Received-SPF: pass (google.com: domain of linux-kernel+bounces-125966-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
From: Hillf Danton <hdanton@sina.com>
To: syzbot <syzbot+1fa663a2100308ab6eab@syzkaller.appspotmail.com>
Cc: linux-kernel@vger.kernel.org,
	syzkaller-bugs@googlegroups.com
Subject: Re: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu
Date: Sun, 31 Mar 2024 14:19:58 +0800
Message-Id: <20240331061958.3340-1-hdanton@sina.com>
In-Reply-To: <000000000000716bb60614acbf37@google.com>
References: 
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

On Wed, 27 Mar 2024 16:27:19 -0700
> syzbot has found a reproducer for the following issue on:
> 
> HEAD commit:    443574b03387 riscv, bpf: Fix kfunc parameters incompatibil..
> git tree:       bpf
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=16ca53c9180000

#syz test https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git  fe46a7dd189e

--- x/kernel/rcu/tree.c
+++ y/kernel/rcu/tree.c
@@ -2957,7 +2957,8 @@ krc_this_cpu_lock(unsigned long *flags)
 
 	local_irq_save(*flags);	// For safely calling this_cpu_ptr().
 	krcp = this_cpu_ptr(&krc);
-	raw_spin_lock(&krcp->lock);
+	while (!raw_spin_trylock(&krcp->lock))
+		;
 
 	return krcp;
 }
--- x/kernel/bpf/lpm_trie.c
+++ y/kernel/bpf/lpm_trie.c
@@ -382,9 +382,10 @@ static long trie_update_elem(struct bpf_
 			trie->n_entries--;
 
 		rcu_assign_pointer(*slot, new_node);
-		kfree_rcu(node, rcu);
 
-		goto out;
+		spin_unlock_irqrestore(&trie->lock, irq_flags);
+		kfree_rcu(node, rcu);
+		return 0;
 	}
 
 	/* If the new node matches the prefix completely, it must be inserted
--