Skip to contentSkip to conversation

x86/apic: Stop the TSC Deadline timer during lapic timer shutdown

PATCH In review Zhang Rui 5 participants 14 messages Oct 9, 2024 – Oct 11, 2024

Subsystems (2)
lore
Share & tools
lore

Zhang Rui

author
#0 lore raw

PATCH v2 1/1 x86/apic: Stop the TSC Deadline timer during lapic timer shutdown

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

According to Intel SDM, for the local APIC timer,
1. "The initial-count register is a read-write register. A write of 0 to the initial-count register effectively stops the local APIC timer, in both one-shot and periodic mode."
2. "In TSC deadline mode, writes to the initial-count register are ignored; and current-count register always reads 0. Instead, timer behavior is controlled using the IA32_TSC_DEADLINE MSR." "In TSC-deadline mode, writing 0 to the IA32_TSC_DEADLINE MSR disarms the local-APIC timer."

Stop the TSC Deadline timer in lapic_timer_shutdown() by writing 0 to MSR_IA32_TSC_DEADLINE.

Fixes 279f1461432c ("x86: apic: Use tsc deadline for oneshot when available")

---
Changes since V1
- improve changelog

---
 arch/x86/kernel/apic/apic.c | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/kernel/apic/apic.c +4 -0
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
 	v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
 	apic_write(APIC_LVTT, v);
 	apic_write(APIC_TMICT, 0);
+
+	if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+		wrmsrl(MSR_IA32_TSC_DEADLINE, 0);
+
 	return 0;
 }
 
Signature · 1 line
2.34.1
Acked-by on v2 1/1
#1 lore raw
Quoted 16 lines · Zhang Rui

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

According to Intel SDM, for the local APIC timer,
1. "The initial-count register is a read-write register. A write of 0 to the initial-count register effectively stops the local APIC timer, in both one-shot and periodic mode."
2. "In TSC deadline mode, writes to the initial-count register are ignored; and current-count register always reads 0. Instead, timer behavior is controlled using the IA32_TSC_DEADLINE MSR." "In TSC-deadline mode, writing 0 to the IA32_TSC_DEADLINE MSR disarms the local-APIC timer."

Quoted 6 lines

Stop the TSC Deadline timer in lapic_timer_shutdown() by writing 0 to MSR_IA32_TSC_DEADLINE.

Fixes 279f1461432c ("x86: apic: Use tsc deadline for oneshot when available")

x86 folks, this is quite nasty, so please make it high-prio.

Alternatively, I can take it through the PM tree.

Quoted 26 lines

---
Changes since V1
- improve changelog

---
 arch/x86/kernel/apic/apic.c | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/kernel/apic/apic.c +4 -1
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
        v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
        apic_write(APIC_LVTT, v);
        apic_write(APIC_TMICT, 0);
+
+       if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+               wrmsrl(MSR_IA32_TSC_DEADLINE, 0);
+
        return 0;
 }
 
--

2.34.1

#2 lore raw
Quoted 3 lines · Rafael J. Wysocki

x86 folks, this is quite nasty, so please make it high-prio.

How much linux-next soak time do you think this needs? We'd ideally like to give it a week in x86/urgent.

#3 lore raw
Quoted 8 lines · Dave Hansen
Quoted 3 lines · Rafael J. Wysocki

x86 folks, this is quite nasty, so please make it high-prio.

How much linux-next soak time do you think this needs? We'd ideally like to give it a week in x86/urgent.

That works, a week in x86/urgent should be fine.

Thank you!

#4 lore raw
In reply to Zhang Rui
Quoted 2 lines · Zhang Rui

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems

Two nits:

Quoted 2 lines

gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

s/gets/get/
s/energy start/Energy Star/

#5 lore raw
Quoted 5 lines · Ricardo Neri
Quoted 2 lines · Zhang Rui

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems

Quoted 7 lines

Two nits:

Quoted 2 lines

gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

s/gets/get/
s/energy start/Energy Star/

Thanks for pointing these out!

#6 lore raw
In reply to Zhang Rui
Quoted 4 lines · Zhang Rui

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

Why haven't we noticed or cared for the last 12 years?

