Skip to contentSkip to conversation

x86/speculation: Disable IBRS when idle

PATCH Superseded Waiman Long 4 participants 15 messages Jun 22, 2023

Subsystems (3)

v3 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v3 cover x86/speculation: Disable IBRS when idle

v3:
- Drop patches 1 ("x86/speculation: Provide a debugfs file to dump SPEC_CTRL MSRs") and 5 ("x86/idle: Disable IBRS entering mwait idle and enable it on wakeup") for now.
- Drop the MSR restoration code in ("x86/idle: Disable IBRS when cpu is offline") as native_play_dead() does not return.
- For patch ("intel_idle: Add ibrs_off module parameter to force disable IBRS"), change the name from "no_ibrs" to "ibrs_off" and document the new parameter in intel_idle.rst.

For Intel processors that need to turn on IBRS to protect against Spectre v2 and Retbleed, the IBRS bit in the SPEC_CTRL MSR affects the performance of the whole core even if only one thread is turning it on when running in the kernel. For user space heavy applications, the performance impact of occasionally turning IBRS on during syscalls shouldn't be significant. Unfortunately, that is not the case when the sibling thread is idling in the kernel. In that case, the performance impact can be significant.

When DPDK is running on an isolated CPU thread processing network packets in user space while its sibling thread is idle. The performance of the busy DPDK thread with IBRS on and off in the sibling idle thread are:

IBRS on IBRS off
------- --------
packets/second: 7.8M 10.4M
avg tsc cycles/packet: 282.26 209.86

This is a 25% performance degradation. The test system is a Intel Xeon 4114 CPU @ 2.20GHz.

Commit bf5835bcdb96 ("intel_idle: Disable IBRS during long idle") disables IBRS when the CPU enters long idle (C6 or below). However, there are existing users out there who have set "intel_idle.max_cstate=1" to decrease latency. Those users won't be able to benefit from this commit. This patch series extends this commit by providing a new "intel_idle.ibrs_off" module parameter to force disable IBRS even when "intel_idle.max_cstate=1" at the expense of increased IRQ response latency. It also includes a commit to allow the disabling of IBRS when a CPU becomes offline.

Waiman Long (3):
x86/idle: Disable IBRS when cpu is offline
intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current
intel_idle: Add ibrs_off module parameter to force disable IBRS

 Documentation/admin-guide/pm/intel_idle.rst | 17 +++++++++++++++-
 arch/x86/kernel/smpboot.c                   | 10 ++++++++++
 drivers/idle/intel_idle.c                   | 22 +++++++++++++++++----
 3 files changed, 44 insertions(+), 5 deletions(-)
Signature · 1 line
2.31.1
#1 lore raw

PATCH v3 1/3 x86/idle: Disable IBRS when cpu is offline

Commit bf5835bcdb96 ("intel_idle: Disable IBRS during long idle") disables IBRS when the CPU enters long idle. However, when a CPU becomes offline, the IBRS bit is still set when X86_FEATURE_KERNEL_IBRS is enabled. That will impact the performance of a sibling CPU. Mitigate this performance impact by clearing all the mitigation bits in SPEC_CTRL MSR when offline. When the CPU is online again, it will be re-initialized and so restoring the SPEC_CTRL value isn't needed.

Add a comment to say that native_play_dead() is a __noreturn function, but it can't be marked as such to avoid confusion about the missing MSR restoration code.

---
 arch/x86/kernel/smpboot.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/x86/kernel/smpboot.c +10 -0
index 352f0ce1ece4..7bc33885518c 100644
@@ -84,6 +84,7 @@
 #include <asm/hw_irq.h>
 #include <asm/stackprotector.h>
 #include <asm/sev.h>
+#include <asm/nospec-branch.h>
 
 /* representing HT siblings of each logical CPU */
 DEFINE_PER_CPU_READ_MOSTLY(cpumask_var_t, cpu_sibling_map);
@@ -1836,8 +1837,17 @@ void __noreturn hlt_play_dead(void)
 	}
 }
 
+/*
+ * naitve_play_dead() is essentially a __noreturn function, but it can't
+ * be marked as such as the compiler may complain about it.
+ */
 void native_play_dead(void)
 {
+	if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS)) {
+		this_cpu_write(x86_spec_ctrl_current, 0);
+		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
+
 	play_dead_common();
 	tboot_shutdown(TB_SHUTDOWN_WFS);
 
