From: "Daniel P. Berrangé" <berrange@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Daniel P. Berrangé" <berrange@redhat.com>,
"Thomas Huth" <thuth@redhat.com>,
"Marc-André Lureau" <marcandre.lureau@redhat.com>
Subject: [PULL 01/26] hw: mark most virtio PCI devices as secure
Date: Mon, 16 Nov 2026 22:21:33 +0000 [thread overview]
Message-ID: <20261116222158.2840471-2-berrange@redhat.com> (raw)
In-Reply-To: <20261116222158.2840471-1-berrange@redhat.com>
These are all intended for use in a virtualization scenario and must
provide a security boundary. This can be done for almost all virtio
PCI devices by modifying the common type register helper.
Exceptions are made for the virtio-gpu-gl, virtio-vga-gl,
virtio-vga-rutatbaga and virtio-gpu-rutabaga devices.
Despite being intended for virtualization use cases, the complexity
of support for 3d acceleration is such that QEMU can't offer security
for these devices at this time.
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
hw/display/virtio-gpu-pci-gl.c | 2 +-
hw/display/virtio-vga-gl.c | 2 +-
hw/display/virtio-vga-rutabaga.c | 2 +-
hw/virtio/virtio-pci.c | 17 ++++++++++++++++-
include/hw/virtio/virtio-pci.h | 10 +++++++++-
5 files changed, 28 insertions(+), 5 deletions(-)
diff --git a/hw/display/virtio-gpu-pci-gl.c b/hw/display/virtio-gpu-pci-gl.c
index c8923795dfa..dc67516b377 100644
--- a/hw/display/virtio-gpu-pci-gl.c
+++ b/hw/display/virtio-gpu-pci-gl.c
@@ -51,7 +51,7 @@ module_kconfig(VIRTIO_PCI);
static void virtio_gpu_gl_pci_register_types(void)
{
- virtio_pci_types_register(&virtio_gpu_gl_pci_info);
+ virtio_pci_types_register_insecure(&virtio_gpu_gl_pci_info);
}
type_init(virtio_gpu_gl_pci_register_types)
diff --git a/hw/display/virtio-vga-gl.c b/hw/display/virtio-vga-gl.c
index 178e4c71725..029d288c035 100644
--- a/hw/display/virtio-vga-gl.c
+++ b/hw/display/virtio-vga-gl.c
@@ -42,7 +42,7 @@ module_kconfig(VIRTIO_VGA);
static void virtio_vga_register_types(void)
{
if (have_vga) {
- virtio_pci_types_register(&virtio_vga_gl_info);
+ virtio_pci_types_register_insecure(&virtio_vga_gl_info);
}
}
diff --git a/hw/display/virtio-vga-rutabaga.c b/hw/display/virtio-vga-rutabaga.c
index 1e07ee0b0d3..8ff715f8a35 100644
--- a/hw/display/virtio-vga-rutabaga.c
+++ b/hw/display/virtio-vga-rutabaga.c
@@ -41,7 +41,7 @@ module_kconfig(VIRTIO_VGA);
static void virtio_vga_register_types(void)
{
if (have_vga) {
- virtio_pci_types_register(&virtio_vga_rutabaga_info);
+ virtio_pci_types_register_insecure(&virtio_vga_rutabaga_info);
}
}
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
index c73f1f546cb..fe2e545c302 100644
--- a/hw/virtio/virtio-pci.c
+++ b/hw/virtio/virtio-pci.c
@@ -2510,7 +2510,9 @@ static void virtio_pci_non_transitional_instance_init(Object *obj)
proxy->disable_modern = false;
}
-void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
+static void
+virtio_pci_types_register_internal(const VirtioPCIDeviceTypeInfo *t,
+ bool secure)
{
char *base_name = NULL;
TypeInfo base_type_info = {
@@ -2527,6 +2529,7 @@ void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
.name = t->generic_name,
.parent = base_type_info.name,
.class_init = virtio_pci_generic_class_init,
+ .secure = secure,
.interfaces = (const InterfaceInfo[]) {
{ INTERFACE_PCIE_DEVICE },
{ INTERFACE_CONVENTIONAL_PCI_DEVICE },
@@ -2562,6 +2565,7 @@ void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
.name = t->non_transitional_name,
.parent = base_type_info.name,
.instance_init = virtio_pci_non_transitional_instance_init,
+ .secure = secure,
.interfaces = (const InterfaceInfo[]) {
{ INTERFACE_PCIE_DEVICE },
{ INTERFACE_CONVENTIONAL_PCI_DEVICE },
@@ -2576,6 +2580,7 @@ void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
.name = t->transitional_name,
.parent = base_type_info.name,
.instance_init = virtio_pci_transitional_instance_init,
+ .secure = secure,
.interfaces = (const InterfaceInfo[]) {
/*
* Transitional virtio devices work only as Conventional PCI
@@ -2590,6 +2595,16 @@ void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
g_free(base_name);
}
+void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t)
+{
