- 22 Apr, 2022 23 commits
-
-
Wander Lairson Costa authored
Note: I am using a small test app + driver located at [0] for the problem description. serco is a driver whose write function dispatches to the serial controller. sertest is a user-mode app that writes n bytes to the serial console using the serco driver. While investigating a bug in the RHEL kernel, I noticed that the serial console throughput is way below the configured speed of 115200 bps in a HP Proliant DL380 Gen9. I was expecting something above 10KB/s, but I got 2.5KB/s. $ time ./sertest -n 2500 /tmp/serco real 0m0.997s user 0m0.000s sys 0m0.997s With the help of the function tracer, I then noticed the serial controller was taking around 410us seconds to dispatch one single byte: $ trace-cmd record -p function_graph -g serial8250_console_write \ ./sertest -n 1 /tmp/serco $ trace-cmd report | serial8250_console_write() { 0.384 us | _raw_spin_lock_irqsave(); 1.836 us | io_serial_in(); 1.667 us | io_serial_out(); | uart_console_write() { | serial8250_console_putchar() { | wait_for_xmitr() { 1.870 us | io_serial_in(); 2.238 us | } 1.737 us | io_serial_out(); 4.318 us | } 4.675 us | } | wait_for_xmitr() { 1.635 us | io_serial_in(); | __const_udelay() { 1.125 us | delay_tsc(); 1.429 us | } ... ... ... 1.683 us | io_serial_in(); | __const_udelay() { 1.248 us | delay_tsc(); 1.486 us | } 1.671 us | io_serial_in(); 411.342 us | } In another machine, I measured a throughput of 11.5KB/s, with the serial controller taking between 80-90us to send each byte. That matches the expected throughput for a configuration of 115200 bps. This patch changes the serial8250_console_write to use the 16550 fifo if available. In my benchmarks I got around 25% improvement in the slow machine, and no performance penalty in the fast machine. Signed-off-by:Wander Lairson Costa <wander@redhat.com> Link: https://lore.kernel.org/r/20220411174841.34936-2-wander@redhat.comSigned-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Duoming Zhou authored
There is a deadlock in sa1100_set_termios(), which is shown below: (Thread 1) | (Thread 2) | sa1100_enable_ms() sa1100_set_termios() | mod_timer() spin_lock_irqsave() //(1) | (wait a time) ... | sa1100_timeout() del_timer_sync() | spin_lock_irqsave() //(2) (wait timer to stop) | ... We hold sport->port.lock in position (1) of thread 1 and use del_timer_sync() to wait timer to stop, but timer handler also need sport->port.lock in position (2) of thread 2. As a result, sa1100_set_termios() will block forever. This patch moves del_timer_sync() before spin_lock_irqsave() in order to prevent the deadlock. Signed-off-by:Duoming Zhou <duoming@zju.edu.cn> Link: https://lore.kernel.org/r/20220417111626.7802-1-duoming@zju.edu.cnSigned-off-by:
-
Junwen Wu authored
The definition of sysrq_key_table's elements, like sysrq_thaw_op and sysrq_showallcpus_op are not consistent with sysrq_ftrace_dump_op, Consistency makes code more readable. Reviewed-by:
Jiri Slaby <jirislaby@kernel.org> Signed-off-by:
Junwen Wu <wudaemon@gmail.com> Link: https://lore.kernel.org/r/20220418153703.97705-1-wudaemon@163.comSigned-off-by:
-
Sherry Sun authored
No need to initialize the count variable in lpuart_copy_rx_to_tty(), so let's remove it here. Reviewed-by:
Sherry Sun <sherry.sun@nxp.com> Link: https://lore.kernel.org/r/20220418021844.29591-1-sherry.sun@nxp.comSigned-off-by:
-
Jiri Slaby authored
Some more serial drivers can be compile-tested under certain circumstances (when building a specific architecture). So allow for that. This reduces the need of zillion mach/subarch-specific configs. And since the 0day bot has only allmodconfig's for some archs, this increases build coverage there too. Note that cpm needs a minor update in the header, so that it drags in at least some defines (CPM2 ones). Signed-off-by:
Jiri Slaby <jslaby@suse.cz> Link: https://lore.kernel.org/r/20220421101708.5640-8-jslaby@suse.czSigned-off-by:
-
Jiri Slaby authored
