- 17 Sep, 2019 39 commits
-
-
Jorge Ramirez-Ortiz authored
Use the bark interrupt as the pre-timeout notifier whenever this interrupt is available. By default, the pretimeout notification shall occur one second earlier than the timeout. Signed-off-by:
Jorge Ramirez-Ortiz <jorge.ramirez-ortiz@linaro.org> Reviewed-by:
Guenter Roeck <linux@roeck-us.net> Link: https://lore.kernel.org/r/20190906205411.31666-2-jorge.ramirez-ortiz@linaro.orgSigned-off-by:
Wim Van Sebroeck <wim@linux-watchdog.org>
-
Oliver Graute authored
Fix only whitespace errors in imx_sc_wdt_probe() Signed-off-by:
Oliver Graute <oliver.graute@kococonnector.com> Reviewed-by:
-
Ivan Mikhaylov authored
The option for the ast2400/2500 to get access to CS0 at runtime. Signed-off-by:
Ivan Mikhaylov <i.mikhaylov@yadro.com> Reviewed-by:
-
Ivan Mikhaylov authored
Set WDT_CLEAR_TIMEOUT_AND_BOOT_CODE_SELECTION into WDT_CLEAR_TIMEOUT_STATUS to clear out boot code source and re-enable access to the primary SPI flash chip while booted via wdt2 from the alternate chip. AST2400 datasheet says: "In the 2nd flash booting mode, all the address mapping to CS0# would be re-directed to CS1#. And CS0# is not accessible under this mode. To access CS0#, firmware should clear the 2nd boot mode register in the WDT2 status register WDT30.bit[1]." Signed-off-by:
-
Chris Packham authored
The orion watchdog can either reset the CPU or generate an interrupt. The interrupt would be useful for debugging as it provides panic() output about the watchdog expiry, however if the interrupt is used the watchdog can't reset the CPU in the event of being stuck in a loop with interrupts disabled or if the CPU is prevented from accessing memory (e.g. an unterminated DMA). The Armada SoCs have spare timers that aren't currently used by the Linux kernel. We can use timer1 to provide a pre-timeout ahead of the watchdog timer and provide the possibility of gathering debug before the reset triggers. Signed-off-by:
Chris Packham <chris.packham@alliedtelesis.co.nz> Reviewed-by:
-
Anson Huang authored
The i.MX7ULP Watchdog Timer (WDOG) module is an independent timer that is available for system use. It provides a safety feature to ensure that software is executing as planned and that the CPU is not stuck in an infinite loop or executing unintended code. If the WDOG module is not serviced (refreshed) within a certain period, it resets the MCU. Add driver support for i.MX7ULP watchdog. Signed-off-by:
Anson Huang <Anson.Huang@nxp.com> Reviewed-by:
-
Anson Huang authored
Add the watchdog bindings for Freescale i.MX7ULP. Signed-off-by:
Rob Herring <rohb@kernel.org> Reviewed-by:
-
Maxime Ripard authored
The Allwinner watchdog has a clock that has been described in some DT, but not all of them. The binding is also completely missing that description. Let's add that property to be consistent. Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Reviewed-by:
Rob Herring <robh@kernel.org> Reviewed-by:
-
Maxime Ripard authored
The Allwinner watchdog has an interrupt, either shared or dedicated depending on the SoC, that has been described in some DT, but not all of them. The binding is also completely missing that description. Let's add that property to be consistent. Signed-off-by:
-
Maxime Ripard authored
The Allwinner SoCs have a watchdog supported in Linux, with a matching Device Tree binding. Now that we have the DT validation in place, let's convert the device tree bindings for that controller over to a YAML schemas. Signed-off-by:
-
Maxime Ripard authored
The watchdogs have a bunch of generic properties that are needed in a device tree. Add a YAML schemas for those. Signed-off-by:
-
Ryan Chen authored
The ast2600 can be supported by the same code as the ast2500. Signed-off-by:
Ryan Chen <ryan_chen@aspeedtech.com> Signed-off-by:
Joel Stanley <joel@jms.id.au> Reviewed-by:
-
Joel Stanley authored
This adds a compatible for the ast2600, a new ASPEED SoC. Signed-off-by:
Andrew Jeffery <andrew@aj.id.au> Reviewed-by:
-
Andrey Smirnov authored
