Since commit 4dec2f11 ("imx-serial: RX DMA startup latency")
the interrupt routine no longer will start rx dma.

imx_dma_rxint no longer needs to be called to try and start dma.
It won't start dma because dma_is_rxing is
already true meaning dma is already started.
Signed-off-by: NTroy Kisky <troy.kisky@boundarydevices.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Since commit 4dec2f11 ("imx-serial: RX DMA startup latency")
the interrupt routine no longer will start rx dma.
So, we no longer need to enable this interrupt to start dma.
Signed-off-by: NTroy Kisky <troy.kisky@boundarydevices.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Canceling hrtimer when holding uart spinlock can deadlock.

CPU0: syscall write
          -> get uart port spinlock
              -> write uart
                  -> start_tx_rs485
                      -> hrtimer_cancel
                          -> wait for hrtimer callback to finish

CPU1: hrtimer IRQ
          -> run hrtimer
              -> em485_handle_stop_tx
                  -> get uart port spinlock

CPU0 is waiting for the hrtimer callback to finish, but the hrtimer
callback running on CPU1 is waiting to get the uart port spinlock.

This deadlock can be avoided by not canceling the hrtimers in these paths.
Setting active_timer=NULL can be done without accessing hrtimer,
and that will effectively cancel operations that would otherwise have been
performed by the hrtimer callback.
Signed-off-by: NTomas Melin <tomas.melin@vaisala.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

With the timer converted to using the new timer_setup()/from_timer() API,
setting the .data field is redundant (and the field will be removed soon),
so drop it.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

During hibernation, freeze/thaw/restore dev_pm_ops are called instead of
suspend/resume. Hook up the hibernation ops. The _noirq parts are
identical, but suspend/resume are replaced with variants that do not
enable wakeup from i.MX UART. There is no need to restore register
contents in thaw_noirq.
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Back in 2010, commit 3b43816f ("ARM: 5933/1: amba-pl011: support
hardware flow control") added an autorts flag to this driver to store
whether automatic hardware flow control is requested.

In 2015, commit 391f93f2 ("serial: core: Rework hw-assisted flow
control support") added a flag with identical functionality to generic
code.

Switch over to that flag and drop the driver-specific one.

Cc: Rabin Vincent <rabin.vincent@stericsson.com>
Cc: Peter Hurley <peter@hurleysoftware.com>
Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This driver's ->rs485_config callback checks if SER_RS485_RTS_ON_SEND
and SER_RS485_RTS_AFTER_SEND have the same value.  If they do, it means
the user has passed in invalid data with the TIOCSRS485 ioctl()
since RTS must have a different polarity when sending and when not
sending.  In this case, rs485 mode is not enabled (the RS485_URA bit
is not set in the RS485 Enable Register) and this is supposed to be
signaled back to the user by clearing the SER_RS485_ENABLED bit in
struct serial_rs485 ... except a missing tilde character is preventing
that from happening.

Fixes: 28e3fb6c ("serial: Add support for Fintek F81216A LPC to 4 UART")
Cc: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Cc: "Ji-Ze Hong (Peter Hong)" <hpeter@gmail.com>
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The final assignment to port is never read, hence it is redundant
and can be removed. Also move the declaration of port to a more
local scope.  Cleans up clang warning:

drivers/tty/serial/serial_core.c:1498:2: warning: Value stored
to 'port' is never read
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The SCSCR register access in sci_break_ctl() is not locked.

sci_start_tx() and sci_set_termios() changes the SCSCR register,
but does not lock sci_port.

Therefore, this patch adds lock during register access.

Also, remove the log output that leads to a double lock.

Some analysis of where locks are not taken is as follows.
It appears that the lock is not taken in:
  - sci_start_tx(), sci_stop_tx()  as this is installed as a callback.
    And all callers of the callback take the lock.
  - start_rx as callers take the lock.
  - stop_rx. this is both installed as a callback and called directly.
    In both cases the caller takes the lock.
Signed-off-by: NTakatoshi Akiyama <takatoshi.akiyama.kj@ps.hitachi-solutions.com>
Signed-off-by: NTakeshi Kihara <takeshi.kihara.df@renesas.com>
Signed-off-by: NSimon Horman <horms+renesas@verge.net.au>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

gcc points out that the length passed into memset here is wrong:

drivers/tty/serial/mvebu-uart.c: In function 'mvebu_uart_probe':
arch/x86/include/asm/string_32.h:324:29: error: 'memset' used with length equal to number of elements without multiplication by element size [-Werror=memset-elt-size]

Moreover, the structure was allocated with kzalloc a few lines earlier,
so that memset is also unnecessary. Let's drop it to shut up the
compiler warning.

