Linux RCU subsystem development
 help / color / mirror / Atom feed
From: Frederic Weisbecker <frederic@kernel.org>
To: Will Deacon <will@kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Michal Hocko <mhocko@kernel.org>,
	Vlastimil Babka <vbabka@suse.cz>,
	"Paul E. McKenney" <paulmck@kernel.org>,
	Neeraj Upadhyay <neeraj.upadhyay@kernel.org>,
	Joel Fernandes <joel@joelfernandes.org>,
	Boqun Feng <boqun.feng@gmail.com>,
	Zqiang <qiang.zhang1211@gmail.com>,
	Uladzislau Rezki <urezki@gmail.com>,
	rcu@vger.kernel.org, Michal Hocko <mhocko@suse.com>
Subject: [PATCH 1/2] arm64: Keep first mismatched 32bits el0 capable CPU online through its callbacks
Date: Mon, 28 Oct 2024 17:54:29 +0100	[thread overview]
Message-ID: <Zx_BxX6573kqNbhP@localhost.localdomain> (raw)
In-Reply-To: <Zx_BDLUDmL-3tLtd@localhost.localdomain>

The first mismatched 32bits el0 capable CPU is designated as the last
resort CPU for compat 32 bits tasks. As such this CPU is forbidden to
go offline.

However this restriction is applied to the device object of the CPU,
which is not easy to revert later if needed because other components may
have forbidden the target to be offline and they are not tracked.

But the task cpu possible mask is going to be made aware of housekeeping
CPUs. In that context, a better 32 bits el0 last resort CPU may be found
later on boot. When that happens, the old fallback can be made
offlineable again.

To make this possible and more flexible, drive the offlineable decision
from the cpuhotplug callbacks themselves.

Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
---
 arch/arm64/kernel/cpufeature.c | 32 ++++++++++++++++++--------------
 1 file changed, 18 insertions(+), 14 deletions(-)

diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 718728a85430..53ee8ce38d5b 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -3591,15 +3591,15 @@ void __init setup_user_features(void)
 	minsigstksz_setup();
 }
 
-static int enable_mismatched_32bit_el0(unsigned int cpu)
-{
-	/*
-	 * The first 32-bit-capable CPU we detected and so can no longer
-	 * be offlined by userspace. -1 indicates we haven't yet onlined
-	 * a 32-bit-capable CPU.
-	 */
-	static int lucky_winner = -1;
+/*
+ * The first 32-bit-capable CPU we detected and so can no longer
+ * be offlined by userspace. -1 indicates we haven't yet onlined
+ * a 32-bit-capable CPU.
+ */
+static int cpu_32bit_unofflineable = -1;
 
+static int mismatched_32bit_el0_online(unsigned int cpu)
+{
 	struct cpuinfo_arm64 *info = &per_cpu(cpu_data, cpu);
 	bool cpu_32bit = id_aa64pfr0_32bit_el0(info->reg_id_aa64pfr0);
 
@@ -3611,7 +3611,7 @@ static int enable_mismatched_32bit_el0(unsigned int cpu)
 	if (cpumask_test_cpu(0, cpu_32bit_el0_mask) == cpu_32bit)
 		return 0;
 
-	if (lucky_winner >= 0)
+	if (cpu_32bit_unofflineable < 0)
 		return 0;
 
 	/*
@@ -3619,16 +3619,20 @@ static int enable_mismatched_32bit_el0(unsigned int cpu)
 	 * 32-bit EL0 online so that is_cpu_allowed() doesn't end up rejecting
 	 * every CPU in the system for a 32-bit task.
 	 */
-	lucky_winner = cpu_32bit ? cpu : cpumask_any_and(cpu_32bit_el0_mask,
-							 cpu_active_mask);
-	get_cpu_device(lucky_winner)->offline_disabled = true;
+	cpu_32bit_unofflineable = cpu_32bit ? cpu : cpumask_any_and(cpu_32bit_el0_mask,
+								    cpu_active_mask);
 	setup_elf_hwcaps(compat_elf_hwcaps);
 	elf_hwcap_fixup();
 	pr_info("Asymmetric 32-bit EL0 support detected on CPU %u; CPU hot-unplug disabled on CPU %u\n",
-		cpu, lucky_winner);
+		cpu, cpu_32bit_unofflineable);
 	return 0;
 }
 
+static int mismatched_32bit_el0_offline(unsigned int cpu)
+{
+	return cpu == cpu_32bit_unofflineable ? -EBUSY : 0;
+}
+
 static int __init init_32bit_el0_mask(void)
 {
 	if (!allow_mismatched_32bit_el0)
@@ -3639,7 +3643,7 @@ static int __init init_32bit_el0_mask(void)
 
 	return cpuhp_setup_state(CPUHP_AP_ONLINE_DYN,
 				 "arm64/mismatched_32bit_el0:online",
-				 enable_mismatched_32bit_el0, NULL);
+				 mismatched_32bit_el0_online, mismatched_32bit_el0_offline);
 }
 subsys_initcall_sync(init_32bit_el0_mask);
 
-- 
2.46.0


  reply	other threads:[~2024-10-28 16:54 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20240926224910.11106-1-frederic@kernel.org>
2024-09-26 22:48 ` [PATCH 11/20] sched: Handle CPU isolation on last resort fallback rq selection Frederic Weisbecker
2024-09-27  7:26   ` Michal Hocko
2024-10-08 10:54   ` Will Deacon
2024-10-08 12:27     ` Frederic Weisbecker
2024-10-15 13:48     ` Frederic Weisbecker
2024-10-28 16:25       ` Will Deacon
2024-10-28 16:51         ` Frederic Weisbecker
2024-10-28 16:54           ` Frederic Weisbecker [this message]
2024-10-28 16:56           ` [PATCH 2/2] sched,arm64: " Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 12/20] kthread: Make sure kthread hasn't started while binding it Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 13/20] kthread: Default affine kthread to its preferred NUMA node Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 16/20] kthread: Implement preferred affinity Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 17/20] rcu: Use kthread preferred affinity for RCU boost Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 18/20] kthread: Unify kthread_create_on_cpu() and kthread_create_worker_on_cpu() automatic format Frederic Weisbecker
2024-09-26 22:49 ` [PATCH 19/20] treewide: Introduce kthread_run_worker[_on_cpu]() Frederic Weisbecker
2024-09-27  5:39   ` Paul E. McKenney
2024-09-26 22:49 ` [PATCH 20/20] rcu: Use kthread preferred affinity for RCU exp kworkers Frederic Weisbecker

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Zx_BxX6573kqNbhP@localhost.localdomain \
    --to=frederic@kernel.org \
    --cc=boqun.feng@gmail.com \
    --cc=joel@joelfernandes.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mhocko@kernel.org \
    --cc=mhocko@suse.com \
    --cc=neeraj.upadhyay@kernel.org \
    --cc=paulmck@kernel.org \
    --cc=peterz@infradead.org \
    --cc=qiang.zhang1211@gmail.com \
    --cc=rcu@vger.kernel.org \
    --cc=tglx@linutronix.de \
    --cc=urezki@gmail.com \
    --cc=vbabka@suse.cz \
    --cc=vincent.guittot@linaro.org \
    --cc=will@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox