Since commit 741c10b096bc ("kernfs: Use RCU to access kernfs_node::name."),
a helper rdt_kn_name() that checks that rdtgroup_mutex is held has been
used for all accesses to the kernfs node name.

rdtgroup_mkdir() uses the name to determine if a valid monitor group is
being created by checking the parent name is "mon_groups". This is done
without holding rdtgroup_mutex, and now triggers the following warning:
| WARNING: suspicious RCU usage
| 6.15.0-rc1 #4465 Tainted: G            E
| -----------------------------
| arch/x86/kernel/cpu/resctrl/internal.h:408 suspicious rcu_dereference_check() usage!
[...]
| Call Trace:
|  <TASK>
|  dump_stack_lvl+0x6c/0xa0
|  lockdep_rcu_suspicious.cold+0x4e/0x96
|  is_mon_groups+0xba/0xd0
|  rdtgroup_mkdir+0x118/0x1970
|  kernfs_iop_mkdir+0xfa/0x1a0
|  vfs_mkdir+0x456/0x760
|  do_mkdirat+0x257/0x310
|  __x64_sys_mkdir+0xd4/0x120
|  do_syscall_64+0x6d/0x150
|  entry_SYSCALL_64_after_hwframe+0x76/0x7e

Creating a control or monitor group calls mkdir_rdt_prepare(), which uses
rdtgroup_kn_lock_live() to take the rdtgroup_mutex.

To avoid taking and dropping the lock, move the check for the monitor group
name and position into mkdir_rdt_prepare() so that it occurs under
rdtgroup_mutex. Hoist is_mon_groups() earlier in the file.

CC: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Fixes: 741c10b096bc ("kernfs: Use RCU to access kernfs_node::name.")
Signed-off-by: James Morse <james.morse@arm.com>
---
 arch/x86/kernel/cpu/resctrl/rdtgroup.c | 48 +++++++++++++++-----------
 1 file changed, 27 insertions(+), 21 deletions(-)

diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
index 93ec829015f1..a81a42cf6656 100644
--- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
+++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
@@ -3553,6 +3553,22 @@ static void mkdir_rdt_prepare_rmid_free(struct rdtgroup *rgrp)
 		free_rmid(rgrp->closid, rgrp->mon.rmid);
 }
 
+/*
+ * We allow creating mon groups only with in a directory called "mon_groups"
+ * which is present in every ctrl_mon group. Check if this is a valid
+ * "mon_groups" directory.
+ *
+ * 1. The directory should be named "mon_groups".
+ * 2. The mon group itself should "not" be named "mon_groups".
+ *   This makes sure "mon_groups" directory always has a ctrl_mon group
+ *   as parent.
+ */
+static bool is_mon_groups(struct kernfs_node *kn, const char *name)
+{
+	return (!strcmp(rdt_kn_name(kn), "mon_groups") &&
+		strcmp(name, "mon_groups"));
+}
+
 static int mkdir_rdt_prepare(struct kernfs_node *parent_kn,
 			     const char *name, umode_t mode,
 			     enum rdt_group_type rtype, struct rdtgroup **r)
@@ -3568,6 +3584,15 @@ static int mkdir_rdt_prepare(struct kernfs_node *parent_kn,
 		goto out_unlock;
 	}
 
+	/*
+	 * Check the parent directory for a monitor group is a "mon_groups"
+	 * directory.
+	 */
+	if (rtype == RDTMON_GROUP && !is_mon_groups(parent_kn, name)) {
+		ret = -EPERM;
+		goto out_unlock;
+	}
+
 	if (rtype == RDTMON_GROUP &&
 	    (prdtgrp->mode == RDT_MODE_PSEUDO_LOCKSETUP ||
 	     prdtgrp->mode == RDT_MODE_PSEUDO_LOCKED)) {
@@ -3751,22 +3776,6 @@ static int rdtgroup_mkdir_ctrl_mon(struct kernfs_node *parent_kn,
 	return ret;
 }
 
-/*
- * We allow creating mon groups only with in a directory called "mon_groups"
- * which is present in every ctrl_mon group. Check if this is a valid
- * "mon_groups" directory.
- *
- * 1. The directory should be named "mon_groups".
- * 2. The mon group itself should "not" be named "mon_groups".
- *   This makes sure "mon_groups" directory always has a ctrl_mon group
- *   as parent.
- */
-static bool is_mon_groups(struct kernfs_node *kn, const char *name)
-{
-	return (!strcmp(rdt_kn_name(kn), "mon_groups") &&
-		strcmp(name, "mon_groups"));
-}
-
 static int rdtgroup_mkdir(struct kernfs_node *parent_kn, const char *name,
 			  umode_t mode)
 {
@@ -3782,11 +3791,8 @@ static int rdtgroup_mkdir(struct kernfs_node *parent_kn, const char *name,
 	if (resctrl_arch_alloc_capable() && parent_kn == rdtgroup_default.kn)
 		return rdtgroup_mkdir_ctrl_mon(parent_kn, name, mode);
 