Fixes: 95f78768 ("serial: mvebu-uart: dissociate RX and TX interrupts")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Variable rts is being assigned but it is never read, hence it can be
removed.  The assignment to param_new to zero is redundant as it is
being updates a few statements later, so remove this redundant
assignment. Cleans up two clang warnings:

drivers/tty/serial/max3100.c:277:3: warning: Value stored to 'rts'
is never read
drivers/tty/serial/max3100.c:439:2: warning: Value stored to 'param_new'
is never read
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The tty_ldisc_receive_buf() helper returns the number of bytes
processed so drop the bogus "not" from the kernel doc comment.

Fixes: 8d082cd3 ("tty: Unify receive_buf() code paths")
Signed-off-by: NJohan Hovold <johan@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Richard Genoud <richard.genoud@gmail.com>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Acked-by: NRichard Genoud <richard.genoud@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list
pointer to all timer callbacks, switch to using the new timer_setup() and
from_timer() to pass the timer pointer explicitly. Moves timer structures
from global to attached to struct cyclades_port.

Cc: Jiri Slaby <jslaby@suse.com>
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: sparclinux@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Tobias Klauser <tklauser@distanz.ch>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Cc: nios2-dev@lists.rocketboards.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Acked-by: NTobias Klauser <tklauser@distanz.ch>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Jiri Slaby <jslaby@suse.com>
Cc: adi-buildroot-devel@lists.sourceforge.net
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Jiri Slaby <jslaby@suse.com>
Cc: Vignesh R <vigneshr@ti.com>
Cc: Ed Blake <ed.blake@imgtec.com>
Cc: Matthias Brugger <mbrugger@suse.com>
Cc: Sean Young <sean@mess.org>
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Reviewed-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Reorder controller initialisation so that in the unlikely event that id
allocation fails, we don't end up releasing id 0 in the destructor.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Serdev currently only supports a single slave device, but the required
sanity checks to prevent further registration attempts were missing.

If a serial-port node has two child nodes with compatible properties,
the OF code would try to register two slave devices using the same id
and name. Driver core will not allow this (and there will be loud
complaints), but the controller's slave pointer would already have been
set to address of the soon to be deallocated second struct
serdev_device. As the first slave device remains registered, this can
lead to later use-after-free issues when the slave callbacks are
accessed.

Note that while the serdev registration helpers are exported, they are
typically only called by serdev core. Any other (out-of-tree) callers
must serialise registration and deregistration themselves.

Fixes: cd6484e1 ("serdev: Introduce new bus for serial attached devices")
Cc: stable <stable@vger.kernel.org>	# 4.11
Cc: Rob Herring <robh@kernel.org>
Signed-off-by: NJohan Hovold <johan@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Add missing error handling for tty-driver open() which may fail (e.g. if
resource allocation fails or if a port is being disconnected).

Note that close() must be called also in case of failed open() and that
the operation sanity check is amended to catch buggy drivers.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The tty-driver open routine is mandatory, but the serdev
tty-port-controller implementation did not treat it as such and would
instead fall back to calling tty_port_open() directly.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Commit 348f9bb3 ("serial: omap: Fix RTS handling") sought to enable
auto RTS upon manual RTS assertion and disable it on deassertion.
However it seems the latter was done incorrectly, it clears all bits in
the Extended Features Register *except* auto RTS.

Fixes: 348f9bb3 ("serial: omap: Fix RTS handling")
Cc: Peter Hurley <peter@hurleysoftware.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The do .. while loop checks for interrupt pending at the
start of the loop and exits if there is none, it then
checks again for this condition at the end of the loop.
Signed-off-by: NMartin Townsend <mtownsend1973@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

UART devices is expected to be enumerated by SerDev subsystem.

During ACPI scan, serial devices behind SPI, I2C or UART buses are not
enumerated, allowing them to be enumerated by their respective parents.

Rename *spi_i2c_slave* to *serial_bus_slave* as this will be used for serial
devices on serial buses (SPI, I2C or UART).

On Macs an empty ResourceTemplate is returned for uart slaves.
Instead the device properties "baud", "parity", "dataBits", "stopBits" are
provided. Add a check for "baud" in acpi_is_serial_bus_slave().
Signed-off-by: NFrédéric Danis <frederic.danis.oss@gmail.com>
Reviewed-by: NSebastian Reichel <sebastian.reichel@collabora.co.uk>
Tested-by: NRonald Tschalär <ronald@innovation.ch>
Tested-by: NPeter Y. Chuang <peteryuchuang@gmail.com>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch allows SerDev module to manage serial devices declared as
attached to an UART in ACPI table.

acpi_serdev_add_device() callback will only take into account entries
without enumerated flag set. This flags is set for all entries during
ACPI scan, except for SPI and I2C serial devices, and for UART with
2nd patch in the series.

