According to our Gadget Framework API documentation,
->set_halt() *must* return -EAGAIN if we have pending
transfers (on either direction) or FIFO isn't empty (on
TX endpoints).

Fix this bug so that the mass storage gadget can be used
without stall=0 parameter.

This patch should be backported to all kernels since v3.2.

Cc: <stable@vger.kernel.org> # v3.2+
Suggested-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Instead of releasing the lock and calling locked
versions of our set_halt() methods, let's hold
the lock all the way through and call unlocked
versions of those functions.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

__dwc3_gadget_ep_set_halt() is the function which
handles the actual halt feature. In order to cope
with some extra cleanup comming as a follow-up patch
let's move the isochronous endpoint check there too.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Use the recently introduced usb_gadget_giveback_request() in favor of
direct invocation of the completion routine.

All places in drivers/usb/ matching "[-.]complete(" were replaced with a
call to usb_gadget_giveback_request(). This was compile-tested with all
ARM drivers enabled and runtime-tested for musb.
Signed-off-by: NMichal Sojka <sojka@merica.cz>
Acked-by: NFelipe Balbi <balbi@ti.com>
Tested-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When we're debugging hard-to-reproduce and time-sensitive
use cases, printk() poses too much overhead. That's when
the kernel's tracing infrastructure comes into play.

This patch implements a few initial tracepoints for the
dwc3 driver. More traces can be added as necessary in order
to ease the task of debugging dwc3.
Reviewed-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

No functional changes, just making sure we're dealing
with unsigned ints.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Those functions are only using within debugging
messages, grouping them into debug.h makes sense.

While at that, also add missing multiple inclusion
guard.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

After commit 2ec2a8be (usb: dwc3: gadget:
always enable IOC on bulk/interrupt transfers)
we created a situation where it was possible to
hang a bulk/interrupt endpoint if we had more
than one pending request in our queue and they
were both started with a single Start Transfer
command.

The problems triggers because we had not enabled
Transfer In Progress event for those endpoints
and we were not able to process early giveback
of requests completed without LST bit set.

Fix the problem by finally enabling Xfer In Progress
event for all endpoint types, except control.

Fixes: 2ec2a8be (usb: dwc3: gadget: always
	enable IOC on bulk/interrupt transfers)
Cc: <stable@vger.kernel.org> # v3.14+
Reported-by: NPratyush Anand <pratyush.anand@st.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

dep->endpoint.desc is checked at the beginning of
dwc3_gadget_ep_queue(), but after that it may be set to NULL
by another thread and then accessed again in dwc3_gadget_ep_queue().
This will lead to kernel oops.

Expand spinlock protection area to aviod race condition.
Signed-off-by: NZhuang Jin Can <jin.can.zhuang@intel.com>
Signed-off-by: NJiebing Li <jiebing.li@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The site-specific OOM messages are unnecessary, because they
duplicate the MM subsystem generic OOM message.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Remove 'start_new' variable from dwc3_endpoint_transfer_complete(),
since this variable has not been used.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In ISOC transfers, when free_slot points to the last TRB (i.e. Link
TRB), and all queued requests meet Missed Interval Isoc error, busy_slot
points to trb0.
	busy_slot->trb0
		   trb1
		   ...
	free_slot->trb31(Link TRB)

After end transfer and receiving the XferNotReady event, trb_left is
caculated as 1 which is wrong, and no TRB will be primed to the
endpoint.

The root cause is free_slot is not increased the same way as busy_slot.
When busy_slot is increased by one, it checks if points to a link TRB
after increasement, but free_slot checks it before increasement.
free_slot should behave the same as busy_slot to make the trb_left
caculation correct.
Reviewed-by: NPratyush Anand <pratyush.anand@st.com>
Signed-off-by: NZhuang Jin Can <jin.can.zhuang@intel.com>
Signed-off-by: NJiebing Li <jiebing.li@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

First usage of ret variable will re-write initial value. Thus, there is no need
to initialize it.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

this makes it slightly easier to read generic CMD
logs. It also helps make sure we're sending proper
parameters for each command.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

this makes it slightly easier to read link state
change interrupt logs.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

That way it's easy for humans looking at dmesg
and humans(?) looking at Databooks.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

so it seems like DWC3 IP doesn't clear stalls
automatically when we disable an endpoint, because
of that, we _must_ make sure stalls are cleared
before clearing the proper bit in DALEPENA register.

