linux-stable.git/drivers/tty, branch v6.6.82

serial: 8250: Fix fifo underflow on flush

2025-02-21T12:57:18+00:00

commit 9e512eaaf8f4008c44ede3dfc0fbc9d9c5118583 upstream.

When flushing the serial port's buffer, uart_flush_buffer() calls
kfifo_reset() but if there is an outstanding DMA transfer then the
completion function will consume data from the kfifo via
uart_xmit_advance(), underflowing and leading to ongoing DMA as the
driver tries to transmit another 2^32 bytes.

This is readily reproduced with serial-generic and amidi sending even
short messages as closing the device on exit will wait for the fifo to
drain and in the underflow case amidi hangs for 30 seconds on exit in
tty_wait_until_sent().  A trace of that gives:

     kworker/1:1-84    [001]    51.769423: bprint:               serial8250_tx_dma: tx_size=3 fifo_len=3
           amidi-763   [001]    51.769460: bprint:               uart_flush_buffer: resetting fifo
 irq/21-fe530000-76    [000]    51.769474: bprint:               __dma_tx_complete: tx_size=3
 irq/21-fe530000-76    [000]    51.769479: bprint:               serial8250_tx_dma: tx_size=4096 fifo_len=4294967293
 irq/21-fe530000-76    [000]    51.781295: bprint:               __dma_tx_complete: tx_size=4096
 irq/21-fe530000-76    [000]    51.781301: bprint:               serial8250_tx_dma: tx_size=4096 fifo_len=4294963197
 irq/21-fe530000-76    [000]    51.793131: bprint:               __dma_tx_complete: tx_size=4096
 irq/21-fe530000-76    [000]    51.793135: bprint:               serial8250_tx_dma: tx_size=4096 fifo_len=4294959101
 irq/21-fe530000-76    [000]    51.804949: bprint:               __dma_tx_complete: tx_size=4096

Since the port lock is held in when the kfifo is reset in
uart_flush_buffer() and in __dma_tx_complete(), adding a flush_buffer
hook to adjust the outstanding DMA byte count is sufficient to avoid the
kfifo underflow.

Fixes: 9ee4b83e51f74 ("serial: 8250: Add support for dmaengine")
Cc: stable 
Signed-off-by: John Keeping 
Link: https://lore.kernel.org/r/20250208124148.1189191-1-jkeeping@inmusicbrands.com
Signed-off-by: Greg Kroah-Hartman

serial: port: Always update ->iotype in __uart_read_properties()

2025-02-21T12:57:18+00:00

commit e8486bd50ecf63c9a1e25271f258a8d959f2672f upstream.

The documentation of the __uart_read_properties() states that
->iotype member is always altered after the function call, but
the code doesn't do that in the case when use_defaults == false
and the value of reg-io-width is unsupported. Make sure the code
follows the documentation.

Note, the current users of the uart_read_and_validate_port_properties()
will fail and the change doesn't affect their behaviour, neither
users of uart_read_port_properties() will be affected since the
alteration happens there even in the current code flow.

Fixes: e894b6005dce ("serial: port: Introduce a common helper to read properties")
Cc: stable 
Signed-off-by: Andy Shevchenko 
Link: https://lore.kernel.org/r/20250124161530.398361-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman

serial: port: Assign ->iotype correctly when ->iobase is set

2025-02-21T12:57:17+00:00

commit 166ac2bba167d575e7146beaa66093bc7c072f43 upstream.

Currently the ->iotype is always assigned to the UPIO_MEM when
the respective property is not found. However, this will not
support the cases when user wants to have UPIO_PORT to be set
or preserved.  Support this scenario by checking ->iobase value
and default the ->iotype respectively.

Fixes: 1117a6fdc7c1 ("serial: 8250_of: Switch to use uart_read_port_properties()")
Fixes: e894b6005dce ("serial: port: Introduce a common helper to read properties")
Cc: stable 
Signed-off-by: Andy Shevchenko 
Link: https://lore.kernel.org/r/20250124161530.398361-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman

tty: xilinx_uartps: split sysrq handling

2025-02-17T08:40:43+00:00

commit b06f388994500297bb91be60ffaf6825ecfd2afe upstream.

lockdep detects the following circular locking dependency:

CPU 0                      CPU 1
========================== ============================
cdns_uart_isr()            printk()
  uart_port_lock(port)       console_lock()
			     cdns_uart_console_write()
                               if (!port->sysrq)
                                 uart_port_lock(port)
  uart_handle_break()
    port->sysrq = ...
  uart_handle_sysrq_char()
    printk()
      console_lock()

The fixed commit attempts to avoid this situation by only taking the
port lock in cdns_uart_console_write if port->sysrq unset. However, if
(as shown above) cdns_uart_console_write runs before port->sysrq is set,
then it will try to take the port lock anyway. This may result in a
deadlock.

