If RX is disabled while there are still unprocessed bytes in RX FIFO,
cdns_uart_handle_rx() called from interrupt handler will get stuck in
the receive loop as read bytes will not get removed from the RX FIFO
and CDNS_UART_SR_RXEMPTY bit will never get set.
Avoid the stuck handler by checking first if RX is disabled. port->lock
protects against race with RX-disabling functions.
This HW behavior was mentioned by Nathan Rossi in 43e98facc4a3 ("tty:
xuartps: Fix RX hang, and TX corruption in termios call") which fixed a
similar issue in cdns_uart_set_termios().
The behavior can also be easily verified by e.g. setting
CDNS_UART_CR_RX_DIS at the beginning of cdns_uart_handle_rx() - the
following loop will then get stuck.
Resetting the FIFO using RXRST would not set RXEMPTY either so simply
issuing a reset after RX-disable would not work.
I observe this frequently on a ZynqMP board during heavy RX load at 1M
baudrate when the reader process exits and thus RX gets disabled.
Fixes: 61ec9016988f ("tty/serial: add support for Xilinx PS UART")
Signed-off-by: Anssi Hannula <anssi.hannula@xxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
---
drivers/tty/serial/xilinx_uartps.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c
index 094f2958cb2b..f0c4f59d9314 100644
--- a/drivers/tty/serial/xilinx_uartps.c
+++ b/drivers/tty/serial/xilinx_uartps.c
@@ -219,6 +219,13 @@ static void cdns_uart_handle_rx(void *dev_id, unsigned int isrstatus)
is_rxbs_support = cdns_uart->quirks & CDNS_UART_RXBS_SUPPORT;
+ /*
+ * RXEMPTY will never be set if RX is disabled as read bytes
+ * will not be removed from the FIFO
+ */
+ if (readl(port->membase + CDNS_UART_CR) & CDNS_UART_CR_RX_DIS)
+ return;
+
while ((readl(port->membase + CDNS_UART_SR) &
CDNS_UART_SR_RXEMPTY) != CDNS_UART_SR_RXEMPTY) {
if (is_rxbs_support)
--
2.17.2