Cc: <stable@vger.kernel.org> # v3.4+
Reported-by: NJohannes Stezenbach <js@sig21.net>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Make dwc3_gadget_resize_tx_fifos() iterate only over IN
endpoints that are actually present, based on the
num_in_eps parameter. This terminates the loop so as to
prevent dereferencing a potential NULL dwc->eps[i] where
i >= (num_in_eps + num_out_eps).
Signed-off-by: NJack Pham <jackp@codeaurora.org>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

We know what "value" is and it upsets static checkers that we appear to
have doubts about it.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

These were cut and paste from the ->disconnect function.

Fixes commit 30d577b9bcc4 ('usb: dwc3: gadget: call gadget driver's
->suspend/->resume')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

we need to pre-start stream transfers otherwise we
will never know when to start them.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When going into bus suspend/resume we _must_
call gadget driver's ->suspend/->resume callbacks
accordingly. This patch implements that very feature
which has been missing forever.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

It's not always we need to force a transfer to be removed
from the core's internal cache. This extra argument will
help differentiating those two cases.
Signed-off-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

During superspeed, HIRD threshold should always
be zero. Curent driver wasn't making sure that
was the case.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

if we have hibernation configured, Databook
instructs us to set KEEP_CONNECT bit together
with RUN_STOP bit, in step 9 of section 12.3.6.1
Initialization for Hibernation Support.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

That argument will be used in later patches when we
have working hibernation support. For now, always
pass it as false.

The idea of this patch is to decrease to size of
following patches and slowly add hibernation building
blocks to the gadget side of dwc3 so that it becomes
very easy to review the actual hibernation code.

[ balbi@ti.com : rewrote patch on top of current
	tree. Added commit log. ]
Signed-off-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This function will be used during hibernation to get
the current link state. It will be needed at least
for Hibernation support.
Signed-off-by: NPaul Zimmerman <paulz@synopsys.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Revision 2.20a of the core has a known issue
which would generate bogus hibernation events
_and_ random failures on USB CV TD.9.23 test
case.

The suggested workaround is to ignore hibernation
events which don't match currently connected
speed.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

by setting IOC always, we can recycle TRBs a
lot sooner at the expense of some increased
CPU load.

The extra load seems to be quite minimal on
OMAP5 devices (instead of 1 IRQ for one MSC
transfer, we get
CONFIG_USB_GADGET_STORAGE_NUM_BUFFERS).
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch adds "maxpacket_limit" to struct usb_ep. This field contains
maximum value of maxpacket supported by driver, and is set in driver probe.
This value should be used by autoconfig() function, because value of field
"maxpacket" is set to value from endpoint descriptor when endpoint becomes
enabled. So when autoconfig() function will be called again for this endpoint,
"maxpacket" value will contain wMaxPacketSize from descriptior instead of
maximum packet size for this endpoint.

For this reason this patch adds new field "maxpacket_limit" which contains
value of maximum packet size (which defines maximum endpoint capabilities).
This value is used in ep_matches() function used by autoconfig().

Value of "maxpacket_limit" should be set in UDC driver probe function, using
usb_ep_set_maxpacket_limit() function, defined in gadget.h. This function
set choosen value to both "maxpacket_limit" and "maxpacket" fields.

This patch modifies UDC drivers by adding support for maxpacket_limit.
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

DWC3 requires epout to have buffer size aligned to MaxPacketSize value.
This patch sets necessary quirk for it.
Signed-off-by: NDavid Cohen <david.a.cohen@linux.intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The dwc3 UDC driver doesn't implement endpoint wedging correctly.
When an endpoint is wedged, the gadget driver should be allowed to
clear the wedge by calling usb_ep_clear_halt().  Only the host is
prevented from resetting the endpoint.

This patch fixes the implementation.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Tested-by: NPratyush Anand <pratyush.anand@st.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

If dwc3_gadget_init_endpoint() fails after allocate some of the eps, we
need to free their memory to avoid leak.
Signed-off-by: NDavid Cohen <david.a.cohen@linux.intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

in order to make our IRQ handler thread easier
to read, we re-factor the inner loop to a separate
function.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

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

Due to that reason, we're reaning it to dwc3_check_event_buf()
instead.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

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

With that feature, we can easily drop IRQF_ONESHOT
from our driver which makes it possible to
properly change IRQ priorities when using RT
patchset *and* it allows us to make use of the
scheduler to choose the proper time to handle
this IRQ thread.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

LPM is enabled in Connection Done interrupt,
there's no need to enable it early.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

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

It will be no-op otherwise.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

no functional changes, just grouping dep
initialization.
Signed-off-by: NFelipe Balbi <balbi@ti.com>

by moving that dev_vdbg() to the internal
__dwc3_gadget_ep_enable() we get the print
even when enable ep0, which calls the internal
function directly.
Signed-off-by: NFelipe Balbi <balbi@ti.com>