- 17 Apr, 2024 1 commit
-
-
Jiri Slaby (SUSE) authored
Pengfei Xu reports, that the -next commit 1788cf6a (tty: serial: switch from circ_buf to kfifo) tries to emit a WARNING and that leads to lockdep errors. Obviously we cannot print anything from uart_put_char()! This WARN_ON_ONCE() was/is a debug aid to check if the condition in uart_put_char() can happen at all. Pengfei Xu confirmed it can. Unlike me and kbuild bot in my queue. Second, I completely forgot about it, so I did not remove it in the final version, nor mentioned it in the commit log. Drop it now as we are all good. And we even have stack traces (and a reproducer)! Signed-off-by:
Jiri Slaby (SUSE) <jirislaby@kernel.org> Reported-by:
Pengfei Xu <pengfei.xu@intel.com> Link: https://lore.kernel.org/r/20240416071904.21440-1-jirislaby@kernel.orgSigned-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
- 16 Apr, 2024 1 commit
-
-
Jiri Slaby (SUSE) authored
Marek reports, that the -next commit 1788cf6a (tty: serial: switch from circ_buf to kfifo) broke meson_uart and qcom_geni_serial. The commit mistakenly advanced the kfifo twice: once by uart_fifo_get()/kfifo_out() and second time by uart_xmit_advance(). To advance the fifo only once, drop the superfluous uart_xmit_advance() from both. To count the TX statistics properly, use uart_fifo_out() in qcom_geni_serial (meson_uart_start_tx() already uses that). I checked all other uses of uart_xmit_advance() and they appear correct: either they are finishing DMA transfers or are after peek/linear_ptr (i.e. they do not advance fifo). Signed-off-by:
Marek Szyprowski <m.szyprowski@samsung.com> Fixes: 1788cf6a ("tty: serial: switch from circ_buf to kfifo") Cc: Neil Armstrong <neil.armstrong@linaro.org> Cc: Kevin Hilman <khilman@baylibre.com> Cc: Jerome Brunet <jbrunet@baylibre.com> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Link: https://lore.kernel.org/r/20240416054825.6211-1-jirislaby@kernel.orgSigned-off-by:
-
- 13 Apr, 2024 1 commit
-
-
Tony Lindgren authored
If CONFIG_SERIAL_CORE_CONSOLE is not set, we get build errors. Let's fix the issue by moving the endif after the serial base console functions. Fixes: 4547cd76 ("serial: 8250: Fix add preferred console for serial8250_isa_init_ports()") Reported-by:
kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202404131416.VJljwvkS-lkp@intel.com/Signed-off-by:
Tony Lindgren <tony@atomide.com> Link: https://lore.kernel.org/r/20240413103343.24231-1-tony@atomide.comSigned-off-by:
-
- 11 Apr, 2024 10 commits
-
-
Sreenath Vijayan authored
When terminal is unresponsive, one cannot use dmesg to view the printk ring buffer messages. Also, syslog services may be disabled, especially on embedded systems, to check the messages after a reboot. In this scenario, replay the messages in printk ring buffer on consoles via sysrq by pressing sysrq+R. The console loglevel will determine which all kernel log messages are displayed. The messages will be displayed only when console_trylock() succeeds. Users could repeat the sysrq key when it fails. If the owner of console subsystem lock is stuck, repeating the key won't work. Suggested-by:
Petr Mladek <pmladek@suse.com> Signed-off-by:
Shimoyashiki Taichi <taichi.shimoyashiki@sony.com> Reviewed-by:
John Ogness <john.ogness@linutronix.de> Signed-off-by:
Sreenath Vijayan <sreenath.vijayan@sony.com> Link: https://lore.kernel.org/r/cc3b9b1aae60a236c6aed1dc7b0ffa2c7cd1f183.1710220326.git.sreenath.vijayan@sony.comSigned-off-by:
-
Sreenath Vijayan authored
Add a generic function console_replay_all() for replaying the kernel log on consoles, in any context. It would allow viewing the logs on an unresponsive terminal via sysrq. Reuse the existing code from console_flush_on_panic() for resetting the sequence numbers, by introducing a new helper function __console_rewind_all(). It is safe to be called under console_lock(). Try to acquire lock on the console subsystem without waiting. If successful, reset the sequence number to oldest available record on all consoles and call console_unlock() which will automatically flush the messages to the consoles. Suggested-by:
-
Andy Shevchenko authored