Check if a serdev device as been allocated during acpi_walk_namespace()
to prevent serdev controller registration instead of the tty-class device.
Signed-off-by: NFrédéric Danis <frederic.danis.oss@gmail.com>
Reviewed-by: NRob Herring <robh@kernel.org>
Reviewed-by: NSebastian Reichel <sebastian.reichel@collabora.co.uk>
Reviewed-by: NJohan Hovold <johan@kernel.org>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Define the missing register offsets and bit fields for the extended
UART port. Add a second driver data structure filled with its port data,
selected with the right compatible (marvell,armada-3700-uart-ext).
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

A3700 boards may have up to two UART ports. Set the new limit to two
maximum UART ports.
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

While the standard UART port can use a single IRQ that 'sums' both RX
and TX interrupts, the extended port cannot and has to use two different
ISR, one for each direction. The standard port also has the hability
to use two separate interrupts (one for each direction).

The logic is then: either there is only one unnamed interrupt on the
standard port and this interrupt must be used for both directions
(this is legacy bindings); or all the interrupts must be described and
named 'uart-sum' (if available), 'uart-rx', 'uart-tx' and two separate
handlers for each direction will be used.
Suggested-by: NAllen Yan <yanwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Pulse interrupts (extended UART only) needs a change of state to trigger
the TX interrupt. In addition to enabling the TX_READY_INT_EN flag,
produce a FIFO state change from 'empty' to 'not full'. For this, write
only one data byte in TX start, making the TX FIFO not empty, and wait
for the TX interrupt to continue the transfer.
Signed-off-by: NAllen Yan <yanwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When receiving data on RX pin before ->uart_startup() is called, some
error bits in the state register could be set up (like BRK_DET).

This is harmless when using only the standard UART (error bits are
read-only), but may procude an endless loop once in the extended UART
RX interrupt handler (error bits must be cleared).

Clear the status register in ->uart_startup() to avoid this situation.
Signed-off-by: NAllen Yan <yanwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Until now, the first UART port baudrate was set by the bootloader.

Add a function allowing to change the baudrate. Changes may be done
from userspace but also at probe time by the kernel. Use the simplest
method: baudrate divisor.

Works for all UART ports until 230400 baud. To achieve higher baudrates,
software should implement the fractional divisor feature that allows
more accuracy for higher rates.
Signed-off-by: NAllen Yan <yanwei@marvell.com>
[<miquel.raynal@free-electrons.com>: changed termios handling]
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The existing UART driver relies on the bootloader to initialize the
port(s). However, the secondary uart port may not be initialized
properly in early boot stage. This patch adds the UART soft reset when
probing, for all ports.
Signed-off-by: NAllen Yan <yanwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There are two UART ports on Armada3700. The second UART is based on the
first one, plus additional features, but it has a different register
layout (some bit fields are also moved inside the registers).

Clearly separate register offsets and bit fields that differ between the
standard and the extended IP. Access them in a generic way. Rename the
defines with the "STD" prefix for future distinction with "EXT" defines.
Point to these defines in the main driver data structure.

The early console only uses the standard port (not extended).
Suggested-by: NWilson Ding <dingwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Until now, the mvebu-uart driver only supported probing a single UART
port. However, some platforms have multiple instances of this UART
controller, and therefore the driver should support multiple ports.

In order to achieve this, we make sure to assign port->line properly,
instead of hardcoding it to zero.
Signed-off-by: NAllen Yan <yanwei@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Acked-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the driver name when requesting an interrupt for consistency.

Avoids possible confusion with DW8250 driver interrupt names in
/proc/interrupts.
Signed-off-by: NYehuda Yitschak <yehuday@marvell.com>
Signed-off-by: NMiquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

pr_err() messages should end with a new-line to avoid other messages
being concatenated.
Signed-off-by: NArvind Yadav <arvind.yadav.cs@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Amend the Serial device bus Kconfig entries to clarify that you most
likely also want to enable TTY port controller support, and make
SERIAL_DEV_CTRL_TTYPORT default to Y (when bus support is enabled).

Note that the TTY port controller is currently the only in-kernel
serdev controller implementation.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NSebastian Reichel <sebastian.reichel@collabora.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.

Cc: Russell King <linux@armlinux.org.uk>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: NKees Cook <keescook@chromium.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>