From patchwork Wed Jan 24 10:57:48 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Fiona Ebner X-Patchwork-Id: 1890187 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@legolas.ozlabs.org Authentication-Results: legolas.ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=nongnu.org (client-ip=209.51.188.17; helo=lists.gnu.org; envelope-from=qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org; receiver=patchwork.ozlabs.org) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-ECDSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by legolas.ozlabs.org (Postfix) with ESMTPS id 4TKgv15ksjz23dy for ; Wed, 24 Jan 2024 21:59:08 +1100 (AEDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1rSaxe-0006zw-IA; Wed, 24 Jan 2024 05:58:26 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rSaxd-0006zb-0B; Wed, 24 Jan 2024 05:58:25 -0500 Received: from proxmox-new.maurer-it.com ([94.136.29.106]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rSaxb-0002xm-5F; Wed, 24 Jan 2024 05:58:24 -0500 Received: from proxmox-new.maurer-it.com (localhost.localdomain [127.0.0.1]) by proxmox-new.maurer-it.com (Proxmox) with ESMTP id DAFF749264; Wed, 24 Jan 2024 11:58:19 +0100 (CET) From: Fiona Ebner To: qemu-devel@nongnu.org Cc: kraxel@redhat.com, m.frank@proxmox.com, marcandre.lureau@redhat.com, berrange@redhat.com, mcascell@redhat.com, qemu-stable@nongnu.org Subject: [PATCH v3 1/2] ui/clipboard: mark type as not available when there is no data Date: Wed, 24 Jan 2024 11:57:48 +0100 Message-Id: <20240124105749.204610-1-f.ebner@proxmox.com> X-Mailer: git-send-email 2.39.2 MIME-Version: 1.0 Received-SPF: pass client-ip=94.136.29.106; envelope-from=f.ebner@proxmox.com; helo=proxmox-new.maurer-it.com X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, T_SCC_BODY_TEXT_LINE=-0.01 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org With VNC, a client can send a non-extended VNC_MSG_CLIENT_CUT_TEXT message with len=0. In qemu_clipboard_set_data(), the clipboard info will be updated setting data to NULL (because g_memdup(data, size) returns NULL when size is 0). If the client does not set the VNC_ENCODING_CLIPBOARD_EXT feature when setting up the encodings, then the 'request' callback for the clipboard peer is not initialized. Later, because data is NULL, qemu_clipboard_request() can be reached via vdagent_chr_write() and vdagent_clipboard_recv_request() and there, the clipboard owner's 'request' callback will be attempted to be called, but that is a NULL pointer. In particular, this can happen when using the KRDC (22.12.3) VNC client. Another scenario leading to the same issue is with two clients (say noVNC and KRDC): The noVNC client sets the extension VNC_FEATURE_CLIPBOARD_EXT and initializes its cbpeer. The KRDC client does not, but triggers a vnc_client_cut_text() (note it's not the _ext variant)). There, a new clipboard info with it as the 'owner' is created and via qemu_clipboard_set_data() is called, which in turn calls qemu_clipboard_update() with that info. In qemu_clipboard_update(), the notifier for the noVNC client will be called, i.e. vnc_clipboard_notify() and also set vs->cbinfo for the noVNC client. The 'owner' in that clipboard info is the clipboard peer for the KRDC client, which did not initialize the 'request' function. That sounds correct to me, it is the owner of that clipboard info. Then when noVNC sends a VNC_MSG_CLIENT_CUT_TEXT message (it did set the VNC_FEATURE_CLIPBOARD_EXT feature correctly, so a check for it passes), that clipboard info is passed to qemu_clipboard_request() and the original segfault still happens. Fix the issue by handling updates with size 0 differently. In particular, mark in the clipboard info that the type is not available. While at it, switch to g_memdup2(), because g_memdup() is deprecated. Cc: qemu-stable@nongnu.org Fixes: CVE-2023-6683 Reported-by: Markus Frank Suggested-by: Marc-André Lureau Signed-off-by: Fiona Ebner Reviewed-by: Marc-André Lureau Tested-by: Markus Frank --- Changes in v3: * Yet another new appraoch, setting available to false when no data is passed in when updating. * Update commit message to focus on the fact that non-extended VNC_MSG_CLIENT_CUT_TEXT messages with len=0 are problematic. ui/clipboard.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/ui/clipboard.c b/ui/clipboard.c index 3d14bffaf8..b3f6fa3c9e 100644 --- a/ui/clipboard.c +++ b/ui/clipboard.c @@ -163,9 +163,15 @@ void qemu_clipboard_set_data(QemuClipboardPeer *peer, } g_free(info->types[type].data); - info->types[type].data = g_memdup(data, size); - info->types[type].size = size; - info->types[type].available = true; + if (size) { + info->types[type].data = g_memdup2(data, size); + info->types[type].size = size; + info->types[type].available = true; + } else { + info->types[type].data = NULL; + info->types[type].size = 0; + info->types[type].available = false; + } if (update) { qemu_clipboard_update(info);