As Jiri rightfully pointed out the current to_max3100_port() macro implementation is fragile in a sense that it expects the variable name to be port, otherwise it blow up the build. Change this to be static inline to prevent bad compilation. Suggested-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by:
Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Link: https://lore.kernel.org/r/20240410141135.1378948-1-andriy.shevchenko@linux.intel.comSigned-off-by:
-
Colin Ian King authored
Variable tmout is being assigned a value that is never read, it is being re-assinged in the following for-loop statement. The assignment is redundant and can be removed. Cleans up clang scan warning: drivers/tty/serial/omap-serial.c:1096:3: warning: Value stored to 'tmout' is never read [deadcode.DeadStores] Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Link: https://lore.kernel.org/r/20240411085537.306020-1-colin.i.king@gmail.comSigned-off-by:
-
Tony Lindgren authored
We need to inline serial_base_add_isa_preferred_console() based on CONFIG_SERIAL_8250_CONSOLE and not based on CONFIG_SERIAL_CORE_CONSOLE. Otherwise we can get the follwoing error as noted by Stephen: ERROR: modpost: "serial_base_add_isa_preferred_console" [drivers/tty/serial/8250/8250.ko] undefined! We also have a duplicate inlined serial_base_add_isa_preferred_console(), in serial_base_bus.c added by the same commit by accident, let's drop it. Fixes: a8b04cfe ("serial: 8250: Add preferred console in serial8250_isa_init_ports()") Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
-
Hugo Villeneuve authored
Since each I2C/SPI probe function can modify sc16is7xx_regcfg at the same time, change structure to be constant and do the required modifications on a local copy. Signed-off-by:
Hugo Villeneuve <hvilleneuve@dimonoff.com> Reviewed-by:
Andy Shevchenko <andy@kernel.org> Link: https://lore.kernel.org/r/20240409154253.3043822-6-hugo@hugovil.comSigned-off-by:
-
Hugo Villeneuve authored
Before, sc16is7xx_lines was checked for a free (zero) bit first, and then later it was set only if UART port registration succeeded. Now that sc16is7xx_lines is shared for the I2C and SPI drivers, there is a possibility that the two drivers can simultaneously try to reserve the same line bit at the same time. To prevent this, make sure line allocation is reserved atomically, and use a new variable to hold the status of UART port registration. Now that we no longer need to search if a bit is set, it is now possible to simplify sc16is7xx_lines allocation by using the IDA framework. Link: https://lore.kernel.org/all/20231212150302.a9ec5d085a4ba65e89ca41af@hugovil.com/ Link: https://lore.kernel.org/all/CAHp75VebCZckUrNraYQj9k=Mrn2kbYs1Lx26f5-8rKJ3RXeh-w@mail.gmail.com/Suggested-by:
Andy Shevchenko <andy.shevchenko@gmail.com> Signed-off-by:
-
Hugo Villeneuve authored
Split the common code from sc16is7xx driver and move the I2C and SPI bus parts into interface-specific source files. sc16is7xx becomes the core functions which can support multiple bus interfaces like I2C and SPI. No functional changes intended. Also simplify and improve Kconfig entries. - Capitalize SPI keyword for consistency - Display list of supported ICs each on a separate line (and aligned) to facilitate locating a specific IC model - Add Manufacturer name at start of description string - Add UART keyword to description string Suggested-by: -
Hugo Villeneuve authored
There is no need to check for previous port registration in sc16is7xx_remove() because if sc16is7xx_probe() succeeded, we are guaranteed to have successfully registered both ports. We can thus unconditionally clear the line allocation bit in sc16is7xx_lines. Signed-off-by:
-
Hugo Villeneuve authored
Replace incorrect include with the proper one for sched_set_fifo() declaration. Fixes: 28d2f209 ("sched,serial: Convert to sched_set_fifo()") Signed-off-by:
-
- 09 Apr, 2024 27 commits
-
-
Lino Sanfilippo authored
In case that no RTS GPIO is available do not use a dedicated nullified serial_rs485 struct to disable RS485 support, but simply delete the SER_RS485_ENABLED flag in the ports rs485_supported struct. This make the structure superfluous and it can be removed. Signed-off-by:
Lino Sanfilippo <l.sanfilippo@kunbus.com> Link: https://lore.kernel.org/r/20240407002709.16224-5-l.sanfilippo@kunbus.comSigned-off-by:
-
Lino Sanfilippo authored
The serial core already checks the RS485 RTS settings for sanity, so remove the superfluous check in serial8250_em485_config(). Signed-off-by:
-
Lino Sanfilippo authored
Variable dmacr is only used if DMA is enabled, so move it into the CONFIG_DMA_ENGINE conditional. Signed-off-by:
-
Lino Sanfilippo authored
Due to earlier code changes function pl011_get_rs485_mode() is now merely a wrapper for uart_get_rs485_mode() which does not add any further functionality. So remove it and instead call uart_get_rs485_mode() directly. Reviewed-by:
Lukas Wunner <lukas@wunner.de> Signed-off-by:
-
Andy Shevchenko authored
Sort the headers in alphabetic order in order to ease the maintenance for this part. Signed-off-by:
-
Andy Shevchenko authored
Instead of using container_of() explicitly, introduce a helper macro. This saves a lot of lines of code. Signed-off-by:
-
Andy Shevchenko authored
The SIMPLE_DEV_PM_OPS() is deprecated, replace it with the DEFINE_SIMPLE_DEV_PM_OPS() and use pm_sleep_ptr() for setting the driver's PM routines. Signed-off-by:
-
Andy Shevchenko authored
MODULE_ALIAS() in most cases is a pure hack to avoid placing ID tables. Replace it with the respective ID tables. Signed-off-by:
-
Andy Shevchenko authored
Switch to use dev_err_probe() to simplify the error path and unify a message template. Signed-off-by:
-
Andy Shevchenko authored
The struct uart_port has a copy of the IRQ that is also stored in the private data structure. Remove the duplication in the latter one. Signed-off-by:
-
Andy Shevchenko authored
Get crystal frequency via device property instead of using a platform data. Signed-off-by:
-
Andy Shevchenko authored
Enable or disable loopback at run-time. Signed-off-by:
-
wangkaiyuan authored
The maple tree register cache is based on a much more modern data structure than the rbtree cache and makes optimisation choices which are probably more appropriate for modern systems than those made by the rbtree cache. Signed-off-by:
wangkaiyuan <wangkaiyuan@inspur.com> Reviewed-by:
-
wangkaiyuan authored
The maple tree register cache is based on a much more modern data structure than the rbtree cache and makes optimisation choices which are probably more appropriate for modern systems than those made by the rbtree cache. Signed-off-by:
-
Andy Shevchenko authored
The driver actually supports more than one chip. Update Kconfig entry to list the supported chips. Signed-off-by:
-
Andy Shevchenko authored
There is no code using max3100_work() before the definition of it. Remove unneeded forward declaration. While at it, move max3100_dowork() and max3100_timeout() down in the code to be after actual max3100_work() implementation. Signed-off-by:
-
Andy Shevchenko authored
Serial core provides a standard timeout for polling modem state. Use it instead of a custom approach. Signed-off-by:
-
Andy Shevchenko authored
While there is no user of this callback in the kernel, it also breaks the relationship in the driver model. The correct implementation should be done via GPIO or regulator framework. Remove custom HW shutdown support for good and, if needed, we will implement it correctly later on. Signed-off-by:
-
Andy Shevchenko authored
There is no user of the struct plat_max3100 outside the driver. Inline its contents into the driver. While at it, drop outdated example in the comment. Signed-off-by:
-
Andy Shevchenko authored
Sparse is not happy about misuse of bitwise types: .../max3100.c:194:13: warning: incorrect type in assignment (different base types) .../max3100.c:194:13: expected unsigned short [addressable] [usertype] etx .../max3100.c:194:13: got restricted __be16 [usertype] .../max3100.c:202:15: warning: cast to restricted __be16 Fix this by choosing proper types for the respective variables. Fixes: 7831d56b ("tty: MAX3100") Signed-off-by:
-
Andy Shevchenko authored