Also, plain language really matters. Is this as simple as: "you close the lid on the laptop and the CPU doesn't power down at all"?

Quoted 9 lines

According to Intel SDM, for the local APIC timer,
1. "The initial-count register is a read-write register. A write of 0 to the initial-count register effectively stops the local APIC timer, in both one-shot and periodic mode."
2. "In TSC deadline mode, writes to the initial-count register are ignored; and current-count register always reads 0. Instead, timer behavior is controlled using the IA32_TSC_DEADLINE MSR." "In TSC-deadline mode, writing 0 to the IA32_TSC_DEADLINE MSR disarms the local-APIC timer."

Is "stopping" and "disarming" the same thing?

Second, while quoting the SDM is great, it would be even better to including the Linux naming for these things. The Linux naming for the APIC registers is completely missing from this changelog. You could say:

"In TSC deadline mode, writes to the initial-count register (APIC_TMICT) are ignored"

which makes it much easier to relate this code:

        apic_write(APIC_TMICT, 0);

back to the SDM language. This is especially true because:

#define APIC_TMICT 0x380

doesn't make it obvious that "ICT" is the "Initial-Count Register". I had to go back to the SDM table to make 100% sure.

This also doesn't ever say which mode the kernel is running in.

Quoted 2 lines

Stop the TSC Deadline timer in lapic_timer_shutdown() by writing 0 to MSR_IA32_TSC_DEADLINE.

This dances around the problem but never comes out and says it:

The CPU package does not go into lower power modes (higher package C-states) unless all local-APIC timers are disabled.

Plus something to connect the old to the new:

On older CPUs, setting APIC_TMICT=0 was sufficient for disabling the local-APIC timer, no matter the timer mode (deadline, one- shot or periodic). But newer CPUs adhere to the strict letter of the law in the SDM and more fully ignore APIC_TMICT when in deadline mode. Those CPUs also don't fully "disable" the timer when IA32_TSC_DEADLINE has passed. They _require_ writing a 0.

Or am I missing something?

Quoted 15 lines
arch/x86/kernel/apic/apic.c +4 -0
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
 	v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
 	apic_write(APIC_LVTT, v);
 	apic_write(APIC_TMICT, 0);
+
+	if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+		wrmsrl(MSR_IA32_TSC_DEADLINE, 0);
+
 	return 0;
 }
 
#7 lore raw
Quoted 8 lines · Dave Hansen
Quoted 4 lines · Zhang Rui

This 12-year-old bug prevents some modern processors from achieving maximum power savings during suspend. For example, Lunar Lake systems gets 0% package C-states during suspend to idle and this causes energy star compliance tests to fail.

Why haven't we noticed or cared for the last 12 years?

Because on the previous platforms the apic_write(APIC_TMICT, 0) in lapic_timer_shutdown() was sufficient even in the TSC deadline mode, or at least no problems with it have ever been reported.

Quoted 2 lines

Also, plain language really matters. Is this as simple as: "you close the lid on the laptop and the CPU doesn't power down at all"?

It will power down somewhat, but not as much as to justify suspending the system. It will just drain the battery at a relatively high rate which will also cause the machine to heat up.

Not a good idea to put it into a bag in this state, for instance.

Quoted 11 lines
Quoted 9 lines

According to Intel SDM, for the local APIC timer,
1. "The initial-count register is a read-write register. A write of 0 to the initial-count register effectively stops the local APIC timer, in both one-shot and periodic mode."
2. "In TSC deadline mode, writes to the initial-count register are ignored; and current-count register always reads 0. Instead, timer behavior is controlled using the IA32_TSC_DEADLINE MSR." "In TSC-deadline mode, writing 0 to the IA32_TSC_DEADLINE MSR disarms the local-APIC timer."

Is "stopping" and "disarming" the same thing?

That is my understanding. If you disarm it and it is not armed again, it will be stopped effectively.

