greybus: loopback: functionally decompose calculation of turn-around times

We have a pattern similar to this over and over again gb->elapsed_nsecs =
timeval_to_ns(&te) - timeval_to_ns(&ts); good software practice dictates we
functionally decompose this. This patch decomposes into
gb_loopback_calc_latency().

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: es-drivers: add outbound timestamp to connection

In order to facilitate grabbing a timestamp that doesn't include greybus
overhead, this patch adds a timestamp right before usb_submit_urb() for
both es1.c and es2.c. Long term the timestmaping of messages like this
probably wants to go away but, for the moment it may have some use to the
firmware people instrumenting the performance of the system.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: add a timestamp kfifo to track connection handoff

For the ES2 test activity it may be beneficial to have a performance
metric that doesn't include any of the greybus stack malloc, workqueues
etc. In order to faciltate, this patch adds a simple kfifo structure to
hold two timestamp values. One timestamp will represent the last reasonable
point a greybus outbound timestamp can be taken, the other timestamp will
represent the first reasonable point an inbound timestamp can be taken. In
order to facilitate this model, tracking the timestamps in the connection
structure appears to be the best place to keep store of this data.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: fix jump label on device_add failure

On device_add() failure in gb_connection_create_range() we jump to
err_remove_ida. Instead we should be jumping to err_free_connection, so
change the flow to accomodate.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: protocol: Remove unnecessary params of gb_protocol_get_version()

Some of the parameters are not really required, drop them.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: protocol: Drop define_get_version support

No more users now, drop it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: preserve major/minor of protocol supported by SVC

These weren't preserved earlier, save them in the connection structure
instead of creating its own fields..

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: vibrator: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: uart: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: spi: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: sdio: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: raw: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: pwm: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: i2c: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: hid: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: gpio: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: control: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: battery: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: audio: Drop get_version support

This is done from a common place now, no need to replicate it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: request protocol version before initializing connection

This can (should) be done from a common place as its required for most
of the protocols. Do it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: Save major/minor supported by module

Save major/minor number supported by the module inside connection
structure, as this can be used later.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: define greybus wide protocol request numbers

Some request numbers (like invalid and get_version) are same across all
protocols. Create common macros for them.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: disconnect connection on failures during initialization

Its possible for connection_init() to fail, in such cases the
disconnected event must be sent to the module.

It is missing currently, fix it.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: error out only for smaller payloads received

!= was used in place of <, while comparing expected and actual payload
size. The module may be running a higher version of the protocol and
might have some extra fields (towards the end) in the structure, and the
AP needs to ignore them.

This also updates the print (expected-payload-size <
actual-payload-size), when the size doesn't match for requests received
by the module. This gives more details required for debugging.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: audio: Add '<' to the print message for short messages

This increases readability a bit more, as it tells which value is actual
size and which one is expected size.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: Remove FIXME for firmware download

Its handled by the firmware driver now, drop the FIXME comment from svc
code.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: vibrator: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: uart: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: spi: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: pwm: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: i2c: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: hid: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: gpio: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: battery: Use (already defined) major/minor macros

We already have macros for these, use them instead of writing fixed
values.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: hid: spell fix (s/infomation/information)

Minor spell fix.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: sdio: error out only for smaller payloads received

!= was used in place of <, while comparing expected and actual payload
size. The module may be running a higher version of the protocol and
might have some extra fields (towards the end) in the structure, and the
AP needs to ignore them.

This also updates the print (expected-payload-size <
actual-payload-size), when the size doesn't match for requests received
by the module. This gives more details required for debugging.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: gpio: Print expected/actual payload size on mismatch

Print (expected-payload-size actual-payload-size), when the size doesn't
match for requests received by the module. This gives more details
required for debugging the issue.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: raw: Print expected/actual payload size on mismatch

Print (expected-payload-size actual-payload-size), when the size doesn't
match for requests received by the module. This gives more details
required for debugging the issue.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: kernel_ver.h: define U32_MAX and U64_MAX

These were not defined, and I just posted patches that use them.

Signed-off-by: Alex Elder <elder@linaro.org>
Tested-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: fix connection cleanup paths

The error paths in gb_loopback_connection_init() are kind of screwed
up--not in proper order, and the label naming convention seems a
little inconsistent.

Fix this, ensuring each error cleans up the setup that's been done
up to that point.  Use the convention that the label indicates the
first thing that needs to be cleaned up.

Reorder the statements in gb_loopback_connection_exit() to match
the order of cleanup in gb_loopback_connection_init().

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: compute average stats on demand only

Stop recording and updating the average every time a sample
is recorded.  Instead, compute it from the sum and count only
when it's required.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: use separate attribute macro for average

Define a separate macro for displaying the average of the samples
collected.  This will be used so we can calculate the average only
when requested, rather than every time a new value gets recorded.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: all read-only attributes are unsigned

The only values passed to the gb_loopback_ro_attr() macro are
unsigned 32-bit values.  So there's no need to pass a "type"
format specifier.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: error is an unsigned attribute

The error count is unsigned, so fix the format specifier used in its
attribute definition.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: use a 32-bit count

The count of statistical samples recorded is currently a 64-bit
value.  32 bits is sufficient, and in fact anything more than
that won't work for the do_div() call it's pass to anyway.  So make
the count field be 32 bits.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: record 32-bit min and max

The minimum and maximum values for stats values are always 32 bits.
Change the type for these fields to reflect this.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: use u32 for stats update

The only values supplied to gb_loopback_update_stats() are 32-bits,
so change the type of the second argument to reflect that.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: use U64_MAX for initialization

Use the largest representable value when initializing the "min"
field when resetting loopback statistics.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: drop unneeded casts for void pointers

There is no need to cast a void pointer to a particular type.
Drop the casts used in this way, mainly in the attribute definition
macros.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: looback: fix two typos

Fix two misspellings.  And add spaces around a '%' operator.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: interface: declare gb_interface_destroy()

Add a public declaration for gb_interface_destroy(), matching
gb_interface_create().

It's not yet used outside "interface.c" but I suppose it
could be, and its scope is currently public.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: greybus_protocols: use only type attributes for message packing

For consistency reasons, use only type attributes for message packing.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: sdio: fix message packing

Add missing packed attributes to prevent implicit structure padding.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: uart: fix message packing

Add missing packed attributes to prevent implicit structure padding.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: fix message packing

Add missing packed attributes to prevent implicit structure padding.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: greybus_protocols: fix typo in comment

Fix misspelled variable name in comment.

Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: operation: Move operation header to greybus_protocols.h

This should be exposed to external users (like gbsim). Move it to
greybus_protocols.h.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: operation: Drop alignment attribute from operation message header

The buffers allocated for message header is already 64 bit aligned and
we have explicit pad bytes in the header structure, to 64 bit align the
operation specific data.

And so there is no need to add the aligned attribute to the operation
message header. Drop it.

Suggested-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Alex Elder <elder@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: disable protocol driver

The USB bridged-PHY protocol driver currently depends on changes to USB
core that are not yet upstream.

Disable for now.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: implement hub_control callback

Implement the hub_control callback.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: fix hc_driver fields

Fix hc_driver description, product_desc and flags fields.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: fix hcd allocation, deregistration and deallocation

Fix allocation, deregistration and deallocation of USB HCD, and update
the hcd_priv helper functions.

The HCD private data was not allocated correctly, something which would
lead to a crash when accessed in hcd_start. The HCD was neither
reregistered or deallocated on connection tear down.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: renumber operation types

Renumber the current operation types.

NOTE: Protocol change.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: usb: clean up driver

Remove unused, broken or unneeded code and constructs.

Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: remove special check for svc cport id

This is required to get things working for now, after the latest revert
of svc protocol is done.

Currently svc's cport id is set to 2 and that hd cport id will be used
for the third connection we make. And that protocol (which is i2c in one
of the cases), may not work as the (dis)connected event isn't sent for
it.

Fix this by getting rid of svc protocol check from (dis)connected
events for now. This must be reverted later, once svc protocol is
included again.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Mark Greer <mgreer@animalcreek.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: revert svc changes to keep things working for a while.

The firmware for the svc changes isn't quite ready, so revert the whole
set of patches in one hunk to get things back to a working state for the
other firmware developers.  The svc patches will be added back in a
separate branch.

Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: warn user if kfifo cannot log all data

The depth of the kfifo used to log the latency data for user-space can be
moved upwards or downward by way of a module parameter. The user may still
specify a test set that's larger than the number of kfifo elements we have
available. If the user specifies more iterations than can be logged give a
warning as feedback and continue with the test.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: provide interface to read all latency data-points

The current loopback code provides the minimum, maximum and average latency
values for a given test set. It would be highly useful for user-space to
have access to each one of the latency metrics in order to graph outliers.

This patch adds a simple character device interface implmenting a read()
interface that allows user-space to read out the saved latency metrics
which have been stored in a kfifo for this purpose.

A module parameter is provided to allow varying the depth of the kfifo in
order to allow a user to capture potentially large data sets. This version
sets the default depth for the kfifo at 8192 dwords.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: make loopback code thread safe

Current code allows a sysfs callback and a kernel worker thread to write
all over and act upon data that could be in the process of being updated by
the other. This patch adds a reasonably coarse mutex to enscure sync
between the two.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: remove redundant timestamping

It is of more interest to graphing system performance to base our
timestamps on the time it takes a greybus_operation_sync() to complete.
Higher level timestamping code is less accurate and not relevant to
throughput and latency characterization.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: initialize svc connection while creating hd

Its really part of initializing the host device and is required for
every 'hd' that is created. Lets move the call to do basic
initialization of svc connection to greybus_create_hd().

Also add a comment to specify why we need to do it that early.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: core: clean up ida memory for host controller

We forgot to free any ida internal structures that were used by this
host controller structure when we free the memory for the controller.
So fix that up by doing so in the release function.

Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>

greybus: remove svc_msg.h

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: es1: create svc connection early enough

The svc connection needs to be ready before creating the URBs, otherwise
the svc version request might come in before the AP was ready to parse
them.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: interface: make gb_interface_destroy() static

The function is only called locally, so mark it static to make sparse
happy.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: remove submit_svc from the host driver

The callback is never used anymore, so remove it from struct
greybus_host_driver as well as from the es1 and es2 drivers.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: greybus.h: remove greybus_svc_in()

The function is gone, remove it from the header file as well.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: add flags and traffic class parameter to connection create op

The AP needs to be able to specify L4 CPort flags and traffic class
parameters on a connection-by-connection basis. Extend the connection
create operation to accept these. Since there's no policy to decide
these, fix them at TC0 with end-to-end-flow control, controlled segment
dropping, and CPort safety valve enabled.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: connection: ask SVC to create connections

Ask the SVC to do all the necessary bits for creating a new connection.
This is skipped for the initial SVC connection.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: create bidirectional routes on hotplug request

Upon receiving a hotplug request, we need to prepare the routing table
to allow packets to flow between the AP interface and the newly detected
interface.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: add route create operation

Implement the SVC Protocol Route Create Operation.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: get rid of old svc-protocol

Its not used anymore as we have more sophisticated svc protocol in
place, lets get rid of earlier code.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: es2: remove svc endpoint message handling

We have switched over to use the "new" svc messages, no more need to
have a special USB endpoint to handle them, they come through the normal
CPort messages.

Based on a patch from Viresh.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: es1: remove svc endpoint message handling

We have switched over to use the "new" svc messages, no more need to
have a special USB endpoint to handle them, they come through the normal
CPort messages.

Based on a patch from Viresh.

Tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: Switch to Cport 0 for svc and control protocol

Initially we fixed it to Cport 2, but its changed to Cport 0 now. Lets
switch that in code as well.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: es1/2: Lets start using svc protocol

All bits and pieces are in place now. Lets start using svc protocol
instead of stuff present in ap.c.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: compare hd_cport_id for matching SVC protocol

SVC protocol isn't per interface, but per entire entire endo. And so svc
cport id must be matched against hd_cport_id, not per-interface cport
id.

Fixes: 82edfac17e5f ("connection: don't send connected/disconnected events for SVC connection")
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: connection: silence warning on unbound protocols

If a protocol was not successfully created, we can't drop the refcount
on it. This might happen for example if the connection fails to bind a
protocol.

Silences a warning on cleanup.

Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: battery: fix panic on operation error

If an operation times out or otherwise returns an error, val->intval
should not be set and an error-code should be returned.

Fixes a panic on unload while receiving -ENOTCONN.

Signed-off-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: interface: localize gb_interface_destroy()

Its not referenced by svc or any other code anymore, lets stop exposing
it to rest of the files.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: remove interface->svc pointer

Its not updated/used anymore, remove it. Also move back the struct
gb_svc to svc.c as its not referenced by external users anymore.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: connection is created only once now

We no longer create a fresh connection on receiving svc-hello message,
but rather update the initial one.

Update 'initial_svc_connection' after the connection is fully
initialized. Look for the partially initialized connection while
removing hd, as hd might be removed before getting svc-hello requests
from svc.

Also update gb_svc_connection_init() to initialize id_map on the first
(and the only) call to connection-init function.

We also can't update connection->bundle->intf->svc, as its a bundle-less
connection. Lets stop updating intf->svc as its not really used.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: svc: fully initialize the partially initialized connection

SVC hello message is received now and we should fully initialize the
partially initialized connection. This can be done by removing and
re-adding the device corresponding to the connection.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Tested-by: Perry Hung <perry@leaflabs.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: convert loopback wake/sleep to a waitqueue

Current code will incrementally poll for gb->type == 0 and sleep.
This type of polling strategy wastes cycles.

This patch changes the sleep strategy by introducing a wait-queue which
waits for gb->type != 0 or kthread_should_stop() to wake-up and work or
to wake-up and terminate.

Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: loopback: add poll support to the iteration_count sysfs file

This adds the ability to poll on "iteration_count" in sysfs and be woken
up when it changes, saving some cycles constantly hammering on the file
waiting for it to change.

Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>

greybus: operation: add completion work queue

Add dedicated bound work queue for operation completions and use the
connection work queues for incoming requests only.

There is no need to keep responses ordered internally or with respect to
requests. Instead allow operations to complete as soon as possible when
a response arrives (or the operation is cancelled).

Note that this also allows synchronous requests to be submitted from
request handlers as responses will no longer be blocked on the same
single-threaded work queue. Similarly, operations can now also be
cancelled from a request handler.

Tested-by: Rui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

greybus: operation: use per-connection work queues

Replace the global operation work queue with per-connection work queues.

There is no need to keep operations strictly ordered across connections,
something which only adds unnecessary latency.

Tested-by: Rui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>