[PATCH 0/2] x86/intel_rdt: Fix cache pseudo-locking error path cleanup

[PATCH 0/2] x86/intel_rdt: Fix cache pseudo-locking error path cleanup

[Reinette Chatre]

Dear Maintainers,

A bug exists in the error handling code during pseudo-lock region creation.
When an error occurs early during pseudo-lock region creation the
pseudo_lock_region struct is not cleaned up properly but remains associated
with the resource group (since it remains in pseudo-locksetup mode).
This partially initialized struct causes problems when other areas
need to obtain resource group data - when partially initialized the
resource group is treated as a pseudo-locked region.

Following is an example of the error being encountered. First a
pseudo-locked region of larger than 4MB is attempted. This fails early
because of lack for support. Since this is not cleaned up properly,
a subsequent attempt fails because it is (incorrectly) believed that
a pseudo-locked region already exists, also the bit_usage file
reports incorrect data.

# mount -t resctrl resctrl /sys/fs/resctrl
# cd /sys/fs/resctrl/
# mkdir p1
# echo 'L3:1=0xffff0' > schemata
# echo pseudo-locksetup > p1/mode
# echo 'L3:1=0xf' > p1/schemata
-bash: echo: write error: Argument list too long
# cat info/last_cmd_status
requested region exceeds maximum size
# echo 'L3:1=0x1' > p1/schemata
-bash: echo: write error: Invalid argument
# cat info/last_cmd_status
pseudo-locked region in hierarchy
# cat info/L3/bit_usage
0=XXSSSSSSSSSSSSSSSSSS;1=XXSSSSSSSSSSSSSSPPPP

After the fixes in this series have been applied:
# mount -t resctrl resctrl /sys/fs/resctrl/
# cd /sys/fs/resctrl/
# mkdir p1
# echo pseudo-locksetup > p1/mode
# echo 'L3:1=0xffff0' > schemata
# echo 'L3:1=0xf' > p1/schemata
-bash: echo: write error: Argument list too long
# cat info/last_cmd_status
requested region exceeds maximum size
# cat info/L3/bit_usage
0=XXSSSSSSSSSSSSSSSSSS;1=XXSSSSSSSSSSSSSS0000
# echo 'L3:1=0x1' > p1/schemata
# cat info/L3/bit_usage
0=XXSSSSSSSSSSSSSSSSSS;1=XXSSSSSSSSSSSSSS000P


Reinette Chatre (2):
  x86/intel_rdt: Move pseudo_lock_region_clear
  x86/intel_rdt: Fix cleanup of plr structure on error

 arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c | 68 ++++++++++++---------
 1 file changed, 40 insertions(+), 28 deletions(-)

-- 
2.17.0

[PATCH 1/2] x86/intel_rdt: Move pseudo_lock_region_clear

[Reinette Chatre]

The pseudo_lock_region_clear() function is moved to earlier in the
file in preparation for its use in functions that currently appear
before it. No functional change.

Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
---
 arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c | 46 ++++++++++-----------
 1 file changed, 23 insertions(+), 23 deletions(-)

diff --git a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
index 6e83f61552a5..1860ec10302d 100644
--- a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
+++ b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
@@ -246,6 +246,29 @@ static int pseudo_lock_cstates_constrain(struct pseudo_lock_region *plr)
 	return ret;
 }
 
+/**
+ * pseudo_lock_region_clear - Reset pseudo-lock region data
+ * @plr: pseudo-lock region
+ *
+ * All content of the pseudo-locked region is reset - any memory allocated
+ * freed.
+ *
+ * Return: void
+ */
+static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
+{
+	plr->size = 0;
+	plr->line_size = 0;
+	kfree(plr->kmem);
+	plr->kmem = NULL;
+	plr->r = NULL;
+	if (plr->d)
+		plr->d->plr = NULL;
+	plr->d = NULL;
+	plr->cbm = 0;
+	plr->debugfs_dir = NULL;
+}
+
 /**
  * pseudo_lock_region_init - Initialize pseudo-lock region information
  * @plr: pseudo-lock region
@@ -318,29 +341,6 @@ static int pseudo_lock_init(struct rdtgroup *rdtgrp)
 	return 0;
 }
 
-/**
- * pseudo_lock_region_clear - Reset pseudo-lock region data
- * @plr: pseudo-lock region
- *
- * All content of the pseudo-locked region is reset - any memory allocated
- * freed.
- *
- * Return: void
- */
-static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
-{
-	plr->size = 0;
-	plr->line_size = 0;
-	kfree(plr->kmem);
-	plr->kmem = NULL;
-	plr->r = NULL;
-	if (plr->d)
-		plr->d->plr = NULL;
-	plr->d = NULL;
-	plr->cbm = 0;
-	plr->debugfs_dir = NULL;
-}
-
 /**
  * pseudo_lock_region_alloc - Allocate kernel memory that will be pseudo-locked
  * @plr: pseudo-lock region
-- 
2.17.0

