Skip to contentSkip to conversation

kprobes: Remove NMI context check

PATCH In review Masami Hiramatsu 2 participants 3 messages Dec 10, 2020 – Dec 11, 2020

Subsystems (2)
lore
Share & tools
lore

Masami Hiramatsu

author maintainer
#0 lore raw

PATCH 1/2 kprobes: Remove NMI context check

commit e03b4a084ea6b0a18b0e874baec439e69090c168 upstream.

The in_nmi() check in pre_handler_kretprobe() is meant to avoid recursion, and blindly assumes that anything NMI is recursive.

However, since commit:

9b38cc704e84 ("kretprobe: Prevent triggering kretprobe from within kprobe_flush_task")

there is a better way to detect and avoid actual recursion.

By setting a dummy kprobe, any actual exceptions will terminate early (by trying to handle the dummy kprobe), and recursion will not happen.

Employ this to avoid the kretprobe_table_lock() recursion, replacing the over-eager in_nmi() check.

---
 kernel/kprobes.c |   16 ++++------------
 1 file changed, 4 insertions(+), 12 deletions(-)
kernel/kprobes.c +4 -12
index e995541d277d..b885d884603d 100644
@@ -1359,7 +1359,8 @@ static void cleanup_rp_inst(struct kretprobe *rp)
 	struct hlist_node *next;
 	struct hlist_head *head;
 
-	/* No race here */
+	/* To avoid recursive kretprobe by NMI, set kprobe busy here */
+	kprobe_busy_begin();
 	for (hash = 0; hash < KPROBE_TABLE_SIZE; hash++) {
 		kretprobe_table_lock(hash, &flags);
 		head = &kretprobe_inst_table[hash];
@@ -1369,6 +1370,8 @@ static void cleanup_rp_inst(struct kretprobe *rp)
 		}
 		kretprobe_table_unlock(hash, &flags);
 	}
+	kprobe_busy_end();
+
 	free_rp_inst(rp);
 }
 NOKPROBE_SYMBOL(cleanup_rp_inst);
@@ -1937,17 +1940,6 @@ static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs)
 	unsigned long hash, flags = 0;
 	struct kretprobe_instance *ri;
 
-	/*
-	 * To avoid deadlocks, prohibit return probing in NMI contexts,
-	 * just skip the probe and increase the (inexact) 'nmissed'
-	 * statistical counter, so that the user is informed that
-	 * something happened:
-	 */
-	if (unlikely(in_nmi())) {
-		rp->nmissed++;
-		return 0;
-	}
-
 	/* TODO: consider to only swap the RA after the last pre_handler fired */
 	hash = hash_ptr(current, KPROBE_HASH_BITS);
 	raw_spin_lock_irqsave(&rp->lock, flags);
author maintainer
#1 lore raw

PATCH 2/2 kprobes: Tell lockdep about kprobe nesting

From: Steven Rostedt (VMware) <[email protected]>

commit 645f224e7ba2f4200bf163153d384ceb0de5462e upstream.

Since the kprobe handlers have protection that prohibits other handlers from executing in other contexts (like if an NMI comes in while processing a kprobe, and executes the same kprobe, it will get fail with a "busy" return). Lockdep is unaware of this protection. Use lockdep's nesting api to differentiate between locks taken in INT3 context and other context to suppress the false warnings.

---
 kernel/kprobes.c |   25 +++++++++++++++++++++----
 1 file changed, 21 insertions(+), 4 deletions(-)
kernel/kprobes.c +21 -4
index b885d884603d..1b7fd1ab8ddc 100644
@@ -1250,7 +1250,13 @@ __acquires(hlist_lock)
 
 	*head = &kretprobe_inst_table[hash];
 	hlist_lock = kretprobe_table_lock_ptr(hash);
-	raw_spin_lock_irqsave(hlist_lock, *flags);
+	/*
+	 * Nested is a workaround that will soon not be needed.
+	 * There's other protections that make sure the same lock
+	 * is not taken on the same CPU that lockdep is unaware of.
+	 * Differentiate when it is taken in NMI context.
+	 */
+	raw_spin_lock_irqsave_nested(hlist_lock, *flags, !!in_nmi());
 }
 NOKPROBE_SYMBOL(kretprobe_hash_lock);
 
@@ -1259,7 +1265,13 @@ static void kretprobe_table_lock(unsigned long hash,
 __acquires(hlist_lock)
 {
 	raw_spinlock_t *hlist_lock = kretprobe_table_lock_ptr(hash);
-	raw_spin_lock_irqsave(hlist_lock, *flags);
+	/*
+	 * Nested is a workaround that will soon not be needed.
+	 * There's other protections that make sure the same lock
+	 * is not taken on the same CPU that lockdep is unaware of.
+	 * Differentiate when it is taken in NMI context.
+	 */
+	raw_spin_lock_irqsave_nested(hlist_lock, *flags, !!in_nmi());
 }
 NOKPROBE_SYMBOL(kretprobe_table_lock);
 
@@ -1942,7 +1954,12 @@ static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs)
 
 	/* TODO: consider to only swap the RA after the last pre_handler fired */
 	hash = hash_ptr(current, KPROBE_HASH_BITS);
-	raw_spin_lock_irqsave(&rp->lock, flags);
+	/*
+	 * Nested is a workaround that will soon not be needed.
+	 * There's other protections that make sure the same lock
+	 * is not taken on the same CPU that lockdep is unaware of.
+	 */
+	raw_spin_lock_irqsave_nested(&rp->lock, flags, 1);
 	if (!hlist_empty(&rp->free_instances)) {
 		ri = hlist_entry(rp->free_instances.first,
 				struct kretprobe_instance, hlist);
@@ -1953,7 +1970,7 @@ static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs)
 		ri->task = current;
 
 		if (rp->entry_handler && rp->entry_handler(ri, regs)) {
-			raw_spin_lock_irqsave(&rp->lock, flags);
+			raw_spin_lock_irqsave_nested(&rp->lock, flags, 1);
 			hlist_add_head(&ri->hlist, &rp->free_instances);
 			raw_spin_unlock_irqrestore(&rp->lock, flags);
 			return 0;