+ virtio_pci_types_register_internal(t, true);
+}
+
+void virtio_pci_types_register_insecure(const VirtioPCIDeviceTypeInfo *t)
+{
+ virtio_pci_types_register_internal(t, false);
+}
+
unsigned virtio_pci_optimal_num_queues(unsigned fixed_queues)
{
/*
diff --git a/include/hw/virtio/virtio-pci.h b/include/hw/virtio/virtio-pci.h
index ee50fb0d551..f033237035a 100644
--- a/include/hw/virtio/virtio-pci.h
+++ b/include/hw/virtio/virtio-pci.h
@@ -239,8 +239,16 @@ typedef struct VirtioPCIDeviceTypeInfo {
const InterfaceInfo *interfaces;
} VirtioPCIDeviceTypeInfo;
-/* Register virtio-pci type(s). @t must be static. */
+/*
+ * Register virtio-pci type(s).
+ * @t must be static.
+ *
+ * Since they serve the virtualization use case, most virtio
+ * devices are expected to be classed as secure. In the rare
+ * scenario where this is not desired, use the second method.
+ */
void virtio_pci_types_register(const VirtioPCIDeviceTypeInfo *t);
+void virtio_pci_types_register_insecure(const VirtioPCIDeviceTypeInfo *t);
/**
* virtio_pci_optimal_num_queues:
--
2.56.1
next prev parent reply other threads:[~2026-11-16 22:23 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-16 22:21 [PULL 00/26] Device security tagging Daniel P. Berrangé
2026-11-16 22:21 ` Daniel P. Berrangé [this message]
2026-11-16 22:21 ` [PULL 02/26] hw: mark all virtio CCW devices as secure Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 03/26] hw: mark all vhost devices a secure Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 04/26] hw: mark most remaining virtio object types as secure Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 05/26] hw/vfio: mark all VFIO object classes " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 06/26] hw/xen: mark all Xen related object types as being secure Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 07/26] hw/net: mark e1000, e1000e, rtl8139 & sPAPR VLAN as secure Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 08/26] hw/usb: mark commonly used USB devices/hosts " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 09/26] hw/watchdog: mark some watchdog devices " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 10/26] hw/scsi: mark spapr SCSI controller " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 11/26] hw/scsi: mark SCSI disk endpoint devices " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 12/26] hw/ide: mark ICH9 and ide-hd/ide-cd " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 13/26] hw: define most common PCI types " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 14/26] hw/pci-host: mark common x86, ppc, arm and s390 PCI hosts " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 15/26] hw/display: mark bochs, ramfb and VGA " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 16/26] hw/tpm: mark all TPM implementations " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 17/26] hw/misc: mark pvpanic, vmcoreinfo " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 18/26] hw/audio: mark Intel HDA devices & codecs " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 19/26] hw/char: mark common serial / console devices " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 20/26] hw/mem: mark nvdimm, pc-dimm & spapr-nvdimm " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 21/26] hw/uefi: mark the EFI vars service " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 22/26] hw/acpi: mark erst, vmclock and vmgenid devices " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 23/26] hw: mark KVM clock and RTC " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 24/26] hw: device AMD, Intel and ARM IOMMUs " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 25/26] hw/input: mark PS/2 and PC Keyboard devices " Daniel P. Berrangé
2026-11-16 22:21 ` [PULL 26/26] hw/i386: mark vmmouse / vmport " Daniel P. Berrangé
2026-11-17 1:58 ` [PULL 00/26] Device security tagging Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261116222158.2840471-2-berrange@redhat.com \
--to=berrange@redhat.com \
--cc=marcandre.lureau@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=thuth@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.