Signature · 1 line
2.31.1
reviewer
#2 lore raw

Re: [PATCH v3 1/3] x86/idle: Disable IBRS when cpu is offline

Quoted 30 lines · Waiman Long

Commit bf5835bcdb96 ("intel_idle: Disable IBRS during long idle") disables IBRS when the CPU enters long idle. However, when a CPU becomes offline, the IBRS bit is still set when X86_FEATURE_KERNEL_IBRS is enabled. That will impact the performance of a sibling CPU. Mitigate this performance impact by clearing all the mitigation bits in SPEC_CTRL MSR when offline. When the CPU is online again, it will be re-initialized and so restoring the SPEC_CTRL value isn't needed.

Add a comment to say that native_play_dead() is a __noreturn function, but it can't be marked as such to avoid confusion about the missing MSR restoration code.

---
 arch/x86/kernel/smpboot.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/x86/kernel/smpboot.c +1 -0
index 352f0ce1ece4..7bc33885518c 100644
@@ -84,6 +84,7 @@
 #include <asm/hw_irq.h>
 #include <asm/stackprotector.h>
 #include <asm/sev.h>
+#include <asm/nospec-branch.h>
 
 /* representing HT siblings of each logical CPU */
 DEFINE_PER_CPU_READ_MOSTLY(cpumask_var_t, cpu_sibling_map);

arch/x86/kernel/smpboot.c void __noreturn hlt_play_dead(void) · line 1841

Quoted 5 lines
arch/x86/kernel/smpboot.c +2 -0
@@ -1836,8 +1837,17 @@ void __noreturn hlt_play_dead(void)
 	}
 }
 
+/*
+ * naitve_play_dead() is essentially a __noreturn function, but it can't

typo: native_play_dead()

Quoted 12 lines
arch/x86/kernel/smpboot.c +7 -0
+ * be marked as such as the compiler may complain about it.
+ */
 void native_play_dead(void)
 {
+	if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS)) {
+		this_cpu_write(x86_spec_ctrl_current, 0);
+		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
+
 	play_dead_common();
 	tboot_shutdown(TB_SHUTDOWN_WFS);
 
Signature · 1 line
~Randy
#3 lore raw

Re: [PATCH v3 1/3] x86/idle: Disable IBRS when cpu is offline

Quoted 35 lines · Randy Dunlap
Quoted 33 lines · Waiman Long

Commit bf5835bcdb96 ("intel_idle: Disable IBRS during long idle") disables IBRS when the CPU enters long idle. However, when a CPU becomes offline, the IBRS bit is still set when X86_FEATURE_KERNEL_IBRS is enabled. That will impact the performance of a sibling CPU. Mitigate this performance impact by clearing all the mitigation bits in SPEC_CTRL MSR when offline. When the CPU is online again, it will be re-initialized and so restoring the SPEC_CTRL value isn't needed.

Add a comment to say that native_play_dead() is a __noreturn function, but it can't be marked as such to avoid confusion about the missing MSR restoration code.

---
  arch/x86/kernel/smpboot.c | 10 ++++++++++
  1 file changed, 10 insertions(+)
arch/x86/kernel/smpboot.c +1 -0
index 352f0ce1ece4..7bc33885518c 100644
@@ -84,6 +84,7 @@
  #include <asm/hw_irq.h>
  #include <asm/stackprotector.h>
  #include <asm/sev.h>
+#include <asm/nospec-branch.h>
 
  /* representing HT siblings of each logical CPU */
  DEFINE_PER_CPU_READ_MOSTLY(cpumask_var_t, cpu_sibling_map);
@@ -1836,8 +1837,17 @@ void __noreturn hlt_play_dead(void)
  	}
  }
 
Quoted 3 lines

arch/x86/kernel/smpboot.c void __noreturn hlt_play_dead(void) · line 1841

Quoted 2 lines
arch/x86/kernel/smpboot.c +2 -0
+/*
+ * naitve_play_dead() is essentially a __noreturn function, but it can't

typo: native_play_dead()

Thanks for spotting the typo. Will fix it in the next version.

