From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=158.69.130.18 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1795207372; b=0EN8d7OWYIyIIqaaPK6MW+JYivkhwPbxgnFJAtYtyisTOCPTJ4/vrkkGpT04ENozUcFE XJbhwnL8tJgzgiz2RvLAMk0CEnFx3Tfysh4u+6YV93uLVrll+8rwqmsfzppBw71gBXFrA gqFWZYjVItfM5eJXX7HYX7R5K5M+ROOh7yzCm3wUpyFs4+n7/DSwSzKT8cGoyIwQLs+wi cyfbkhi2M3rf3jYqakqqUxg/GCbMdnpVxpFPbf4gZ7i+BxifGeJNiGNNcFcDk2oTcPKGG ZxJnbhbGLdefPz9DVeKQSK4N8LEQlQp2z2eqcrpOsuTHkYDO9JXM4MOiVqdEqQEtNxQ== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1795207372; h=DMARC-Filter:DKIM-Signature:Received:From:To:Cc:Subject:Date: Message-Id:X-Mailer:In-Reply-To:References:MIME-Version: Content-Transfer-Encoding; bh=vE+uNjXsB2iA/fb83lE/XKLUhCwUuFaXjUJLCJJWSCI=; b=I870YI4skev3rZUC2ihgBUXa1ez8wKW/E6JoVxlgdN5yqRZIW53psRqzqUISztLbXOxD mG1fGGwD9mtaJgH6pvTChDHSirjW2lvCSWEjxh0R/vGeOJH6k3ad/Oak8uHf9VaDLSZ8T ED81hbvWWLccBXmmH/4UWv8sXy55S2DiulJo61YPRdSk8f2AdHkg/TKvTtnYV5vYxqNir Z52iLSxlQDZCLEMn7mDUlSsPQyVH5il3kXu9IWbFXhg8ZaCwrU/jmSt3uopc9pRSB9QcG csgHrdTH2wmx1xKHVZ2GtVL8gnq2IcQNRjMgK4ODr8AHHAQo5xVBGSRTbzHqMsfx6Rg== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass header.from=efficios.com; spf=pass smtp.mailfrom=efficios.com; dkim=pass header.d=efficios.com header.i=@efficios.com header.a=rsa-sha256 header.s=smtpout1 header.b=J0Ebh9eN; arc=none smtp.remote-ip=158.69.130.18 Received: from smtpout.efficios.com (smtpout.efficios.com [158.69.130.18]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id E9E8A60A6F4C; Fri, 20 Nov 2026 20:42:45 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org E9E8A60A6F4C Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=efficios.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=efficios.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=efficios.com header.i=@efficios.com header.a=rsa-sha256 header.s=smtpout1 header.b=J0Ebh9eN DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=efficios.com; s=smtpout1; t=1795207360; bh=vE+uNjXsB2iA/fb83lE/XKLUhCwUuFaXjUJLCJJWSCI=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=J0Ebh9eNqdwgHF3sJSdkQVj4yXGzfbxfno1u1Q16gCnI6Y08MQyRzk8zpJSY4o/wY w/MY4tlTAbdVE55FzRPZ3Hw69XyvyA65MZtZxg8PQjmR7rDp+Er7r3qlR5BnM6LEb0 7wPILlp+/BM7rXLblMye6TG1u1EXb7FpDYQ6L9RoG9dwNOkRrPH0KZphCaiUD+Ek4T UESTEXhSuMQPEv1saN3QyAjUaK+W6CiUTKZUQ6JkZBNV9v2Q1xBfldR7hgW3aQcKcR PM5D30EqvkaSmldtwB729MUDNSqbgQx/HHsmhXMiwvPKLXkbBO7PdDjEhem50/9grD 0DCMi8+sQiSYg== Received: from thinkos.internal.efficios.com (mtl.efficios.com [216.120.195.104]) by smtpout.efficios.com (Postfix) with ESMTPSA id 4jWS2S3r6qzV52; Fri, 20 Nov 2026 15:42:40 -0500 (EST) From: Mathieu Desnoyers To: juri.lelli@redhat.com, peterz@infradead.org, mingo@redhat.com, vincent.guittot@linaro.org Cc: Mathieu Desnoyers , tglx@kernel.org, will@kernel.org, boqun@kernel.org, dietmar.eggemann@arm.com, rostedt@goodmis.org, bsegall@google.com, mgorman@suse.de, vschneid@redhat.com, kprateek.nayak@amd.com, luca.abeni@santannapisa.it, mathieu.poirier@linaro.org, longman@redhat.com, ridong.chen@linux.dev, tj@kernel.org, hannes@cmpxchg.org, mkoutny@suse.com, shuah@kernel.org, bigeasy@linutronix.de, clrkwllms@kernel.org, parri.andrea@gmail.com, m.szyprowski@samsung.com, kayracizmeci@gmail.com, weihengqiuu@gmail.com, linux-kernel@vger.kernel.org, cgroups@vger.kernel.org, linux-kselftest@vger.kernel.org, linux-rt-devel@lists.linux.dev, stable@vger.kernel.org