From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33719) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fsA4n-0002IN-UU for qemu-devel@nongnu.org; Tue, 21 Aug 2018 13:04:22 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fsA4X-0006rA-Qj for qemu-devel@nongnu.org; Tue, 21 Aug 2018 13:04:11 -0400 Received: from mail-wm0-x235.google.com ([2a00:1450:400c:c09::235]:36956) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1fsA4W-0006Eo-Rw for qemu-devel@nongnu.org; Tue, 21 Aug 2018 13:04:01 -0400 Received: by mail-wm0-x235.google.com with SMTP id n11-v6so3495222wmc.2 for ; Tue, 21 Aug 2018 10:03:37 -0700 (PDT) Sender: Paolo Bonzini From: Paolo Bonzini Date: Tue, 21 Aug 2018 19:02:02 +0200 Message-Id: <1534870966-9287-31-git-send-email-pbonzini@redhat.com> In-Reply-To: <1534870966-9287-1-git-send-email-pbonzini@redhat.com> References: <1534870966-9287-1-git-send-email-pbonzini@redhat.com> Subject: [Qemu-devel] [PULL 30/74] i386: Fix arch_query_cpu_model_expansion() leak List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Eduardo Habkost From: Eduardo Habkost Reported by Coverity: Error: RESOURCE_LEAK (CWE-772): [#def439] qemu-2.12.0/target/i386/cpu.c:3179: alloc_fn: Storage is returned from allocation function "qdict_new". qemu-2.12.0/qobject/qdict.c:34:5: alloc_fn: Storage is returned from allocation function "g_malloc0". qemu-2.12.0/qobject/qdict.c:34:5: var_assign: Assigning: "qdict" = "g_malloc0(4120UL)". qemu-2.12.0/qobject/qdict.c:37:5: return_alloc: Returning allocated memory "qdict". qemu-2.12.0/target/i386/cpu.c:3179: var_assign: Assigning: "props" = storage returned from "qdict_new()". qemu-2.12.0/target/i386/cpu.c:3217: leaked_storage: Variable "props" going out of scope leaks the storage it points to. This was introduced by commit b8097deb359b ("i386: Improve query-cpu-model-expansion full mode"). The leak is only theoretical: if ret->model->props is set to props, the qapi_free_CpuModelExpansionInfo() call will free props too in case of errors. The only way for this to not happen is if we enter the default branch of the switch statement, which would never happen because all CpuModelExpansionType values are being handled. It's still worth to change this to make the allocation logic easier to follow and make the Coverity error go away. To make everything simpler, initialize ret->model and ret->model->props earlier in the function. While at it, remove redundant check for !prop because prop is always initialized at the beginning of the function. Fixes: b8097deb359bbbd92592b9670adfe9e245b2d0bd Signed-off-by: Eduardo Habkost Message-Id: <20180816183509.8231-1-ehabkost@redhat.com> Signed-off-by: Paolo Bonzini --- target/i386/cpu.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/target/i386/cpu.c b/target/i386/cpu.c index 4e4fe8f..f24295e 100644 --- a/target/i386/cpu.c +++ b/target/i386/cpu.c @@ -3880,6 +3880,9 @@ arch_query_cpu_model_expansion(CpuModelExpansionType type, } props = qdict_new(); + ret->model = g_new0(CpuModelInfo, 1); + ret->model->props = QOBJECT(props); + ret->model->has_props = true; switch (type) { case CPU_MODEL_EXPANSION_TYPE_STATIC: @@ -3900,15 +3903,9 @@ arch_query_cpu_model_expansion(CpuModelExpansionType type, goto out; } - if (!props) { - props = qdict_new(); - } x86_cpu_to_dict(xc, props); - ret->model = g_new0(CpuModelInfo, 1); ret->model->name = g_strdup(base_name); - ret->model->props = QOBJECT(props); - ret->model->has_props = true; out: object_unref(OBJECT(xc)); -- 1.8.3.1