- 10 Jun, 2022 40 commits
-
-
Ilpo Järvinen authored
When tty is not read from, XON/XOFF may get stuck into an intermediate buffer. As those characters are there to do software flow-control, it is not very useful. In the case where neither end reads from ttys, the receiving ends might not be able receive the XOFF characters and just keep sending more data to the opposite direction. This problem is almost guaranteed to occur with DMA which sends data in large chunks. If TTY is slow to process characters, that is, eats less than given amount in receive_buf, invoke lookahead for the rest of the chars to process potential XON/XOFF characters. We need to keep track of how many characters have been processed by the lookahead to avoid processing the flow control char again on the normal path. Bookkeeping occurs parallel on two layers (tty_buffer and n_tty) to avoid passing the lookahead_count through the whole call chain. When a flow-control char is processed, two things must occur: a) it must not be treated as normal char b) if not yet processed, flow-control actions need to be taken The return value of n_tty_receive_char_flow_ctrl() tells caller a), and b) is kept internal to n_tty_receive_char_flow_ctrl(). If characters were previous looked ahead, __receive_buf() makes two calls to the appropriate n_tty_receive_buf_* function. First call is made with lookahead_done=true for the characters that were subject to lookahead earlier and then with lookahead=false for the new characters. Either of the calls might be skipped when it has no characters to handle. Reported-by:
Gilles Buloz <gilles.buloz@kontron.com> Reviewed-by:
Andy Shevchenko <andy.shevchenko@gmail.com> Signed-off-by:
Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Link: https://lore.kernel.org/r/20220606153652.63554-2-ilpo.jarvinen@linux.intel.comSigned-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization. When em485 init fails, there are two possible paths of entry: 1) uart_rs485_config (init path) that fully clears port->rs485 on error. 2) ioctl path with a pre-existing, valid port->rs485 unto which the kernel falls back on error and port->rs485 should therefore be kept untouched. The temporary rs485 struct is not returned to userspace in case of error so its flag don't matter. ...Thus SER_RS485_ENABLED clearing on error can/should be dropped. Signed-off-by: -
Ilpo Järvinen authored
Serial core handles serial_rs485 assignment. It is safe to remove this assignment because sc16is7xx_reg_proc() takes port.lock at start (and sc16is7xx_reconf_rs485() would too). Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 assignment. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 assignment. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization. Signed-off-by:
-
Ilpo Järvinen authored
The driver provides different rs485_supported for the case where RTS is not available making it unnecessary to handle it in imx_uart_rs485_config. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization. Remove custom sanitization from lpuart_config_rs485. This change loses dev_err when SER_RS485_RX_DURING_TX is set due to incorrect configuration. Other drivers do not do similar prinout for full-duplex case and it should be done in serial core if it is desirable to notify on this condition. Personally, I doesn't see it important because the kernel gracefully downgrades to half-duplex. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization and copying rs485 struct. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization and copying rs485 struct. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization and copying rs485 struct. Signed-off-by:
-
Ilpo Järvinen authored
Serial core handles serial_rs485 sanitization and rs485 struct assignment. As serial_rs485 is already clear for the non-RS485 case by serial core, there no need to clear flags in the driver. Signed-off-by:
-
Ilpo Järvinen authored
In order to be add new flags more cleanly and safely, return -EINVAL from TIOCSRS485 ioctl for the flags bits which are not among the current legacy ones. This might cause a regression for userspace as those non-flag bits do not currently trigger -EINVAL. However, it would only occur if the userspace is sending garbage bits so perhaps we'll get away with this change. Signed-off-by:
-
Ilpo Järvinen authored
When SER_RS485_ENABLED is not set, having any other flag/field set in serial_rs485 struct does not have an effect different from not having them set. Thus, make the serial_rs485 struct also match the behavior for all flags, not just SER_RS485_ENABLED. Some drivers do similar clearing of rs485 struct in their rs485_config() already, but not all. This change makes the behavior consistent across drivers. Don't try to validate rs485 struct further when no RS485 is requested, this silences some bogus warnings. This change has (minor) userspace visible impact. Signed-off-by:
-
Ilpo Järvinen authored
