Deadlock issue is seen when enabling CONFIG_PROVE_LOCKING=Y, and uart
console as wakeup source. Deadlock occurs when resuming from low power
mode if system is waked up via usart console.
The deadlock is triggered 100% when also disabling console suspend prior
to go to suspend.

Simplified call stack, deadlock condition:
- stm32_console_write <-- spin_lock already held
- print_circular_bug
- pm_wakeup_dev_event <-- triggers lockdep as seen above
- stm32_receive_chars
- stm32_interrupt <-- wakeup via uart console, takes the lock

So, revisit spin_lock in stm32-usart driver:
- there is no need to hold the lock to access ICR (atomic clear of status
  flags)
- only hold the lock inside stm32_receive_chars() routine (no need to
  call pm_wakeup_dev_event with lock held)
- keep stm32_transmit_chars() routine called with lock held

Fixes: 48a6092f ("serial: stm32-usart: Add STM32 USART Driver")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Link: https://lore.kernel.org/r/20210304162308.8984-6-erwan.leray@foss.st.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

TX and RX FIFO thresholds may be cleared after suspend/resume, depending
on the low power mode.

Those configurations (done in startup) are not effective for UART console,
as:
- the reference manual indicates that FIFOEN bit can only be written when
  the USART is disabled (UE=0)
- a set_termios (where UE is set) is requested firstly for console
  enabling, before the startup.

Fixes: 84872dc4 ("serial: stm32: add RX and TX FIFO flush")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210304162308.8984-5-erwan.leray@foss.st.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Incorrect characters are observed on console during boot. This issue occurs
when init/main.c is modifying termios settings to open /dev/console on the
rootfs.

This patch adds a waiting loop in set_termios to wait for TX shift register
empty (and TX FIFO if any) before stopping serial port.

Fixes: 48a6092f ("serial: stm32-usart: Add STM32 USART Driver")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210304162308.8984-4-erwan.leray@foss.st.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

RX is configured, but usart is not enabled in startup function.
Kernel documentation specifies that startup should enable the port for
reception.
Fix the startup by enabling usart for reception.

Fixes: 84872dc4 ("serial: stm32: add RX and TX FIFO flush")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210304162308.8984-3-erwan.leray@foss.st.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The probe and remove orders are wrong as the uart_port is registered
before saving device data in the probe, and unregistered after DMA
resource deallocation in the remove. uart_port registering should be
done at the end of probe and unregistering should be done at the begin of
remove to avoid resource allocation issues.

Fix probe and remove orders. This enforce resource allocation occur at
proper time.
Terminate both DMA rx and tx transfers before removing device.

Move pm_runtime after uart_remove_one_port() call in remove() to keep the
probe error path.

Fixes: 34891872 ("serial: stm32: adding dma support")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210304162308.8984-2-erwan.leray@foss.st.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The Samsung SoC UART nodes have usually DMA so dtschema has to reflect
this to fix dtbs_check warnings like:

  arch/arm/boot/dts/exynos4210-smdkv310.dt.yaml: serial@13800000:
    'dma-names', 'dmas' do not match any of the regexes: 'pinctrl-[0-9]+'
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210212163905.70171-1-krzk@kernel.orgSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In case of error, the function devm_platform_get_and_ioremap_resource()
returns ERR_PTR() and never returns NULL. The NULL test in the return
value check should be replaced with IS_ERR().

Fixes: 1da81e55 ("drivers/tty/serial: add LiteUART driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Link: https://lore.kernel.org/r/20210305034929.3234352-1-weiyongjun1@huawei.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Software node is always created for additional device
properties. If the properties are constant, the software
node can also be constant.
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20210304081311.17340-1-heikki.krogerus@linux.intel.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Earlycon support is identical to S3C2410, but Apple SoCs also need
MMIO mapped as nGnRnE. This is handled generically for normal drivers
including the normal UART path here, but earlycon uses fixmap and
runs before that scaffolding is ready.

Since this is the only case where we need this fix, it makes more
sense to do it here in the UART driver instead of introducing a
whole fdt nonposted-mmio resolver just for earlycon/fixmap.
Suggested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-26-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Apple SoCs are a distant descendant of Samsung designs and use yet
another variant of their UART style, with different interrupt handling.

In particular, this variant has the following differences with existing
ones:

* It includes a built-in interrupt controller with different registers,
  using only a single platform IRQ

* Internal interrupt sources are treated as edge-triggered, even though
  the IRQ output is level-triggered. This chiefly affects the TX IRQ
  path: the driver can no longer rely on the TX buffer empty IRQ
  immediately firing after TX is enabled, but instead must prime the
  FIFO with data directly.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-25-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Apple mobile devices originally used Samsung SoCs (starting with the
S5L8900), and their current in-house SoCs continue to use compatible
UART peripherals. We'll call this UART variant apple,s5l-uart.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210304213902.83903-24-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This picks up the non-posted I/O mode needed for Apple platforms to
work properly.

This removes the request/release functions, which are no longer
necessary, since devm_ioremap_resource takes care of that already. Most
other drivers already do it this way, anyway.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-23-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* Split out s3c24xx_serial_tx_chars from s3c24xx_serial_tx_irq,
  where only the latter acquires the port lock. This will be necessary
  on platforms which have edge-triggered IRQs, as we need to call
  s3c24xx_serial_tx_chars to kick off transmission from outside IRQ
  context, with the port lock held.

