- 27 Sep, 2016 3 commits
-
-
Anirudha Sarangi authored
The existing interrupt handling logic has following issues. - Upon a parity error with default configuration, the control never comes out of the ISR thereby hanging Linux. - The error handling logic around framing and parity error are buggy. There are chances that the errors will never be captured. This patch ensures that the status registers are cleared on all cases so that a hang situation never arises. Signed-off-by:
Anirudha Sarangi <anirudh@xilinx.com> Signed-off-by:
Michal Simek <michal.simek@xilinx.com> [stelford@cadence.com: cherry picked from https://github.com/Xilinx/linux-xlnx commit ac297e20d399850d7a8e373b6eccf2e183c15165 with manual conflict resolution] Signed-off-by:
Scott Telford <stelford@cadence.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Arnd Bergmann authored
Building this driver with a 64-bit dma_addr_t type results in a compiler warning: drivers/tty/serial/stm32-usart.c: In function 'stm32_of_dma_rx_probe': drivers/tty/serial/stm32-usart.c:746:20: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] drivers/tty/serial/stm32-usart.c: In function 'stm32_of_dma_tx_probe': drivers/tty/serial/stm32-usart.c:818:20: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] While the type conversion here is harmless, this hints at a different problem: we pass an __iomem pointer into a DMA engine, which expects a phys_addr_t. This happens to work because stm32 has no MMU and ioremap() is an identity mapping here, but it's still an incorrect API use. Using dma_addr_t is doubly wrong here, because that would be the result of dma_map_single() rather than the physical address. Using the mapbase instead fixes multiple issues: - the warning is gone - we don't go through ioremap in error - the cast is gone, making it use the correct resource_size_t/phys_addr_t type in the process. Fixes: 34891872 ("serial: stm32: adding dma support") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Reviewed-by:
Gerald Baeza <gerald.baeza@st.com> Signed-off-by:
-
Alexey Starikovskiy authored
The problem with previous code was it rounded values in wrong place and produced wrong baud rate in some cases. Signed-off-by:
Alexey Starikovskiy <aystarik@gmail.com> [nicolas.ferre@atmel.com: port to newer kernel and add commit log] Signed-off-by:
Nicolas Ferre <nicolas.ferre@atmel.com> Reviewed-by:
Boris Brezillon <boris.brezillon@free-electrons.com> Reviewed-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
-
- 23 Sep, 2016 6 commits
-
-
Eric W. Biederman authored
In 99.99% of the cases only root in a user namespace can mount /dev/pts and in those cases the owner of /dev/pts/ptmx will remain root.root In the oddball case where someone else has CAP_SYS_ADMIN this code modifies the /dev/pts mount code to use current_fsuid and current_fsgid as the values to use when creating the /dev/ptmx inode. As is done when any other file is created. This is a code simplification, and it allows running without a root user entirely. Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Eric W. Biederman authored
devpts does not and never will have anything to sync so don't bother calling sync_filesystems on remount. Signed-off-by:
-
Eric W. Biederman authored
Signed-off-by:
-
Eric W. Biederman authored
Now that all of the work of setting up a superblock has been moved to devpts_fill_super simplify devpts_mount by calling mount_nodev instead of rolling mount_nodev by hand. Signed-off-by:
-
Eric W. Biederman authored
The code makes more sense here and things are just clearer. Signed-off-by:
-
Eric W. Biederman authored
Signed-off-by:
-
- 22 Sep, 2016 22 commits
-
-
Anirudha Sarangi authored
The patch makes changes not to enable parity error interrupt. With the current implementation, each parity error results in two distinct interrupts (almost always). The first one is normal parity error interrupt with no data in the fifo and the second one is a proper Rx interrupt with the received data in the fifo. By disabling parity error interrupt we still ensure handling of parity errors as for the Rx fifo interrupt the parity error still shows up in the interrupt status register. Considering the fact that the by default INPCK and IGNPAR are not set, this is the optimal implementation for parity error handling. Signed-off-by:
-
Alexandre TORGUE authored
Signed-off-by:
Alexandre TORGUE <alexandre.torgue@st.com> Signed-off-by:
-
Alexandre TORGUE authored
Signed-off-by:
-
Alexandre TORGUE authored
This patch adds dma mode support for rx and tx with pio mode as fallback in case of dma error. Signed-off-by:
-
Alexandre TORGUE authored
Signed-off-by:
-
Alexandre TORGUE authored
