Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 23 Apr 2020 10:46:26 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc:     mingo@kernel.org, linux-kernel@vger.kernel.org, tglx@linutronix.de,
        rostedt@goodmis.org, qais.yousef@arm.com, juri.lelli@redhat.com,
        vincent.guittot@linaro.org, bsegall@google.com, mgorman@suse.de,
        rafael.j.wysocki@intel.com
Subject: Re: [PATCH 04/23] sched,acpi_pad: Convert to sched_set_fifo*()
Message-ID: <20200423084626.GT20713@hirez.programming.kicks-ass.net>
References: <20200422112719.826676174@infradead.org>
 <20200422112831.455977635@infradead.org>
 <c620bb30-eeee-336b-f8d6-a98e903a48e2@arm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <c620bb30-eeee-336b-f8d6-a98e903a48e2@arm.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Apr 22, 2020 at 06:45:36PM +0200, Dietmar Eggemann wrote:
> On 22/04/2020 13:27, Peter Zijlstra wrote:
> > Because SCHED_FIFO is a broken scheduler model (see previous patches)
> > take away the priority field, the kernel can't possibly make an
> > informed decision.
> > 
> > In this case, use fifo_low, because it only cares about being above
> > SCHED_NORMAL. Effectively no change in behaviour.
> > 
> > XXX: this driver is still complete crap; why isn't it using proper
> > idle injection or at the very least play_idle() ?
> > 
> > Cc: rafael.j.wysocki@intel.com
> > Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> > Reviewed-by: Ingo Molnar <mingo@kernel.org>
> > ---
> >  drivers/acpi/acpi_pad.c |    3 +--
> >  1 file changed, 1 insertion(+), 2 deletions(-)
> > 
> > --- a/drivers/acpi/acpi_pad.c
> > +++ b/drivers/acpi/acpi_pad.c
> > @@ -136,12 +136,11 @@ static unsigned int idle_pct = 5; /* per
> >  static unsigned int round_robin_time = 1; /* second */
> >  static int power_saving_thread(void *data)
> >  {
> > -	struct sched_param param = {.sched_priority = 1};
> >  	int do_sleep;
> >  	unsigned int tsk_index = (unsigned long)data;
> >  	u64 last_jiffies = 0;
> >  
> > -	sched_setscheduler(current, SCHED_RR, &param);
> 
> I was wondering what happened to the SCHED_RR cases but as I can see now
> they are handled here and in the next patch.

Oh right; I completely forgot to mention that in the Changelog didn't I
:-(

In this case, this driver is a broken piece of crap and doing fake idle
with RR is just plain idiotic. Also note the WARNs in play_idle().

Also, rjw, what was the point of renmaing play_idle() to
play_idle_precise() if there is only one anyway? The changelog talks
about adding play_idle_precise() but the patch (rightfully) doesn't add
another version but replaces the existing one. But why change the name?!