pic32_uart contains this: #ifdef CONFIG_SERIAL_PIC32_CONSOLE ... console_initcall(pic32_console_init); ... core_initcall(pic32_late_console_init); ... #endif ... arch_initcall(pic32_uart_init); When the driver is built as module, all three above become module_init(). So if SERIAL_PIC32_CONSOLE is set while SERIAL_PIC32=m, it results in the following build error: In file included from include/linux/device/driver.h:21, from include/linux/device.h:32, from include/linux/platform_device.h:13, from drivers/tty/serial/pic32_uart.c:12: include/linux/module.h:131:49: error: redefinition of '__inittest' So make sure SERIAL_PIC32_CONSOLE can be set only when SERIAL_PIC32=y -- similar as for other drivers. Signed-off-by: -
Jiri Slaby authored
The code wants to know if the circ buffer is empty, so use the proper macro. No functional change intended, just saner function name used for that use case. Cc: Andy Gross <agross@kernel.org> Cc: Bjorn Andersson <bjorn.andersson@linaro.org> Signed-off-by:
-
Jiri Slaby authored
struct uart_port::membase is declared as a pointer. So it should be initialized by NULL, not zero constant. Signed-off-by:
-
Jiri Slaby authored
Cache port->state->xmit into a local variable (xmit) in cdns_uart_handle_tx(). This reduces length of some lines there significantly. I.e. makes the code more readable. Signed-off-by:
-
Jiri Slaby authored
Return from the true branch of the 'if'. This saves one indentation level and makes the code more readable. The two comments about what obvious code does are removed too. Signed-off-by:
-
Tom Rix authored
Smatch reports this issue sunplus-uart.c:501:26: warning: symbol 'sunplus_console_ports' was not declared. Should it be static? sunplus_console_ports is only used in sunplus-uart.c so change its storage-class specifier to static Reviewed-by:
Tom Rix <trix@redhat.com> Link: https://lore.kernel.org/r/20220421152505.1531507-1-trix@redhat.comSigned-off-by:
-
Yu Tu authored
Make UART driver compatible with S4 SOC UART. Meanwhile, the S4 SOC UART uses 12MHz as the clock source for baud rate calculations. Reviewed-by:
Neil Armstrong <narmstrong@baylibre.com> Signed-off-by:
Yu Tu <yu.tu@amlogic.com> Link: https://lore.kernel.org/r/20220422111320.19234-3-yu.tu@amlogic.comSigned-off-by:
-
Yu Tu authored
tty: serial: meson: Add a 12MHz internal clock rate to calculate baud rate in order to meet the baud rate requirements of special BT modules A /2 divider over XTAL was introduced since G12A, and is preferred to be used over the still present /3 divider since it provides much closer frequencies vs the request baudrate. Especially the BT module uses 3Mhz baud rate. 8Mhz calculations can lead to baud rate bias, causing some problems. Reviewed-by:
-
Jiri Slaby authored
Some members of struct icom_port are completely unused or only set and never read. Remove all those. Signed-off-by:
-
Jiri Slaby authored
uart_ops::release_port() and uart_ops::request_port() are not required by the serial layer. So no need to define empty ones. Remove them. Signed-off-by:
-
Jiri Slaby authored
Use list_for_each_entry() helper instead of explicit combo of list_for_each() and list_entry(). Note that pos is used as a reference point in list_add_tail() in icom_alloc_adapter(). This functionality remains as with an empty list, cur_adapter_entry->icom_adapter_entry is still the list head. This simplifies the code a bit. Signed-off-by:
-
Jiri Slaby authored
The baud rates are unsigned constants. So mark them as such. Not only it makes sense, but they are passed also to uart_get_baud_rate() and that expects unsigned int as baud rates on input. Signed-off-by:
-
Jiri Slaby authored
Signed-off-by:
-
Jiri Slaby authored
There is no point keeping the header content separated. The header was not even protected against double inclusion. So move the content to the appropriate source file. Signed-off-by:
-
Jiri Slaby authored
There is a lot of sparse warnings: .../icom.c:228:30: warning: cast from restricted __le16 .../icom.c:232:66: warning: incorrect type in assignment (different base types) .../icom.c:232:66: expected unsigned int [usertype] leBuffer .../icom.c:232:66: got restricted __le32 [usertype] .../icom.c:237:30: warning: cast from restricted __le16 ... .../icom.c:1228:22: warning: cast from restricted __le16 And they are correct. So sort them all out by using proper __leXX and uXX types and the right direction of conversion: le16_to_cpu() instead of cpu_to_le16(), where appropriate. Signed-off-by:
-
Jiri Slaby authored
