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 2A9551119B for ; Thu, 24 Aug 2023 14:54:20 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 9D9B2C433C8; Thu, 24 Aug 2023 14:54:19 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1692888860; bh=jltj/H5xCx1bUjyJEiFAiQ+aEbwSGHQMsOAppHVkWqg=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=DlkYOtsgY8ZBo7fSawvLZlYmZnP6uiiuJDBU3N0vghHpgOZPKQ49JE1w3N56FVF3b rak5ThVs0tZ+g/JqjwnyTXMNo8M512ouIvSA+jEYEIh7IxclosEdsakqBZNI/0gaDF qGbFKgge7EuwrEh30Vn9M73acBnEofYilK+ubwA0= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Wolfram Sang , "Michael S. Tsirkin" , Sasha Levin Subject: [PATCH 5.15 067/139] virtio-mmio: dont break lifecycle of vm_dev Date: Thu, 24 Aug 2023 16:49:50 +0200 Message-ID: <20230824145026.568164294@linuxfoundation.org> X-Mailer: git-send-email 2.42.0 In-Reply-To: <20230824145023.559380953@linuxfoundation.org> References: <20230824145023.559380953@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Wolfram Sang [ Upstream commit 55c91fedd03d7b9cf0c5199b2eb12b9b8e95281a ] vm_dev has a separate lifecycle because it has a 'struct device' embedded. Thus, having a release callback for it is correct. Allocating the vm_dev struct with devres totally breaks this protection, though. Instead of waiting for the vm_dev release callback, the memory is freed when the platform_device is removed. Resulting in a use-after-free when finally the callback is to be called. To easily see the problem, compile the kernel with CONFIG_DEBUG_KOBJECT_RELEASE and unbind with sysfs. The fix is easy, don't use devres in this case. Found during my research about object lifetime problems. Fixes: 7eb781b1bbb7 ("virtio_mmio: add cleanup for virtio_mmio_probe") Signed-off-by: Wolfram Sang Message-Id: <20230629120526.7184-1-wsa+renesas@sang-engineering.com> Signed-off-by: Michael S. Tsirkin Signed-off-by: Sasha Levin --- drivers/virtio/virtio_mmio.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/virtio/virtio_mmio.c b/drivers/virtio/virtio_mmio.c index fe696aafaed86..f4d43d60d710f 100644 --- a/drivers/virtio/virtio_mmio.c +++ b/drivers/virtio/virtio_mmio.c @@ -572,9 +572,8 @@ static void virtio_mmio_release_dev(struct device *_d) struct virtio_device *vdev = container_of(_d, struct virtio_device, dev); struct virtio_mmio_device *vm_dev = to_virtio_mmio_device(vdev); - struct platform_device *pdev = vm_dev->pdev; - devm_kfree(&pdev->dev, vm_dev); + kfree(vm_dev); } /* Platform device */ @@ -585,7 +584,7 @@ static int virtio_mmio_probe(struct platform_device *pdev) unsigned long magic; int rc; - vm_dev = devm_kzalloc(&pdev->dev, sizeof(*vm_dev), GFP_KERNEL); + vm_dev = kzalloc(sizeof(*vm_dev), GFP_KERNEL); if (!vm_dev) return -ENOMEM; -- 2.40.1