Skip to contentSkip to conversation

x86/resctrl: Check monitoring static key in MBM overflow handler

PATCH Superseded Xiaochen Shen 2 participants 4 messages Dec 11, 2019 – Jan 17, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/resctrl: Check monitoring static key in MBM overflow handler

Currently, there are three static keys in resctrl file system:
rdt_mon_enable_key and rdt_alloc_enable_key indicate if monitoring feature and allocation feature is enabled respectively. rdt_enable_key is enabled when either monitoring feature or allocation feature is enabled.

If no monitoring feature is supported (either hardware doesn't support monitoring feature or the feature is disabled by kernel command line option "rdt="), rdt_enable_key is still enabled but rdt_mon_enable_key is disabled.

MBM is a monitoring feature. MBM overflow handler intends to check if monitoring feature is not enabled for fast return. So the accurate check here is to check rdt_mon_enable_key instead of rdt_enable_key.

Fixes e33026831bdb ("x86/intel_rdt/mbm: Handle counter overflow")
---
 arch/x86/kernel/cpu/resctrl/internal.h | 1 +
 arch/x86/kernel/cpu/resctrl/monitor.c  | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/resctrl/internal.h +1 -0
index e49b77283924..181c992f448c 100644
@@ -57,6 +57,7 @@ static inline struct rdt_fs_context *rdt_fc2context(struct fs_context *fc)
 }
 
 DECLARE_STATIC_KEY_FALSE(rdt_enable_key);
+DECLARE_STATIC_KEY_FALSE(rdt_mon_enable_key);
 
 /**
  * struct mon_evt - Entry in the event list of a resource
arch/x86/kernel/cpu/resctrl/monitor.c +2 -2
index 397206f23d14..773124b0e18a 100644
@@ -514,7 +514,7 @@ void mbm_handle_overflow(struct work_struct *work)
 
 	mutex_lock(&rdtgroup_mutex);
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		goto out_unlock;
 
 	d = get_domain_from_cpu(cpu, &rdt_resources_all[RDT_RESOURCE_L3]);
@@ -543,7 +543,7 @@ void mbm_setup_overflow_handler(struct rdt_domain *dom, unsigned long delay_ms)
 	unsigned long delay = msecs_to_jiffies(delay_ms);
 	int cpu;
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		return;
 	cpu = cpumask_any(&dom->cpu_mask);
 	dom->mbm_work_cpu = cpu;
Signature · 1 line
1.8.3.1
#1 lore raw

Hi,

I am sorry I sent out a wrong version of patch.
Please ignore this patch. I will resend soon.

Quoted 59 lines · Xiaochen Shen

Currently, there are three static keys in resctrl file system:
rdt_mon_enable_key and rdt_alloc_enable_key indicate if monitoring feature and allocation feature is enabled respectively. rdt_enable_key is enabled when either monitoring feature or allocation feature is enabled.

If no monitoring feature is supported (either hardware doesn't support monitoring feature or the feature is disabled by kernel command line option "rdt="), rdt_enable_key is still enabled but rdt_mon_enable_key is disabled.

MBM is a monitoring feature. MBM overflow handler intends to check if monitoring feature is not enabled for fast return. So the accurate check here is to check rdt_mon_enable_key instead of rdt_enable_key.

Fixes e33026831bdb ("x86/intel_rdt/mbm: Handle counter overflow")
---
  arch/x86/kernel/cpu/resctrl/internal.h | 1 +
  arch/x86/kernel/cpu/resctrl/monitor.c  | 4 ++--
  2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/resctrl/internal.h +1 -0
index e49b77283924..181c992f448c 100644
@@ -57,6 +57,7 @@ static inline struct rdt_fs_context *rdt_fc2context(struct fs_context *fc)
  }
 
  DECLARE_STATIC_KEY_FALSE(rdt_enable_key);
+DECLARE_STATIC_KEY_FALSE(rdt_mon_enable_key);
 
  /**
   * struct mon_evt - Entry in the event list of a resource
arch/x86/kernel/cpu/resctrl/monitor.c +2 -2
index 397206f23d14..773124b0e18a 100644
@@ -514,7 +514,7 @@ void mbm_handle_overflow(struct work_struct *work)
 
  	mutex_lock(&rdtgroup_mutex);
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
  		goto out_unlock;
 
  	d = get_domain_from_cpu(cpu, &rdt_resources_all[RDT_RESOURCE_L3]);
@@ -543,7 +543,7 @@ void mbm_setup_overflow_handler(struct rdt_domain *dom, unsigned long delay_ms)
  	unsigned long delay = msecs_to_jiffies(delay_ms);
  	int cpu;
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
  		return;
  	cpu = cpumask_any(&dom->cpu_mask);
  	dom->mbm_work_cpu = cpu;
Signature · 2 lines
Best regards,
Xiaochen
#2 lore raw

PATCH RESEND 1/1 x86/resctrl: Check monitoring static key in MBM overflow handler

In reply to Xiaochen Shen

Currently, there are three static keys in resctrl file system:
rdt_mon_enable_key and rdt_alloc_enable_key indicate if monitoring feature and allocation feature is enabled respectively. rdt_enable_key is enabled when either monitoring feature or allocation feature is enabled.