Cheers,
Longman

Quoted 13 lines
Quoted 13 lines
arch/x86/kernel/smpboot.c +7 -0
+ * be marked as such as the compiler may complain about it.
+ */
  void native_play_dead(void)
  {
+	if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS)) {
+		this_cpu_write(x86_spec_ctrl_current, 0);
+		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
+
  	play_dead_common();
  	tboot_shutdown(TB_SHUTDOWN_WFS);
 
#4 lore raw

Re: [PATCH v3 1/3] x86/idle: Disable IBRS when cpu is offline

In reply to Waiman Long
Quoted 4 lines · Waiman Long

+/*
+ * naitve_play_dead() is essentially a __noreturn function, but it can't
+ * be marked as such as the compiler may complain about it.
+ */

FWIW, we could in theory do so by marking the smp_ops.play_dead function pointer as __noreturn, but it would be tricky to teach objtool how to understand that.

arch/x86/kernel/smpboot.c void __noreturn hlt_play_dead(void) · line 1849

Quoted 6 lines

void native_play_dead(void)
{

arch/x86/kernel/smpboot.c +4 -0
+	if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS)) {
+		this_cpu_write(x86_spec_ctrl_current, 0);
+		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}

Can update_spec_ctrl() be used instead?

Signature · 1 line
Josh
#5 lore raw

Re: [PATCH v3 1/3] x86/idle: Disable IBRS when cpu is offline

Quoted 8 lines · Josh Poimboeuf
Quoted 4 lines · Waiman Long

+/*
+ * naitve_play_dead() is essentially a __noreturn function, but it can't
+ * be marked as such as the compiler may complain about it.
+ */

FWIW, we could in theory do so by marking the smp_ops.play_dead function pointer as __noreturn, but it would be tricky to teach objtool how to understand that.

I added the comment here because I had taken out the MSR restoration code. We can always replace that later on if there is a better way to do that.

Quoted 8 lines

arch/x86/kernel/smpboot.c void __noreturn hlt_play_dead(void) · line 1849

Quoted 7 lines
  void native_play_dead(void)
  {
arch/x86/kernel/smpboot.c +4 -0
+	if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS)) {
+		this_cpu_write(x86_spec_ctrl_current, 0);
+		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}

Can update_spec_ctrl() be used instead?

Yes, the code is similar to what has been done in update_spec_ctrl(). Using it, however, will require exporting the function either by putting it into a public header or making it a global function.

Cheers,
Longman

Quoted 1 line
#6 lore raw

PATCH v3 2/3 intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

In reply to Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

---
 drivers/idle/intel_idle.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/idle/intel_idle.c +6 -2
index aa2d19db2b1d..07fa23707b3c 100644
@@ -181,13 +181,17 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
 	u64 spec_ctrl = spec_ctrl_current();
 	int ret;
 
-	if (smt_active)
+	if (smt_active) {
+		__this_cpu_write(x86_spec_ctrl_current, 0);
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
 
 	ret = __intel_idle(dev, drv, index);
 
-	if (smt_active)
+	if (smt_active) {
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__this_cpu_write(x86_spec_ctrl_current, spec_ctrl);
+	}
 
 	return ret;
 }
Signature · 1 line
2.31.1
#7 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Is this fixing an actual bug or is there some other reason for doing this?

Quoted 17 lines
---
 drivers/idle/intel_idle.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/idle/intel_idle.c +2 -1
index aa2d19db2b1d..07fa23707b3c 100644
@@ -181,13 +181,17 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
 	u64 spec_ctrl = spec_ctrl_current();
 	int ret;
 
-	if (smt_active)
+	if (smt_active) {
+		__this_cpu_write(x86_spec_ctrl_current, 0);

drivers/idle/intel_idle.c static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev, · line 194

Quoted 10 lines
drivers/idle/intel_idle.c +4 -1
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
 
 	ret = __intel_idle(dev, drv, index);
 
-	if (smt_active)
+	if (smt_active) {
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__this_cpu_write(x86_spec_ctrl_current, spec_ctrl);
+	}

More candidates for update_spec_ctrl()?

Signature · 1 line
Josh
#8 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

Quoted 33 lines · Josh Poimboeuf
Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Is this fixing an actual bug or is there some other reason for doing this?