[tip:x86/cache] x86/intel_rdt: Move pseudo_lock_region_clear()

[tip-bot for Reinette Chatre]

Commit-ID:  ce730f1cc1255be152c879a2bc5f295d341d8036
Gitweb:     https://git.kernel.org/tip/ce730f1cc1255be152c879a2bc5f295d341d8036
Author:     Reinette Chatre <reinette.chatre@intel.com>
AuthorDate: Sat, 30 Jun 2018 22:03:02 -0700
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Tue, 3 Jul 2018 08:38:39 +0200

x86/intel_rdt: Move pseudo_lock_region_clear()

The pseudo_lock_region_clear() function is moved to earlier in the file in
preparation for its use in functions that currently appear before it. No
functional change.

Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: fenghua.yu@intel.com
Cc: tony.luck@intel.com
Cc: vikas.shivappa@linux.intel.com
Cc: gavin.hindman@intel.com
Cc: jithu.joseph@intel.com
Cc: dave.hansen@intel.com
Cc: hpa@zytor.com
Link: https://lkml.kernel.org/r/ef098ec2a45501e23792289bff80ae3152141e2f.1530421026.git.reinette.chatre@intel.com

---
 arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c | 46 ++++++++++++++---------------
 1 file changed, 23 insertions(+), 23 deletions(-)

diff --git a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
index 6e83f61552a5..1860ec10302d 100644
--- a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
+++ b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
@@ -246,6 +246,29 @@ out_err:
 	return ret;
 }
 
+/**
+ * pseudo_lock_region_clear - Reset pseudo-lock region data
+ * @plr: pseudo-lock region
+ *
+ * All content of the pseudo-locked region is reset - any memory allocated
+ * freed.
+ *
+ * Return: void
+ */
+static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
+{
+	plr->size = 0;
+	plr->line_size = 0;
+	kfree(plr->kmem);
+	plr->kmem = NULL;
+	plr->r = NULL;
+	if (plr->d)
+		plr->d->plr = NULL;
+	plr->d = NULL;
+	plr->cbm = 0;
+	plr->debugfs_dir = NULL;
+}
+
 /**
  * pseudo_lock_region_init - Initialize pseudo-lock region information
  * @plr: pseudo-lock region
@@ -318,29 +341,6 @@ static int pseudo_lock_init(struct rdtgroup *rdtgrp)
 	return 0;
 }
 
-/**
- * pseudo_lock_region_clear - Reset pseudo-lock region data
- * @plr: pseudo-lock region
- *
- * All content of the pseudo-locked region is reset - any memory allocated
- * freed.
- *
- * Return: void
- */
-static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
-{
-	plr->size = 0;
-	plr->line_size = 0;
-	kfree(plr->kmem);
-	plr->kmem = NULL;
-	plr->r = NULL;
-	if (plr->d)
-		plr->d->plr = NULL;
-	plr->d = NULL;
-	plr->cbm = 0;
-	plr->debugfs_dir = NULL;
-}
-
 /**
  * pseudo_lock_region_alloc - Allocate kernel memory that will be pseudo-locked
  * @plr: pseudo-lock region

[PATCH 2/2] x86/intel_rdt: Fix cleanup of plr structure on error

[Reinette Chatre]

When a resource group enters pseudo-locksetup mode a pseudo_lock_region
is associated with it. When the user writes to the resource group's
schemata file the CBM of the requested pseudo-locked region is entered
into the pseudo_lock_region struct. If any part of pseudo-lock region
creation fails the resource group will remain in pseudo-locksetup mode
with the pseudo_lock_region associated with it.

In case of failure during pseudo-lock region creation care needs to be
taken to ensure that the pseudo_lock_region struct associated with the
resource group is cleared from any pseudo-locking data - especially the
CBM. This is because the existence of a pseudo_lock_region struct with a
CBM is significant in other areas of the code, for example, the display
of bit_usage and initialization of a new resource group.

Fix the error path of pseudo-lock region creation to ensure that the
pseudo_lock_region struct is cleared at each error exit.

Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
---
 arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c | 22 ++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
index 1860ec10302d..8fd79c281ee6 100644
--- a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
+++ b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
@@ -290,6 +290,7 @@ static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
 static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 {
 	struct cpu_cacheinfo *ci;
+	int ret;
 	int i;
 
 	/* Pick the first cpu we find that is associated with the cache. */
@@ -298,7 +299,8 @@ static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 	if (!cpu_online(plr->cpu)) {
 		rdt_last_cmd_printf("cpu %u associated with cache not online\n",
 				    plr->cpu);
-		return -ENODEV;
+		ret = -ENODEV;
+		goto out_region;
 	}
 
 	ci = get_cpu_cacheinfo(plr->cpu);
@@ -312,8 +314,11 @@ static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 		}
 	}
 
