]> git.kernelconcepts.de Git - karo-tx-linux.git/commit
clocksource/drivers/arm_arch_timer: Avoid infinite recursion when ftrace is enabled
authorDing Tianhong <dingtianhong@huawei.com>
Thu, 10 Aug 2017 02:52:45 +0000 (10:52 +0800)
committerDaniel Lezcano <daniel.lezcano@linaro.org>
Fri, 11 Aug 2017 14:01:43 +0000 (16:01 +0200)
commitadb4f11e0a8f4e29900adb2b7af28b6bbd5c1fa4
tree4c94934dec69bf5c73623de49c60a67f9287a289
parent599dc457c79bde8bd4fe8bbb2ba1f30ef3d7a5c8
clocksource/drivers/arm_arch_timer: Avoid infinite recursion when ftrace is enabled

On platforms with an arch timer erratum workaround, it's possible for
arch_timer_reg_read_stable() to recurse into itself when certain
tracing options are enabled, leading to stack overflows and related
problems.

For example, when PREEMPT_TRACER and FUNCTION_GRAPH_TRACER are
selected, it's possible to trigger this with:

$ mount -t debugfs nodev /sys/kernel/debug/
$ echo function_graph > /sys/kernel/debug/tracing/current_tracer

The problem is that in such cases, preempt_disable() instrumentation
attempts to acquire a timestamp via trace_clock(), resulting in a call
back to arch_timer_reg_read_stable(), and hence recursion.

This patch changes arch_timer_reg_read_stable() to use
preempt_{disable,enable}_notrace(), which avoids this.

This problem is similar to the fixed by upstream commit 96b3d28bf4
("sched/clock: Prevent tracing recursion in sched_clock_cpu()").

Fixes: 6acc71ccac71 ("arm64: arch_timer: Allows a CPU-specific erratum to only affect a subset of CPUs")
Signed-off-by: Ding Tianhong <dingtianhong@huawei.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
arch/arm64/include/asm/arch_timer.h