Quoted 22 lines
---
 drivers/idle/intel_idle.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/idle/intel_idle.c +3 -1
index aa2d19db2b1d..07fa23707b3c 100644
@@ -181,13 +181,17 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
 	u64 spec_ctrl = spec_ctrl_current();
 	int ret;
 
-	if (smt_active)
+	if (smt_active) {
+		__this_cpu_write(x86_spec_ctrl_current, 0);
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
 
 	ret = __intel_idle(dev, drv, index);
 
Quoted 7 lines

drivers/idle/intel_idle.c static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev, · line 194

Quoted 5 lines
drivers/idle/intel_idle.c +3 -1
-	if (smt_active)
+	if (smt_active) {
 		native_wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__this_cpu_write(x86_spec_ctrl_current, spec_ctrl);
+	}

More candidates for update_spec_ctrl()?

Both this and the play_dead case can't use update_spec_ctrl() because RCU isn't there anymore and all that is noinstr. Additionally, both sites rely on preemption being off already, where update_spec_ctrl() can't do that.

That said, I suppose one could write it like so:

static __always_inline __update_spec_ctrl(u64 val)

{
	__this_cpu_write(x86_spec_ctrl_current, val);
	native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
}

static void update_spec_ctrl(u64 val)

{
	preempt_disable();
	__update_spec_ctrl(val);
	preempt_enable();
}

And then you can use __update_spec_ctrl(). But that would need a wee audit of using native_wrmsrl() in all places, probably ok, IIRC Xen wasn't using our IBRS stuff anyway.

#9 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

Quoted 43 lines · Peter Zijlstra
Quoted 29 lines · Josh Poimboeuf
Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Is this fixing an actual bug or is there some other reason for doing this?

Quoted 19 lines
---
  drivers/idle/intel_idle.c | 8 ++++++--
  1 file changed, 6 insertions(+), 2 deletions(-)
drivers/idle/intel_idle.c +3 -1
index aa2d19db2b1d..07fa23707b3c 100644
@@ -181,13 +181,17 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
  	u64 spec_ctrl = spec_ctrl_current();
  	int ret;
 
-	if (smt_active)
+	if (smt_active) {
+		__this_cpu_write(x86_spec_ctrl_current, 0);
  		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
 
Quoted 8 lines

drivers/idle/intel_idle.c static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev, · line 194

Quoted 7 lines
drivers/idle/intel_idle.c +3 -1
  	ret = __intel_idle(dev, drv, index);
 
-	if (smt_active)
+	if (smt_active) {
  		native_wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__this_cpu_write(x86_spec_ctrl_current, spec_ctrl);
+	}

More candidates for update_spec_ctrl()?

Both this and the play_dead case can't use update_spec_ctrl() because RCU isn't there anymore and all that is noinstr. Additionally, both sites rely on preemption being off already, where update_spec_ctrl() can't do that.

Quoted 7 lines

That said, I suppose one could write it like so:

static __always_inline __update_spec_ctrl(u64 val)

{
	__this_cpu_write(x86_spec_ctrl_current, val);
	native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
}

We can put this into asm/nospec-branch.h since x86_spec_ctrl_current is defined there as well.

Cheers,
Longman

#10 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

In reply to Peter Zijlstra
Quoted 4 lines · Peter Zijlstra

Both this and the play_dead case can't use update_spec_ctrl() because RCU isn't there anymore and all that is noinstr. Additionally, both sites rely on preemption being off already, where update_spec_ctrl() can't do that.

Oh, I might be wrong about the preemption thing; it doesn't make sense to do wrmsr with preemption on, so it could be simpler.

Quoted 18 lines

That said, I suppose one could write it like so:

static __always_inline __update_spec_ctrl(u64 val)

{
	__this_cpu_write(x86_spec_ctrl_current, val);
	native_wrmsrl(MSR_IA32_SPEC_CTRL, val);
}

static void update_spec_ctrl(u64 val)

{
	preempt_disable();
	__update_spec_ctrl(val);
	preempt_enable();
}

And then you can use __update_spec_ctrl(). But that would need a wee audit of using native_wrmsrl() in all places, probably ok, IIRC Xen wasn't using our IBRS stuff anyway.

#11 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

In reply to Josh Poimboeuf
Quoted 10 lines · Josh Poimboeuf
Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Is this fixing an actual bug or is there some other reason for doing this?

No actual bug, he did this for his debugfs file -- which is no longer part of the series. With that on, you can observe the x86_spec_ctrl_current value while idle.

