[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241125134505.GC38837@noisy.programming.kicks-ass.net>
Date: Mon, 25 Nov 2024 14:45:05 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Patryk Wlazlyn <patryk.wlazlyn@...ux.intel.com>
Cc: x86@...nel.org, linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
rafael.j.wysocki@...el.com, len.brown@...el.com,
artem.bityutskiy@...ux.intel.com, dave.hansen@...ux.intel.com,
tglx@...utronix.de, gautham.shenoy@....com
Subject: Re: [RFC PATCH v4 5/8] x86/smp native_play_dead: Prefer
cpuidle_play_dead() over mwait_play_dead()
On Mon, Nov 25, 2024 at 02:20:25PM +0100, Patryk Wlazlyn wrote:
> The algorithm based on cpuid leaf 0x5 in the mwait_play_dead(),
> for looking up the mwait hint for the deepest cstate may calculate the
> wrong hint on recent Intel x86 platforms.
>
> Prefer cpuidle_play_dead() over the generic mwait_play_dead() loop and
> only fallback to the later in case of missing appropriate enter_dead()
> handler.
So 7/8 should go first, such that picking cpuidle doesn't end up in some
weird state.
>
> Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@...ux.intel.com>
> ---
> arch/x86/kernel/smpboot.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
> index d0464c7a0af5..2627b56fb9bc 100644
> --- a/arch/x86/kernel/smpboot.c
> +++ b/arch/x86/kernel/smpboot.c
> @@ -1413,9 +1413,10 @@ void native_play_dead(void)
> play_dead_common();
> tboot_shutdown(TB_SHUTDOWN_WFS);
>
> + /* The first successful play_dead() will not return */
> + cpuidle_play_dead();
> mwait_play_dead();
> - if (cpuidle_play_dead())
> - hlt_play_dead();
> + hlt_play_dead();
> }
>
> #else /* ... !CONFIG_HOTPLUG_CPU */
> --
> 2.47.0
>
Powered by blists - more mailing lists