diff mbox series

[v1] ACPI: OSL: Use usleep_range() in acpi_os_sleep()

Message ID 5857066.DvuYhMxLoT@rjwysocki.net (mailing list archive)
State Handled Elsewhere, archived
Headers show
Series [v1] ACPI: OSL: Use usleep_range() in acpi_os_sleep() | expand

Commit Message

Rafael J. Wysocki Dec. 5, 2024, 11:24 a.m. UTC
From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Subject: [PATCH v0.1] ACPI: OSL: Use usleep_range() in acpi_os_sleep()

As stated by Len in [1], the extra delay added by msleep() to the
sleep time value passed to it can be significant, roughly between
1.5 ns on systems with HZ = 1000 and as much as 15 ms on systems with
HZ = 100, which is hardly acceptable, at least for small sleep time
values.

msleep(5) on the default HZ = 250 in Ubuntu on a modern PC takes about
12 ms.  This results in over 800 ms of spurious system resume delay on
systems such as the Dell XPS-13-9300, which use ASL Sleep(5ms) in a
tight loop.

Address this by using usleep_range() in acpi_os_sleep() instead of
msleep().  For short sleep times this is a no brainer, but even for
long sleeps usleep_range() should be preferred because timer wheel
timers are optimized for cancelation before they expire and this
particular timer is not going to be canceled.

Add at least 50 us on top of the requested sleep time in case the
timer can be subject to coalescing, which is consistent with what's
done in user space in this context [2], but for sleeps longer than 5 ms
use 1% of the requested sleep time for this purpose.

The rationale here is that longer sleeps don't need that much of a timer
precision as a rule and making the timer a more likely candidate for
coalescing in these cases is generally desirable.  It starts at 5 ms so
that the delta between the requested sleep time and the effective
deadline is a contiuous function of the former.

Link: https://lore.kernel.org/linux-pm/c7db7e804c453629c116d508558eaf46477a2d73.1731708405.git.len.brown@intel.com/ [1]
Link: https://lore.kernel.org/linux-pm/
CAJvTdK=Q1kwWA6Wxn8Zcf0OicDEk6cHYFAvQVizgA47mXu63+g@mail.gmail.com/ [2]
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=216263
Reported-by: Len Brown <lenb@kernel.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Tested-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

The previous RFC version of this patch is here:

https://lore.kernel.org/linux-pm/5839859.DvuYhMxLoT@rjwysocki.net/

The difference between this version and the RFC is a changelog update
suggested by Len, the Closes: tag and the tags from Hans and Mario.

---
 drivers/acpi/osl.c |   22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

Comments

Len Brown Dec. 11, 2024, 4:14 a.m. UTC | #1
On Thu, Dec 5, 2024 at 7:24 AM Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
>
> Add at least 50 us on top of the requested sleep time in case the
> timer can be subject to coalescing, which is consistent with what's
> done in user space in this context [2], but for sleeps longer than 5 ms
> use 1% of the requested sleep time for this purpose.
>
> The rationale here is that longer sleeps don't need that much of a timer
> precision as a rule and making the timer a more likely candidate for
> coalescing in these cases is generally desirable.  It starts at 5 ms so
> that the delta between the requested sleep time and the effective
> deadline is a contiuous function of the former.

timerslack_ns defaults to 50,000 ns.

So when a user invokes nanosleep(50ms), they get slacked out to 50.050 ms

With this patch, if the AML BIOS programmer invokes Sleep(50ms),
it gets slacked out to 50.500 ms -- a 10x longer slack period.

I have not seen an explanation for why these cases should be treated
differently.

Len Brown, Intel Open Source Technology Center
diff mbox series

Patch

Index: linux-pm/drivers/acpi/osl.c
===================================================================
--- linux-pm.orig/drivers/acpi/osl.c
+++ linux-pm/drivers/acpi/osl.c
@@ -607,7 +607,27 @@  acpi_status acpi_os_remove_interrupt_han
 
 void acpi_os_sleep(u64 ms)
 {
-	msleep(ms);
+	u64 usec = ms * USEC_PER_MSEC, delta_us = 50;
+
+	/*
+	 * Use a hrtimer because the timer wheel timers are optimized for
+	 * cancelation before they expire and this timer is not going to be
+	 * canceled.
+	 *
+	 * Set the delta between the requested sleep time and the effective
+	 * deadline to at least 50 us in case there is an opportunity for timer
+	 * coalescing.
+	 *
+	 * Moreover, longer sleeps can be assumed to need somewhat less timer
+	 * precision, so sacrifice some of it for making the timer a more likely
+	 * candidate for coalescing by setting the delta to 1% of the sleep time
+	 * if it is above 5 ms (this value is chosen so that the delta is a
+	 * continuous function of the sleep time).
+	 */
+	if (ms > 5)
+		delta_us = (USEC_PER_MSEC / 100) * ms;
+
+	usleep_range(usec, usec + delta_us);
 }
 
 void acpi_os_stall(u32 us)