Arguably that's not even inconsistent because we disable/enable the thing with IRQs disabled, so nothing can observe the difference.

#12 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

Quoted 8 lines · Peter Zijlstra
Quoted 7 lines · Josh Poimboeuf
Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Quoted 5 lines
Quoted 2 lines

Is this fixing an actual bug or is there some other reason for doing this?

No actual bug, he did this for his debugfs file -- which is no longer part of the series. With that on, you can observe the x86_spec_ctrl_current value while idle.

Right. That is the main reason as I want the SPEC_CTRL MSRs value to be observable by some external mean:-)

Regards,
Longman

#13 lore raw

Re: [PATCH v3 2/3] intel_idle: Sync up the SPEC_CTRL MSR value to x86_spec_ctrl_current

In reply to Josh Poimboeuf
Quoted 9 lines · Josh Poimboeuf
Quoted 6 lines · Waiman Long

When intel_idle_ibrs() is called, it modifies the SPEC_CTRL MSR to 0 in order disable IBRS. However, the new MSR value isn't reflected in x86_spec_ctrl_current which is at odd with the other code that keep track of its state in that percpu variable. Fix that by updating x86_spec_ctrl_current percpu value to always match the content of the SPEC_CTRL MSR.

Is this fixing an actual bug or is there some other reason for doing this?

It is not a bug per se. It is mainly to make the per cpu variable more up to date.

Quoted 20 lines
Quoted 20 lines
---
  drivers/idle/intel_idle.c | 8 ++++++--
  1 file changed, 6 insertions(+), 2 deletions(-)
drivers/idle/intel_idle.c +3 -1
index aa2d19db2b1d..07fa23707b3c 100644
@@ -181,13 +181,17 @@ static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev,
  	u64 spec_ctrl = spec_ctrl_current();
  	int ret;
 
-	if (smt_active)
+	if (smt_active) {
+		__this_cpu_write(x86_spec_ctrl_current, 0);
  		native_wrmsrl(MSR_IA32_SPEC_CTRL, 0);
+	}
 
Quoted 8 lines

drivers/idle/intel_idle.c static __cpuidle int intel_idle_ibrs(struct cpuidle_device *dev, · line 194

Quoted 7 lines
drivers/idle/intel_idle.c +3 -1
  	ret = __intel_idle(dev, drv, index);
 
-	if (smt_active)
+	if (smt_active) {
  		native_wrmsrl(MSR_IA32_SPEC_CTRL, spec_ctrl);
+		__this_cpu_write(x86_spec_ctrl_current, spec_ctrl);
+	}

More candidates for update_spec_ctrl()?

I don't think we can use update_spec_ctrl() here simply because of the noinstr requirement. See commit 9b461a6faae7 ("cpuidle, intel_idle: Fix CPUIDLE_FLAG_IBRS").

Cheers,
Longman

#14 lore raw

PATCH v3 3/3 intel_idle: Add ibrs_off module parameter to force disable IBRS

In reply to Waiman Long

Commit bf5835bcdb96 ("intel_idle: Disable IBRS during long idle") disables IBRS when the cstate is 6 or lower. However, there are some use cases where a customer may want to use max_cstate=1 to lower latency. Such use cases will suffer from the performance degradation caused by the enabling of IBRS in the sibling idle thread. Add a "ibrs_off" module parameter to force disable IBRS and the CPUIDLE_FLAG_IRQ_ENABLE flag if set.

In the case of a Skylake server with max_cstate=1, this new ibrs_off option will likely increase the IRQ response latency as IRQ will now be disabled.

When running SPECjbb2015 with cstates set to C1 on a Skylake system.

First test when the kernel is booted with: "intel_idle.ibrs_off"
max-jOPS = 117828, critical-jOPS = 66047

Then retest when the kernel is booted without the "intel_idle.ibrs_off" added.
max-jOPS = 116408, critical-jOPS = 58958

That means booting with "intel_idle.ibrs_off" improves performance by:
max-jOPS: 1.2%, which could be considered noise range.
critical-jOPS: 12%, which is definitely a solid improvement.

The admin-guide/pm/intel_idle.rst file is updated to add a description about the new "ibrs_off" module parameter.

