linux/drivers/clocksource, branch v6.3

Merge tag 'timers-core-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2023-02-21T17:45:13Z

Pull timer updates from Thomas Gleixner: "Updates for timekeeping, timers and clockevent/source drivers: Core: - Yet another round of improvements to make the clocksource watchdog more robust: - Relax the clocksource-watchdog skew criteria to match the NTP criteria. - Temporarily skip the watchdog when high memory latencies are detected which can lead to false-positives. - Provide an option to enable TSC skew detection even on systems where TSC is marked as reliable. Sigh! - Initialize the restart block in the nanosleep syscalls to be directed to the no restart function instead of doing a partial setup on entry. This prevents an erroneous restart_syscall() invocation from corrupting user space data. While such a situation is clearly a user space bug, preventing this is a correctness issue and caters to the least suprise principle. - Ignore the hrtimer slack for realtime tasks in schedule_hrtimeout() to align it with the nanosleep semantics. Drivers: - The obligatory new driver bindings for Mediatek, Rockchip and RISC-V variants. - Add support for the C3STOP misfeature to the RISC-V timer to handle the case where the timer stops in deeper idle state. - Set up a static key in the RISC-V timer correctly before first use. - The usual small improvements and fixes all over the place" * tag 'timers-core-2023-02-20' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (30 commits) clocksource/drivers/timer-sun4i: Add CLOCK_EVT_FEAT_DYNIRQ clocksource/drivers/em_sti: Mark driver as non-removable clocksource/drivers/sh_tmu: Mark driver as non-removable clocksource/drivers/riscv: Patch riscv_clock_next_event() jump before first use clocksource/drivers/timer-microchip-pit64b: Add delay timer clocksource/drivers/timer-microchip-pit64b: Select driver only on ARM dt-bindings: timer: sifive,clint: add comaptibles for T-Head's C9xx dt-bindings: timer: mediatek,mtk-timer: add MT8365 clocksource/drivers/riscv: Get rid of clocksource_arch_init() callback clocksource/drivers/sh_cmt: Mark driver as non-removable clocksource/drivers/timer-microchip-pit64b: Drop obsolete dependency on COMPILE_TEST clocksource/drivers/riscv: Increase the clock source rating clocksource/drivers/timer-riscv: Set CLOCK_EVT_FEAT_C3STOP based on DT dt-bindings: timer: Add bindings for the RISC-V timer device RISC-V: time: initialize hrtimer based broadcast clock event device dt-bindings: timer: rk-timer: Add rktimer for rv1126 time/debug: Fix memory leak with using debugfs_lookup() clocksource: Enable TSC watchdog checking of HPET and PMTMR only when requested posix-timers: Use atomic64_try_cmpxchg() in __update_gt_cputime() clocksource: Verify HPET and PMTMR when TSC unverified ...

Merge tag 'clocksource.2023.02.06b' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu into timers/core

2023-02-13T18:28:48Z

Pull clocksource watchdog changes from Paul McKenney: o Improvements to clocksource-watchdog console messages. o Loosening of the clocksource-watchdog skew criteria to match those of NTP (500 parts per million, relaxed from 400 parts per million). If it is good enough for NTP, it is good enough for the clocksource watchdog. o Suspend clocksource-watchdog checking temporarily when high memory latencies are detected. This avoids the false-positive clock-skew events that have been seen on production systems running memory-intensive workloads. o On systems where the TSC is deemed trustworthy, use it as the watchdog timesource, but only when specifically requested using the tsc=watchdog kernel boot parameter. This permits clock-skew events to be detected, but avoids forcing workloads to use the slow HPET and ACPI PM timers. These last two timers are slow enough to cause systems to be needlessly marked bad on the one hand, and real skew does sometimes happen on production systems running production workloads on the other. And sometimes it is the fault of the TSC, or at least of the firmware that told the kernel to program the TSC with the wrong frequency. o Add a tsc=revalidate kernel boot parameter to allow the kernel to diagnose cases where the TSC hardware works fine, but was told by firmware to tick at the wrong frequency. Such cases are rare, but they really have happened on production systems. Link: https://lore.kernel.org/r/20230210193640.GA3325193@paulmck-ThinkPad-P17-Gen-1

clocksource/drivers/timer-sun4i: Add CLOCK_EVT_FEAT_DYNIRQ

2023-02-13T12:10:17Z

Add CLOCK_EVT_FEAT_DYNIRQ to allow the IRQ could be runtime set affinity to the cores that needs wake up, otherwise saying core0 has to send IPI to wakeup core1. With CLOCK_EVT_FEAT_DYNIRQ set, when broadcast timer could wake up the cores, IPI is not needed. After enabling this feature, especially the scene where cpuidle is enabled can benefit. Signed-off-by: Yangtao Li Link: https://lore.kernel.org/r/20230209040239.24710-1-frank.li@vivo.com Signed-off-by: Daniel Lezcano