+	ret = -1;
 	rdt_last_cmd_puts("unable to determine cache line size\n");
-	return -1;
+out_region:
+	pseudo_lock_region_clear(plr);
+	return ret;
 }
 
 /**
@@ -365,16 +370,23 @@ static int pseudo_lock_region_alloc(struct pseudo_lock_region *plr)
 	 */
 	if (plr->size > KMALLOC_MAX_SIZE) {
 		rdt_last_cmd_puts("requested region exceeds maximum size\n");
-		return -E2BIG;
+		ret = -E2BIG;
+		goto out_region;
 	}
 
 	plr->kmem = kzalloc(plr->size, GFP_KERNEL);
 	if (!plr->kmem) {
 		rdt_last_cmd_puts("unable to allocate memory\n");
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out_region;
 	}
 
-	return 0;
+	ret = 0;
+	goto out;
+out_region:
+	pseudo_lock_region_clear(plr);
+out:
+	return ret;
 }
 
 /**
-- 
2.17.0

[tip:x86/cache] x86/intel_rdt: Fix cleanup of plr structure on error

[tip-bot for Reinette Chatre]

Commit-ID:  546d3c74277398a3d76d059bd2db47186bb47fc8
Gitweb:     https://git.kernel.org/tip/546d3c74277398a3d76d059bd2db47186bb47fc8
Author:     Reinette Chatre <reinette.chatre@intel.com>
AuthorDate: Sat, 30 Jun 2018 22:03:03 -0700
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Tue, 3 Jul 2018 08:38:39 +0200

x86/intel_rdt: Fix cleanup of plr structure on error

When a resource group enters pseudo-locksetup mode a pseudo_lock_region is
associated with it. When the user writes to the resource group's schemata
file the CBM of the requested pseudo-locked region is entered into the
pseudo_lock_region struct. If any part of pseudo-lock region creation fails
the resource group will remain in pseudo-locksetup mode with the
pseudo_lock_region associated with it.

In case of failure during pseudo-lock region creation care needs to be
taken to ensure that the pseudo_lock_region struct associated with the
resource group is cleared from any pseudo-locking data - especially the
CBM. This is because the existence of a pseudo_lock_region struct with a
CBM is significant in other areas of the code, for example, the display of
bit_usage and initialization of a new resource group.

Fix the error path of pseudo-lock region creation to ensure that the
pseudo_lock_region struct is cleared at each error exit.

Fixes: 018961ae5579 ("x86/intel_rdt: Pseudo-lock region creation/removal core")
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: fenghua.yu@intel.com
Cc: tony.luck@intel.com
Cc: vikas.shivappa@linux.intel.com
Cc: gavin.hindman@intel.com
Cc: jithu.joseph@intel.com
Cc: dave.hansen@intel.com
Cc: hpa@zytor.com
Link: https://lkml.kernel.org/r/49b4782f6d204d122cee3499e642b2772a98d2b4.1530421026.git.reinette.chatre@intel.com

---
 arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
index 1860ec10302d..8fd79c281ee6 100644
--- a/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
+++ b/arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c
@@ -290,6 +290,7 @@ static void pseudo_lock_region_clear(struct pseudo_lock_region *plr)
 static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 {
 	struct cpu_cacheinfo *ci;
+	int ret;
 	int i;
 
 	/* Pick the first cpu we find that is associated with the cache. */
@@ -298,7 +299,8 @@ static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 	if (!cpu_online(plr->cpu)) {
 		rdt_last_cmd_printf("cpu %u associated with cache not online\n",
 				    plr->cpu);
-		return -ENODEV;
+		ret = -ENODEV;
+		goto out_region;
 	}
 
 	ci = get_cpu_cacheinfo(plr->cpu);
@@ -312,8 +314,11 @@ static int pseudo_lock_region_init(struct pseudo_lock_region *plr)
 		}
 	}
 
+	ret = -1;
 	rdt_last_cmd_puts("unable to determine cache line size\n");
-	return -1;
+out_region:
+	pseudo_lock_region_clear(plr);
+	return ret;
 }
 
 /**
@@ -365,16 +370,23 @@ static int pseudo_lock_region_alloc(struct pseudo_lock_region *plr)
 	 */
 	if (plr->size > KMALLOC_MAX_SIZE) {
 		rdt_last_cmd_puts("requested region exceeds maximum size\n");
-		return -E2BIG;
+		ret = -E2BIG;
+		goto out_region;
 	}
 
 	plr->kmem = kzalloc(plr->size, GFP_KERNEL);
 	if (!plr->kmem) {
 		rdt_last_cmd_puts("unable to allocate memory\n");
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out_region;
 	}
 
-	return 0;
+	ret = 0;
+	goto out;
+out_region:
+	pseudo_lock_region_clear(plr);
+out:
+	return ret;
 }
 
 /**