Fix this by splitting sysrq handling into two parts. We use the prepare
helper under the port lock and defer handling until we release the lock.

Fixes: 74ea66d4ca06 ("tty: xuartps: Improve sysrq handling")
Signed-off-by: Sean Anderson 
Cc: stable@vger.kernel.org # c980248179d: serial: xilinx_uartps: Use port lock wrappers
Acked-by: John Ogness 
Link: https://lore.kernel.org/r/20250110213822.2107462-1-sean.anderson@linux.dev
Signed-off-by: Sean Anderson 
Signed-off-by: Greg Kroah-Hartman

serial: sh-sci: Do not probe the serial port if its slot in sci_ports[] is in use

2025-02-17T08:40:24+00:00

commit 9f7dea875cc7f9c1a56a5c688290634a59cd1420 upstream.

In the sh-sci driver, sci_ports[0] is used by earlycon. If the earlycon is
still active when sci_probe() is called and the new serial port is supposed
to map to sci_ports[0], return -EBUSY to prevent breaking the earlycon.

This situation should occurs in debug scenarios, and users should be
aware of the potential conflict.

Fixes: 0b0cced19ab1 ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
Cc: stable@vger.kernel.org
Signed-off-by: Claudiu Beznea 
Link: https://lore.kernel.org/r/20250116182249.3828577-4-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Greg Kroah-Hartman

serial: sh-sci: Drop __initdata macro for port_cfg

2025-02-17T08:40:24+00:00

commit eaeee4225dba30bef4d424bdf134a07b7f423e8b upstream.

The port_cfg object is used by serial_console_write(), which serves as
the write function for the earlycon device. Marking port_cfg as __initdata
causes it to be freed after kernel initialization, resulting in earlycon
becoming unavailable thereafter. Remove the __initdata macro from port_cfg
to resolve this issue.

Fixes: 0b0cced19ab1 ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
Cc: stable@vger.kernel.org
Reviewed-by: Geert Uytterhoeven 
Signed-off-by: Claudiu Beznea 
Fixes: 0b0cced19ab15c9e ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
Link: https://lore.kernel.org/r/20250116182249.3828577-2-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Greg Kroah-Hartman

serial: 8250: Adjust the timeout for FIFO mode

2025-02-08T08:52:27+00:00

[ Upstream commit d91f98be26510f5f81ec66425bb0306d1ccd571a ]

After a console has written a record into UART_TX, it uses
wait_for_xmitr() to wait until the data has been sent out before
returning. However, wait_for_xmitr() will timeout after 10ms,
regardless if the data has been transmitted or not.

For single bytes, this timeout is sufficient even at very slow
baud rates, such as 1200bps. However, when FIFO mode is used,
there may be 64 bytes pushed into the FIFO at once. At a baud
rate of 115200bps, the 10ms timeout is still sufficient. But
when using lower baud rates (such as 57600bps), the timeout
is _not_ sufficient. This causes longer lines to be cut off,
resulting in lost and horribly misformatted output on the
console.

When using FIFO mode, take the number of bytes into account to
determine an appropriate maximum timeout. Increasing the timeout
does not affect performance since ideally the timeout never
occurs.

Fixes: 8f3631f0f6eb ("serial/8250: Use fifo in 8250 console driver")
Signed-off-by: John Ogness 
Reviewed-by: Andy Shevchenko 
Reviewed-by: Wander Lairson Costa 
Reviewed-by: Petr Mladek 
Link: https://lore.kernel.org/r/20250107212702.169493-2-john.ogness@linutronix.de
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Sasha Levin

of: remove internal arguments from of_property_for_each_u32()

2025-02-08T08:51:51+00:00

[ Upstream commit 9722c3b66e21ff08aec570d02a97d331087fd70f ]

The of_property_for_each_u32() macro needs five parameters, two of which
are primarily meant as internal variables for the macro itself (in the
for() clause). Yet these two parameters are used by a few drivers, and this
can be considered misuse or at least bad practice.