Keep the clock enabled at the end of stm32_init_port but disable it in stm32_serial_remove. Note that stm32_pm function is there to manage the clock at runtime. Signed-off-by:
-
Alexandre TORGUE authored
"st,hw-flow-ctrl" property is documented in device tree binding whereas "auto-flow-control" was used in the code. The driver is now aligned with the binding name "st,hw-flow-ctrl". Signed-off-by:
-
Alexandre TORGUE authored
Signed-off-by:
-
Alexandre TORGUE authored
Signed-off-by:
-
Alexandre TORGUE authored
This adds documentation of device tree bindings for the STM32 USART Signed-off-by:
Maxime Coquelin <mcoquelin.stm32@gmail.com> Signed-off-by:
-
Alexandre TORGUE authored
Register offset management rework to support both stm32f4 (default) and stm32f7. Driver rework to ensure same functional level on both stm32f4 and stm32f7: no new feature in this version yet. Signed-off-by:
-
Nava kishore Manne authored
After issuing the reset, driver is not checking the rx and tx reset done status. So, modified driver to wait for the reset done status. Signed-off-by:
Nava kishore Manne <navam@xilinx.com> Signed-off-by:
-
Nava kishore Manne authored
This patch adds RXBS register access support for zynqmp. To avoid the corner error conditions it will consider only RXBS[2:0] bits while checking the error conditions (Parity,Framing and BRAK). Signed-off-by:
-
Jiri Olsa authored
First loop in wait_for_xmitr could also trigger NMI watchdog in case reading from the port is slow: PID: 0 TASK: ffffffff819c1460 CPU: 0 COMMAND: "swapper/0" #0 [ffff88019f405e58] crash_nmi_callback at ffffffff8104d382 #1 [ffff88019f405e68] nmi_handle at ffffffff8168ead9 #2 [ffff88019f405eb0] do_nmi at ffffffff8168ec53 #3 [ffff88019f405ef0] end_repeat_nmi at ffffffff8168df13 [exception RIP: delay_tsc+50] RIP: ffffffff81325642 RSP: ffff88019f403bb0 RFLAGS: 00000083 RAX: 00000000000005c8 RBX: ffffffff81f83000 RCX: 0000024e4fb88a8b RDX: 0000024e4fb89053 RSI: 0000000000000000 RDI: 00000000000007d1 RBP: ffff88019f403bb0 R8: 000000000000000a R9: 0000000000000000 R10: 0000000000000000 R11: ffff88019f403ad6 R12: 000000000000250f R13: 0000000000000020 R14: ffffffff81d360c7 R15: 0000000000000047 ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018 --- <NMI exception stack> --- #4 [ffff88019f403bb0] delay_tsc at ffffffff81325642 #5 [ffff88019f403bb8] __const_udelay at ffffffff813255a8 #6 [ffff88019f403bc8] wait_for_xmitr at ffffffff81404390 #7 [ffff88019f403bf0] serial8250_console_putchar at ffffffff8140455c #8 [ffff88019f403c10] uart_console_write at ffffffff813ff00a #9 [ffff88019f403c40] serial8250_console_write at ffffffff814044ae #10 [ffff88019f403c88] call_console_drivers.constprop.15 at ffffffff81086b01 #11 [ffff88019f403cb0] console_unlock at ffffffff8108842f #12 [ffff88019f403ce8] vprintk_emit at ffffffff81088834 #13 [ffff88019f403d58] vprintk_default at ffffffff81088ba9 #14 [ffff88019f403d68] printk at ffffffff8167f034 Adding touch_nmi_watchdog call to the first loop as well. Reported-by:Chunyu Hu <chuhu@redhat.com> Signed-off-by:
Jiri Olsa <jolsa@kernel.org> Signed-off-by:
-
Jan Kiszka authored
Less magic that only requires comments. Signed-off-by:
Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by:
-
Shawn Guo authored
Commit 0e125a5f ("tty: amba-pl011: define flag register bits for ZTE device") changes earlycon function pl011_putc() to use a pointer to uart_amba_port. This causes a regression when earlycon is enabled, because uart_amba_port is not available yet at earlycon time. Let's revert the change on pl011_putc() to fix the regression. The earlycon support for ZTE device can probably be added later by declaring a new earlycon setup function with a vendor specific compatible. Reported-by:
Sudeep Holla <sudeep.holla@arm.com> Fixes: 0e125a5f ("tty: amba-pl011: define flag register bits for ZTE device") Signed-off-by:
Shawn Guo <shawn.guo@linaro.org> Tested-by:
-
Wei Yongjun authored
Commit 5d7519df ("serial: mxs-auart: Disable clock on error path") try to disable clock on error path, but still missing the clk_set_rate() error handling path. Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Reviewed-by:
Fabio Estevam <fabio.estevam@nxp.com> Signed-off-by:
-
Adam Borowski authored
For now, these fall back to regular (dark) colors. It'd be tempting to replace blink with bright backgrounds, as permitted by CGA/VGA -- we already muck with the other programmable bit (foreground brightness vs 512 character font). This would bring vgacon in line with fbcon, which doesn't support blink anywhere but on some drivers renders that bit as bright background. If that is done, this commit should be amended to be one of ways of setting that bit. Signed-off-by:
Adam Borowski <kilobyte@angband.pl> Signed-off-by:
-
Adam Borowski authored
These codes are supported by all major terminals, thus they occasionally see some use despite being redundant with \e[38;5;(x+8)m or (less exactly) \e[1;3(x)m. Signed-off-by:
-
Adam Borowski authored
Some guy went on a patching spree, adding 24-bit colour support all around: https://gist.github.com/XVilka/8346728Signed-off-by:
-
Adam Borowski authored
All other uses of vc_npar are inclusive (save for < NPAR) which raises eyebrows, so let's at least do so consistently. Signed-off-by:
-
Adam Borowski authored
This makes it show up on UBSAN: perl -e 'for (0..15) {my @x=("0")x$_;push @x,qw(38 2 64 128 192 4);printf "\e[%smAfter %d zeroes.\e[0m\n", join(";",@x[0..($_+5<15?$_+5:15)]), $_}' Seems harmless: if you can programmatically read attributes of a vt character (/dev/vcsa*), multiple probes can obtain parts of vt_mode then lowest byte (5th on 64-bit big-endian) of a pointer. Signed-off-by:
-
- 15 Sep, 2016 7 commits
-
-
Kees Cook authored
The ptmx_fops structure is only changed during init, so mark it as such. Signed-off-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
-
Rob Herring authored
Commit 761ed4a9 ("tty: serial_core: convert uart_close to use tty_port_close") created a case where a port used for a console does not get shutdown on tty closing. Then a call to uart_tx_stopped() segfaults because the tty is NULL. This could be fixed to restore old behavior, but we also want to allow tty_ports to work without a tty attached. So this change to allow a NULL tty_struct is needed either way. Fixes: 761ed4a9 ("tty: serial_core: convert uart_close to use tty_port_close") Reported-by:
kernel test robot <xiaolong.ye@intel.com> Signed-off-by:
Rob Herring <robh@kernel.org> Signed-off-by:
-
Fabien Lahoudere authored
dmaengine_terminate_all() is deprecated and should be replaced by dmaengine_terminate_sync() in non-atomic context or dmaengine_terminate_async() with dmaengine_synchronize(). See commit b36f09c3 ("dmaengine: Add transfer termination synchronization support") Signed-off-by:
Fabien Lahoudere <fabien.lahoudere@collabora.co.uk> Signed-off-by:
-
Andy Shevchenko authored
The commit 761ed4a9 ("tty: serial_core: convert uart_close to use tty_port_close") refactored uart_close() to use tty_port_close(). At the same time it introduced a potential NULL pointer dereference. Rearrange the code to avoid kernel crash. Fixes: 761ed4a9 ("tty: serial_core: convert uart_close to use tty_port_close") Cc: Rob Herring <robh@kernel.org> Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by:
-
Richard Genoud authored
Since commit 18dfef9c ("serial: atmel: convert to irq handling provided mctrl-gpio"), interrupts from GPIOs are not disabled any more when the serial port is closed, leading to an oops when the one of the input pin is toggled (CTS/DSR/DCD/RNG). This is only the case if those pins are used as GPIOs, i.e. declared like that: usart1: serial@f8020000 { /* CTS and DTS will be handled by GPIO */ status = "okay"; rts-gpios = <&pioB 17 GPIO_ACTIVE_LOW>; cts-gpios = <&pioB 16 GPIO_ACTIVE_LOW>; dtr-gpios = <&pioB 14 GPIO_ACTIVE_LOW>; dsr-gpios = <&pioC 31 GPIO_ACTIVE_LOW>; rng-gpios = <&pioB 12 GPIO_ACTIVE_LOW>; dcd-gpios = <&pioB 15 GPIO_ACTIVE_LOW>; }; That's because modem interrupts used to be freed in atmel_shutdown(). After commit 18dfef9c ("serial: atmel: convert to irq handling provided mctrl-gpio"), this code was just removed. Calling atmel_disable_ms() disables the interrupts and everything works fine again. Tested on at91sam9g35-cm (This patch doesn't apply on -stable kernels, fixes for 4.4 and 4.7 will be sent after this one is applied.) Signed-off-by:
Richard Genoud <richard.genoud@gmail.com> Fixes: 18dfef9c ("serial: atmel: convert to irq handling provided mctrl-gpio") Acked-by:
-
Wei Yongjun authored
The function lpuart_start_rx_dma() is called from several places, in some of which, such as lpuart_startup(), a lock be held here, so we should use GFP_ATOMIC when a lock is held. Fixes: 5887ad43 ("tty: serial: fsl_lpuart: Use cyclic DMA for Rx") Signed-off-by:
-
Liu Xiang authored
When dev.of_node is not NULL, we also need to set IRQF_TRIGGER_FALLING flag, otherwise it may cause uncertain interrupts. Signed-off-by:
Liu Xiang <liu.xiang6@zte.com.cn> Signed-off-by:
-
- 13 Sep, 2016 2 commits
-
-
Wei Yongjun authored
Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR. Generated by: scripts/coccinelle/api/ptr_ret.cocci Signed-off-by:
-
Fabio Estevam authored
We should disable the previously acquired clock when enabling s->clk fails. Signed-off-by:
-
