linux/drivers/clocksource, branch v3.13

clocksource: cadence_ttc: Fix mutex taken inside interrupt context

2013-12-30T10:32:24Z

When the kernel is compiled with: CONFIG_HIGH_RES_TIMERS=no CONFIG_HZ_PERIODIC=yes CONFIG_DEBUG_ATOMIC_SLEEP=yes The following WARN appears: WARNING: CPU: 1 PID: 0 at linux/kernel/mutex.c:856 mutex_trylock+0x70/0x1fc() DEBUG_LOCKS_WARN_ON(in_interrupt()) Modules linked in: CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.12.0-xilinx-dirty #93 [] (unwind_backtrace+0x0/0x11c) from [] (show_stack+0x10/0x14) [] (show_stack+0x10/0x14) from [] (dump_stack+0x7c/0xc0) [] (dump_stack+0x7c/0xc0) from [] (warn_slowpath_common+0x60/0x84) [] (warn_slowpath_common+0x60/0x84) from [] (warn_slowpath_fmt+0x2c/0x3c) [] (warn_slowpath_fmt+0x2c/0x3c) from [] (mutex_trylock+0x70/0x1fc) [] (mutex_trylock+0x70/0x1fc) from [] (clk_prepare_lock+0xc/0xe4) [] (clk_prepare_lock+0xc/0xe4) from [] (clk_get_rate+0xc/0x44) [] (clk_get_rate+0xc/0x44) from [] (ttc_set_mode+0x34/0x78) [] (ttc_set_mode+0x34/0x78) from [] (clockevents_set_mode+0x28/0x5c) [] (clockevents_set_mode+0x28/0x5c) from [] (tick_broadcast_on_off+0x190/0x1c0) [] (tick_broadcast_on_off+0x190/0x1c0) from [] (clockevents_notify+0x58/0x1ac) [] (clockevents_notify+0x58/0x1ac) from [] (cpuidle_setup_broadcast_timer+0x20/0x24) [] (cpuidle_setup_broadcast_timer+0x20/0x24) from [] (generic_smp_call_function_single_interrupt+0) [] (generic_smp_call_function_single_interrupt+0xe0/0x130) from [] (handle_IPI+0x88/0x118) [] (handle_IPI+0x88/0x118) from [] (gic_handle_irq+0x58/0x60) [] (gic_handle_irq+0x58/0x60) from [] (__irq_svc+0x44/0x78) Exception stack(0xef099fa0 to 0xef099fe8) 9fa0: 00000001 ef092100 00000000 ef092100 ef098000 00000015 c0399f2c c0579d74 9fc0: 0000406a 413fc090 00000000 00000000 00000000 ef099fe8 c00666ec c000f46c 9fe0: 20000113 ffffffff [] (__irq_svc+0x44/0x78) from [] (arch_cpu_idle+0x34/0x3c) [] (arch_cpu_idle+0x34/0x3c) from [] (cpu_startup_entry+0xa8/0x10c) [] (cpu_startup_entry+0xa8/0x10c) from [<000085a4>] (0x85a4) We are in an interrupt context (IPI) and we are calling clk_get_rate in the set_mode function which in turn ends up by getting a mutex... Even if that does not hang, it is a potential kernel deadlock. It is not allowed to call clk_get_rate() from interrupt context. To avoid such calls the timer input frequency is stored in the driver's data struct which makes it accessible to the driver in any context. [dlezcano] completed the changelog with the WARN trace and added a more detailed description. Tested on zync zc702. Acked-by: Daniel Lezcano Tested-by: Daniel Lezcano Signed-off-by: Soren Brinkmann Signed-off-by: Daniel Lezcano

clocksource: dw_apb_timer_of: Fix support for dts binding "snps,dw-apb-timer"

2013-12-10T18:49:18Z

In commit 620f5e1cbf (dts: Rename DW APB timer compatible strings), both "snps,dw-apb-timer-sp" and "snps,dw-apb-timer-osc" were deprecated in place of "snps,dw-apb-timer". But the driver also needs to be udpated in order to support this new binding "snps,dw-apb-timer". Signed-off-by: Dinh Nguyen Signed-off-by: Daniel Lezcano

clocksource: dw_apb_timer_of: Fix read_sched_clock

2013-12-10T18:49:18Z

The read_sched_clock should return the ~value because the clock is a countdown implementation. read_sched_clock() should be the same as __apbt_read_clocksource(). Signed-off-by: Dinh Nguyen Signed-off-by: Daniel Lezcano

clocksource: sunxi: Stop timer from ticking before enabling interrupts

2013-12-10T18:41:28Z

The sun4i timer can still be ticking when we enable the interrupt. If another timer is actually used (A7 architected timer, for example), odds are that the interrupt will eventually fire with the event_handler pointer being NULL. The obvious fix it to stop the timer before registering the interrupt. Observed and tested on sun7i (cubietruck). Cc: Daniel Lezcano Acked-by: Maxime Ripard Signed-off-by: Marc Zyngier Signed-off-by: Daniel Lezcano

