In some condition we receive the break interrupt but nothing is putted
in the Rx FIFO and the correspondend bit in the status register is not
set. Thus, no-one clear the interrupt and the handler will be called
forever.

This patch clear the break interrupt as soon as it occurs. Then, if the
break character '\0' is putted in the fifo we will manage it.

We can also unmask the Break interrupt but its bit in ISR is still set
on break. So I think is better to keep the registers clean.
Signed-off-by: NFederico Vaga <federico.vaga@cern.ch>
Acked-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NFederico Vaga <federico.vaga@cern.ch>
Acked-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NFederico Vaga <federico.vaga@cern.ch>
Acked-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The request for an IRQ handler must be done after whole configuration. This
was not the case for this driver which request the IRQ in the middle of
the configuration. Sometimes, it happens that something is not completely
configured, we recieve an interrupt thus we stumble into troubles in the
IRQ handler.
Signed-off-by: NFederico Vaga <federico.vaga@cern.ch>
Acked-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In some conditions (echo or particular sequence of special
characters), on buffer push, the tty layer calls the write operation
while we are holding the spinlock. This means deadlock within the same
process on kernels version < 3.12. It seems not a problem on recent
kernel, but the patch still valid as locking optimization.

The protected variables by the spinlock are: xmit_buf, nb_bytes,
pointer_read and pointer_write. So, this patch reduces the locked area
in the IRQ handler only to these variables. Most of the code inside the
locked area in the IRQ handler is not protected elsewhere; it means
that it is not protected at all.
Signed-off-by: NFederico Vaga <federico.vaga@cern.ch>
Acked-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Delete successive tests to the same location.  The code tested the result
of a previous allocation, that itself was already tested.  It is changed to
test the result of the most recent allocation.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@s exists@
local idexpression y;
expression x,e;
@@

*if ( \(x == NULL\|IS_ERR(x)\|y != 0\) )
 { ... when forall
   return ...; }
... when != \(y = e\|y += e\|y -= e\|y |= e\|y &= e\|y++\|y--\|&y\)
    when != \(XT_GETPAGE(...,y)\|WMI_CMD_BUF(...)\)
*if ( \(x == NULL\|IS_ERR(x)\|y != 0\) )
 { ... when forall
   return ...; }
// </smpl>
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Cc: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There was a bug in the code when managing a GE IP-OCTAL-485 board. The RX would
be enabled but we have a wrong state in the rx_enable flag.

Then, if the user changes the terminal settings, RX would not be enabled again.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Now, we start converting tty buffer functions to actually use
tty_port. This will allow us to get rid of the need of tty in many
call sites. Only tty_port will needed and hence no more
tty_port_tty_get in those paths.

Now, the one where most of tty_port_tty_get gets removed:
tty_flip_buffer_push.

IOW we also closed all the races in drivers not using tty_port_tty_get
at all yet.

Also we move tty_flip_buffer_push declaration from include/linux/tty.h
to include/linux/tty_flip.h to all others while we are changing it
anyway.
Signed-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Now, we start converting tty buffer functions to actually use
tty_port. This will allow us to get rid of the need of tty in many
call sites. Only tty_port will needed and hence no more
tty_port_tty_get in those paths.

tty_insert_flip_char is the next one to proceed. This one is used all
over the code, so the patch is huge.
Signed-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Added shutdown callback to disable RX and TX when there is no other client
accesing the device.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Thus, we don't enable RX when a termios setup has been called, as it could be
disabled previously.

As the control registers (Rx, Tx flags specifically) cannot be read from the
device, we keep this info in rx_enable.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The function is already called in ipoctal_irq_rx()
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We protect important data such as TX buffer pointer, nb_bytes counter and status
registers of the device, from accessing several times at the same time.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Due to the IRQ processing, we can generate another IRQ that can come before we
end the previous one, so we lost it. E.g. when transmitting a character.

To allow the processing in SMP machines, we ack the IRQ at the beginning of the
IRQ handler.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

RX is enabled when the tty port is open, so no need to do it in initialization
time: it can allow the device to receive characters but no TTY client is
listening to them.

It produced an infinite number of IRQ as RxFIFO is not read to clear that
IRQ in the device, so it is still pending.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Depending of the device, it disables the TX mode in different places when there
is no more data to transmit.

This patch reorder them and disable the TX mode in the same place.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In case of several characters present in RxFIFO, they will have the flag of the
previous one, no matter if the actual character was received properly or not.

This patch fixes this bug.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Don't block the TTY client when sending characters.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Trying to setup the pppd server to use ipoctal's serial ports, it says the ports
are busy the first time. If the operation is repeated, a kernel bug due to a
dereference of a NULL pointer appears.

Removing the one-access-only setup from the driver, removes this kernel bug.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

It is an unsigned int so that check is pointless.
Signed-off-by: NAlberto Garcia <agarcia@igalia.com>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In order to transmit data, the driver enables Tx and sleeps until
*board_write is set to 1 by the interrupt handler.

It can happen, though, that the data is sent even before the process
is asleep. In this case *board_write must be set to 1 anyway,
otherwise we will be waiting for a condition that will never be true.
Signed-off-by: NAlberto Garcia <agarcia@igalia.com>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Its contents are merged into ipack.h. So this file is not needed.

Doing that, it simplifies the ipack-related driver development.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move ipack header files to include/linux/ directory where they belong.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The ipack subsystem is cleaned up enough to now move out of the staging
tree, and into drivers/ipack.

Cc: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Cc: Jens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Adapt the copyright clause to CERN guidelines, as it has the copyright over
them.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

After commit "TTY: move tty buffers to tty_port", the tty buffers are
not freed in some drivers. This is because tty_port_destructor is not
called whenever a tty_port is freed. This was an assumption I counted
with but was unfortunately untrue. So fix the drivers to fulfil this
assumption.

To be sure, the TTY buffers (and later some stuff) are gone along with
the tty_port, we have to call tty_port_destroy at tear-down places.
This is mostly where the structure containing a tty_port is freed.
This patch does exactly that -- put tty_port_destroy at those places.
Signed-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There also is a MEM16 space.  This will make it clear which one is
which, once support for MEM16 space is added.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the regular ioremap functions and their managed counterparts instead
of the ones provided through IPack callbacks.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

By doing so we can remove ipoctal_check_model() and we also no longer need
to map the IPACK_ID_SPACE.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The ordering was wrong.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

It is replicating information contained in ipack_device->bus->bus_nr.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove count_wr and the assigment of nb_bytes = 0 in that function as is
useless. Now it returns the count of the characters actually sent.

There is other nb_bytes = 0 deleted that has a duplicate a few lines before.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Clear the BREAK flag from the ISR register. Doing that, we avoid to read
the same condition for the next character received.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The field irq currently is identical to the slot number.  It does not seem to
have any real use.  The number is written to hardware in ipoctal but it seems
the value that is written does not matter.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Now the IPack device acknowledges its own IRQ.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The RX FIFO has a size of 3 characters. Check if when we are picking the
oldest one, we have more to read.
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use tty_dev->dev's drdata to associate struct ipocal_channel to the
respective tty_struct.
Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NJens Taprogge <jens.taprogge@taprogge.org>
Signed-off-by: NSamuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>