From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 892685BF93; Mon, 18 Dec 2023 14:03:38 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b="F8/XAFK6" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 11BA6C433C8; Mon, 18 Dec 2023 14:03:37 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1702908218; bh=trbIOKhCetAVdmhf2FShdphHIJ7EeGCf3qsjRYZbPjY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=F8/XAFK6CjTRs5eki4d1ZMdIGgVH/D/eNxXudnCSDgQiuyh+CTYbaVjzUeQ6vrwDH j36OoeG9HHkrCLu7WExaNMVqRna0Ur0BoYwJ3YYHcZRkb/b0oE74Nem1B9XwepIbgI 27/7XSqleurXdCf+IPFM3Wiof6CtqOIyfZVMGP8c= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Li Nan , Josef Bacik , Jens Axboe , Sasha Levin Subject: [PATCH 6.6 097/166] nbd: fix null-ptr-dereference while accessing nbd->config Date: Mon, 18 Dec 2023 14:51:03 +0100 Message-ID: <20231218135109.325084543@linuxfoundation.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20231218135104.927894164@linuxfoundation.org> References: <20231218135104.927894164@linuxfoundation.org> User-Agent: quilt/0.67 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: stable@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 6.6-stable review patch. If anyone has any objections, please let me know. ------------------ From: Li Nan [ Upstream commit c2da049f419417808466c529999170f5c3ef7d3d ] Memory reordering may occur in nbd_genl_connect(), causing config_refs to be set to 1 while nbd->config is still empty. Opening nbd at this time will cause null-ptr-dereference. T1 T2 nbd_open nbd_get_config_unlocked nbd_genl_connect nbd_alloc_and_init_config //memory reordered refcount_set(&nbd->config_refs, 1) // 2 nbd->config ->null point nbd->config = config // 1 Fix it by adding smp barrier to guarantee the execution sequence. Signed-off-by: Li Nan Reviewed-by: Josef Bacik Link: https://lore.kernel.org/r/20231116162316.1740402-4-linan666@huaweicloud.com Signed-off-by: Jens Axboe Signed-off-by: Sasha Levin --- drivers/block/nbd.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index daaf8805e876c..3f03cb3dc33cc 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -397,8 +397,16 @@ static u32 req_to_nbd_cmd_type(struct request *req) static struct nbd_config *nbd_get_config_unlocked(struct nbd_device *nbd) { - if (refcount_inc_not_zero(&nbd->config_refs)) + if (refcount_inc_not_zero(&nbd->config_refs)) { + /* + * Add smp_mb__after_atomic to ensure that reading nbd->config_refs + * and reading nbd->config is ordered. The pair is the barrier in + * nbd_alloc_and_init_config(), avoid nbd->config_refs is set + * before nbd->config. + */ + smp_mb__after_atomic(); return nbd->config; + } return NULL; } @@ -1559,7 +1567,15 @@ static int nbd_alloc_and_init_config(struct nbd_device *nbd) init_waitqueue_head(&config->conn_wait); config->blksize_bits = NBD_DEF_BLKSIZE_BITS; atomic_set(&config->live_connections, 0); + nbd->config = config; + /* + * Order refcount_set(&nbd->config_refs, 1) and nbd->config assignment, + * its pair is the barrier in nbd_get_config_unlocked(). + * So nbd_get_config_unlocked() won't see nbd->config as null after + * refcount_inc_not_zero() succeed. + */ + smp_mb__before_atomic(); refcount_set(&nbd->config_refs, 1); return 0; -- 2.43.0