The serial core expects the UART driver to transmit x_char
(START/STOP chars) even if tx is stopped and before data already
in the tx ring buffer if possible. Also, sending x_char must
not cause additional data in the tx ring buffer to transmit
if tx is stopped.

Cause x_char to be transmitted before any other data is sent.
Auto-stop tx if the tx ring buffer is empty or tx should be stopped.
Only perform one write wakeup if tx ring buffer space is below
threshold.

x_char handling in DMA mode is still broken; add FIXME.
Signed-off-by: NPeter Hurley <peter@hurleysoftware.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix breakage introduced by
commit c557d392,
'serial: Test for no tx data on tx restart'.
Reported-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NPeter Hurley <peter@hurleysoftware.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Commit 717f3bba,
'serial_core: Fix conditional start_tx on ring buffer not empty'
exposes an incorrect assumption in several drivers' start_tx methods;
the tx ring buffer can, in fact, be empty when restarting tx while
performing flow control.

Affected drivers:
sunsab.c
ip22zilog.c
pmac_zilog.c
sunzilog.c
m32r_sio.c
imx.c

Other in-tree serial drivers either are not affected or already
test for empty tx ring buffer before transmitting.

Test for empty tx ring buffer in start_tx() method, after transmitting
x_char (if applicable).
Reported-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Cc: Seth Bollinger <sethb@digi.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: NPeter Hurley <peter@hurleysoftware.com>
Cc: stable <stable@vger.kernel.org> # 3.15
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This reverts commit e2f27866.

Huang reports that this patch is broken and should be reverted.

Cc: Huang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch disables the receiver ready interrupt for imx_stop_rx.
It reduces the interrupt numbers when the uart is going to close
or suspend.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The DMA wait queue makes the code very complicated:
  For RX, the @->stop_rx hook does not really stop the RX;
  For TX, the @->stop_tx hook does not really stop the TX.

The above make the imx_shutdown has to wait the RX/TX DMA to be finished.

In order to make code more simple, this patch removes the DMA wait queue.
By calling the dmaengine_terminate_all, this patch makes the RX stops
immediately after we call the @->stop_rx hook, so does the TX.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In the imx_startup(), we will reset the uart port which will reset all
the FIFOs, including the URXD.

So the code to clear the RX FIFO is redundant. Just remove it.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Current code resets the uart port only when it supports the irda mode.
In actually, we also need to reset the uart port in the non-irda mode.
A hang was caught in the following case:

    UART A transmits data to the other end. But the transmission maybe
    terminated. In some corner case, the TX FIFO maybe not empty.

The kernel will hang at the imx_set_termios():
	............................................................
	while (!(readl(sport->port.membase + USR2) & USR2_TXDC))
		barrier();
	............................................................

This patch resets the uart port all the time in the imx_startup().
And fix the hang.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Bit 7 of UCR3 is described in the i.MX reference manuals (with the exception
of i.MX1) as follows:

ADNIMP: Autobaud Detection Not Improved-. Disables new features of
	autobaud detection (See Baud Rate Automatic Detection
	Protocol, for more details).

	0 Autobaud detection new features selected
	1 Keep old autobaud detection mechanism

The "new features" mechanism occasionally cause the receiver to get out of sync
and continuously produce received characters of '\xff'.

In order to reproduce the problem:

$ stty -F /dev/ttymxc0 19200

- Change the terminal baudrate to 19200

- Type in the console and it should look good

- Change the terminal baudrate back to 115200

- Type 'b' in the console, then a stream of garbage characters is seen.

Also rename the bit definition as per the reference manual.

Tested on mx6q, mx6dl, mx6solo and mx53.

Based on a patch from Eric Nelson for U-boot.
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
Tested-by: NEric Nelson <eric.nelson@boundarydevices.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use devm_ioremap_resource() in order to make the code simpler and
it gives proper codes on errors.
Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The current code only wakes up the processes when the circle
buffer has less data then the WAKEUP_CHARS.

But sometimes, the circle buffer may has data more then the WAKEUP_CHARS,
in such case, the processes will hang.

This patch makes it always wakes up the processes in the TX callback.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Add the loopback mode support for imx uart driver.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

I worried that the delay of the sdma_run_channel0() maybe too long for
interrupt context, so I added the workqueues for RX/TX DMA.