If no monitoring feature is supported (either hardware doesn't support monitoring feature or the feature is disabled by kernel command line option "rdt="), rdt_enable_key is still enabled but rdt_mon_enable_key is disabled.

MBM is a monitoring feature. MBM overflow handler intends to check if monitoring feature is not enabled for fast return. So the accurate check here is to check rdt_mon_enable_key instead of rdt_enable_key.

Fixes e33026831bdb ("x86/intel_rdt/mbm: Handle counter overflow")
---
 arch/x86/kernel/cpu/resctrl/internal.h | 1 +
 arch/x86/kernel/cpu/resctrl/monitor.c  | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/resctrl/internal.h +1 -0
index e49b77283924..181c992f448c 100644
@@ -57,6 +57,7 @@ static inline struct rdt_fs_context *rdt_fc2context(struct fs_context *fc)
 }
 
 DECLARE_STATIC_KEY_FALSE(rdt_enable_key);
+DECLARE_STATIC_KEY_FALSE(rdt_mon_enable_key);
 
 /**
  * struct mon_evt - Entry in the event list of a resource
arch/x86/kernel/cpu/resctrl/monitor.c +2 -2
index 397206f23d14..773124b0e18a 100644
@@ -514,7 +514,7 @@ void mbm_handle_overflow(struct work_struct *work)
 
 	mutex_lock(&rdtgroup_mutex);
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		goto out_unlock;
 
 	d = get_domain_from_cpu(cpu, &rdt_resources_all[RDT_RESOURCE_L3]);
@@ -543,7 +543,7 @@ void mbm_setup_overflow_handler(struct rdt_domain *dom, unsigned long delay_ms)
 	unsigned long delay = msecs_to_jiffies(delay_ms);
 	int cpu;
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		return;
 	cpu = cpumask_any(&dom->cpu_mask);
 	dom->mbm_work_cpu = cpu;
Signature · 1 line
1.8.3.1
merge bot Applied
#3 lore raw

[tip: x86/cache] x86/resctrl: Check monitoring static key in the MBM overflow handler

The following commit has been merged into the x86/cache branch of tip:

The following commit has been merged into the x86/cache branch of tip:

Commit-ID:     536a0d8e79fb928f2735db37dda95682b6754f9a
Gitweb:        https://git.kernel.org/tip/536a0d8e79fb928f2735db37dda95682b6754f9a
Author:        Xiaochen Shen <[email protected]>
AuthorDate:    Thu, 12 Dec 2019 04:05:05 +08:00
Committer:     Borislav Petkov <[email protected]>
CommitterDate: Fri, 17 Jan 2020 19:32:32 +01:00

x86/resctrl: Check monitoring static key in the MBM overflow handler

Currently, there are three static keys in the resctrl file system:
rdt_mon_enable_key and rdt_alloc_enable_key indicate if the monitoring feature and the allocation feature are enabled, respectively. The rdt_enable_key is enabled when either the monitoring feature or the allocation feature is enabled.

If no monitoring feature is present (either hardware doesn't support a monitoring feature or the feature is disabled by the kernel command line option "rdt="), rdt_enable_key is still enabled but rdt_mon_enable_key is disabled.

MBM is a monitoring feature. The MBM overflow handler intends to check if the monitoring feature is not enabled for fast return.

So check the rdt_mon_enable_key in it instead of the rdt_enable_key as former is the more accurate check.

[ bp: Massage commit message. ]

Fixes e33026831bdb ("x86/intel_rdt/mbm: Handle counter overflow")
---
 arch/x86/kernel/cpu/resctrl/internal.h | 1 +
 arch/x86/kernel/cpu/resctrl/monitor.c  | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/resctrl/internal.h +1 -0
index e49b772..181c992 100644
@@ -57,6 +57,7 @@ static inline struct rdt_fs_context *rdt_fc2context(struct fs_context *fc)
 }
 
 DECLARE_STATIC_KEY_FALSE(rdt_enable_key);
+DECLARE_STATIC_KEY_FALSE(rdt_mon_enable_key);
 
 /**
  * struct mon_evt - Entry in the event list of a resource
arch/x86/kernel/cpu/resctrl/monitor.c +2 -2
index 397206f..773124b 100644
@@ -514,7 +514,7 @@ void mbm_handle_overflow(struct work_struct *work)
 
 	mutex_lock(&rdtgroup_mutex);
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		goto out_unlock;
 
 	d = get_domain_from_cpu(cpu, &rdt_resources_all[RDT_RESOURCE_L3]);
@@ -543,7 +543,7 @@ void mbm_setup_overflow_handler(struct rdt_domain *dom, unsigned long delay_ms)
 	unsigned long delay = msecs_to_jiffies(delay_ms);
 	int cpu;
 
-	if (!static_branch_likely(&rdt_enable_key))
+	if (!static_branch_likely(&rdt_mon_enable_key))
 		return;
 	cpu = cpumask_any(&dom->cpu_mask);
 	dom->mbm_work_cpu = cpu;