Update checked I2C functionality mask to reflect all of the SMBus primitives used by this driver. Signed-off-by:
Andrey Smirnov <andrew.smirnov@gmail.com> Cc: Chris Healy <cphealy@gmail.com> Cc: Guenter Roeck <linux@roeck-us.net> Cc: Rick Ramstetter <rick@anteaterllc.com> Cc: linux-watchdog@vger.kernel.org Cc: linux-kernel@vger.kernel.org Reviewed-by:
-
Andrey Smirnov authored
There's only one user of ziirave_firm_write_block_data(), so we may as well inline it. Signed-off-by:
-
Andrey Smirnov authored
Bootloader firmware expects the following traffic for DOWNLOAD_END: S Addr Wr [A] 0x10 [A] P using ziirave_firm_write_byte() will result in S Addr Wr [A] 0x10 [A] 0x01 [A] 0x01 [A] P which happens to work because firmware will ignore any extra bytes sent. Fix this by converting the code to use i2c_smbus_write_byte() instead. Signed-off-by:
-
Andrey Smirnov authored
Bootloader firmware doesn't implement DOWNLOAD_START or DOWNLOAD_PACKET in a non-blocking way. It will stretch the clock of the first status byte read until the operation is complete. Polling for the status is not really necessary since it will always succed on the first try. Replace polling code with a simple single byte read to simplify things. Signed-off-by:
-
Andrey Smirnov authored
Bootloader firmware expects the following traffic for RESET_PROCESSOR: S Addr Wr [A] 0x0b [A] 0x01 [A] P using ziirave_firm_write_byte() will result in S Addr Wr [A] 0x0b [A] 0x01 [A] 0x01 [A] P which happens to work because firmware will ignore any extra bytes and expected magic value matches byte count sent by i2c_smbus_write_block_data(). Fix this by converting the code to use i2c_smbus_write_byte_data() instead. Signed-off-by:
-
Andrey Smirnov authored
Bootloader firmware expects the following traffic for DOWNLOAD_END: S Addr Wr [A] 0x11 [A] P using ziirave_firm_write_byte() will result in S Addr Wr [A] 0x11 [A] 0x01 [A] 0x01 [A] P which happens to work because firmware will ignore any extra bytes sent. Fix this by converting the code to use i2c_smbus_write_byte() instead. Signed-off-by:
-
Andrey Smirnov authored
Bootloader firmware expects the following traffic for JUMP_TO_BOOTLOADER: S Addr Wr [A] 0x0c [A] 0x01 [A] P using ziirave_firm_write_byte() will result in S Addr Wr [A] 0x0c [A] 0x01 [A] 0x01 [A] P which happens to work because firmware will ignore any extra bytes and expected magic value matches byte count sent by i2c_smbus_write_block_data(). Fix this by converting the code to use i2c_smbus_write_byte_data() instead. Signed-off-by:
-
Andrey Smirnov authored
Fix misleading error message in ziirave_wdt_init_duration(). Saying "unable to set ..." implies that an attempt at communication with watchdog device has taken palce and was not successful. In this case, however, all it indicates is that no reset pulse duration was specified either via kernel parameter or Device Tree. Re-phase the log message to be more clear about benign nature of this event. Signed-off-by:
-
Andrey Smirnov authored
Bootloader code will ignore any attempts to write data to any flash area outside of [ZIIRAVE_FIRM_FLASH_MEMORY_START; ZIIRAVE_FIRM_FLASH_MEMORY_END]. Firmware update code already have an appropriate check to skip those areas when validating updated firmware. Firmware programming code, OTOH, does not and will needlessly send no-op I2C traffic. Add an appropriate check to __ziirave_firm_write_pkt() so as to save all of that wasted effort. While at it, normalize all of the address handling code to use full 32-bit address in units of bytes and convert it to an appropriate value only in places where that is necessary. Signed-off-by:
-
Andrey Smirnov authored
We only compare first 'len' bytes of read firmware, so we don't need to read more that that. Signed-off-by:
-
Andrey Smirnov authored
Ihex_next_binrec() will return NULL if next record's 'len' is zero, so explicit checks for that in the driver are unnecessary. Drop them. Signed-off-by:
-
Andrey Smirnov authored
Instead of doing this explicitly use put_unaligned_le16() to place 16-bit address value into command payload. Signed-off-by:
-
Andrey Smirnov authored
