Re: [PATCH v2 0/4] ACPI / watchdog: Add support for WDAT (Watchdog Action Table)
On Tuesday, September 20, 2016 03:30:50 PM Mika Westerberg wrote: > Hi, > > The WDAT (Watchdog Action Table) is a special ACPI table introduced by > Microsoft [1] that abstracts the watchdog hardware from the OS. Windows > uses this table for its watchdog implementation instead of a native iTCO > driver. > > Microsoft re-licensed the WDAT specification to be under Microsoft > Community Promise license [2] so it should be fine to use it in Linux. > > This series brings WDAT table support to Linux. > > When the driver is enabled and we find out that there is a WDAT table, the > driver will take over the native iTCO watchdog driver. Main advantage in > this is that we do not need to change the native iTCO driver whenever the > hardware changes. For example in Skylake iTCO moved to sit behind SMBus and > the NO_REBOOT bit was hidden behind P2SB (Primary to Sideband). In addition > we can expect this to be tested much better by OEMs who typically validate > that Windows works fine on their hardware/firmware. > > Patch [1/4] adds ACPI enumeration support and the driver itself. It also > introduces acpi_has_watchdog() which can be used to check if we should use > ACPI watchdog or native one. > > Patches [2-4/4] prevent creation of the native iTCO platform device if we > detect that the ACPI watchdog (WDAT) should be used instead. > > The previous version of the series can be found in [3]. > > Changes from v1: > * Moved wdat_wdt.c to live under drivers/watchdog > * Added checks for timer_period, min_count and max_count > * Use min_hw_heartbeat_ms and max_hw_heartbeat_ms instead of > min/max_timeout > * Instead of stopping the watchdog set WDOG_HW_RUNNING > * Switched to use devm_watchdog_register_device() and dropped > wdat_wdt_remove() > * Do not ping watchdog in resume() > * Added review tag from Guenter Roeck to patches [2-4/4]. > > [1] http://msdn.microsoft.com/en-us/windows/hardware/gg463320.aspx > [2] https://msdn.microsoft.com/en-us/openspecifications/dn646766.aspx > [3] http://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1230607.html > > Mika Westerberg (4): > ACPI / watchdog: Add support for WDAT hardware watchdog > mfd: lpc_ich: Do not create iTCO watchdog when WDAT table exists > i2c: i801: Do not create iTCO watchdog when WDAT table exists > platform/x86: intel_pmc_ipc: Do not create iTCO watchdog when WDAT > table exists > > drivers/acpi/Kconfig | 3 + > drivers/acpi/Makefile| 1 + > drivers/acpi/acpi_watchdog.c | 123 > drivers/acpi/internal.h | 10 + > drivers/acpi/scan.c | 1 + > drivers/i2c/busses/i2c-i801.c| 4 +- > drivers/mfd/lpc_ich.c| 4 + > drivers/platform/x86/intel_pmc_ipc.c | 12 +- > drivers/watchdog/Kconfig | 13 + > drivers/watchdog/Makefile| 1 + > drivers/watchdog/wdat_wdt.c | 525 > +++ > include/linux/acpi.h | 6 + > 12 files changed, 698 insertions(+), 5 deletions(-) > create mode 100644 drivers/acpi/acpi_watchdog.c > create mode 100644 drivers/watchdog/wdat_wdt.c I'm queing up this series for 4.9. Please let me know if there are any objections. Thanks, Rafael
Re: [PATCH v2 0/4] ACPI / watchdog: Add support for WDAT (Watchdog Action Table)
On Tuesday, September 20, 2016 03:30:50 PM Mika Westerberg wrote: > Hi, > > The WDAT (Watchdog Action Table) is a special ACPI table introduced by > Microsoft [1] that abstracts the watchdog hardware from the OS. Windows > uses this table for its watchdog implementation instead of a native iTCO > driver. > > Microsoft re-licensed the WDAT specification to be under Microsoft > Community Promise license [2] so it should be fine to use it in Linux. > > This series brings WDAT table support to Linux. > > When the driver is enabled and we find out that there is a WDAT table, the > driver will take over the native iTCO watchdog driver. Main advantage in > this is that we do not need to change the native iTCO driver whenever the > hardware changes. For example