From: "Paul E. McKenney" <paulmck@kernel.org>
To: rcu@vger.kernel.org
Cc: linux-kernel@vger.kernel.org, kernel-team@fb.com,
rostedt@goodmis.org, "Paul E. McKenney" <paulmck@kernel.org>,
Neeraj Upadhyay <quic_neeraju@quicinc.com>,
Eric Dumazet <edumazet@google.com>,
Alexei Starovoitov <ast@kernel.org>,
Andrii Nakryiko <andrii@kernel.org>,
Martin KaFai Lau <kafai@fb.com>, KP Singh <kpsingh@kernel.org>
Subject: [PATCH rcu 23/32] rcu-tasks: Scan running tasks for RCU Tasks Trace readers
Date: Mon, 20 Jun 2022 15:54:02 -0700 [thread overview]
Message-ID: <20220620225411.3842519-23-paulmck@kernel.org> (raw)
In-Reply-To: <20220620225402.GA3842369@paulmck-ThinkPad-P17-Gen-1>
A running task might be within an RCU Tasks Trace read-side critical
section for any length of time, but will not be placed on any of the
per-CPU rcu_tasks_percpu structure's ->rtp_blkd_tasks lists. Therefore
any RCU Tasks Trace grace-period processing that does not scan the full
task list must interact with the running tasks.
This commit therefore causes the rcu_tasks_trace_pregp_step() function
to IPI each CPU in order to place the corresponding task on the holdouts
list and to record whether or not it was in an RCU Tasks Trace read-side
critical section. Yes, it is possible to avoid adding it to that list
if it is not a reader, but that would prevent the system from remembering
that this task was in a quiescent state. Which is why the running tasks
are unconditionally added to the holdout list.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: KP Singh <kpsingh@kernel.org>
---
kernel/rcu/tasks.h | 51 ++++++++++++++++++++++++++++++++++++----------
1 file changed, 40 insertions(+), 11 deletions(-)
diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
index 1aa6a24a9bc2b..a8f95864c921a 100644
--- a/kernel/rcu/tasks.h
+++ b/kernel/rcu/tasks.h
@@ -14,7 +14,7 @@
struct rcu_tasks;
typedef void (*rcu_tasks_gp_func_t)(struct rcu_tasks *rtp);
-typedef void (*pregp_func_t)(void);
+typedef void (*pregp_func_t)(struct list_head *hop);
typedef void (*pertask_func_t)(struct task_struct *t, struct list_head *hop);
typedef void (*postscan_func_t)(struct list_head *hop);
typedef void (*holdouts_func_t)(struct list_head *hop, bool ndrpt, bool *frptp);
@@ -661,7 +661,7 @@ static void rcu_tasks_wait_gp(struct rcu_tasks *rtp)
struct task_struct *t;
set_tasks_gp_state(rtp, RTGS_PRE_WAIT_GP);
- rtp->pregp_func();
+ rtp->pregp_func(&holdouts);
/*
* There were callbacks, so we need to wait for an RCU-tasks
@@ -791,7 +791,7 @@ static void rcu_tasks_wait_gp(struct rcu_tasks *rtp)
// disabling.
/* Pre-grace-period preparation. */
-static void rcu_tasks_pregp_step(void)
+static void rcu_tasks_pregp_step(struct list_head *hop)
{
/*
* Wait for all pre-existing t->on_rq and t->nvcsw transitions
@@ -1449,24 +1449,48 @@ static void trc_wait_for_one_reader(struct task_struct *t,
}
}
-/* Do first-round processing for the specified task. */
-static void rcu_tasks_trace_pertask(struct task_struct *t,
- struct list_head *hop)
+/*
+ * Initialize for first-round processing for the specified task.
+ * Return false if task is NULL or already taken care of, true otherwise.
+ */
+static bool rcu_tasks_trace_pertask_prep(struct task_struct *t, bool notself)
{
// During early boot when there is only the one boot CPU, there
// is no idle task for the other CPUs. Also, the grace-period
// kthread is always in a quiescent state. In addition, just return
// if this task is already on the list.
- if (unlikely(t == NULL) || t == current || !list_empty(&t->trc_holdout_list))
- return;
+ if (unlikely(t == NULL) || (t == current && notself) || !list_empty(&t->trc_holdout_list))
+ return false;
rcu_st_need_qs(t, 0);
t->trc_ipi_to_cpu = -1;
- trc_wait_for_one_reader(t, hop);
+ return true;
+}
+
+/* Do first-round processing for the specified task. */
+static void rcu_tasks_trace_pertask(struct task_struct *t, struct list_head *hop)
+{
+ if (rcu_tasks_trace_pertask_prep(t, true))
+ trc_wait_for_one_reader(t, hop);
+}
+
+/*
+ * Get the current CPU's current task on the holdout list.
+ * Calls to this function must be serialized.
+ */
+static void rcu_tasks_trace_pertask_handler(void *hop_in)
+{
+ struct list_head *hop = hop_in;
+ struct task_struct *t = current;
+
+ // Pull in the currently running task, but only if it is currently
+ // in an RCU tasks trace read-side critical section.
