Fix null-pointer dereference on failure to look up irq due to missing
error handling.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Make sure to verify the length of incoming requests before trying to
parse the request buffer, which can even be NULL on empty requests.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Include the operation id as well as the received and expected size
(from header) when reporting incomplete messages.

This information is useful when debugging communication errors.

Also invert the size test to match the error message and increase
readability.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Use dev_err whenever we have a connection for more informative error
messages.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Fix the payload size of incoming requests, which should not include the
operation message-header size.

When creating requests we pass the sizes of request and response
payloads and greybus core allocates buffers and adds the required
headers. Specifically, the payload sizes do not include the
message-header size.

This is currently not the case for incoming requests however, something
which prevents protocol drivers from implementing appropriate input
verification and could lead to random data being treated as a valid
message in case of a short request.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Incoming operations are created without a response message. If a
protocol driver fails to send a response, or if the operation were to be
cancelled before it has been fully processed, we get a null-pointer
dereference when the operation is released.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Incoming operations are created without a response message. If an
operation were to be cancelled before it has been fully processed (e.g.
on connection destroy), we would get a null-pointer dereference in
gb_operation_cancel.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Make sure to return a proper response in case we get a request we do not
recognise.

This fixes an infinite loop and use-after-free bug, where the freed
operations structure would get re-added to the work queue indefinitely.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Make sure to drop the operation reference when sending the request fails
to avoid leaking the operation structures.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Fix use-after-free when sending responses due to reference imbalance.

Make sure to take a reference to the operation when sending responses.
This reference is dropped in greybus_data_sent when the message has been
sent, while the initial reference is dropped in gb_operation_work after
processing the corresponding request.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Fix up obsolete comments referring to null callback pointers for
synchronous operations, and make sure a callback is always provided when
sending a request.

Also document that the callback is responsible for dropping the initial
(and not necessarily final) reference to the operation.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Fix typo in comment.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Add missing EXPORT_SYMBOL_GPL for gb_operation_response_alloc,
gb_operation_result, gb_operation_get, gb_operation_request_send and
gb_operation_cancel, which are all supposed to be accessible from
protocol handlers.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

And this is the warning I was getting on kernel version > 3.14

  CC [M]  greybus/connection.o
  In file included from
  include/asm-generic/gpio.h:4:0,
	from arch/arm/include/asm/gpio.h:9,
	from include/linux/gpio.h:48,
	from greybus/kernel_ver.h:59,
	from greybus/connection.c:12:
  include/linux/kernel.h:35:0: warning: "U16_MAX" redefined

kernel_ver.h is taking care of defining U16_MAX only if is not defined earlier,
but it is often included as the first .h file. <linux/kernel.h> might be
included later, which always defines it, unconditionally. And so this warning.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This brings the debug log functionality of es1.c to es2.c, along with
some other minor cleanups.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Reviewed-by: Alex Elder <elder@linaro.org>

This removes the error checking for debugfs initialization as we really
don't care if it failed or not.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Reviewed-by: Alex Elder <elder@linaro.org>

These clever macros were fine for early development, but they're
more of a distraction now.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This is an old patch that I neglected to send out.  It's cleaning
up a couple things that got committed before I had a chance to
comment on them.

In operation.c there is a "FIXME" comment that is easily proven
wrong by inspection.

In gb_protocol_put(), there is another wrong "FIXME" comment as
well.  We can also use our cached copies of the protocol major
and minor version number in another spot.  And balance that
out by using a cached copy of the protocol id.
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Merge the es1 log file fixes into master.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

In order to decrement the reference count of module on failures, we must call
put_device(module->dev). This was missing for one of the error cases, fix it.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

As mentioned in kernel coding guidelines.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

identify_descriptor() doesn't return 0 anymore and so we don't need to check the
returned value against 0.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

We are calculating descriptors expected size differently based on the type of
descriptor, that's fine but at few places we aren't taking size of the header
into account. And that looks wrong.

Lets make sure it is atleast as big as descriptor's header.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

If you write 0 to the debugfs file, the log will stop being updated.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

One function shouldn't do two different things depending on a parameter
passed to it, so split usb_log_enable() into usb_log_disable()
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Add a blank line in apb1_log_enable_read() to make checkpatch happy.

Oh, and it makes the code more readable too...
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

No need to duplicate built-in functions that the kernel has, so have the
core kernel parse the userspace string.  Saves us an allocation and
makes the logic simpler.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

decimal is not octal, so use the proper mode settings for the debugfs
files.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

We aren't changing these pointers, so mark them read-only as that is the
preferred way.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

The function can, and even expects NULL, so don't check.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Asm is only for when you are doing arch-specific stuff, which we aren't
doing here.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

No trailing spaces or spaces before tabs are allowed.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

It's "const char __user *buf", not "char __user *buf".

'make check' is your friend.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

On AP module (form factor), we don't have access to APBridge JTAG or UART.
But sometime, we still need to get log from APBridge. Add a new request in control endpoint
to get APBridge logs.
Logs can be accessed using debugfs (greybus/apb1_log).
Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

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

Add error messages on failures to deactivate, set and get operation
handlers as any errors would not be detected by the upper layers (either
because the callbacks are declared void or expected to return a boolean
value).
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Add warning and refuse to set output value for pin configured as input,
as the result of such an operation is undefined.

Remove incorrect todo-comment suggesting that the driver could
implicitly switch direction as part of the call.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Clean up allocation of line-state array.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Use the more informative dev_err and dev_warn for messages, and make the
messages more uniform.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Make sure to propagate any errors detected up the call chain.

This specifically means that we will detect failed connection init,
something which is now handled more gracefully by greybus core.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>