Quoted 30 lines

Second, while quoting the SDM is great, it would be even better to including the Linux naming for these things. The Linux naming for the APIC registers is completely missing from this changelog. You could say:

"In TSC deadline mode, writes to the initial-count register (APIC_TMICT) are ignored"

which makes it much easier to relate this code:

        apic_write(APIC_TMICT, 0);

back to the SDM language. This is especially true because:

#define APIC_TMICT 0x380

doesn't make it obvious that "ICT" is the "Initial-Count Register". I had to go back to the SDM table to make 100% sure.

This also doesn't ever say which mode the kernel is running in.

Quoted 2 lines

Stop the TSC Deadline timer in lapic_timer_shutdown() by writing 0 to MSR_IA32_TSC_DEADLINE.

This dances around the problem but never comes out and says it:

The CPU package does not go into lower power modes (higher package C-states) unless all local-APIC timers are disabled.

Plus something to connect the old to the new:

Quoted 8 lines

On older CPUs, setting APIC_TMICT=0 was sufficient for disabling the local-APIC timer, no matter the timer mode (deadline, one- shot or periodic). But newer CPUs adhere to the strict letter of the law in the SDM and more fully ignore APIC_TMICT when in deadline mode. Those CPUs also don't fully "disable" the timer when IA32_TSC_DEADLINE has passed. They _require_ writing a 0.

Or am I missing something?

No, you are right.

We need a new version of the patch with a better changelog.

Quoted 17 lines
Quoted 15 lines
arch/x86/kernel/apic/apic.c +4 -0
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
      v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
      apic_write(APIC_LVTT, v);
      apic_write(APIC_TMICT, 0);
+
+     if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+             wrmsrl(MSR_IA32_TSC_DEADLINE, 0);
+
      return 0;
 }
 
#8 lore raw
In reply to Zhang Rui

arch/x86/kernel/apic/apic.c static int lapic_timer_shutdown(struct clock_event_device *evt) · line 446

Quoted 11 lines · Zhang Rui
arch/x86/kernel/apic/apic.c +3 -0
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
 	v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
 	apic_write(APIC_LVTT, v);
 	apic_write(APIC_TMICT, 0);
+
+	if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+		wrmsrl(MSR_IA32_TSC_DEADLINE, 0);

One last thing, and this is a super nit. We presumably have the actual APIC_LVTT value (v) sitting in a register already. Is there any difference logically between a X86_FEATURE_TSC_DEADLINE_TIMER check and an APIC_LVTT check for APIC_LVT_TIMER_TSCDEADLINE?

I suspect this will generate more compact code:

	if (v & APIC_LVT_TIMER_TSCDEADLINE)
		wrmsrl(MSR_IA32_TSC_DEADLINE, 0);

Does it have any downsides?

Oh, and how hot is this path? Is this wrmsr() going to matter? I presume it's pretty cheap because it's one of the special architecturally non-serializing WRMSRs.

#9 lore raw
Quoted 19 lines · Dave Hansen

arch/x86/kernel/apic/apic.c static int lapic_timer_shutdown(struct clock_event_device *evt) · line 446

Quoted 11 lines · Zhang Rui
arch/x86/kernel/apic/apic.c +3 -0
index 6513c53c9459..d1006531729a 100644
@@ -441,6 +441,10 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
      v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
      apic_write(APIC_LVTT, v);
      apic_write(APIC_TMICT, 0);
+
+     if (boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+             wrmsrl(MSR_IA32_TSC_DEADLINE, 0);

One last thing, and this is a super nit. We presumably have the actual APIC_LVTT value (v) sitting in a register already. Is there any difference logically between a X86_FEATURE_TSC_DEADLINE_TIMER check and an APIC_LVTT check for APIC_LVT_TIMER_TSCDEADLINE?

Quoted 6 lines

I suspect this will generate more compact code:

if (v & APIC_LVT_TIMER_TSCDEADLINE)
wrmsrl(MSR_IA32_TSC_DEADLINE, 0);

Does it have any downsides?

I don't see any.

Quoted 3 lines

Oh, and how hot is this path? Is this wrmsr() going to matter? I presume it's pretty cheap because it's one of the special architecturally non-serializing WRMSRs.

lapic_timer_shutdown() is called under a raw spin lock in ___tick_broadcast_oneshot_control(), so it better not take too much time or PREEMPT_RT might be unhappy. I'm not sure how often that happens, though.