Sanitize serial_rs485 struct before calling into rs485_setup. The drivers provide supported_rs485 to help sanitization of the fields. If neither of SER_RS485_RTS_ON_SEND or SER_RS485_RTS_AFTER_SEND supported, don't pretend they can be set to sane settings but clear them both instead. If only one of them is supported it may look tempting to use the one driver supports to set the other, however, the userspace does not have that information readily available so it wouldn't be helpful. While adjusting the documentation, remove also the claim that TIOCGRS485 would call driver specific code. In reality, it does nothing else than copies the stored serial_rs485 structure from uart_port to userspace. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. This driver does not support delay_rts_after_send but the pre-existing behavior is to return -EINVAL if delay_rts_after_send is non-zero. In contrast, other drivers that do not support delay_rts_after_send either zero delay_rts_after_send or do not care (leave the inaccurate value). As changing this would cause userspace visible impact, the change is not attempted here. But perhaps it should be still tried (maybe nobody finds that kind of API oddity significant)? Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. In the case where RTS is lacking, RS485 cannot be enabled so provide zero rs485_supported for that case. Perhaps it would make sense to not provide rs485_config() at all in that case but such a change would have userspace visible impact/change in behavior so this patch does not attempt it. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. In the case where RTS is lacking, RS485 cannot be enabled so provide zero rs485_supported for that case. Perhaps it would make sense to not provide rs485_config() at all in that case but such a change would have userspace visible impact/change in behavior so this patch does not attempt it. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
8250_of uses em485, fill in rs485_supported accordingly. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Differentiate based on which port is in question. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. Signed-off-by:
-
Ilpo Järvinen authored
Add information on supported serial_rs485 features. When the driver is using em485, take advantage of serial8250_em485_supported. Signed-off-by:
-
Ilpo Järvinen authored
bcm2835aux uses em485, fill in rs485_supported accordingly. Signed-off-by:
-
Ilpo Järvinen authored
Create serial8250_em485_supported for the serial_rs485 features supported by the em485 framework. Signed-off-by:
-
Ilpo Järvinen authored
Preparing to move serial_rs485 struct sanitization into serial core, each driver has to provide what fields/flags it supports. This information is pointed into by rs485_supported. Signed-off-by:
-
Ilpo Järvinen authored
Add uart_sanitize_serial_rs485() function for sanitizing serial_rs485 structure fields. No functional changes. Signed-off-by:
-
Ilpo Järvinen authored
A few serial drivers make a call to rs485_config() themselves (all these seem to relate to init). Convert them all to use a common helper which makes it easy to make adjustments on tasks related to it as serial_rs485 struct sanitization is going to be added. In pci_fintek_setup() (in 8250_pci.c), the rs485_config() call was made with NULL, however, it can be changed to pass uart_port's rs485 struct. No other callers should pass NULL into rs485_config() so the NULL check can now be eliminated. Signed-off-by:
-
Ilpo Järvinen authored
Make sure LSR flags are preserved in dw8250_tx_wait_empty(). This function is called from a low-level out function and therefore cannot call serial_lsr_in() as it would lead to infinite recursion. It is borderline if the flags need to be saved here at all since this code relates to writing LCR register which usually implies no important characters should be arriving. Fixes: 914eaf93 ("serial: 8250_dw: Allow TX FIFO to drain before writing to UART_LCR") Reviewed-by:
-
Ilpo Järvinen authored
dw8250_handle_irq() reads LSR under a few conditions, convert both to use serial_lsr_in() in order to preserve LSR flags properly across reads. Fixes: 424d7918 ("serial: 8250_dw: Avoid "too much work" from bogus rx timeout interrupt") Fixes: aa63d786 ("serial: 8250: dw: Add support for DMA flow controlling devices") Cc: Douglas Anderson <dianders@chromium.org> Cc: Phil Edworthy <phil.edworthy@renesas.com> Cc: Miquel Raynal <miquel.raynal@bootlin.com> Reviewed-by:
-
Ilpo Järvinen authored
serial8250_rx_chars() has max_count based character limit. If it triggers, the function returns the old LSR value (and it has never returned only flags which were not handled). Adjust the comment to match behavior and warn about which flags can be depended on. I'd have moved LSR read before LSR read and used serial_lsr_in() also here but I came across an old discussion about the topic. That discussion generated commit d22f8f10 ("serial: 8250: Fix lost rx state") so I left the code as it is (it works as long as the callers only use a subset of the LSR flags which holds true today) and changed the comment instead. Link: https://www.spinics.net/lists/linux-serial/msg16220.htmlReviewed-by:
-