-	/*
-	 * If RDT monitoring is supported and the parent directory is a valid
-	 * "mon_groups" directory, add a monitoring subdirectory.
-	 */
-	if (resctrl_arch_mon_capable() && is_mon_groups(parent_kn, name))
+	/* Else, attempt to add a monitoring subdirectory. */
+	if (resctrl_arch_mon_capable())
 		return rdtgroup_mkdir_mon(parent_kn, name, mode);
 
 	return -EPERM;
-- 
2.39.5

* James Morse <james.morse@arm.com> wrote:

> +	/*
> +	 * Check the parent directory for a monitor group is a "mon_groups"
> +	 * directory.
> +	 */

This sentence probably wants to be a bit less linguistically ambiguous, 
via something like:

  s/Check the parent directory for a monitor group is a "mon_groups" directory.
   /Check that the parent directory for a monitor group is a "mon_groups" directory.

With that:

  Acked-by: Ingo Molnar <mingo@kernel.org>

Thanks,

	Ingo

Hi James,

On 4/7/25 5:46 AM, James Morse wrote:
> Since commit 741c10b096bc ("kernfs: Use RCU to access kernfs_node::name."),
> a helper rdt_kn_name() that checks that rdtgroup_mutex is held has been
> used for all accesses to the kernfs node name.
> 
> rdtgroup_mkdir() uses the name to determine if a valid monitor group is
> being created by checking the parent name is "mon_groups". This is done
> without holding rdtgroup_mutex, and now triggers the following warning:
> | WARNING: suspicious RCU usage
> | 6.15.0-rc1 #4465 Tainted: G            E
> | -----------------------------
> | arch/x86/kernel/cpu/resctrl/internal.h:408 suspicious rcu_dereference_check() usage!
> [...]
> | Call Trace:
> |  <TASK>
> |  dump_stack_lvl+0x6c/0xa0
> |  lockdep_rcu_suspicious.cold+0x4e/0x96
> |  is_mon_groups+0xba/0xd0
> |  rdtgroup_mkdir+0x118/0x1970
> |  kernfs_iop_mkdir+0xfa/0x1a0
> |  vfs_mkdir+0x456/0x760
> |  do_mkdirat+0x257/0x310
> |  __x64_sys_mkdir+0xd4/0x120
> |  do_syscall_64+0x6d/0x150
> |  entry_SYSCALL_64_after_hwframe+0x76/0x7e
> 
> Creating a control or monitor group calls mkdir_rdt_prepare(), which uses
> rdtgroup_kn_lock_live() to take the rdtgroup_mutex.
> 
> To avoid taking and dropping the lock, move the check for the monitor group
> name and position into mkdir_rdt_prepare() so that it occurs under
> rdtgroup_mutex. Hoist is_mon_groups() earlier in the file.
> 
> CC: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> Fixes: 741c10b096bc ("kernfs: Use RCU to access kernfs_node::name.")
> Signed-off-by: James Morse <james.morse@arm.com>
> ---

Thank you very much.

Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>

Reinette

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

Commit-ID:     45c2e30bbd64d75559b99f3a5c455129a7a34b06
Gitweb:        https://git.kernel.org/tip/45c2e30bbd64d75559b99f3a5c455129a7a34b06
Author:        James Morse <james.morse@arm.com>
AuthorDate:    Mon, 07 Apr 2025 13:46:37 +01:00
Committer:     Borislav Petkov (AMD) <bp@alien8.de>
CommitterDate: Wed, 09 Apr 2025 11:35:08 +02:00

x86/resctrl: Fix rdtgroup_mkdir()'s unlocked use of kernfs_node::name

Since

  741c10b096bc ("kernfs: Use RCU to access kernfs_node::name.")

a helper rdt_kn_name() that checks that rdtgroup_mutex is held has been used
for all accesses to the kernfs node name.

rdtgroup_mkdir() uses the name to determine if a valid monitor group is being
created by checking the parent name is "mon_groups". This is done without
holding rdtgroup_mutex, and now triggers the following warning:

  | WARNING: suspicious RCU usage
  | 6.15.0-rc1 #4465 Tainted: G            E
  | -----------------------------
  | arch/x86/kernel/cpu/resctrl/internal.h:408 suspicious rcu_dereference_check() usage!
  [...]
  | Call Trace:
  |  <TASK>
  |  dump_stack_lvl
  |  lockdep_rcu_suspicious.cold
  |  is_mon_groups
  |  rdtgroup_mkdir
  |  kernfs_iop_mkdir
  |  vfs_mkdir
  |  do_mkdirat
  |  __x64_sys_mkdir
  |  do_syscall_64
  |  entry_SYSCALL_64_after_hwframe

Creating a control or monitor group calls mkdir_rdt_prepare(), which uses
rdtgroup_kn_lock_live() to take the rdtgroup_mutex.

