit microframe, the START TRANSFER
command will pass and the transfer will start at the scheduled time, if
it is off by 1, the command will still pass, but the transfer will start
2 seconds in the future. For all other conditions, the START TRANSFER
command will fail with bus-expiry.

In order to workaround this issue, we can test for the correct
combination of BIT[15:14] by sending START TRANSFER commands with
different values of BIT[15:14]: 'b00, 'b01, 'b10, and 'b11. Each
combination is 2^14 uframe apart (or 2 seconds). 4 seconds into the
future will result in a bus-expiry status. As the result, within the 4
possible combinations for BIT[15:14], there will be 2 successful and 2
failure START COMMAND status. One of the 2 successful command status
will result in a 2-second delay start. The smaller BIT[15:14] value is
the correct combination.

Since there are only 4 outcomes and the results are ordered, we can
simply test 2 START TRANSFER commands with BIT[15:14] combinations 'b00
and 'b01 to deduce the smaller successful combination.

Let test0 = test status for combination 'b00 and test1 = test status for
'b01 of BIT[15:14]. The correct combination is as follow:

if test0 fails and test1 passes, BIT[15:14] is 'b01
if test0 fails and test1 fails, BIT[15:14] is 'b10
if test0 passes and test1 fails, BIT[15:14] is 'b11
if test0 passes and test1 passes, BIT[15:14] is 'b00

Synopsys STAR 9001202023: Wrong microframe number for isochronous IN
endpoints.
Signed-off-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

This reverts commit ffb80fc6.

Turns out that commit is wrong. Host controllers are allowed to use
Clear Feature HALT as means to sync data toggle between host and
periperal.

Cc: <stable@vger.kernel.org>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

When chaining ISOC TRBs together, only the first ISOC TRB should be of
type ISOC_FIRST, all others should be of type ISOC. This patch fixes
that.

Fixes: c6267a51 ("usb: dwc3: gadget: align transfers to wMaxPacketSize")
Cc: <stable@vger.kernel.org> # v4.11+
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Current check for the last extra TRB for zero and unaligned transfers
does not account for isoc OUT. The last TRB of the Buffer Descriptor for
isoc OUT transfers will be retired with HWO=0. As a result, we won't
return early. The req->remaining will be updated to include the BUFSIZ
count of the extra TRB, and the actual number of transferred bytes
calculation will be wrong.

To fix this, check whether it's a short or zero packet and the last TRB
chain bit to return early.

Fixes: c6267a51 ("usb: dwc3: gadget: align transfers to wMaxPacketSize")
Cc: <stable@vger.kernel.org>
Signed-off-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

When operating in USB 2.0 speeds (HS/FS), if GUSB2PHYCFG.ENBLSLPM or
GUSB2PHYCFG.SUSPHY is set, it must be cleared before issuing an endpoint
command.

Current implementation only save and restore GUSB2PHYCFG.SUSPHY
configuration. We must save and clear both GUSB2PHYCFG.ENBLSLPM and
GUSB2PHYCFG.SUSPHY settings. Restore them after the command is
completed.

DWC_usb3 3.30a and DWC_usb31 1.90a programming guide section 3.2.2
Signed-off-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Fix kernel-doc warning:

../drivers/usb/dwc3/gadget.c:510: warning: Excess function parameter 'dwc' description in 'dwc3_gadget_start_config'
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Cc: Felipe Balbi <balbi@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

endpoint->maxpacket is not updated after setting the
usb_set_maxpacket_limit() on endpoint enable. The MaxPacketSize can be
different than the endpoint->maxpacket_limit.

DWC3 has been consistently using MaxPacketSize from the endpoint's
descriptor, so let's keep it consistent and use the MaxPacketSize from
the endpoint's descriptor instead.
Signed-off-by: NThinh Nguyen <thinhn@synopsys.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

dwc3_ep_dequeue() waits for completion of End Transfer command using
wait_event_lock_irq(), which will release the dwc3->lock while waiting
and reacquire after completion. This allows a potential race condition
with ep_disable() which also removes all requests from started_list
and pending_list.

The check for NULL r->trb should catch this but currently it exits to
the wrong 'out1' label which calls dwc3_gadget_giveback(). Since its
list entry was already removed, if CONFIG_DEBUG_LIST is enabled a
'list_del corruption' bug is thrown since its next/prev pointers are
already LIST_POISON1/2. If r->trb is NULL it should simply exit to
'out0'.

Fixes: cf3113d8 ("usb: dwc3: gadget: properly increment dequeue pointer on ep_dequeue")
Cc: stable@vger.kernel.org # v4.12+
Signed-off-by: NMayank Rana <mrana@codeaurora.org>
Signed-off-by: NJack Pham <jackp@codeaurora.org>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Fixes the following sparse warning:

drivers/usb/dwc3/gadget.c:169:6: warning:
 symbol 'dwc3_gadget_del_and_unmap_request' was not declared. Should it be static?
Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We dont' need to touch req->direction or req->epnum from
ep_queue(). It's enough that we initialize both fields from
alloc_request() and just keep them for the entire lifetime of the
request.

No functional changes.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Instead of *always* calling dwc3_gadget_ep_get_transfer_index() after
sending a Start Transfer command, we can call it once from
dwc3_send_gadget_ep_cmd() itself.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Instead of returning resource index number just to assign it to a
field inside 'dep' which was passed as argument, we can assing
dep->resource_index from inside dwc3_gadget_ep_get_transfer_index()
itself.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We will *always* start transfer to the next uFrame number aligned to
dep->interval.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

In case we have many started requests and one of them in the middle is
completed with Missed Isoc, let's not End Transfer as that would
result in us loosing (possibly) many more intervals.

Instead, let's allow the controller to go through its list of started
requests.
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>