Now that the kernel uses C11 to build, these two parameters can be avoided
by declaring them internally, thus changing this pattern:

  struct property *prop;
  const __be32 *p;
  u32 val;

  of_property_for_each_u32(np, "xyz", prop, p, val) { ... }

to this:

  u32 val;

  of_property_for_each_u32(np, "xyz", val) { ... }

However two variables cannot be declared in the for clause even with C11,
so declare one struct that contain the two variables we actually need. As
the variables inside this struct are not meant to be used by users of this
macro, give the struct instance the noticeable name "_it" so it is visible
during code reviews, helping to avoid new code to use it directly.

Most usages are trivially converted as they do not use those two
parameters, as expected. The non-trivial cases are:

 - drivers/clk/clk.c, of_clk_get_parent_name(): easily doable anyway
 - drivers/clk/clk-si5351.c, si5351_dt_parse(): this is more complex as the
   checks had to be replicated in a different way, making code more verbose
   and somewhat uglier, but I refrained from a full rework to keep as much
   of the original code untouched having no hardware to test my changes

All the changes have been build tested. The few for which I have the
hardware have been runtime-tested too.

Reviewed-by: Andre Przywara  # drivers/clk/sunxi/clk-simple-gates.c, drivers/clk/sunxi/clk-sun8i-bus-gates.c
Acked-by: Bartosz Golaszewski  # drivers/gpio/gpio-brcmstb.c
Acked-by: Nicolas Ferre  # drivers/irqchip/irq-atmel-aic-common.c
Acked-by: Jonathan Cameron  # drivers/iio/adc/ti_am335x_adc.c
Acked-by: Uwe Kleine-König  # drivers/pwm/pwm-samsung.c
Acked-by: Richard Leitner  # drivers/usb/misc/usb251xb.c
Acked-by: Mark Brown  # sound/soc/codecs/arizona.c
Reviewed-by: Richard Fitzgerald  # sound/soc/codecs/arizona.c
Acked-by: Michael Ellerman  # arch/powerpc/sysdev/xive/spapr.c
Acked-by: Stephen Boyd  # clk
Signed-off-by: Luca Ceresoli 
Acked-by: Lee Jones 
Link: https://lore.kernel.org/r/20240724-of_property_for_each_u32-v3-1-bea82ce429e2@bootlin.com
Signed-off-by: Rob Herring (Arm) 
Stable-dep-of: 28fa3291cad1 ("clk: fix an OF node reference leak in of_clk_get_parent_name()")
Signed-off-by: Sasha Levin

serial: sc16is7xx: use device_property APIs when configuring irda mode

2025-02-08T08:51:50+00:00

[ Upstream commit 1a0a2a1e57aa8dcdae25229f6e95cf6bd4817faf ]

Convert driver to be property provider agnostic and allow it to be
used on non-OF platforms.

Signed-off-by: Hugo Villeneuve 
Link: https://lore.kernel.org/r/20230927160153.2717788-2-hugo@hugovil.com
Signed-off-by: Greg Kroah-Hartman 
Stable-dep-of: 28fa3291cad1 ("clk: fix an OF node reference leak in of_clk_get_parent_name()")
Signed-off-by: Sasha Levin

tty: serial: 8250: Fix another runtime PM usage counter underflow

2025-01-17T12:36:21+00:00

commit ed2761958ad77e54791802b07095786150eab844 upstream.

The commit f9b11229b79c ("serial: 8250: Fix PM usage_count for console
handover") fixed one runtime PM usage counter balance problem that
occurs because .dev is not set during univ8250 setup preventing call to
pm_runtime_get_sync(). Later, univ8250_console_exit() will trigger the
runtime PM usage counter underflow as .dev is already set at that time.

Call pm_runtime_get_sync() to balance the RPM usage counter also in
serial8250_register_8250_port() before trying to add the port.

Reported-by: Borislav Petkov (AMD) 
Fixes: bedb404e91bb ("serial: 8250_port: Don't use power management for kernel console")
Cc: stable 
Tested-by: Borislav Petkov (AMD) 
Signed-off-by: Ilpo Järvinen 
Link: https://lore.kernel.org/r/20241210170120.2231-1-ilpo.jarvinen@linux.intel.com
Signed-off-by: Greg Kroah-Hartman