To avoid taking and dropping the lock, move the check for the monitor group
name and position into mkdir_rdt_prepare() so that it occurs under
rdtgroup_mutex. Hoist is_mon_groups() earlier in the file.

  [ bp: Massage. ]

Fixes: 741c10b096bc ("kernfs: Use RCU to access kernfs_node::name.")
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Reinette Chatre <reinette.chatre@intel.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20250407124637.2433230-1-james.morse@arm.com
---
 arch/x86/kernel/cpu/resctrl/rdtgroup.c | 48 ++++++++++++++-----------
 1 file changed, 27 insertions(+), 21 deletions(-)

diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
index 93ec829..cc4a541 100644
--- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
+++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
@@ -3553,6 +3553,22 @@ static void mkdir_rdt_prepare_rmid_free(struct rdtgroup *rgrp)
 		free_rmid(rgrp->closid, rgrp->mon.rmid);
 }
 
+/*
+ * We allow creating mon groups only with in a directory called "mon_groups"
+ * which is present in every ctrl_mon group. Check if this is a valid
+ * "mon_groups" directory.
+ *
+ * 1. The directory should be named "mon_groups".
+ * 2. The mon group itself should "not" be named "mon_groups".
+ *   This makes sure "mon_groups" directory always has a ctrl_mon group
+ *   as parent.
+ */
+static bool is_mon_groups(struct kernfs_node *kn, const char *name)
+{
+	return (!strcmp(rdt_kn_name(kn), "mon_groups") &&
+		strcmp(name, "mon_groups"));
+}
+
 static int mkdir_rdt_prepare(struct kernfs_node *parent_kn,
 			     const char *name, umode_t mode,
 			     enum rdt_group_type rtype, struct rdtgroup **r)
@@ -3568,6 +3584,15 @@ static int mkdir_rdt_prepare(struct kernfs_node *parent_kn,
 		goto out_unlock;
 	}
 
+	/*
+	 * Check that the parent directory for a monitor group is a "mon_groups"
+	 * directory.
+	 */
+	if (rtype == RDTMON_GROUP && !is_mon_groups(parent_kn, name)) {
+		ret = -EPERM;
+		goto out_unlock;
+	}
+
 	if (rtype == RDTMON_GROUP &&
 	    (prdtgrp->mode == RDT_MODE_PSEUDO_LOCKSETUP ||
 	     prdtgrp->mode == RDT_MODE_PSEUDO_LOCKED)) {
@@ -3751,22 +3776,6 @@ out_unlock:
 	return ret;
 }
 
-/*
- * We allow creating mon groups only with in a directory called "mon_groups"
- * which is present in every ctrl_mon group. Check if this is a valid
- * "mon_groups" directory.
- *
- * 1. The directory should be named "mon_groups".
- * 2. The mon group itself should "not" be named "mon_groups".
- *   This makes sure "mon_groups" directory always has a ctrl_mon group
- *   as parent.
- */
-static bool is_mon_groups(struct kernfs_node *kn, const char *name)
-{
-	return (!strcmp(rdt_kn_name(kn), "mon_groups") &&
-		strcmp(name, "mon_groups"));
-}
-
 static int rdtgroup_mkdir(struct kernfs_node *parent_kn, const char *name,
 			  umode_t mode)
 {
@@ -3782,11 +3791,8 @@ static int rdtgroup_mkdir(struct kernfs_node *parent_kn, const char *name,
 	if (resctrl_arch_alloc_capable() && parent_kn == rdtgroup_default.kn)
 		return rdtgroup_mkdir_ctrl_mon(parent_kn, name, mode);
 
-	/*
-	 * If RDT monitoring is supported and the parent directory is a valid
-	 * "mon_groups" directory, add a monitoring subdirectory.
-	 */
-	if (resctrl_arch_mon_capable() && is_mon_groups(parent_kn, name))
+	/* Else, attempt to add a monitoring subdirectory. */
+	if (resctrl_arch_mon_capable())
 		return rdtgroup_mkdir_mon(parent_kn, name, mode);
 
 	return -EPERM;