kernel.git/drivers/tty/serial/imx.c, branch linux-4.17.y

serial: imx: fix cached UCR2 read on software reset

2018-04-23T07:53:36Z

To reset the UART the SRST needs be cleared (low active). According to the documentation the bit will remain active for 4 module clocks until it is cleared (set to 1). Hence the real register need to be read in case the cached register indicates that the SRST bit is zero. This bug lead to wrong baudrate because the baud rate register got restored before reset completed in imx_flush_buffer. Fixes: 3a0ab62f43de ("serial: imx: implement shadow registers for UCRx and UFCR") Signed-off-by: Stefan Agner Reviewed-by: Fabio Estevam Reviewed-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: warn user when using unsupported configuration

2018-04-23T07:53:36Z

When using half-duplex mode (which disables receiver during txing) the RTS signal cannot be driven low during transmission when using i.MX UART RTS/CTS control. This seems to be a limitation of the i.MX UART IP: The RTS (CTS_B) signal is controlled by the receiver. When the receiver is disabled, the signal stays in UART logic idle state which is high... If SER_RS485_RTS_ON_SEND is used, RTS needs to be high active during transmission. Since this is the default state of the RTS (CTS_B) signal when the receiver is off, half-duplex mode in this configuration works fine. However, a low-active RTS signal (flag SER_RS485_RTS_ON_SEND not set) cannot be generated when the receiver is turned off. Print an error if the user selects this unsupported configuration (both SER_RS485_RTS_ON_SEND and SER_RS485_RX_DURING_TX unset) and configure the closest working configuration (set the SER_RS485_RX_DURING_TX flag). Signed-off-by: Stefan Agner Acked-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

Merge 4.16-rc6 into tty-next

2018-03-20T10:27:18Z

We want the serial/tty fixes in here as well. Signed-off-by: Greg Kroah-Hartman

serial: imx: consistently use imx_uart_ as prefix for all functions

2018-03-09T18:21:02Z

Having a fixed prefix helps at several places. It ensures that another driver doesn't use the same function name which confuses the linker and tools like ctags. It simplifies working with function tracing and dynamic printk() support which can filter on function names. And last but not least it helps the human source code reader to understand if a given function belongs to a driver or a more general part of the kernel. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: don't prepare to send if no data is available

2018-03-09T18:21:02Z

serial_core might call the .start_tx callback without any data being available to send. In this case return early instead of going through all the setup needed for sending which might include disabling RX in RS485 half-duplex mode. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: Fix handling of TC irq in combination with DMA

2018-03-09T18:21:02Z

When using RS485 half duplex the Transmitter Complete irq is needed to determine the moment when the transmitter can be disabled. When using DMA this irq must only be enabled when DMA has completed to transfer all data. Otherwise the CPU might busily trigger this irq which is not properly handled and so the also pending irq for the DMA transfer cannot trigger. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: Also enable the aging timer in PIO mode

2018-03-09T18:21:01Z

This allows to increase the RX waterlevel which allows to delay the RRDY irq. The desired effect is that less irqs are needed to handle characters and so reduce irq count of the system. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: ensure that RX irqs are off if RX is off

2018-03-09T18:21:01Z

Make sure that UCR1.RXDMAEN and UCR1.ATDMAEN (for the DMA case) and UCR1.RRDYEN (for the PIO case) are off iff UCR1.RXEN is disabled. This ensures that the fifo isn't read with RX disabled which results in an exception. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: Stop to receive in .stop_rx()

2018-03-09T18:21:01Z

When the UART is used in DMA mode, .stop_rx() does nothing if the port isn't suspended. This is wrong as .stop_rx() should stop receiving characters unconditionally. When the port is about to be closed the DMA channel is stopped in .shutdown(), so this isn't necessary to be in .stop_rx() here, too. Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman

serial: imx: setup fifo waterlevel before enabling aging timer

2018-03-09T18:21:01Z

The aging timer fires if there are characters in the RX fifo but the water level isn't reached yet. Make sure that the waterlevel is configured before the aging timer is enabled to trigger a DMA request (UCR1_ATDMAEN). Signed-off-by: Uwe Kleine-König Signed-off-by: Greg Kroah-Hartman