Subject: [PATCH v10 37/40] sched/deadline: Keep the start of the period across a change of dl_deadline Date: Fri, 20 Nov 2026 15:42:14 -0500 Message-Id: <3360961974ba9401f7d6d4d7d5fd513db04c9fc5.1795207233.git.mathieu.desnoyers@efficios.com> X-Mailer: git-send-email 2.39.5 In-Reply-To: References: X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list MIME-Version: 1.0 Content-Transfer-Encoding: 8bit dl_next_period() works out the start of the next period of a deadline entity as deadline - dl_deadline + dl_period. Commit 5ac69d37784b ("sched/deadline: Make sure the replenishment timer fires in the next period") added it. The replenishment timer of a task with a constrained deadline then fires at its next period, not at its deadline. Commit df8eac8cafce ("sched/deadline: Throttle a constrained deadline task activated after the deadline") uses it too. It throttles a constrained task that wakes between its deadline and its next period. The formula assumes that the deadline was set with the current dl_deadline. But sched_setattr() can change dl_deadline during a period, and the deadline stays. The start of the next period then moves by the old dl_deadline minus the new one: - A shorter dl_deadline moves it later. A task that wakes after its period ended is throttled until that later time. For an implicit task made constrained, that is up to dl_period - dl_deadline. - A longer dl_deadline moves it earlier. A constrained task that wakes between its deadline and its real next period is not throttled. It gets a new runtime before its next period. Fix this by recording the start of each period in period_start, where the deadline is set. dl_next_period() and dl_in_constrained_gap() use it. The Sashiko AI review of this series found this. The case wake_after_deadline_shrunk of deadline_bw_test, added later in this series, tests the shorter dl_deadline. It fails without this patch. Fixes: 5ac69d37784b ("sched/deadline: Make sure the replenishment timer fires in the next period") Signed-off-by: Mathieu Desnoyers Cc: stable@vger.kernel.org --- include/linux/sched.h | 1 + kernel/sched/deadline.c | 25 +++++++++++++++++++------ 2 files changed, 20 insertions(+), 6 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 491ddeea7d65..244e57fa8c21 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -661,6 +661,7 @@ struct sched_dl_entity { */ s64 runtime; /* Remaining runtime for this instance */ u64 deadline; /* Absolute deadline for this instance */ + u64 period_start; /* Absolute start of this instance */ unsigned int flags; /* Specifying the scheduler behaviour */ /* diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index b9c206a59c16..e9b3d3ec71d7 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -89,7 +89,8 @@ static inline int on_dl_rq(struct sched_dl_entity *dl_se) */ static inline void dl_start_own_period(struct sched_dl_entity *dl_se) { - dl_se->deadline = rq_clock(rq_of_dl_se(dl_se)) + dl_se->dl_deadline; + dl_se->period_start = rq_clock(rq_of_dl_se(dl_se)); + dl_se->deadline = dl_se->period_start + dl_se->dl_deadline; dl_se->runtime = dl_se->dl_runtime; } @@ -103,7 +104,7 @@ static inline bool dl_in_constrained_gap(struct sched_dl_entity *dl_se) u64 now = rq_clock(rq_of_dl_se(dl_se)); return !dl_is_implicit(dl_se) && dl_time_before(dl_se->deadline, now) && - dl_time_before(now, dl_se->deadline - dl_se->dl_deadline + dl_se->dl_period); + dl_time_before(now, dl_se->period_start + dl_se->dl_period); } #ifdef CONFIG_RT_MUTEXES @@ -1260,7 +1261,8 @@ static inline void replenish_dl_new_period(struct sched_dl_entity *dl_se, struct rq *rq) { /* for non-boosted task, pi_of(dl_se) == dl_se */ - dl_se->deadline = rq_clock(rq) + pi_of(dl_se)->dl_deadline; + dl_se->period_start = rq_clock(rq); + dl_se->deadline = dl_se->period_start + pi_of(dl_se)->dl_deadline; dl_se->runtime = pi_of(dl_se)->dl_runtime; dl_note_donor_refill(dl_se); @@ -1384,6 +1386,7 @@ static void replenish_dl_entity(struct sched_dl_entity *dl_se) WARN_ON_ONCE(!is_dl_boosted(dl_se)); if (!dl_se->dl_deadline) { dl_se->deadline = rq_clock(rq); + dl_se->period_start = dl_se->deadline; dl_se->runtime = 0; dl_note_donor_refill(dl_se); } @@ -1406,7 +1409,8 @@ static void replenish_dl_entity(struct sched_dl_entity *dl_se) */ if (dl_se->dl_deadline == 0 || (dl_se->dl_defer_armed && dl_entity_overflow(dl_se, rq_clock(rq)))) { - dl_se->deadline = rq_clock(rq) + deadline; + dl_se->period_start = rq_clock(rq); + dl_se->deadline = dl_se->period_start + deadline; dl_se->runtime = runtime; dl_note_donor_refill(dl_se); } @@ -1425,6 +1429,7 @@ static void replenish_dl_entity(struct sched_dl_entity *dl_se) */ while (dl_se->runtime <= 0) { dl_se->deadline += period; + dl_se->period_start += period; dl_se->runtime += runtime; } @@ -1658,9 +1663,14 @@ static void update_dl_entity(struct sched_dl_entity *dl_se) } } +/* + * Return the start of the next period of @dl_se. Do not work it out from + * the deadline and dl_deadline: sched_setattr() can change dl_deadline + * during the period. + */ static inline u64 dl_next_period(struct sched_dl_entity *dl_se) { - return dl_se->deadline - dl_se->dl_deadline + dl_se->dl_period; + return dl_se->period_start + dl_se->dl_period; } #ifdef CONFIG_RT_MUTEXES @@ -1728,7 +1738,7 @@ static int start_dl_timer(struct sched_dl_entity *dl_se) WARN_ON_ONCE(!dl_se->dl_throttled); act = ns_to_ktime(dl_se->deadline - dl_se->runtime); } else { - /* act = deadline - rel-deadline + period */ + /* act = the start of the next period */ act = ns_to_ktime(dl_next_period(dl_se)); } @@ -2603,6 +2613,7 @@ int dl_server_apply_params(struct sched_dl_entity *dl_se, u64 runtime, u64 perio dl_se->runtime = 0; dl_se->deadline = 0; + dl_se->period_start = 0; dl_se->dl_bw = to_ratio(dl_se->dl_period, dl_se->dl_runtime); dl_se->dl_density = to_ratio(dl_se->dl_deadline, dl_se->dl_runtime); @@ -4848,6 +4859,7 @@ static void switched_from_dl(struct rq *rq, struct task_struct *p) */ if (!p->dl.dl_runtime) { p->dl.deadline = 0; + p->dl.period_start = 0; p->dl.runtime = 0; } @@ -5394,6 +5406,7 @@ static void __dl_clear_params(struct sched_dl_entity *dl_se) dl_se->dl_density = 0; dl_se->deadline = 0; + dl_se->period_start = 0; dl_se->runtime = 0; dl_se->dl_throttled = 0; -- 2.47.3