linux-kernel - Re: [RFC 1/3] softirq: Use preempt_latency

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190529093056.GA146079@google.com>
Date:   Wed, 29 May 2019 05:30:56 -0400
From:   Joel Fernandes <joel@...lfernandes.org>
To:     Daniel Bristot de Oliveira <bristot@...hat.com>
Cc:     linux-kernel@...r.kernel.org, williams@...hat.com,
        daniel@...stot.me, "Steven Rostedt (VMware)" <rostedt@...dmis.org>,
        Ingo Molnar <mingo@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
        Matthias Kaehlcke <mka@...omium.org>,
        Frederic Weisbecker <frederic@...nel.org>,
        Yangtao Li <tiny.windzz@...il.com>,
        Tommaso Cucinotta <tommaso.cucinotta@...tannapisa.it>
Subject: Re: [RFC 1/3] softirq: Use preempt_latency_stop/start to trace
 preemption

On Tue, May 28, 2019 at 05:16:22PM +0200, Daniel Bristot de Oliveira wrote:
> prempt_disable/enable tracepoints occurs only in the preemption
> enabled <-> disable transition. As preempt_latency_stop() and
> preempt_latency_start() already do this control, avoid code
> duplication by using these functions in the softirq code as well.
> 
> RFC: Should we move preempt_latency_start/preempt_latency_stop
> to a trace source file... or even a header?

Yes, I think so. Also this patch changes CALLER_ADDR0 passed to the
tracepoint because there's one more level of a non-inlined function call
in the call chain right?  Very least the changelog should document this
change in functional behavior, IMO.

Looks like a nice change otherwise, thanks!


> Signed-off-by: Daniel Bristot de Oliveira <bristot@...hat.com>
> Cc: "Steven Rostedt (VMware)" <rostedt@...dmis.org>
> Cc: Ingo Molnar <mingo@...hat.com>
> Cc: Peter Zijlstra <peterz@...radead.org>
> Cc: Thomas Gleixner <tglx@...utronix.de>
> Cc: "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
> Cc: Matthias Kaehlcke <mka@...omium.org>
> Cc: "Joel Fernandes (Google)" <joel@...lfernandes.org>
> Cc: Frederic Weisbecker <frederic@...nel.org>
> Cc: Yangtao Li <tiny.windzz@...il.com>
> Cc: Tommaso Cucinotta <tommaso.cucinotta@...tannapisa.it>
> Cc: linux-kernel@...r.kernel.org
> ---
>  kernel/sched/core.c |  4 ++--
>  kernel/softirq.c    | 13 +++++--------
>  2 files changed, 7 insertions(+), 10 deletions(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 874c427742a9..8c0b414e45dc 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -3152,7 +3152,7 @@ static inline void sched_tick_stop(int cpu) { }
>   * If the value passed in is equal to the current preempt count
>   * then we just disabled preemption. Start timing the latency.
>   */
> -static inline void preempt_latency_start(int val)
> +void preempt_latency_start(int val)
>  {
>  	if (preempt_count() == val) {
>  		unsigned long ip = get_lock_parent_ip();
> @@ -3189,7 +3189,7 @@ NOKPROBE_SYMBOL(preempt_count_add);
>   * If the value passed in equals to the current preempt count
>   * then we just enabled preemption. Stop timing the latency.
>   */
> -static inline void preempt_latency_stop(int val)
> +void preempt_latency_stop(int val)
>  {
>  	if (preempt_count() == val)
>  		trace_preempt_on(CALLER_ADDR0, get_lock_parent_ip());
> diff --git a/kernel/softirq.c b/kernel/softirq.c
> index 2c3382378d94..c9ad89c3dfed 100644
> --- a/kernel/softirq.c
> +++ b/kernel/softirq.c
> @@ -108,6 +108,8 @@ static bool ksoftirqd_running(unsigned long pending)
>   * where hardirqs are disabled legitimately:
>   */
>  #ifdef CONFIG_TRACE_IRQFLAGS
> +extern void preempt_latency_start(int val);
> +extern void preempt_latency_stop(int val);
>  void __local_bh_disable_ip(unsigned long ip, unsigned int cnt)
>  {
>  	unsigned long flags;
> @@ -130,12 +132,8 @@ void __local_bh_disable_ip(unsigned long ip, unsigned int cnt)
>  		trace_softirqs_off(ip);
>  	raw_local_irq_restore(flags);
>  
> -	if (preempt_count() == cnt) {
> -#ifdef CONFIG_DEBUG_PREEMPT
> -		current->preempt_disable_ip = get_lock_parent_ip();
> -#endif
> -		trace_preempt_off(CALLER_ADDR0, get_lock_parent_ip());
> -	}
> +	preempt_latency_start(cnt);
> +
>  }
>  EXPORT_SYMBOL(__local_bh_disable_ip);
>  #endif /* CONFIG_TRACE_IRQFLAGS */
> @@ -144,8 +142,7 @@ static void __local_bh_enable(unsigned int cnt)
>  {
>  	lockdep_assert_irqs_disabled();
>  
> -	if (preempt_count() == cnt)
> -		trace_preempt_on(CALLER_ADDR0, get_lock_parent_ip());
> +	preempt_latency_stop(cnt);
>  
>  	if (softirq_count() == (cnt & SOFTIRQ_MASK))
>  		trace_softirqs_on(_RET_IP_);
> -- 
> 2.20.1
>