The removal of the last MAX3100 device triggers the removal of the driver. However, code doesn't update the respective global variable and after insmod — rmmod — insmod cycle the kernel oopses: max3100 spi-PRP0001:01: max3100_probe: adding port 0 BUG: kernel NULL pointer dereference, address: 0000000000000408 ... RIP: 0010:serial_core_register_port+0xa0/0x840 ... max3100_probe+0x1b6/0x280 [max3100] spi_probe+0x8d/0xb0 Update the actual state so next time UART driver will be registered again. Hugo also noticed, that the error path in the probe also affected by having the variable set, and not cleared. Instead of clearing it move the assignment after the successfull uart_register_driver() call. Fixes: 7831d56b ("tty: MAX3100") Signed-off-by:
-
Andy Shevchenko authored
uart_handle_cts_change() has to be called with port lock taken, Since we run it in a separate work, the lock may not be taken at the time of running. Make sure that it's taken by explicitly doing that. Without it we got a splat: WARNING: CPU: 0 PID: 10 at drivers/tty/serial/serial_core.c:3491 uart_handle_cts_change+0xa6/0xb0 ... Workqueue: max3100-0 max3100_work [max3100] RIP: 0010:uart_handle_cts_change+0xa6/0xb0 ... max3100_handlerx+0xc5/0x110 [max3100] max3100_work+0x12a/0x340 [max3100] Fixes: 7831d56b ("tty: MAX3100") Signed-off-by:
-
Bastien Curutchet authored
The UART's input clock rate can change at runtime but this is not handled by the driver. Add a clock_notifier callback that updates the divisors when the input clock is updated. The serial8250_update_uartclk() is used to do so. PRE_RATE_CHANGE and ABORT_RATE_CHANGE notifications are ignored, only the POST_RATE_CHANGE is used. Not using PRE_RATE_CHANGE notification can result in a few corrupted bytes during frequency transitions but, IMHO, it can be acceptable in many use cases. It has been tested on a DAVINCI/OMAP-L138 processor. Signed-off-by:
Bastien Curutchet <bastien.curutchet@bootlin.com> Reviewed-by:
Herve Codina <herve.codina@bootlin.com> Link: https://lore.kernel.org/r/20240405120552.35991-1-bastien.curutchet@bootlin.comSigned-off-by:
-
Kanak Shilledar authored
Convert the Actions Semi Owl UART to newer DT schema. Created DT schema based on the .txt file which had `compatible`, `reg` and `interrupts` as the required properties. This binding is used by Actions S500, S700 and S900 SoC. S700 and S900 use the same UART compatible string. Reviewed-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by:
Kanak Shilledar <kanakshilledar111@protonmail.com> Link: https://lore.kernel.org/r/20240405080235.11563-1-kanakshilledar@gmail.comSigned-off-by:
-
Uwe Kleine-König authored
There are considerations to drop platform_driver_probe() as a concept that isn't relevant any more today. It comes with an added complexity that makes many users hold it wrong. (E.g. this driver should have marked the driver struct with __refdata to prevent the below mentioned false positive section mismatch warning.) This fixes a W=1 build warning: WARNING: modpost: drivers/tty/serial/pmac_zilog: section mismatch in reference: pmz_driver+0x8 (section: .data) -> pmz_detach (section: .exit.text) Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Link: https://lore.kernel.org/r/5ea3174616abc9fa256f115b4fb175d289ac1754.1711748999.git.u.kleine-koenig@pengutronix.deTested-by:
Michael Ellerman <mpe@ellerman.id.au> Signed-off-by:
-
Uwe Kleine-König authored
As described in the added code comment, a reference to .exit.text is ok for drivers registered via module_platform_driver_probe(). Make this explicit to prevent the following section mismatch warning WARNING: modpost: drivers/tty/amiserial: section mismatch in reference: amiga_serial_driver+0x8 (section: .data) -> amiga_serial_remove (section: .exit.text) that triggers on an allmodconfig W=1 build. Signed-off-by:
-
Uwe Kleine-König authored
This change strips $abs_srctree of the input file containing the character mapping table in the generated output. The motivation for this change is Yocto emitting a build warning WARNING: linux-lxatac-6.7-r0 do_package_qa: QA Issue: File /usr/src/debug/linux-lxatac/6.7-r0/drivers/tty/vt/consolemap_deftbl.c in package linux-lxatac-src contains reference to TMPDIR So this change brings us one step closer to make the build result reproducible independent of the build path. Signed-off-by:
-