Instead of zeroing out all of the packet and then overwriting a significant portion of those zeros via memcpy(), zero out only a portion of the packet that is known to not contain any data. Signed-off-by:
-
Andrey Smirnov authored
Both memset() and ziirave_firm_write_block_data() expect length in bytes as an argument, not a number of elements in array. It just happens that in this particular case both values are equal. Modify the code to use sizeof() instead. Signed-off-by:
-
Andrey Smirnov authored
Zeros don't contribute anything to checksum value, so we can skip unused portion of the packet when calculating its checksum. Signed-off-by:
-
Andrey Smirnov authored
We don't need to check for packet length more than once, so drop the extra check in ziirave_firm_upload(). While at it move the check at the very start of __ziirave_firm_write_pkt(), as to not waste any time preparing a packet we'll never use. Signed-off-by:
-
Andrey Smirnov authored
There no reason why ziirave_firm_write_pkt() has to take firmware data via 'struct ihex_binrec' and it can just take address, data pointer and data length as individual arguments. Make this change to allow us to drastically simplify handling page crossing case by removing all of the extra code required to split 'struct ihex_binrec' into two. Signed-off-by:
-
Andrey Smirnov authored
Log bootloader/firmware info during probe. This information is available via sysfs already, but it's really helpful to have this in kernel log during startup as well. Signed-off-by:
-
Andrey Smirnov authored
Reprogramming bootloader on watchdog MCU will result in reported default timeout value of "0". That in turn will be unnecessarily rejected by the driver as invalid device (-ENODEV). Simplify probe to read stored timeout value, set it to a sane default if it is bogus, and then program that value unconditionally. Signed-off-by:
-
Andrey Smirnov authored
Add more error logging to ziirave_firm_upload() for diagnostics. Signed-off-by:
-
Andrey Smirnov authored
The driver is quite silent in case of probe failure, which makes it more difficult to diagnose problem from the kernel log. Add logging to all of the silent error paths ziirave_wdt_probe() to improve that. Signed-off-by:
-
Andrey Smirnov authored
Add missing newline. Signed-off-by:
-
Anson Huang authored
An error message is already displayed by watchdog_register_device() when failed, so no need to have error log again for failure of calling devm_watchdog_register_device(). Signed-off-by:
-
Rasmus Villemoes authored
Converting from ms to s requires dividing by 1000, not multiplying. So this is currently taking the smaller of new_timeout and 1.28e8, i.e. effectively new_timeout. The driver knows what it set max_hw_heartbeat_ms to, so use that value instead of doing a division at run-time. FWIW, this can easily be tested by booting into a busybox shell and doing "watchdog -t 5 -T 130 /dev/watchdog" - without this patch, the watchdog fires after 130&127 == 2 seconds. Fixes: b07e228e "watchdog: imx2_wdt: Fix set_timeout for big timeout values" Cc: stable@vger.kernel.org # 5.2 plus anything the above got backported to Signed-off-by:
Rasmus Villemoes <linux@rasmusvillemoes.dk> Reviewed-by:
-
Arnd Bergmann authored
The cpwd_compat_ioctl() contains a bogus mutex that dates back to a leftover BKL instance. Simplify the implementation by using the new compat_ptr_ioctl() helper function that will do the right thing for all calls here. Note that WIOCSTART/WIOCSTOP don't take any arguments, so the compat_ptr() conversion is not needed here, but it also doesn't hurt. Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Reviewed-by:
-
Guenter Roeck authored
Commit f7a94db4 ("s390/watchdog: use watchdog API") converted the driver to use the watchdog API, but some includes as well as MODULE_ALIAS_MISCDEV() were missed. Cc: Philipp Hachtmann <phacht@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by:
-
- 16 Sep, 2019 1 commit
-
-
Arnd Bergmann authored
The ARM w90x900 platform is getting removed, so this driver is obsolete Signed-off-by:
-