Integrate both the to_icom_adapter() macro and icom_kref_release() wrapper into icom_remove_adapter(). (And keep it icom_kref_release() name.) It makes the code easier to follow without complex indirections. Signed-off-by:
-
Jiri Slaby authored
In icom, there is an ICOM_PORT macro to perform upcasts from struct uart_port to struct icom_port. It's not completely safe and it works only because the first member of icom_port is uart_port. Nowadays, we use container_of for such an upcast instead. So introduce a helper (to_icom_port()) with container_of in it and convert all the ICOM_PORT users to the new helper. Apart from the code and type safety, it's also clear what icom_port (the variable) is. Unlike with the old ICOM_PORT (the macro with the cast). Signed-off-by:
-
Jiri Slaby authored
It's unused, so remove the macro. Signed-off-by:
-
- 21 Apr, 2022 1 commit
-
-
Randy Dunlap authored
The moxa-smartio.rst file was moved to a different location, so update the MAINTAINERS file entry for it. Quietens one build warning: MAINTAINERS:29715: WARNING: unknown document: ../driver-api/serial/moxa-smartio Fixes: 5b437ae9 ("Documentation: tty: move moxa-smartio.rst to tty") Cc: Jiri Slaby <jirislaby@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Reviewed-by:
Randy Dunlap <rdunlap@infradead.org> Link: https://lore.kernel.org/r/20220420235545.17081-1-rdunlap@infradead.orgSigned-off-by:
-
- 20 Apr, 2022 4 commits
-
-
Daniel Starke authored
n_gsm is based on the 3GPP 07.010 and its newer version is the 3GPP 27.010. See https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=1516 The changes from 07.010 to 27.010 are non-functional. Therefore, I refer to the newer 27.010 here. Chapter 5.2.1.2 describes the encoding of the address field within the frame header. It is made up of the DLCI address, command/response (CR) bit and EA bit. Use the predefined CR value instead of a plain 2 in alignment to the remaining code and to make the encoding obvious. Signed-off-by:
Daniel Starke <daniel.starke@siemens.com> Link: https://lore.kernel.org/r/20220420101346.3315-3-daniel.starke@siemens.comSigned-off-by:
-
Daniel Starke authored
Remove commented out code as it is never used and if anyone accidentally turned it on, it would be broken. Signed-off-by:
-
Ilpo Järvinen authored
This partially reverts commit f6f58610. The code added by that commit containted math overflow for 32-bit archs. In addition, the approach used in it is unnecessarily complicated requiring a dedicated timer just for notemt. A simpler approach for providing UART_CAP_NOTEMT already exists (patches 1-2): https://lore.kernel.org/linux-serial/20220411083321.9131-3-ilpo.jarvinen@linux.intel.com/T/#u Thus, simply revert the UART_CAP_NOTEMT change for now. There were two driver changes within the patch series adding UART_CAP_NOTEMT taking advantage of the newly added flag. This does not revert the driver changes and therefore also UART_CAP_NOTEMT define has to remain. UART_CAP_NOTEMT remains no-op until support is again added. Fixes: f6f58610 ("serial: 8250: Handle UART without interrupt on TEMT using em485") Signed-off-by:
Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Link: https://lore.kernel.org/r/5f874142-fb1f-bff7-f33-fac823e65e2e@linux.intel.comSigned-off-by:
-
Jiri Slaby authored
The commit below causes the following warnings: Documentation/driver-api/index.rst:14: WARNING: toctree contains reference to nonexisting document 'driver-api/tty' Documentation/driver-api/tty/index.rst: WARNING: document isn't included in any toctree This is because I omitted the "/index" part from the link. That part is needed for directories and I was confused by the surrounding links which are links to files. Fix it in this patch. Fixes: b96cd8b0 ("Documentation: move tty to driver-api") Cc: Jonathan Corbet <corbet@lwn.net> Reported-by:
-
- 15 Apr, 2022 12 commits
-
-
Sherry Sun authored
Now fsl_lpuart driver use both of_alias_get_id() and ida_simple_get() in .probe(), which has the potential bug. For example, when remove the lpuart7 alias in dts, of_alias_get_id() will return error, then call ida_simple_get() to allocate the id 0 for lpuart7, this may confilct with the lpuart4 which has alias 0. aliases { ... serial0 = &lpuart4; serial1 = &lpuart5; serial2 = &lpuart6; serial3 = &lpuart7; } So remove the ida_simple_get() in .probe(), return an error directly when calling of_alias_get_id() fails, which is consistent with other uart drivers behavior. Fixes: 3bc3206e ("serial: fsl_lpuart: Remove the alias node dependence") Signed-off-by: -