in Skylake iTCO moved to sit behind SMBus and > the NO_REBOOT bit was hidden behind P2SB (Primary to Sideband). In addition > we can expect this to be tested much better by OEMs who typically validate > that Windows works fine on their hardware/firmware. > > Patch [1/4] adds ACPI enumeration support and the driver itself. It also > introduces acpi_has_watchdog() which can be used to check if we should use > ACPI watchdog or native one. > > Patches [2-4/4] prevent creation of the native iTCO platform device if we > detect that the ACPI watchdog (WDAT) should be used instead. > > The previous version of the series can be found in [3]. > > Changes from v1: > * Moved wdat_wdt.c to live under drivers/watchdog > * Added checks for timer_period, min_count and max_count > * Use min_hw_heartbeat_ms and max_hw_heartbeat_ms instead of > min/max_timeout > * Instead of stopping the watchdog set WDOG_HW_RUNNING > * Switched to use devm_watchdog_register_device() and dropped > wdat_wdt_remove() > * Do not ping watchdog in resume() > * Added review tag from Guenter Roeck to patches [2-4/4]. > > [1] http://msdn.microsoft.com/en-us/windows/hardware/gg463320.aspx > [2] https://msdn.microsoft.com/en-us/openspecifications/dn646766.aspx > [3] http://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1230607.html > > Mika Westerberg (4): > ACPI / watchdog: Add support for WDAT hardware watchdog > mfd: lpc_ich: Do not create iTCO watchdog when WDAT table exists > i2c: i801: Do not create iTCO watchdog when WDAT table exists > platform/x86: intel_pmc_ipc: Do not create iTCO watchdog when WDAT > table exists > > drivers/acpi/Kconfig | 3 + > drivers/acpi/Makefile| 1 + > drivers/acpi/acpi_watchdog.c | 123 > drivers/acpi/internal.h | 10 + > drivers/acpi/scan.c | 1 + > drivers/i2c/busses/i2c-i801.c| 4 +- > drivers/mfd/lpc_ich.c| 4 + > drivers/platform/x86/intel_pmc_ipc.c | 12 +- > drivers/watchdog/Kconfig | 13 + > drivers/watchdog/Makefile| 1 + > drivers/watchdog/wdat_wdt.c | 525 > +++ > include/linux/acpi.h | 6 + > 12 files changed, 698 insertions(+), 5 deletions(-) > create mode 100644 drivers/acpi/acpi_watchdog.c > create mode 100644 drivers/watchdog/wdat_wdt.c I'm queing up this series for 4.9. Please let me know if there are any objections. Thanks, Rafael
[PATCH v2 0/4] ACPI / watchdog: Add support for WDAT (Watchdog Action Table)
Hi, The WDAT (Watchdog Action Table) is a special ACPI table introduced by Microsoft [1] that abstracts the watchdog hardware from the OS. Windows uses this table for its watchdog implementation instead of a native iTCO driver. Microsoft re-licensed the WDAT specification to be under Microsoft Community Promise license [2] so it should be fine to use it in Linux. This series brings WDAT table support to Linux. When the driver is enabled and we find out that there is a WDAT table, the driver will take over the native iTCO watchdog driver. Main advantage in this is that we do not need to change the native iTCO driver whenever the hardware changes. For example in Skylake iTCO moved to sit behind SMBus and the NO_REBOOT bit was hidden behind P2SB (Primary to Sideband). In addition we can expect this to be tested much better by OEMs who typically validate that Windows works fine on their hardware/firmware. Patch [1/4] adds ACPI enumeration support and the driver itself. It also introduces acpi_has_watchdog() which can be used to check if we should use ACPI watchdog or native one. Patches [2-4/4] prevent creation of the native iTCO platform device if we detect that the ACPI watchdog (WDAT) should be used instead. The previous version of the series can be found in [3]. Changes from v1: * Moved wdat_wdt.c to live under drivers/watchdog * Added checks for timer_period, min_count and max_count * Use min_hw_heartbeat_ms and max_hw_heartbeat_ms instead of min/max_timeout * Instead of stopping the watchdog set WDOG_HW_RUNNING * Switched to use devm_watchdog_register_device() and dropped