clocksource: clksrc-of: Do not drop unheld reference on device node

2013-12-10T18:41:27Z

When booting a recent kernel on ARM with OF_DYNAMIC enabled, the kernel warns about the following: [ 0.000000] ERROR: Bad of_node_put() on /timer@50004600 [ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.12.0-rc5-next-20131017-00077-gedfd827-dirty #406 [ 0.000000] [] (unwind_backtrace+0x0/0xf4) from [] (show_stack+0x10/0x14) [ 0.000000] [] (show_stack+0x10/0x14) from [] (dump_stack+0x9c/0xc8) [ 0.000000] [] (dump_stack+0x9c/0xc8) from [] (of_node_release+0x90/0x9c) [ 0.000000] [] (of_node_release+0x90/0x9c) from [] (of_find_matching_node_and_match+0x78/0xb4) [ 0.000000] [] (of_find_matching_node_and_match+0x78/0xb4) from [] (clocksource_of_init+0x60/0x70) [ 0.000000] [] (clocksource_of_init+0x60/0x70) from [] (start_kernel+0x1f4/0x33c) [ 0.000000] [] (start_kernel+0x1f4/0x33c) from [<80008074>] (0x80008074) This is caused by clocksource_of_init() dropping a reference on the device node that it never took. The reference taken by the loop is implicitly dropped on subsequent iterations. See the implementation of and the comment on top of the of_find_matching_node_and_match() function for reference (no pun intended). Signed-off-by: Thierry Reding Signed-off-by: Daniel Lezcano Cc: Uwe Kleine-König

clocksource: armada-370-xp: Register sched_clock after the counter reset

2013-12-10T18:41:27Z

This commit registers the sched_clock _after_ the counter reset (instead of before). This removes the timestamp 'jump' in kernel log messages. Before this change: [ 0.000000] sched_clock: 32 bits at 25MHz, resolution 40ns, wraps every 171798691800ns [ 0.000000] Initializing Coherency fabric [ 0.000000] Aurora cache controller enabled [ 0.000000] l2x0: 16 ways, CACHE_ID 0x00000100, AUX_CTRL 0x1a696b12, Cache size: 1024 kB [ 163.507447] Calibrating delay loop... 1325.05 BogoMIPS (lpj=662528) [ 163.521419] pid_max: default: 32768 minimum: 301 [ 163.526185] Mount-cache hash table entries: 512 [ 163.531095] CPU: Testing write buffer coherency: ok After this change: [ 0.000000] sched_clock: 32 bits at 25MHz, resolution 40ns, wraps every 171798691800ns [ 0.000000] Initializing Coherency fabric [ 0.000000] Aurora cache controller enabled [ 0.000000] l2x0: 16 ways, CACHE_ID 0x00000100, AUX_CTRL 0x1a696b12, Cache size: 1024 kB [ 0.016849] Calibrating delay loop... 1325.05 BogoMIPS (lpj=662528) [ 0.030820] pid_max: default: 32768 minimum: 301 [ 0.035588] Mount-cache hash table entries: 512 [ 0.040500] CPU: Testing write buffer coherency: ok Signed-off-by: Ezequiel Garcia Signed-off-by: Daniel Lezcano Acked-by: Jason Cooper

clocksource: time-efm32: Select CLKSRC_MMIO

2013-12-10T18:41:26Z

The time-efm32 driver uses the clocksource MMIO functions. Thus it needs to select CLKSRC_MMIO in Kconfig. Signed-off-by: Axel Lin Signed-off-by: Daniel Lezcano Acked-by: Uwe Kleine-König

clocksource: arm_arch_timer: Hide eventstream Kconfig on non-ARM

2013-11-20T23:06:52Z

Pavel Machek reports that this config is exposed on x86 where the ARM architected timers aren't even present. Make it depend on the ARM architected timers being selected so that non-ARM builds aren't asked about it. Reported-by: Pavel Machek Reviewed-by: Pavel Machek Signed-off-by: Stephen Boyd Signed-off-by: Daniel Lezcano

clocksource: sh_tmu: Add clk_prepare/unprepare support

2013-11-20T20:55:24Z

Prepare the clock at probe time, as there is no other appropriate place in the driver where we're allowed to sleep. Cc: Daniel Lezcano Cc: linux-kernel@vger.kernel.org Signed-off-by: Laurent Pinchart Acked-by: Simon Horman Signed-off-by: Daniel Lezcano

clocksource: sh_tmu: Release clock when sh_tmu_register() fails

2013-11-20T20:55:23Z

Fix the probe error path to release the clock resource when the sh_tmu_register() call fails. Cc: Daniel Lezcano Cc: linux-kernel@vger.kernel.org Signed-off-by: Laurent Pinchart Acked-by: Simon Horman Signed-off-by: Daniel Lezcano