---
 Documentation/admin-guide/pm/intel_idle.rst | 17 ++++++++++++++++-
 drivers/idle/intel_idle.c                   | 14 ++++++++++++--
 2 files changed, 28 insertions(+), 3 deletions(-)
Documentation/admin-guide/pm/intel_idle.rst +16 -1
index b799a43da62e..8604e6d1fe2c 100644
@@ -170,7 +170,7 @@ and ``idle=nomwait``.  If any of them is present in the kernel command line, the
 ``MWAIT`` instruction is not allowed to be used, so the initialization of
 ``intel_idle`` will fail.
 
-Apart from that there are four module parameters recognized by ``intel_idle``
+Apart from that there are five module parameters recognized by ``intel_idle``
 itself that can be set via the kernel command line (they cannot be updated via
 sysfs, so that is the only way to change their values).
 
@@ -216,6 +216,21 @@ are ignored).
 The idle states disabled this way can be enabled (on a per-CPU basis) from user
 space via ``sysfs``.
 
+The ``ibrs_off`` module parameter is a boolean flag (default to false). It is
+used to control if IBRS (Indirect Branch Restricted Speculation) should be
+turned off, if set, when the CPU enters an idle state.  This flag will not
+affect CPUs that are using Enhanced IBRS which can remain on with little
+performance impact.
+
+For some CPUs, IBRS will be selected as mitigation for Spectre v2 and Retbleed
+security vulnerabilities by default.  Leaving the IBRS mode on while idling may
+have a performance impact on its sibling CPU.  The IBRS mode will be turned off
+by default when the CPU enters into a deep idle state, but not in some
+shallower ones.  Setting the ``ibrs_off`` module parameter will force the IBRS
+mode to off when the CPU is in any one of the available idle states.  This may
+help performance of a sibling CPU at the expense of a slightly higher wakeup
+latency for the idle CPU.
+
 
 .. _intel-idle-core-and-package-idle-states:
 
drivers/idle/intel_idle.c +12 -2
index 07fa23707b3c..8bbf1e6d845c 100644
@@ -69,6 +69,7 @@ static int max_cstate = CPUIDLE_STATE_MAX - 1;
 static unsigned int disabled_states_mask __read_mostly;
 static unsigned int preferred_states_mask __read_mostly;
 static bool force_irq_on __read_mostly;
+static bool ibrs_off __read_mostly;
 
 static struct cpuidle_device __percpu *intel_idle_cpuidle_devices;
 
@@ -1907,12 +1908,15 @@ static void __init intel_idle_init_cstates_icpu(struct cpuidle_driver *drv)
 			WARN_ON_ONCE(state->flags & CPUIDLE_FLAG_IRQ_ENABLE);
 			state->enter = intel_idle_xstate;
 		} else if (cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS) &&
-			   state->flags & CPUIDLE_FLAG_IBRS) {
+			  ((state->flags & CPUIDLE_FLAG_IBRS) || ibrs_off)) {
 			/*
 			 * IBRS mitigation requires that C-states are entered
 			 * with interrupts disabled.
 			 */
-			WARN_ON_ONCE(state->flags & CPUIDLE_FLAG_IRQ_ENABLE);
+			if (ibrs_off && (state->flags & CPUIDLE_FLAG_IRQ_ENABLE))
+				state->flags &= ~CPUIDLE_FLAG_IRQ_ENABLE;
+			else
+				WARN_ON_ONCE(state->flags & CPUIDLE_FLAG_IRQ_ENABLE);
 			state->enter = intel_idle_ibrs;
 		} else if (state->flags & CPUIDLE_FLAG_IRQ_ENABLE) {
 			state->enter = intel_idle_irq;
@@ -2165,3 +2169,9 @@ MODULE_PARM_DESC(preferred_cstates, "Mask of preferred idle states");
  * 'CPUIDLE_FLAG_INIT_XSTATE' and 'CPUIDLE_FLAG_IBRS' flags.
  */
 module_param(force_irq_on, bool, 0444);
+/*
+ * Force the disabling of IBRS when X86_FEATURE_KERNEL_IBRS is on and
+ * CPUIDLE_FLAG_IRQ_ENABLE isn't set.
+ */
+module_param(ibrs_off, bool, 0444);
+MODULE_PARM_DESC(ibrs_off, "Disable IBRS when idle");
Signature · 1 line
2.31.1