wdat_wdt_remove() * Do not ping watchdog in resume() * Added review tag from Guenter Roeck to patches [2-4/4]. [1] http://msdn.microsoft.com/en-us/windows/hardware/gg463320.aspx [2] https://msdn.microsoft.com/en-us/openspecifications/dn646766.aspx [3] http://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1230607.html Mika Westerberg (4): ACPI / watchdog: Add support for WDAT hardware watchdog mfd: lpc_ich: Do not create iTCO watchdog when WDAT table exists i2c: i801: Do not create iTCO watchdog when WDAT table exists platform/x86: intel_pmc_ipc: Do not create iTCO watchdog when WDAT table exists drivers/acpi/Kconfig | 3 + drivers/acpi/Makefile| 1 + drivers/acpi/acpi_watchdog.c | 123 drivers/acpi/internal.h | 10 + drivers/acpi/scan.c | 1 + drivers/i2c/busses/i2c-i801.c| 4 +- drivers/mfd/lpc_ich.c| 4 + drivers/platform/x86/intel_pmc_ipc.c | 12 +- drivers/watchdog/Kconfig | 13 + drivers/watchdog/Makefile| 1 + drivers/watchdog/wdat_wdt.c | 525 +++ include/linux/acpi.h | 6 + 12 files changed, 698 insertions(+), 5 deletions(-) create mode 100644 drivers/acpi/acpi_watchdog.c create mode 100644 drivers/watchdog/wdat_wdt.c -- 2.9.3
[PATCH v2 0/4] ACPI / watchdog: Add support for WDAT (Watchdog Action Table)
Hi, The WDAT (Watchdog Action Table) is a special ACPI table introduced by Microsoft [1] that abstracts the watchdog hardware from the OS. Windows uses this table for its watchdog implementation instead of a native iTCO driver. Microsoft re-licensed the WDAT specification to be under Microsoft Community Promise license [2] so it should be fine to use it in Linux. This series brings WDAT table support to Linux. When the driver is enabled and we find out that there is a WDAT table, the driver will take over the native iTCO watchdog driver. Main advantage in this is that we do not need to change the native iTCO driver whenever the hardware changes. For example in Skylake iTCO moved to sit behind SMBus and the NO_REBOOT bit was hidden behind P2SB (Primary to Sideband). In addition we can expect this to be tested much better by OEMs who typically validate that Windows works fine on their hardware/firmware. Patch [1/4] adds ACPI enumeration support and the driver itself. It also introduces acpi_has_watchdog() which can be used to check if we should use ACPI watchdog or native one. Patches [2-4/4] prevent creation of the native iTCO platform device if we detect that the ACPI watchdog (WDAT) should be used instead. The previous version of the series can be found in [3]. Changes from v1: * Moved wdat_wdt.c to live under drivers/watchdog * Added checks for timer_period, min_count and max_count * Use min_hw_heartbeat_ms and max_hw_heartbeat_ms instead of min/max_timeout * Instead of stopping the watchdog set WDOG_HW_RUNNING * Switched to use devm_watchdog_register_device() and dropped wdat_wdt_remove() * Do not ping watchdog in resume() * Added review tag from Guenter Roeck to patches [2-4/4]. [1] http://msdn.microsoft.com/en-us/windows/hardware/gg463320.aspx [2] https://msdn.microsoft.com/en-us/openspecifications/dn646766.aspx [3] http://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1230607.html Mika Westerberg (4): ACPI / watchdog: Add support for WDAT hardware watchdog mfd: lpc_ich: Do not create iTCO watchdog when WDAT table exists i2c: i801: Do not create iTCO watchdog when WDAT table exists platform/x86: intel_pmc_ipc: Do not create iTCO watchdog when WDAT table exists drivers/acpi/Kconfig | 3 + drivers/acpi/Makefile| 1 + drivers/acpi/acpi_watchdog.c | 123 drivers/acpi/internal.h | 10 + drivers/acpi/scan.c | 1 + drivers/i2c/busses/i2c-i801.c| 4 +- drivers/mfd/lpc_ich.c| 4 + drivers/platform/x86/intel_pmc_ipc.c | 12 +- drivers/watchdog/Kconfig | 13 + drivers/watchdog/Makefile| 1 + drivers/watchdog/wdat_wdt.c | 525 +++ include/linux/acpi.h | 6 + 12 files changed, 698 insertions(+), 5 deletions(-) create mode 100644 drivers/acpi/acpi_watchdog.c create mode 100644 drivers/watchdog/wdat_wdt.c -- 2.9.3