But tested with bluetooth device, I find that the delay of sdma_run_channel0()
is about 8us (tested in imx6dl sabreauto board). I think the delay
is acceptable.

This patch removes the RX/TX workqueues for DMA, it makes the
code more clear.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the dmaengine_tx_status to simplify the code, do not change any logic.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When the (xmit->tail > xmit->head) is true and the xmit->head is 0,
we only need one DMA scatterlist in actually. Current code uses two DMA
scatterlists in this case, this is obviously wrong.

This patch fixes it.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Assume the following situation:

  If the DMA is enabled, and the a TX DMA operation is working,
But we have not issued the TX DMA operation (or we have issued the
TX DMA operation with dma_async_issue_pending(), but the DMA has not
started to move the data from the memory to the TXFIFO).

At this time, we may get the wrong status of the transmitter when we
call the imx_tx_empty. In such situation, only check the USR2_TXDC
does not enough for us.

This patch checks the DMA's situation, and return 0 when the TX DMA is
working.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The current driver does not implement the flush_buffer hook for
uart_ops. When we enable the DMA for the driver, and test it with Bluetooth,
we may meet the following bug for TX:

    [1] User application may call the flush operation at any time.
        The uart_flush_buffer() calls the uart_circ_clear() to set
        the xmit->head and xmit->tail with 0.

    [2] The TX DMA callback can be called at any time too.
        The dma_tx_call() will update the xmit->tail.

    If [2] occurs just after the [1], we will get the wrong xmit->tail.

This patch implements the flush_buffer hook to fix this issue.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This reverts commit f7d2c0bb, as it
causes build errors when the driver is built as a module.
Reported-by: NGuenter Roeck <linux@roeck-us.net>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This resolves a warning where resource_size_t is larger than void *:

  drivers/tty/serial/imx.c:1542:6: warning: cast to pointer from integer
  	of different size [-Wint-to-pointer-cast]

Since iomem_base is a void *, casting to unsigned long is safe.

It's unclear to me that this comparison is truly needed, but it's there
on several other drivers as well.
Signed-off-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Commit f7d2c0bb (serial: i.MX: evaluate linux,stdout-path property) introduced
the following sparse warning:

drivers/tty/serial/imx.c:1916:77: warning: Using plain integer as NULL pointer

Pass NULL as the last argument of add_preferred_console() instead of zero.
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
Acked-by: NSascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The slave_config is not initialized, so some of its fields may
be set with random data which may causes the failure in the following
dmaengine_prep_slave_sg().

This patch fixes this issue.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

devicetrees may have the linux,stdout-path property to specify the
console. This patch adds support to the i.MX serial driver for this.
Signed-off-by: NSascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We only enable the DMA support when the following are meet:

  [1] The uart port supports the hardware flow control(CTS/RTS).
      (Some uart port does not support the CTS/RTS.)

  [2] The application enables the CTS/RTS.

  [3] The Soc is imx6q.
      For the sdma's firmware limit, we do not support the DMA except
      the imx6q platform.

  [4] The uart is not used as a console.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

It was not possible to set custom-baudrates like 62500.
Signed-off-by: NHubert Feurstein <h.feurstein@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We will add the DMA support for the imx uart. For the firmware's limit,
only the imx6 serial chips (including the imx6q, imx6dl, imx6sl) can
support the DMA.

This patch adds the necessary macro and helper to distinguish the
imx6q uart from the other imx uart. Other chips can use the "fsl,imx6q-uart"
to enable the DMA support.

This patch also replaces the check "is_imx21_uart()" with "!is_imx1_uart()".
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The uart_console() check makes the clocks(clk_per and clk_ipg) opened
even when we close the console uart.

This patch enable/disable the clocks in imx_console_write(),
so we can keep the clocks closed when the console uart is closed.

Also remove the clock enable/disable oprations in the probe, we do not
need them any more.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d063
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Acked-by: NBarry Song <baohua.song@csr.com>
Acked-by: NTony Prisk <linux@prisktech.co.nz>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Since commit ab78029e (drivers/pinctrl: grab default handles from device core),
we can rely on device core for setting the default pins. Compile tested only.

