linux-kernel - [PATCH v2 2/2] regulator: core: Sleep (not delay) in set

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220420141511.v2.2.If0fc61a894f537b052ca41572aff098cf8e7e673@changeid>
Date:   Wed, 20 Apr 2022 14:22:13 -0700
From:   Brian Norris <briannorris@...omium.org>
To:     Liam Girdwood <lgirdwood@...il.com>,
        Mark Brown <broonie@...nel.org>
Cc:     linux-kernel@...r.kernel.org, Matthias Kaehlcke <mka@...omium.org>,
        Brian Norris <briannorris@...omium.org>
Subject: [PATCH v2 2/2] regulator: core: Sleep (not delay) in set_voltage()

These delays can be relatively large (e.g., hundreds of microseconds to
several milliseconds on RK3399 Gru systems). Per
Documentation/timers/timers-howto.rst, that should usually use a
sleeping delay. Let's use the existing regulator delay helper to handle
both large and small delays appropriately. This avoids burning a bunch
of CPU time and hurting scheduling latencies when hitting regulators a
lot (e.g., during cpufreq).

The sleep vs. delay issue choice has been made differently over time --
early versions of RK3399 Gru PWM-regulator support used usleep_range()
in pwm-regulator.c. More of this got moved into the regulator core,
in commits like:

73e705bf81ce regulator: core: Add set_voltage_time op

At the same time, the sleep turned into a delay.

It's OK to sleep in _regulator_do_set_voltage(), as we aren't in an
atomic context. (All our callers grab various mutexes already.)

I avoid using fsleep() because it uses a usleep_range() of [N to N*2],
and usleep_range() very commonly biases to the high end of the range. We
don't want to double the expected delay, especially for long delays.

Signed-off-by: Brian Norris <briannorris@...omium.org>
Reviewed-by: Matthias Kaehlcke <mka@...omium.org>
---

Changes in v2:
 * Add Matthias's "Reviewed-by"
 * Use regulator helper instead of fsleep()
 * Reorder patch 1&2

 drivers/regulator/core.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 8b188e56fd0b..0279ff687e1d 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -3548,12 +3548,7 @@ static int _regulator_do_set_voltage(struct regulator_dev *rdev,
 	}

 	/* Insert any necessary delays */
-	if (delay >= 1000) {
-		mdelay(delay / 1000);
-		udelay(delay % 1000);
-	} else if (delay) {
-		udelay(delay);
-	}
+	_regulator_delay_helper(delay);

 	if (best_val >= 0) {
 		unsigned long data = best_val;
-- 
2.36.0.rc0.470.gd361397f0d-goog