CC-ing qemu-devel with a patch solely for libvirt doesn't make sense.
Also 'libvirt-security' list is private and is is intended as a first
contact list for stuff to be embargoed. It makes little sense to include
it when posting to the public 'libvir-list'.
On Thu, Sep 14, 2023 at 09:44:13 +0000, Миронов Сергей Владимирович wrote:
> There are no checks, virDomainChrSourceDefNew can return 0.
s/0/NULL
While very technically true, realistically that can't happen any more.
'virObjectNew' always returns a valid pointer or abort()s, and
VIR_CLASS_NEW can return 0 on programming errors.
Thus this is not a security issue.
> Return value of a function 'virDomainChrSourceDefNew'
>
> is dereferenced at qemu_hotplug.c without checking for NULL,
>
> but it is usually checked for this function.
Remove the extra empty lines please.
>
>
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
>
>
> Fixes: 1f85f0967b ("ci: jobs.sh: Add back '--no-suite syntax-check --print-errorlogs'")
^^ This makes no sense. The commit you are referencing is changing a
shell script.
>
> Signed-off-by: Sergey Mironov <mironov@xxxxxxxxxx>
>
> ---
> src/qemu/qemu_hotplug.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
> index 177ca87d11..09e16c2c7e 100644
> --- a/src/qemu/qemu_hotplug.c
> +++ b/src/qemu/qemu_hotplug.c
> @@ -3207,6 +3207,8 @@ qemuDomainAttachFSDevice(virQEMUDriver *driver,
> qemuAssignDeviceFSAlias(vm->def, fs);
>
> chardev = virDomainChrSourceDefNew(priv->driver->xmlopt);
> + if (chardev == NULL)
> + goto cleanup;
> chardev->type = VIR_DOMAIN_CHR_TYPE_UNIX;
> chardev->data.nix.path = qemuDomainGetVHostUserFSSocketPath(priv, fs);
> --
> 2.31.1