Acked-by: Linus Walleij <linus.walleij@linaro.org> (personally at LCE13)
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Since commit 0c375501 (serial: imx: enable the clocks for console), the
imx_startup() function calls clk_prepare_enable conditionally, so we
need to call clk_disable_unprepare inside imx_shutdown() under the same
condition to avoid unbalanced clock calls.

This avoids the following warning:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 70 at drivers/clk/clk.c:780 __clk_disable+0x68/0x84()
Modules linked in:
CPU: 0 PID: 70 Comm: mountall Not tainted 3.10.0-rc4-next-20130607+ #435
Backtrace:
[<800116a4>] (dump_backtrace+0x0/0x10c) from [<80011844>] (show_stack+0x18/0x1c)
 r6:8069f4e8 r5:0000030c r4:00000000 r3:00000000
[<8001182c>] (show_stack+0x0/0x1c) from [<8053bce0>] (dump_stack+0x78/0x94)
[<8053bc68>] (dump_stack+0x0/0x94) from [<80023df8>] (warn_slowpath_common+0x6c/0x8c)
 r4:00000000 r3:00000000
[<80023d8c>] (warn_slowpath_common+0x0/0x8c) from [<80023e3c>] (warn_slowpath_null+0x24/0x2c)
 r8:bf2ed008 r7:bfaa9810 r6:000f0013 r5:bf824b80 r4:bf824b80
[<80023e18>] (warn_slowpath_null+0x0/0x2c) from [<8041af84>] (__clk_disable+0x68/0x84)
[<8041af1c>] (__clk_disable+0x0/0x84) from [<8041b098>] (clk_disable+0x20/0x2c)
 r4:600f0013 r3:00000001
[<8041b078>] (clk_disable+0x0/0x2c) from [<802c93e8>] (imx_shutdown+0xbc/0xec)
 r5:bf824b80 r4:bfaa9810
[<802c932c>] (imx_shutdown+0x0/0xec) from [<802c63a0>] (uart_port_shutdown+0x34/0x40)
 r5:bf86f860 r4:bfaa9810
[<802c636c>] (uart_port_shutdown+0x0/0x40) from [<802c68c0>] (uart_shutdown+0x98/0xc4)
 r4:bf86f800 r3:00000000
[<802c6828>] (uart_shutdown+0x0/0xc4) from [<802c7514>] (uart_close+0x5c/0x198)
 r7:bfaa9810 r6:bf274400 r5:bf86f86c r4:bf86f800
[<802c74b8>] (uart_close+0x0/0x198) from [<802ac648>] (tty_release+0xf8/0x500)
[<802ac550>] (tty_release+0x0/0x500) from [<800c5a30>] (__fput+0x9c/0x208)
[<800c5994>] (__fput+0x0/0x208) from [<800c5bac>] (____fput+0x10/0x14)
[<800c5b9c>] (____fput+0x0/0x14) from [<80040234>] (task_work_run+0xb4/0xec)
[<80040180>] (task_work_run+0x0/0xec) from [<80029238>] (do_exit+0x2b0/0x920)
 r8:8000e144 r7:000000f8 r6:bf306300 r5:00000000 r4:bfac1180
[<80028f88>] (do_exit+0x0/0x920) from [<80029a4c>] (do_group_exit+0x50/0xd4)
 r7:000000f8
[<800299fc>] (do_group_exit+0x0/0xd4) from [<80029ae8>] (__wake_up_parent+0x0/0x28)
 r7:000000f8 r6:00000001 r5:0006f7ae r4:0006f79a
[<80029ad0>] (SyS_exit_group+0x0/0x18) from [<8000dfc0>] (ret_fast_syscall+0x0/0x30)
---[ end trace 16d080eb7efea4e9 ]---
Reported-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
Tested-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The console's clocks are disabled after the uart driver is probed.
It makes that we can see less log from the console now
(though we still can get all the log by the `dmesg`).

So enable the clocks for console, and we can see all the log again.