clocksource/drivers/em_sti: Mark driver as non-removable

2023-02-13T12:10:17Z

The comment in the remove callback suggests that the driver is not supposed to be unbound. However returning an error code in the remove callback doesn't accomplish that. Instead set the suppress_bind_attrs property (which makes it impossible to unbind the driver via sysfs). The only remaining way to unbind a em_sti device would be module unloading, but that doesn't apply here, as the driver cannot be built as a module. Also drop the useless remove callback. Signed-off-by: Uwe Kleine-König Link: https://lore.kernel.org/r/20230207193010.469495-1-u.kleine-koenig@pengutronix.de Signed-off-by: Daniel Lezcano

clocksource/drivers/sh_tmu: Mark driver as non-removable

2023-02-13T12:10:17Z

The comment in the remove callback suggests that the driver is not supposed to be unbound. However returning an error code in the remove callback doesn't accomplish that. Instead set the suppress_bind_attrs property (which makes it impossible to unbind the driver via sysfs). The only remaining way to unbind a sh_tmu device would be module unloading, but that doesn't apply here, as the driver cannot be built as a module. Also drop the useless remove callback. Signed-off-by: Uwe Kleine-König Link: https://lore.kernel.org/r/20230207193614.472060-1-u.kleine-koenig@pengutronix.de Signed-off-by: Daniel Lezcano

clocksource/drivers/riscv: Patch riscv_clock_next_event() jump before first use

2023-02-13T12:10:17Z

A static key is used to select between SBI and Sstc timer usage in riscv_clock_next_event(), but currently the direction is resolved after cpuhp_setup_state() is called (which sets the next event). The first event will therefore fall through the sbi_set_timer() path; this breaks Sstc-only systems. So, apply the jump patching before first use. Fixes: 9f7a8ff6391f ("RISC-V: Prefer sstc extension if available") Signed-off-by: Matt Evans Reviewed-by: Palmer Dabbelt Acked-by: Palmer Dabbelt Reviewed-by: Anup Patel Link: https://lore.kernel.org/r/CDDAB2D0-264E-42F3-8E31-BA210BEB8EC1@rivosinc.com Signed-off-by: Daniel Lezcano

clocksource/drivers/timer-microchip-pit64b: Add delay timer

2023-02-13T12:10:17Z

Add delay timer. Signed-off-by: Claudiu Beznea Link: https://lore.kernel.org/r/20230203130537.1921608-3-claudiu.beznea@microchip.com Signed-off-by: Daniel Lezcano

clocksource/drivers/timer-microchip-pit64b: Select driver only on ARM

2023-02-13T12:10:17Z

Microchip PIT64B is currently available on ARM based devices. Thus select it only for ARM. This allows implementing delay timer. Signed-off-by: Claudiu Beznea Link: https://lore.kernel.org/r/20230203130537.1921608-2-claudiu.beznea@microchip.com Signed-off-by: Daniel Lezcano

clocksource/drivers/riscv: Get rid of clocksource_arch_init() callback

2023-02-13T12:10:17Z

Having a clocksource_arch_init() callback always sets vdso_clock_mode to VDSO_CLOCKMODE_ARCHTIMER if GENERIC_GETTIMEOFDAY is enabled, this is required for the riscv-timer. This works for platforms where just riscv-timer clocksource is present. On platforms where other clock sources are available we want them to register with vdso_clock_mode set to VDSO_CLOCKMODE_NONE. On the Renesas RZ/Five SoC OSTM block can be used as clocksource [0], to avoid multiple clock sources being registered as VDSO_CLOCKMODE_ARCHTIMER move setting of vdso_clock_mode in the riscv-timer driver instead of doing this in clocksource_arch_init() callback as done similarly for ARM/64 architecture. [0] drivers/clocksource/renesas-ostm.c Signed-off-by: Lad Prabhakar Tested-by: Samuel Holland Reviewed-by: Conor Dooley Reviewed-by: Samuel Holland Link: https://lore.kernel.org/r/20221229224601.103851-1-prabhakar.mahadev-lad.rj@bp.renesas.com Signed-off-by: Daniel Lezcano

clocksource/drivers/sh_cmt: Mark driver as non-removable

2023-02-13T12:10:17Z

The comment in the remove callback suggests that the driver is not supposed to be unbound. However returning an error code in the remove callback doesn't accomplish that. Instead set the suppress_bind_attrs property (which makes it impossible to unbind the driver via sysfs). The only remaining way to unbind a sh_cmt device would be module unloading, but that doesn't apply here, as the driver cannot be built as a module. Also drop the useless remove callback. Signed-off-by: Uwe Kleine-König Link: https://lore.kernel.org/r/20230123220221.48164-1-u.kleine-koenig@pengutronix.de Signed-off-by: Daniel Lezcano