Al Cooper authored
When flow control is enabled, the UART should set RTS to false during suspend to stop incoming data. Currently, the suspend routine sets the mctrl register in the uart to zero, but leaves the shadow version in the uart_port struct alone so that resume can restore it. This causes a problem later in suspend when serial8250_do_shutdown() is called which uses the shadow mctrl register to clear some additional bits but ends up restoring RTS. The solution is to clear RTS from the shadow version before serial8250_do_shutdown() is called and restore it after. Signed-off-by:
Al Cooper <alcooperx@comcast.net> Link: https://lore.kernel.org/r/20220324145620.41573-1-alcooperx@gmail.comSigned-off-by:
-
Eric Tremblay authored
Signed-off-by:
Eric Tremblay <etremblay@distech-controls.com> Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Link: https://lore.kernel.org/r/20220330104642.229507-4-u.kleine-koenig@pengutronix.deSigned-off-by:
-
Eric Tremblay authored
The Freescale variant of the 16550A doesn't have an interrupt on TEMT available when using the FIFO mode. Signed-off-by:
-
Eric Tremblay authored
Introduce the UART_CAP_NOTEMT capability. The capability indicates that the UART doesn't have an interrupt available on TEMT. In the case where the device does not support it, we calculate the maximum time it could take for the transmitter to empty the shift register. When we get in the situation where we get the THRE interrupt, we check if the TEMT bit is set. If it's not, we start the a timer and recall __stop_tx() after the delay. The transmit sequence is a bit modified when the capability is set. The new timer is used between the last interrupt(THRE) and a potential stop_tx timer. Signed-off-by:
Giulio Benetti <giulio.benetti@micronovasrl.com> [moved to use added UART_CAP_TEMT] Signed-off-by:
Heiko Stuebner <heiko.stuebner@theobroma-systems.com> [moved to use added UART_CAP_NOTEMT, improve timeout] Signed-off-by:
-
Miaoqian Lin authored
platform_get_resource() may fail and return NULL, so we should better check it's return value to avoid a NULL pointer dereference. Fixes: 54da3e38 ("serial: 8250_aspeed_vuart: use UPF_IOREMAP to set up register mapping") Signed-off-by:
Miaoqian Lin <linmq006@gmail.com> Link: https://lore.kernel.org/r/20220404143842.16960-1-linmq006@gmail.comSigned-off-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already assigns the passed serial_rs485 struct to the uart port. So remove the assignment from the drivers rs485_config() function to avoid redundancy. Reviewed-by:
Claudiu Beznea <claudiu.beznea@microchip.com> Acked-by:
Richard Genoud <richard.genoud@gmail.com> Signed-off-by:
Lino Sanfilippo <LinoSanfilippo@gmx.de> Link: https://lore.kernel.org/r/20220410104642.32195-10-LinoSanfilippo@gmx.deSigned-off-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already ensures that only one of both options RTS on send or RTS after send is set. It also assigns the passed serial_rs485 struct to the uart port. So remove the check and the assignment from the drivers rs485_config() function to avoid redundancy. Signed-off-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already assigns the passed serial_rs485 struct to the uart port. So remove the assignment in the drivers rs485_config() function to avoid reduncancy. Acked-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already nullifies the padding field of the passed serial_rs485 struct before returning it to userspace. Doing the same in the drivers rs485_config() function is redundant, so remove the concerning memset in this function. Signed-off-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already clamps the RTS delays. It also assigns the passed serial_rs485 struct to the uart port. So remove these tasks from the drivers rs485_config() function to avoid redundancy. Signed-off-by:
-
Lino Sanfilippo authored
In uart_set_rs485_config() the serial core already ensures that only one of both options RTS on send or RTS after send is set. So remove this check from the drivers rs485_config() function to avoid redundancy. Signed-off-by:
-