This patch also disables the sport->clk_per when we fail to enable
the sport->clk_ipg;
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When CONFIG_SERIAL_IMX_CONSOLE is not selected the following build warnings
appear:

drivers/tty/serial/imx.c:274:13: warning: 'imx_port_ucrs_save' defined but not used [-Wunused-function]
drivers/tty/serial/imx.c:283:13: warning: 'imx_port_ucrs_restore' defined but not used [-Wunused-function]

imx_port_ucrs_save() and imx_port_ucrs_restore() are only used when
CONFIG_CONSOLE_POLL or CONFIG_SERIAL_IMX_CONSOLE are selected, so protect these
functions declaration with a proper ifdef.
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Current code opens the clocks when the uart driver is probed.
This will wastes some power if several uarts are enabled, but not really
used.

So close these clocks for uart, and enable the clocks only when
the uart is used.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The uart works in the DCE mode by default, but sometime we need it
works at the DTE mode.

This patch adds the support for the DTE mode.
Signed-off-by: NHuang Shijie <b32955@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We only want to enable hardware flow control if RTS/CTS pins
are connected.
Signed-off-by: NLucas Stach <l.stach@pengutronix.de>
Signed-off-by: NMarkus Pargmann <mpa@pengutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch enables the overrun error (ORE) interrupt and increases the
counter in case of overrun.
Signed-off-by: NAlexander Stein <alexander.stein@systec-electronic.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Commit 677fe555 (serial: imx: Fix recursive locking bug) introduces an
uninitialized variable warning as below.

  CC      drivers/tty/serial/imx.o
drivers/tty/serial/imx.c: In function ‘imx_console_write’:
include/linux/spinlock.h:340:2: warning: ‘flags’ may be used uninitialized in this function [-Wuninitialized]
drivers/tty/serial/imx.c:1214:16: note: ‘flags’ was declared here

Initialize the variable to suppress the warning.
Signed-off-by: NShawn Guo <shawn.guo@linaro.org>
Acked-by: NDirk Behme <dirk.behme@de.bosch.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

commit 9ec1882d (tty: serial: imx: console write routing is unsafe
on SMP) introduced a recursive locking bug in imx_console_write().

The callchain is:

imx_rxint()
  spin_lock_irqsave(&sport->port.lock,flags);
  ...
  uart_handle_sysrq_char();
    sysrq_function();
      printk();
        imx_console_write();
          spin_lock_irqsave(&sport->port.lock,flags); <--- DEAD

The bad news is that the kernel debugging facilities can dectect the
problem, but the printks never surface on the serial console for
obvious reasons.

There is a similar issue with oops_in_progress. If the kernel crashes
we really don't want to be stuck on the lock and unable to tell what
happened.

In general most UP originated drivers miss these checks and nobody
ever notices because CONFIG_PROVE_LOCKING seems to be still ignored by
a large number of developers.

The solution is to avoid locking in the sysrq case and trylock in the
oops_in_progress case.

This scheme is used in other drivers as well and it would be nice if
we could move this to a common place, so the usual copy/paste/modify
bugs can be avoided.

Now there is another issue with this scheme:

CPU0 	    	     	 CPU1
printk()
			 rxint()
			   sysrq_detection() -> sets port->sysrq
			 return from interrupt
  console_write()
     if (port->sysrq)
     	avoid locking

port->sysrq is reset with the next receive character. So as long as
the port->sysrq is not reset and this can take an endless amount of
time if after the break no futher receive character follows, all
console writes happen unlocked.

While the current writer is protected against other console writers by
the console sem, it's unprotected against open/close or other
operations which fiddle with the port. That's what the above mentioned
commit tried to solve.

That's an issue in all drivers which use that scheme and unfortunately
there is no easy workaround. The only solution is to have a separate
indicator port->sysrq_cpu. uart_handle_sysrq_char() then sets it to
smp_processor_id() before calling into handle_sysrq() and resets it to
-1 after that. Then change the locking check to:

     if (port->sysrq_cpu == smp_processor_id())
     	 locked = 0;
     else if (oops_in_progress)
         locked = spin_trylock_irqsave(port->lock, flags);
     else
  	 spin_lock_irqsave(port->lock, flags);

That would force all other cpus into the spin_lock path. Problem
solved, but that's way beyond the scope of this fix and really wants
to be implemented in a common function which calls the uart specific
write function to avoid another gazillion of hard to debug
copy/paste/modify bugs.
Reported-and-tested-by: NTim Sander <tim@krieglstein.org>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: stable <stable@vger.kernel.org>  # 3.6+
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>