Johan Hovold [Tue, 17 Mar 2015 09:55:52 +0000 (10:55 +0100)]
greybus: connection: fix oops after failed init
Make sure not to call connection_exit for connections that have never
been initialised (e.g. due to failure to init).
This fixes oopses due to null-dereferences and use-after-free in
connection_exit callbacks (e.g. trying to remove a gpio-chip that has
never been added) when the bundle and interface are ultimately
destroyed.
Johan Hovold [Mon, 2 Mar 2015 08:55:26 +0000 (09:55 +0100)]
greybus: connection: fix locking in gb_hd_connection_find
Fix unconditional re-enabling of interrupts in gb_hd_connection_find,
which can be called with local interrupts disabled from the USB
completion handler.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com> Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
When it receive an interrupt, the function gb_gpio_request_recv doesn't
use the good gpio number to get the irq number. Then, the expected irq is never fired.
Johan Hovold [Mon, 2 Mar 2015 11:34:40 +0000 (12:34 +0100)]
greybus: operation: fix locking issues
Fix unconditional re-enabling of interrupts in various operation
functions that can all be called with local interrupts disabled from USB
completion handlers.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Johan Hovold [Fri, 13 Feb 2015 06:58:04 +0000 (14:58 +0800)]
greybus: gpio: fix memory leaks at init and exit
Fix three related memory leaks in the init an exit callbacks, where the
gpio-lines array was never freed at all and the controller data wasn't
freed in the init error path.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Matt Porter [Tue, 17 Feb 2015 15:48:23 +0000 (10:48 -0500)]
greybus: gpio: add interrupt handling support
Adds gpio interrupt handling support using an irqchip/irqdomain
instantiation inside the GB GPIO driver. This implementation works
on older kernels such as 3.10 that do not have the gpiolib irqchip
helpers. Any line on a Greybus gpiochip may be configured as an
interrupt. Once configured, IRQ event messages received from a
module fire off the registered interrupt handler.
Signed-off-by: Matt Porter <mporter@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Johan Hovold [Fri, 13 Feb 2015 03:28:09 +0000 (11:28 +0800)]
greybus: bundle: fix sleep-while-atomic in gb_bundle_destroy
Make sure to release the spin lock protecting the interface bundle lists
before tearing down the connections and removing the bundle device,
which are operations that may sleep.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com> Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: es1.c: wait until the last possible minute to start the svc messages
When initializing the USB device, we were starting up the svc message
queue before the cport urbs were allocated. This might not be an issue
for "slower" machines, but not having any allocated urbs for a cport
might be an issue if we were to handle svc messages.
So wait until everything is properly initialized and allocated before
starting the svc urb.
SVC messages come in in an "order", so don't mess them up by processing
them out of order. Fix this by making our work queue ordered, which
should keep everything in line.
Reported-by: Perry Hung <perry@leaflabs.com> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Viresh Kumar [Thu, 22 Jan 2015 06:40:38 +0000 (12:10 +0530)]
greybus: i2c: fix name conflict between function and struct: gb_i2c_transfer_request
'gb_i2c_transfer_request' is the name given to a function and a struct. Though
we don't get any compilation errors/warnings about it, but the names should be
unique.
Viresh Kumar [Thu, 22 Jan 2015 07:12:39 +0000 (12:42 +0530)]
greybus: i2c: fix name conflict between function and struct: gb_i2c_transfer_response
'gb_i2c_transfer_response' is the name given to a function and a struct. Though
we don't get any compilation errors/warnings about it, but the names should be
unique.
Alexandre Bailon [Thu, 22 Jan 2015 07:23:37 +0000 (15:23 +0800)]
greybus: protocol.c: fix a kernel panic caused by __gb_protocol_register
__gb_protocol_register check if the protocol is not already registred,
and then register it. It register in existing->lists but at this point,
existing is always NULL (we exist just before if not).
Use gb_protocols instead.
Viresh Kumar [Wed, 21 Jan 2015 10:40:41 +0000 (16:10 +0530)]
greybus: Remove "gb-" prefix from .c files
Some files are still prefixed with "gb-" with the reasoning that the modules
would be named so, i.e. gb-*.ko. But this can be done by playing a bit in
Makefile instead and keep uniform naming of .c files.
Viresh Kumar [Wed, 21 Jan 2015 10:40:40 +0000 (16:10 +0530)]
greybus: Remove "-gb" suffix from .c files
Some files are prefixed with "gb-" and some are suffixed with "-gb". The
rationale behind the first one is that the modules would be named so, i.e.
gb-*.ko. But there is no reason to keep the "-gb" suffix in the second case.
Rui Miguel Silva [Tue, 20 Jan 2015 16:38:44 +0000 (16:38 +0000)]
greybus: es1: release urb on error path
if error is return when submiting the urb, we need to make sure to release the
urb from the pool, or from the dinamicly allocated. As in it, factor out the free
code and create the free_urb function.