* Rename s3c24xx_serial_rx_chars to s3c24xx_serial_rx_irq for
  consistency with the above. All it does now is call two other
  functions anyway.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-22-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This decouples the TTY layer PORT_ types, which are exposed to
userspace, from the driver-internal flag of what kind of port this is.

This removes s3c24xx_serial_has_interrupt_mask, which was just checking
for a specific type anyway.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-21-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Instead of patching a single global ops structure depending on the port
type, use a separate s3c64xx_serial_ops for the S3C64XX type. This
allows us to mark the structures as const.

Also split out s3c64xx_serial_shutdown into a separate function now that
we have a separate ops structure; this avoids excessive branching
control flow and mirrors s3c64xx_serial_startup. tx_claimed and
rx_claimed are only used in the S3C24XX functions.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-19-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This simplifies the code by removing the only distinction between the
S3C2410 and S3C2440 codepaths.
Signed-off-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210304213902.83903-20-marcan@marcan.stSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

When SERIAL_IMX_EARLYCON is enabled, and SERIAL_CORE is disabled,
Kbuild gives the following warning:

WARNING: unmet direct dependencies detected for SERIAL_EARLYCON
  Depends on [n]: TTY [=y] && HAS_IOMEM [=y] && SERIAL_CORE [=n]
  Selected by [y]:
  - SERIAL_IMX_EARLYCON [=y] && TTY [=y] && HAS_IOMEM [=y] && (ARCH_MXC [=y] || COMPILE_TEST [=n]) && OF [=y]

This is because SERIAL_IMX_EARLYCON selects SERIAL_EARLYCON,
without selecting SERIAL_CORE, despite that config option depending
on SERIAL_CORE.
Signed-off-by: Julian Braha <julianbraha@gmail.com>
Link: https://lore.kernel.org/r/20210225044248.4745-1-julianbraha@gmail.comSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The tty line disciplines don't expect tty_operations::write_room to
return negative values. Fix the five drivers which violate this.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-44-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The tty line disciplines don't expect tty_operations::chars_in_buffer to
return negative values. Fix the two drivers which violate this.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-43-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There is no need for 'else' when the 'if' part already returned. This
makes tty_chars_in_buffer similar to tty_write_room too.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-42-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dp, f, and i are only duplicates of gsmld_receive_buf's parameters. Use
the parameters directly (cp, fp, and count) and delete these local
variables.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-41-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Forward declarations make the code larger and rewrites harder. Harder as
they are often omitted from global changes. Remove forward declarations
which are not really needed, i.e. the definition of the function is
before its first use.

Cc: linuxppc-dev@lists.ozlabs.org
Reviewed-by: Tyrel Datwyler <tyreld@linux.ibm.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-40-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Forward declarations make the code larger and rewrites harder. Harder as
they are often omitted from global changes. Remove forward declarations
which are not really needed, i.e. the definition of the function is
before its first use.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-39-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ptychar was not const, so mark it as such. And move this variable to the
only place where it's used.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-36-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Now that noone checks the return value, switch the return type of
tty_unregister_driver to void. We can do that as we always return zero.

Generally, drivers are not allowed to call tty_unregister_driver while
there are open devices.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-35-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These drivers check tty_unregister_driver return value. But they don't
handle a failure correctly (they free the driver in any case). So stop
checking tty_unregister_driver return value and remove also the prints.

In the next patch, tty_unregister_driver's return type will be switched
to void.

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Cc: Jiri Kosina <jikos@kernel.org>
Cc: David Sterba <dsterba@suse.com>
Reviewed-by: Max Filippov <jcmvbkbc@gmail.com>
Acked-by: David Sterba <dsterba@suse.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-34-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

To fix the warning:
warning: no previous prototype for 'rs_init'

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-33-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Use DEFINE_TIMER and avoid runtime initialization of the serial_timer.

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-32-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These are likely taken over from amiserial. iss doesn't do anything of
that.

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-31-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

We already do tty_port_link_device in rs_init, so we don't need to
reassign a port to tty->port. It would be too late in tty::ops::open
anyway.

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-30-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There is no need to print the information during module load. Neither to
print some artificial version. So drop these strings and a print.

Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: linux-xtensa@linux-xtensa.org
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-29-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

All these functions are called with a valid tty pointer, no need to
check that.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: sparclinux@vger.kernel.org
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-28-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Do not define global variables for driver/device names, use the strings
directly.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: sparclinux@vger.kernel.org
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-27-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The version number is artificial, no need to dump it to logs during
initialization.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: <sparclinux@vger.kernel.org>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-26-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These are used only in this unit, so make them static.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-25-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These are either commented out or can be printed by tracing mechanisms.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-24-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove useless prints from init and exit functions. The version is
artificial anyway.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-23-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

It is unused.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-22-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The code for char+error handling in jsm_input was complete mess of
letters. Introduce 3 new local variables and use them with care.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-21-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Casts of 'void *' pointer are superfluous. So remove them.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux390@de.ibm.com
Cc: linux-s390@vger.kernel.org
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Link: https://lore.kernel.org/r/20210302062214.29627-20-jslaby@suse.czSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>