serial: 8250_of: clear stuck empty-FIFO RX-timeout on LPC32xx
Subsystems (2)
PATCH v4 1/1 serial: 8250_of: clear stuck empty-FIFO RX-timeout on LPC32xx
The NXP LPC32xx UART (PORT_LPC3220) can latch an RX character-timeout interrupt while the RX FIFO is empty: IIR reports UART_IIR_RX_TIMEOUT (0x0c) but LSR.DR is clear. A character timeout is only cleared by reading RHR, but serial8250_rx_chars() reads RHR only when LSR.DR is set, so nothing ever clears the condition. The interrupt is level-triggered and re-fires immediately, so on a single-core ARM926 the resulting interrupt storm livelocks the CPU.
It is reproducible when userspace repeatedly opens the front-panel port (ttyS1): serial8250_do_set_termios() re-enables interrupts on unlock and the handler then spins forever with iir=0xcc lsr=0x60 ier=0x05, tripping the soft-lockup detector in serial8250_handle_irq_locked().
LPC32xx has no dedicated 8250 glue driver, it's driven by the generic 8250_of. Add a hardware specific handle_irq for PORT_LPC3220, wired up in of_platform_serial_setup() the same way fsl8250_handle_irq is installed. The handler follows dw8250_handle_irq(): on an RX timeout with an empty FIFO (LSR.DR and LSR.BI clear) it does one throwaway RHR read to clear the condition, then calls serial8250_handle_irq_locked(). No real received data is ever discarded, and it is a no-op on healthy UARTs which never report a timeout with DR clear.
This is the same class of bug already worked around in other 8250 drivers;
see commit 424d79183af0 ("serial: 8250_dw: Avoid "too much work" from bogus rx timeout interrupt") which reports the identical iir=0xcc/lsr=0x60. See alsoUART_RX_TIMEOUT_QUIRK in 8250_omap, and the note in 8250_bcm7271.
---
Changes in v4:
- Add MODULE_IMPORT_NS("SERIAL_8250") so 8250_of.c builds as a module. Now calls the namespaced serial8250_handle_irq_locked(). Fixes the modular build break reported by Greg KH. Reproduced the failure with CONFIG_SERIAL_OF_PLATFORM=m and confirmed v4 builds cleanly
Changes in v3:
- Move the workaround out of serial8250_handle_irq_locked() into an LPC32xx-specific ->handle_irq in 8250_of.c
- Skip the dummy read when LSR.BI is set, matching dw8250_handle_irq
Changes in v2:
- Drop the PORT_LPC3220 gate and handle the spurious RX timeout generically
v3: https://lore.kernel.org/linux-serial/[email protected]/
v2: https://lore.kernel.org/linux-serial/[email protected]/
v1: https://lore.kernel.org/linux-serial/[email protected]/
drivers/tty/serial/8250/8250_of.c | 38 +++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+)
drivers/tty/serial/8250/8250_of.c +38 -0
@@ -82,6 +82,40 @@ static int of_platform_serial_clk_notifier_cb(struct notifier_block *nb, unsignereturn NOTIFY_DONE ; } +static int lpc32xx_handle_irq (struct uart_port *port) +{ +struct uart_8250_port *up =up_to_u8250p (port); +unsigned int iir; +u16 status; + +guard (serial8250_rpm)(up); + + iir =serial_port_in (port,UART_IIR ); +if (iir &UART_IIR_NO_INT ) +return 0 ; + +guard (uart_port_lock_check_sysrq_irqsave)(port); + +/* +* The LPC32xx UART can assert an RX character-timeout interrupt while +* the RX FIFO is empty: IIR reports UART_IIR_RX_TIMEOUT but LSR.DR is +* clear. The timeout is only cleared by reading RHR, but the core RX +* path skips that read when the FIFO is empty, so the level-triggered +* IRQ re-fires forever and livelocks this single-core SoC. Do one +* throwaway RHR read to clear it; a healthy UART never reports a +* timeout with DR/BI clear, so no received data is ever discarded. +*/ +if ((iir &0x3f ) ==UART_IIR_RX_TIMEOUT ) { + status =serial_lsr_in (up); +if (!(status & (UART_LSR_DR |UART_LSR_BI ))) +serial_port_in (port,UART_RX ); + } + +serial8250_handle_irq_locked (port, iir); + +return 1 ; +} +/* * Fill a struct uart_port for a given device node */
@@ -185,6 +219,9 @@ static int of_platform_serial_setup(struct platform_device *ofdev,case PORT_NPCM : ret =npcm_setup (port);break ; +case PORT_LPC3220 : + port->handle_irq = lpc32xx_handle_irq; +break ;default :/* Nothing to do */ ret =0 ;
@@ -381,6 +418,7 @@ static struct platform_driver of_platform_serial_driver = {module_platform_driver (of_platform_serial_driver); +MODULE_IMPORT_NS ("SERIAL_8250" );MODULE_AUTHOR ("Arnd Bergmann <[email protected]>" );MODULE_LICENSE ("GPL" );MODULE_DESCRIPTION ("Serial Port driver for Open Firmware platform devices" );
base-commit: da7b5fd4e17f8e44c5590f2d603c01d499f056e6
Signature · 1 line
2.25.1
Quoted 22 lines · Ryan Wilbur
The NXP LPC32xx UART (
PORT_LPC3220) can latch an RX character-timeout interrupt while the RX FIFO is empty: IIR reportsUART_IIR_RX_TIMEOUT(0x0c) but LSR.DR is clear. A character timeout is only cleared by reading RHR, butserial8250_rx_chars()reads RHR only when LSR.DR is set, so nothing ever clears the condition. The interrupt is level-triggered and re-fires immediately, so on a single-core ARM926 the resulting interrupt storm livelocks the CPU.It is reproducible when userspace repeatedly opens the front-panel port (ttyS1):
serial8250_do_set_termios()re-enables interrupts on unlock and the handler then spins forever with iir=0xcc lsr=0x60 ier=0x05, tripping the soft-lockup detector inserial8250_handle_irq_locked().LPC32xx has no dedicated 8250 glue driver, it's driven by the generic 8250_of. Add a hardware specific
handle_irqforPORT_LPC3220, wired up inof_platform_serial_setup()the same wayfsl8250_handle_irqis installed. The handler followsdw8250_handle_irq(): on an RX timeout with an empty FIFO (LSR.DR and LSR.BI clear) it does one throwaway RHR read to clear the condition, then callsserial8250_handle_irq_locked(). No real received data is ever discarded, and it is a no-op on healthy UARTs which never report a timeout with DR clear.
Quoted 4 lines
This is the same class of bug already worked around in other 8250 drivers;
see commit424d79183af0("serial: 8250_dw: Avoid "too much work" from bogus rx timeout interrupt") which reports the identical iir=0xcc/lsr=0x60. See alsoUART_RX_TIMEOUT_QUIRKin 8250_omap, and the note in 8250_bcm7271.
As long as this quirk is not anymore in the generic 8250 code, I'm fine with it
Signature · 2 lines
With Best Regards, Andy Shevchenko