Signed-off-by: Rui Miguel Silva <rmfrfs@gmail.com> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Perry Hung [Wed, 14 Jan 2015 21:19:26 +0000 (16:19 -0500)]
greybus: gb_operation: replace timeout workqueue
If an operation is issued and the response never comes back,
gb_operation_timeout() cancels the operation but never wakes up the
waiter in gb_operation_request_send().
This patch removes the timeout workqueue and changes the request wait to
wait_for_completion_interruptible_timeout(), with timeout set to
OPERATION_TIMEOUT_DEFAULT.
Signed-off-by: Perry Hung <perry@leaflabs.com> Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: Move the es1_ap_desc.c file to Documentation directory
This .c file isn't needed by the kernel driver, it's there for firmware
developers only, so just move it into the Documentation directory to
reduce confusion.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
Alexandre Bailon [Wed, 14 Jan 2015 09:23:49 +0000 (10:23 +0100)]
greybus: i2c-gb: fix bad message size in gb_i2c
The data_in_size variable was set to 1 for the status byte.
But now, the status byte has move to header. Then, the status byte
is "allocated" twice and cause bad message size error.
greybus: sysfs: put a \n at the end of all sysfs files
Right now some sysfs attributes have \n and some do not, so fix that and
put \n at the end of all of them to make it easier to parse things
properly in userspace.
We want to be able to "blame" a protocol for things at times, so give
them a name we can refer to them by. Announce when they are added or
removed from the system so we have a chance to know what is going on
in the kernel logs.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: module: get rid of global list of modules
Use the list that the driver core keeps of our structure, no need to
duplicate it with a local list as well. This gets rid of a static lock
too, always a nice thing to do.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
This bundles together the existing GP Bridged PHY protocols that were
part of the Greybus core: USB, UART, SDIO, PWM, and GPIO. This is now a
stand-alone kernel module. More logic will be moving here in the future
to handle bridged devices.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: vibrator-gb: move vibrator protocol to a stand-alone module.
We can't use the gb_protocol_driver() macro here as we need to do some
init and exit logic when loading and removing, so "open code" the module
init and exit functions.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: protocol: add a module owner to a protocol
Now that protocols can be in a module, we need to reference count them
to lock them into memory so they can't be removed while in use. So add
a module owner structure, and have it automatically be assigned when
registering the protocol.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: i2c-gb: split out into a stand-alone kernel module.
This splits the i2c-gb protocol into a stand-alone kernel module.
It's not going to stay in this fashion for long, this was done to test
the "can a protcol be loaded later" logic. Future refactoring is going
to move the gpbridge protocols to a separate kernel module, where this
protocol is going to live.
But for now, split it out, it is good to test with, and shows a bug in
gbsim at the moment.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: protocol: split binding of prototcols to connections out of init
When adding a new protocol to the system, walk all bundles and try to
hook up any connections that do not have a protocol already. This sets
the stage to allow for protocols to be loaded at any time, not just
before the device is seen in the system.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: protocol: switch gb_protocol_register() to return an int
We will want to return this value as a return value for module_init()
and bool does not play well with module_init(). So make it a "real"
error value and return int and fix up all callers of the function.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: interface: remove global manifest_descs list
The list was global and had no locking. It's not like we were ever
parsing more than one manifest at the same time right now, but we might
in the future. And we really want this to be local to the interface
itself, for future work redoing how to bind protocols to bundles, so
move the list to the interface structure.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Modules in the greybus system sit above the interface, so insert them
early in the sysfs tree. We dynamically create them when we have an
interface that references a module, as we don't get a "module create"
message directly. They also dynamically go away when the last interface
associated with a module is removed.
Naming scheme for modules/interfaces/bundles/connections is bumped up by
one ':', and now looks like the following:
We still have some "confusion" about interface ids and module ids, which
will be cleaned up later when the svc control protocol changes die down,
right now we just name a module after the interface as we don't have any
modules that have multiple interfaces in our systems.
This is really a list of interfaces, not modules, so rename it so that
we don't get confused when we really do add modules to the whole system
later on.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
It's a lot of renaming, some structures got renamed and also some
fields, but the goal was to rename things to make sense with the new
naming of how the system is put together in the 'driver model' view.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: es1-ap-usb: don't protest when getting -EPROTO USB errors
-EPROTO happens when devices are starting to go away in a system, or
there is something wrong on the USB connection. Either way, it's safe
to resubmit the urb for this error, don't complain to userspace about
this, as the user will see this for every device removed, which looks
scary, but means nothing.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: bundle: rename interface.[c|h] to bundle.[c|h]
We are renameing the "interface" term to "bundle" so rename the files
before we start changing structure names to make it easier for people to
see what really is happening in the changes.
