Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933491AbcK2PDP (ORCPT ); Tue, 29 Nov 2016 10:03:15 -0500 Received: from gloria.sntech.de ([95.129.55.99]:47315 "EHLO gloria.sntech.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933947AbcK2PBX (ORCPT ); Tue, 29 Nov 2016 10:01:23 -0500 From: Heiko =?ISO-8859-1?Q?St=FCbner?= To: Alexander Kochetkov Cc: linux-kernel@vger.kernel.org, devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-rockchip@lists.infradead.org, Thomas Gleixner , Mark Rutland , Rob Herring , Russell King , Caesar Wang , Huang Tao , Daniel Lezcano Subject: Re: [PATCH v3 13/13] clocksource/drivers/rockchip_timer: Prevent ftrace recursion Date: Tue, 29 Nov 2016 16:01:09 +0100 Message-ID: <7837485.H9nfDmPu0F@diego> User-Agent: KMail/4.14.10 (Linux/4.6.0-1-amd64; KDE/4.14.22; x86_64; ; ) In-Reply-To: <1480427118-5126-14-git-send-email-al.kochet@gmail.com> References: <1480343486-25539-1-git-send-email-al.kochet@gmail.com> <1480427118-5126-1-git-send-email-al.kochet@gmail.com> <1480427118-5126-14-git-send-email-al.kochet@gmail.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2217 Lines: 63 Am Dienstag, 29. November 2016, 16:45:18 schrieb Alexander Kochetkov: > Currently rockchip_timer can be used as a scheduler clock. We properly > marked rk_timer_sched_clock_read() as notrace but we then call another > function rk_timer_counter_read() that _wasn't_ notrace. > > Having a traceable function in the sched_clock() path leads to a recursion > within ftrace and a kernel crash. > > Fix this by adding an extra notrace function to keep other users of > rk_timer_counter_read() traceable. > > Signed-off-by: Alexander Kochetkov you introduced the issue yourself in patch 11/13. In general any patch should never leave the kernel in a worse state than it was before, so no patch should ever introduce known issues itself. In that line of thought, don't patches 10+11 introduce warnings about unused functions as well when applied without patch 12? Maybe merge them? Heiko > --- > drivers/clocksource/rockchip_timer.c | 9 +++++++-- > 1 file changed, 7 insertions(+), 2 deletions(-) > > diff --git a/drivers/clocksource/rockchip_timer.c > b/drivers/clocksource/rockchip_timer.c index 1af80a0..a127822 100644 > --- a/drivers/clocksource/rockchip_timer.c > +++ b/drivers/clocksource/rockchip_timer.c > @@ -87,7 +87,7 @@ static void rk_timer_update_counter(u64 cycles, struct > rk_timer *timer) writel_relaxed(upper, timer->base + TIMER_LOAD_COUNT1); > } > > -static u64 rk_timer_counter_read(struct rk_timer *timer) > +static u64 notrace _rk_timer_counter_read(struct rk_timer *timer) > { > u64 counter; > u32 lower; > @@ -106,6 +106,11 @@ static u64 rk_timer_counter_read(struct rk_timer > *timer) return counter; > } > > +static u64 rk_timer_counter_read(struct rk_timer *timer) > +{ > + return _rk_timer_counter_read(timer); > +} > + > static void rk_timer_interrupt_clear(struct rk_timer *timer) > { > writel_relaxed(1, timer->base + TIMER_INT_STATUS); > @@ -168,7 +173,7 @@ static u64 notrace rk_timer_sched_clock_read(void) > { > struct rk_clocksource *_cs = &cs_timer; > > - return ~rk_timer_counter_read(&_cs->timer); > + return ~_rk_timer_counter_read(&_cs->timer); > } > > static int __init rk_timer_init(struct device_node *np, u32 ctrl_reg)