Also tick_program_event() calls it to stop the tick, but it is assumed that this may take time AFAICS.

#10 lore raw

How about something like the completely untested attached patch?

IMNHO, it improves on what was posted here because it draws a parallel with an AMD erratum and also avoids writes to APIC_TMICT that would get ignored anyway.

Subject: x86/apic: Always explicitly disarm TSC-deadline timer

New processors have become pickier about the local APIC timer state before entering low power modes. These low power modes are used (for example) when you close your laptop lid and suspend. If you put your laptop in a bag in this unnecessarily-high-power state, it is likely to get quite toasty while it quickly sucks the battery dry.

The problem boils down to some CPUs' inability to power down until the kernel fully disables the local APIC timer. The current kernel code works in one-shot and periodic modes but does not work for deadline mode. Deadline mode has been the supported and preferred mode on Intel CPUs for over a decade and uses an MSR to drive the timer instead of an APIC register.

Disable the TSC Deadline timer in lapic_timer_shutdown() by writing to MSR_IA32_TSC_DEADLINE when in TSC-deadline mode. Also avoid writing to the initial-count register (APIC_TMICT) which is ignored in TSC-deadline mode.

Note: The APIC_LVTT|=APIC_LVT_MASKED operation should theoretically be enough to tell the hardware that the timer will not fire in any of the timer modes. But mitigating AMD erratum 411[1] also requires clearing out APIC_TMICT. Solely setting APIC_LVT_MASKED is also ineffective in practice on Intel Lunar Lake systems, which is the motivation for this change.

1. 411 Processor May Exit Message-Triggered C1E State Without an Interrupt if Local APIC Timer Reaches Zero - https://www.amd.com/content/dam/amd/en/documents/archived-tech-docs/revision-guides/41322_10h_Rev_Gd.pdf

														    //
arch/x86/kernel/apic/apic.c +13 -1
index 6513c53c9459e..5436a4083065d 100644
@@ -440,7 +440,19 @@ static int lapic_timer_shutdown(struct clock_event_device *evt)
 	v = apic_read(APIC_LVTT);
 	v |= (APIC_LVT_MASKED | LOCAL_TIMER_VECTOR);
 	apic_write(APIC_LVTT, v);
-	apic_write(APIC_TMICT, 0);
+
+	/*
+	 * Setting APIC_LVT_MASKED should be enough to tell the
+	 * hardware that this timer will never fire. But AMD
+	 * erratum 411 and some Intel CPU behavior circa 2024
+	 * say otherwise. Time for belt and suspenders programming,
+	 * mask the timer and zero the counter registers:
+	 */
+	if (v & APIC_LVT_TIMER_TSCDEADLINE)
+		wrmsrl(MSR_IA32_TSC_DEADLINE, 0);
+	else
+		apic_write(APIC_TMICT, 0);
+
 	return 0;
 }
 
author
#11 lore raw

Hi, Dave,

Quoted 7 lines · Dave Hansen

How about something like the completely untested attached patch?

IMNHO, it improves on what was posted here because it draws a parallel
with an AMD erratum and also avoids writes to APIC_TMICT that would get
ignored anyway.

Thanks a lot for the rewrite. The patch looks great to me.

I will test it on our test boxes to make sure skipping the APIC_TMICT write in TSC deadline mode does not bring unexpected impact to the old machines.

thanks,
rui

Reviewed-by
#12 lore raw
In reply to Dave Hansen
Quoted 6 lines · Dave Hansen

How about something like the completely untested attached patch?

IMNHO, it improves on what was posted here because it draws a parallel with an AMD erratum and also avoids writes to APIC_TMICT that would get ignored anyway.

Please feel free to add

to this one when it's ready.

Tested-by Srinivas Pandruvada
#13 lore raw
Quoted 11 lines · Rafael J. Wysocki
Quoted 8 lines · Dave Hansen

How about something like the completely untested attached patch?

IMNHO, it improves on what was posted here because it draws a parallel
with an AMD erratum and also avoids writes to APIC_TMICT that would get
ignored anyway.

Quoted 5 lines

Please feel free to add

to this one when it's ready.

Tested on Lunar Lake for the new patch