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com> Reviewed-by: Alex Elder <elder@linaro.org>
greybus: driver matching: Greybus drivers bind to interface blocks, not modules
Because of this, rename greybus_module_id to greybus_interface_block_id.
We still need to add a way for a "class" driver to be bound to an
interface, but for now, all we really need is the vendor/product pair as
the GP Bridge interface block is going to be our main user.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: interface_block: move sysfs files into the interface_block.c file
No need to keep these out in sysfs.c, move them into the
interface_block.c file so that we can see them easier, and remove some
variable definitions by taking advantage of the attribute group macro.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: interface_block: rename the sysfs files to not have 'module' in them
The sysfs files for an interface block should not have 'module' in them.
This was a hold-over from when we thought we were going to have
all attributes of a "module" in one directory. Remove the prefix as
it's not needed, and is confusing considering modules can not have
strings or any of these attributes.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Rename struct gb_module to struct gb_interface_block
It's a complex rename, some functions got their name changed where
needed, but primarily this change is focused on the structure and where
it is used. Future changes will clean up the remaining usages of the
term "module" in individual changes, this one spanned the whole
subsystem so do it all at once.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
greybus: interface_block: rename module.[c|h] to interface_block.[c|h]
"modules" in the driver model here, are really "interface blocks" as
that is what they are physically tied to. So rename the files before we
start changing the code to make it obvious what is going on.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
We removed the module version from the spec, so remove them from the
code as well. It's still in the manifest as we need to sync with gbsim
/ firmware when we do that, which will happen sometime in the next
weeks.
Reviewed-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Thu, 11 Dec 2014 22:48:38 +0000 (16:48 -0600)]
greybus: switch cport id used for sends
In talking with Perry today I learned that the CPort id expected to
supplied over the HSIC interface to the APB is different from the
way I understood it.
My understanding was that the CPort id to supply always specified
the CPort id on the other end of a connection. However, Perry says
the mapping between local CPort id and remote CPort id (and device
id) is done by the host UniPro interface.
So whether sending or receiving data, the CPort id that the Greybus
code should supply to the AP Bridge is the one representing the AP
side of a connection.
This patch fixes this. The receive side already used that CPort id;
it's only the sending code that needed to be changed.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 10 Dec 2014 20:50:48 +0000 (14:50 -0600)]
greybus: ENODEV can be an expected error too
When probing for i2c devices, a read transfer operation can be used.
In this case, it is expected that some devices will not be found, so
ENODEV is an expected failure. Don't issue a warning if the return
value is -ENODEV.
Note: I anticipate we might have to be more precise in identifying
this specific case, but for now this eliminates a bogus warning when
probing i2c devices.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 10 Dec 2014 14:43:33 +0000 (08:43 -0600)]
greybus: define GB_OP_NONEXISTENT
The i2c protocol needs a way to indicate an i2c device doesn't exist
(which is not necessarily an error). Define GB_OP_NONEXISTENT to
indicate this, and updating the status<->errno mapping functions
accordingly.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 3 Dec 2014 18:27:46 +0000 (12:27 -0600)]
greybus: record type in operation structure
I've gone back and forth on this, but now that I'm looking at
asynchronous operations I know that the asynchronous callback will
want to know what type of operation it is handling, and right now
that's only available in the message header.
So record an operation's type in the operation structure, and use
it in a few spots where the header type was being used previously.
Pass the type to gb_operation_create_incoming() so it can fill
it in after the operation has been created.
Clean up the crap comments above the definition of the operation
structure.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 3 Dec 2014 18:27:45 +0000 (12:27 -0600)]
greybus: use null pointer for empty payload
Currently message->payload always points to the address immediately
following the header in a message. If the payload length is 0, this
is not a valid pointer.
Change the code to assign a null pointer to the payload in this
case. I have verified that no code dereferences the payload pointer
unless the payload is known to have non-zero size.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 3 Dec 2014 18:27:44 +0000 (12:27 -0600)]
greybus: only record message payload size
An asynchronous operation will want to know how big the response
message it receives is. Rather than require the sender to record
that information, expose a new field "payload_size" available to
the protocol code for this purpose.
An operation message consists of a header and a payload. The size
of the message can be derived from the size of the payload, so
record only the payload size and not the size of the whole message.
Reorder the fields in a message structure.
Update the description of the message header structure.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Alex Elder [Wed, 3 Dec 2014 18:27:43 +0000 (12:27 -0600)]
greybus: don't let i2c code assume non-null payload pointer
This is in preparation for an upcoming patch, which makes the
payload pointer be NULL when a message has zero bytes of payload.
It ensures a null payload pointer never gets dereferenced. To do
this we pass the response structure to gb_i2c_transfer_response()
rather than just its data, and if it's null, returning immediately.
Rearrange the logic in gb_i2c_transfer_operation() a bit.
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
projects / karo-tx-linux.git / log