+ if (rcu_tasks_trace_pertask_prep(t, false))
+ trc_add_holdout(t, hop);
}
/* Initialize for a new RCU-tasks-trace grace period. */
-static void rcu_tasks_trace_pregp_step(void)
+static void rcu_tasks_trace_pregp_step(struct list_head *hop)
{
int cpu;
@@ -1474,9 +1498,14 @@ static void rcu_tasks_trace_pregp_step(void)
for_each_possible_cpu(cpu)
WARN_ON_ONCE(per_cpu(trc_ipi_to_cpu, cpu));
- // Disable CPU hotplug across the tasklist scan.
+ // Disable CPU hotplug across the CPU scan.
// This also waits for all readers in CPU-hotplug code paths.
cpus_read_lock();
+
+ // These smp_call_function_single() calls are serialized to
+ // allow safe access to the hop list.
+ for_each_online_cpu(cpu)
+ smp_call_function_single(cpu, rcu_tasks_trace_pertask_handler, hop, 1);
}
/*
--
2.31.1.189.g2e36527f23
next prev parent reply other threads:[~2022-06-20 22:56 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-20 22:54 [PATCH rcu 0/32] RCU Tasks updates for v5.20 Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 01/32] rcu-tasks: Check for abandoned callbacks Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 02/32] rcu-tasks: Split rcu_tasks_one_gp() from rcu_tasks_kthread() Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 03/32] rcu-tasks: Move synchronize_rcu_tasks_generic() down Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 04/32] rcu-tasks: Drive synchronous grace periods from calling task Paul E. McKenney
2022-09-01 10:36 ` Sascha Hauer
2022-09-01 17:27 ` Paul E. McKenney
2022-09-01 17:33 ` Paul E. McKenney
2022-09-02 11:52 ` Sascha Hauer
2022-09-02 12:04 ` Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 05/32] rcu-tasks: Merge state into .b.need_qs and atomically update Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 06/32] rcu-tasks: Remove rcu_tasks_trace_postgp() wait for counter Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 07/32] rcu-tasks: Make trc_read_check_handler() fetch ->trc_reader_nesting only once Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 08/32] rcu-tasks: Idle tasks on offline CPUs are in quiescent states Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 09/32] rcu-tasks: Handle idle tasks for recently offlined CPUs Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 10/32] rcu-tasks: RCU Tasks Trace grace-period kthread has implicit QS Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 11/32] rcu-tasks: Make rcu_note_context_switch() unconditionally call rcu_tasks_qs() Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 12/32] rcu-tasks: Simplify trc_inspect_reader() QS logic Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 13/32] rcu-tasks: Add slow-IPI indicator to RCU Tasks Trace stall warnings Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 14/32] rcu-tasks: Flag offline CPUs in " Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 15/32] rcu-tasks: Make RCU Tasks Trace stall warnings print full .b.need_qs field Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 16/32] rcu-tasks: Make RCU Tasks Trace stall warning handle idle offline tasks Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 17/32] rcu-tasks: Add data structures for lightweight grace periods Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 18/32] rcu-tasks: Track blocked RCU Tasks Trace readers Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 19/32] rcu-tasks: Untrack blocked RCU Tasks Trace at reader end Paul E. McKenney
2022-06-20 22:53 ` [PATCH rcu 20/32] rcu-tasks: Add blocked-task indicator to RCU Tasks Trace stall warnings Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 21/32] rcu-tasks: Move rcu_tasks_trace_pertask() before rcu_tasks_trace_pregp_step() Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 22/32] rcu-tasks: Avoid rcu_tasks_trace_pertask() duplicate list additions Paul E. McKenney
2022-06-20 22:54 ` Paul E. McKenney [this message]
2022-06-20 22:54 ` [PATCH rcu 24/32] rcu-tasks: Pull in tasks blocked within RCU Tasks Trace readers Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 25/32] rcu-tasks: Stop RCU Tasks Trace from scanning idle tasks Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 26/32] rcu-tasks: Stop RCU Tasks Trace from scanning full tasks list Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 27/32] rcu-tasks: Maintain a count of tasks blocking RCU Tasks Trace grace period Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 28/32] rcu-tasks: Eliminate RCU Tasks Trace IPIs to online CPUs Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 29/32] rcu-tasks: Disable and enable CPU hotplug in same function Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 30/32] rcu-tasks: Update comments Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 31/32] rcu-tasks: Be more patient for RCU Tasks boot-time testing Paul E. McKenney
2022-06-20 22:54 ` [PATCH rcu 32/32] rcu-tasks: Use delayed_work to delay rcu_tasks_verify_self_tests() Paul E. McKenney
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=20220620225411.3842519-23-paulmck@kernel.org \
--to=paulmck@kernel.org \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=edumazet@google.com \
--cc=kafai@fb.com \
--cc=kernel-team@fb.com \
--cc=kpsingh@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=quic_neeraju@quicinc.com \
--cc=rcu@vger.kernel.org \
--cc=rostedt@goodmis.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