* [PULL 01/24] meson: cleanup win32 library detection
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 02/24] target/i386: fix TB exit logic in gen_movl_seg() when writing to SS Paolo Bonzini
` (22 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel
As pointed out by Akihiko Odaki, all Win32 libraries in MinGW have lowercase
names. This means that on (case-insensitive) Windows you can use the mixed-case
names suggested by Microsoft or all-lowercase names, while on Linux you need to
make them lowercase.
QEMU was already using lowercase names, so there is no need to test the
mixed-case name version of libSynchronization. Remove the unnecessary test
and while at it make all the tests use "required: true".
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
meson.build | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/meson.build b/meson.build
index 34729c2a3dd..ed60be2a2d0 100644
--- a/meson.build
+++ b/meson.build
@@ -843,15 +843,12 @@ host_dsosuf = '.so'
if host_os == 'windows'
midl = find_program('midl', required: false)
widl = find_program('widl', required: false)
- pathcch = cc.find_library('pathcch')
- synchronization = cc.find_library('Synchronization', required: false)
- if not synchronization.found()
- # The library name is lowercase on mingw
- synchronization = cc.find_library('synchronization', required: true)
- endif
- socket = cc.find_library('ws2_32')
- winmm = cc.find_library('winmm')
+ # MinGW uses lowercase for library names
+ pathcch = cc.find_library('pathcch', required: true)
+ synchronization = cc.find_library('synchronization', required: true)
+ socket = cc.find_library('ws2_32', required: true)
+ winmm = cc.find_library('winmm', required: true)
win = import('windows')
version_res = win.compile_resources('version.rc',
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 02/24] target/i386: fix TB exit logic in gen_movl_seg() when writing to SS
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
2025-06-20 16:40 ` [PULL 01/24] meson: cleanup win32 library detection Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 03/24] hw: Fix type constant for DTB files Paolo Bonzini
` (21 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Mark Cave-Ayland, Peter Maydell
From: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Before commit e54ef98c8a ("target/i386: do not trigger IRQ shadow for LSS"), any
write to SS in gen_movl_seg() would cause a TB exit. The changes introduced by
this commit were intended to restrict the DISAS_EOB_INHIBIT_IRQ exit to the case
where inhibit_irq is true, but missed that a DISAS_EOB_NEXT exit can still be
required when writing to SS and inhibit_irq is false.
Comparing the PE(s) && !VM86(s) section with the logic in x86_update_hflags(), we
can see that the DISAS_EOB_NEXT exit is still required for the !CODE32 case when
writing to SS in gen_movl_seg() because any change to the SS flags can affect
hflags. Similarly we can see that the existing CODE32 case is still correct since
a change to any of DS, ES and SS can affect hflags. Finally for the
gen_op_movl_seg_real() case an explicit TB exit is not needed because the segment
register selector does not affect hflags.
Update the logic in gen_movl_seg() so that a write to SS with inhibit_irq set to
false where PE(s) && !VM86(s) will generate a DISAS_EOB_NEXT exit along with the
inline comment. This has the effect of allowing Win98SE to boot in QEMU once
again.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: e54ef98c8a ("target/i386: do not trigger IRQ shadow for LSS")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2987
Link: https://lore.kernel.org/r/20250611130315.383151-1-mark.cave-ayland@ilande.co.uk
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/tcg/translate.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c
index 0fcddc2ec08..0cb87d02012 100644
--- a/target/i386/tcg/translate.c
+++ b/target/i386/tcg/translate.c
@@ -2033,8 +2033,11 @@ static void gen_movl_seg(DisasContext *s, X86Seg seg_reg, TCGv src, bool inhibit
tcg_gen_trunc_tl_i32(sel, src);
gen_helper_load_seg(tcg_env, tcg_constant_i32(seg_reg), sel);
- /* For move to DS/ES/SS, the addseg or ss32 flags may change. */
- if (CODE32(s) && seg_reg < R_FS) {
+ /*
+ * For moves to SS, the SS32 flag may change. For CODE32 only, changes
+ * to SS, DS and ES may change the ADDSEG flags.
+ */
+ if (seg_reg == R_SS || (CODE32(s) && seg_reg < R_FS)) {
s->base.is_jmp = DISAS_EOB_NEXT;
}
} else {
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 03/24] hw: Fix type constant for DTB files
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
2025-06-20 16:40 ` [PULL 01/24] meson: cleanup win32 library detection Paolo Bonzini
2025-06-20 16:40 ` [PULL 02/24] target/i386: fix TB exit logic in gen_movl_seg() when writing to SS Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 04/24] pc-bios/dtb/meson: Prefer target name to be outfile, not infile Paolo Bonzini
` (20 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Bernhard Beschow
From: Bernhard Beschow <shentey@gmail.com>
Commit fcb1ad456c58 ("system/datadir: Add new type constant for DTB files")
introduced a new type constant for DTB files and converted the boards with
bundled device trees to use it. Convert the other boards for consistency.
Fixes: fcb1ad456c58 ("system/datadir: Add new type constant for DTB files")
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250610204131.2862-2-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/arm/boot.c | 2 +-
hw/ppc/e500.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/arm/boot.c b/hw/arm/boot.c
index 79afb51b8a5..64040504a1e 100644
--- a/hw/arm/boot.c
+++ b/hw/arm/boot.c
@@ -527,7 +527,7 @@ int arm_load_dtb(hwaddr addr, const struct arm_boot_info *binfo,
if (binfo->dtb_filename) {
char *filename;
- filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, binfo->dtb_filename);
+ filename = qemu_find_file(QEMU_FILE_TYPE_DTB, binfo->dtb_filename);
if (!filename) {
fprintf(stderr, "Couldn't open dtb file %s\n", binfo->dtb_filename);
goto fail;
diff --git a/hw/ppc/e500.c b/hw/ppc/e500.c
index 6899802bedd..723c97fad2e 100644
--- a/hw/ppc/e500.c
+++ b/hw/ppc/e500.c
@@ -408,7 +408,7 @@ static int ppce500_load_device_tree(PPCE500MachineState *pms,
if (dtb_file) {
char *filename;
- filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, dtb_file);
+ filename = qemu_find_file(QEMU_FILE_TYPE_DTB, dtb_file);
if (!filename) {
goto out;
}
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 04/24] pc-bios/dtb/meson: Prefer target name to be outfile, not infile
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (2 preceding siblings ...)
2025-06-20 16:40 ` [PULL 03/24] hw: Fix type constant for DTB files Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 05/24] rust: qemu_api: introduce MaybeUninit field projection Paolo Bonzini
` (19 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Bernhard Beschow
From: Bernhard Beschow <shentey@gmail.com>
Makes this custom_target() usage consistent with other ones in QEMU.
Fixes: 6e0dc9d2a88a ("meson: compile bundled device trees")
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250610204131.2862-3-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
pc-bios/dtb/meson.build | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/pc-bios/dtb/meson.build b/pc-bios/dtb/meson.build
index 7a71835bca7..993032949fe 100644
--- a/pc-bios/dtb/meson.build
+++ b/pc-bios/dtb/meson.build
@@ -9,7 +9,7 @@ dtc = find_program('dtc', required: false)
if dtc.found()
foreach out : dtbs
f = fs.replace_suffix(out, '.dts')
- custom_target(f,
+ custom_target(out,
build_by_default: have_system,
input: files(f),
output: out,
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 05/24] rust: qemu_api: introduce MaybeUninit field projection
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (3 preceding siblings ...)
2025-06-20 16:40 ` [PULL 04/24] pc-bios/dtb/meson: Prefer target name to be outfile, not infile Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 06/24] rust: hpet: fully initialize object during instance_init Paolo Bonzini
` (18 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
Add a macro that makes it possible to convert a MaybeUninit<> into
another MaybeUninit<> for a single field within it. Furthermore, it is
possible to use the resulting MaybeUninitField<> in APIs that take the
parent object, such as memory_region_init_io().
This allows removing some of the undefined behavior from instance_init()
functions, though this may not be the definitive implementation.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/qemu-api/meson.build | 1 +
rust/qemu-api/src/lib.rs | 1 +
rust/qemu-api/src/uninit.rs | 85 +++++++++++++++++++++++++++++++++++++
3 files changed, 87 insertions(+)
create mode 100644 rust/qemu-api/src/uninit.rs
diff --git a/rust/qemu-api/meson.build b/rust/qemu-api/meson.build
index cac8595a148..33653b4a28e 100644
--- a/rust/qemu-api/meson.build
+++ b/rust/qemu-api/meson.build
@@ -28,6 +28,7 @@ _qemu_api_rs = static_library(
'src/qom.rs',
'src/sysbus.rs',
'src/timer.rs',
+ 'src/uninit.rs',
'src/vmstate.rs',
'src/zeroable.rs',
],
diff --git a/rust/qemu-api/src/lib.rs b/rust/qemu-api/src/lib.rs
index 93902fc94bc..c78198f0f41 100644
--- a/rust/qemu-api/src/lib.rs
+++ b/rust/qemu-api/src/lib.rs
@@ -27,6 +27,7 @@
pub mod qom;
pub mod sysbus;
pub mod timer;
+pub mod uninit;
pub mod vmstate;
pub mod zeroable;
diff --git a/rust/qemu-api/src/uninit.rs b/rust/qemu-api/src/uninit.rs
new file mode 100644
index 00000000000..04123b4ae99
--- /dev/null
+++ b/rust/qemu-api/src/uninit.rs
@@ -0,0 +1,85 @@
+//! Access fields of a [`MaybeUninit`]
+
+use std::{
+ mem::MaybeUninit,
+ ops::{Deref, DerefMut},
+};
+
+pub struct MaybeUninitField<'a, T, U> {
+ parent: &'a mut MaybeUninit<T>,
+ child: *mut U,
+}
+
+impl<'a, T, U> MaybeUninitField<'a, T, U> {
+ #[doc(hidden)]
+ pub fn new(parent: &'a mut MaybeUninit<T>, child: *mut U) -> Self {
+ MaybeUninitField { parent, child }
+ }
+
+ /// Return a constant pointer to the containing object of the field.
+ ///
+ /// Because the `MaybeUninitField` remembers the containing object,
+ /// it is possible to use it in foreign APIs that initialize the
+ /// child.
+ pub fn parent(f: &Self) -> *const T {
+ f.parent.as_ptr()
+ }
+
+ /// Return a mutable pointer to the containing object.
+ ///
+ /// Because the `MaybeUninitField` remembers the containing object,
+ /// it is possible to use it in foreign APIs that initialize the
+ /// child.
+ pub fn parent_mut(f: &mut Self) -> *mut T {
+ f.parent.as_mut_ptr()
+ }
+}
+
+impl<'a, T, U> Deref for MaybeUninitField<'a, T, U> {
+ type Target = MaybeUninit<U>;
+
+ fn deref(&self) -> &MaybeUninit<U> {
+ // SAFETY: self.child was obtained by dereferencing a valid mutable
+ // reference; the content of the memory may be invalid or uninitialized
+ // but MaybeUninit<_> makes no assumption on it
+ unsafe { &*(self.child.cast()) }
+ }
+}
+
+impl<'a, T, U> DerefMut for MaybeUninitField<'a, T, U> {
+ fn deref_mut(&mut self) -> &mut MaybeUninit<U> {
+ // SAFETY: self.child was obtained by dereferencing a valid mutable
+ // reference; the content of the memory may be invalid or uninitialized
+ // but MaybeUninit<_> makes no assumption on it
+ unsafe { &mut *(self.child.cast()) }
+ }
+}
+
+/// ```
+/// #[derive(Debug)]
+/// struct S {
+/// x: u32,
+/// y: u32,
+/// }
+///
+/// # use std::mem::MaybeUninit;
+/// # use qemu_api::{assert_match, uninit_field_mut};
+///
+/// let mut s: MaybeUninit<S> = MaybeUninit::zeroed();
+/// uninit_field_mut!(s, x).write(5);
+/// let s = unsafe { s.assume_init() };
+/// assert_match!(s, S { x: 5, y: 0 });
+/// ```
+#[macro_export]
+macro_rules! uninit_field_mut {
+ ($container:expr, $($field:tt)+) => {{
+ let container__: &mut ::std::mem::MaybeUninit<_> = &mut $container;
+ let container_ptr__ = container__.as_mut_ptr();
+
+ // SAFETY: the container is not used directly, only through a MaybeUninit<>,
+ // so the safety is delegated to the caller and to final invocation of
+ // assume_init()
+ let target__ = unsafe { std::ptr::addr_of_mut!((*container_ptr__).$($field)+) };
+ $crate::uninit::MaybeUninitField::new(container__, target__)
+ }};
+}
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 06/24] rust: hpet: fully initialize object during instance_init
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (4 preceding siblings ...)
2025-06-20 16:40 ` [PULL 05/24] rust: qemu_api: introduce MaybeUninit field projection Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 07/24] rust: qom: introduce ParentInit Paolo Bonzini
` (17 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
The array of BqlRefCell<HPETTimer> is not initialized yet at the
end of instance_init. In particular, the "state" field is NonNull
and therefore it is invalid to have it as zero bytes.
Note that MaybeUninit is necessary because assigning to self.timers[index]
would trigger Drop of the old value.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/hw/timer/hpet/src/device.rs | 42 +++++++++++++++++++-------------
1 file changed, 25 insertions(+), 17 deletions(-)
diff --git a/rust/hw/timer/hpet/src/device.rs b/rust/hw/timer/hpet/src/device.rs
index 735b2fbef79..340ca1d355d 100644
--- a/rust/hw/timer/hpet/src/device.rs
+++ b/rust/hw/timer/hpet/src/device.rs
@@ -4,6 +4,7 @@
use std::{
ffi::{c_int, c_void, CStr},
+ mem::MaybeUninit,
pin::Pin,
ptr::{addr_of_mut, null_mut, NonNull},
slice::from_ref,
@@ -25,6 +26,7 @@
qom_isa,
sysbus::{SysBusDevice, SysBusDeviceImpl},
timer::{Timer, CLOCK_VIRTUAL, NANOSECONDS_PER_SECOND},
+ uninit_field_mut,
vmstate::VMStateDescription,
vmstate_fields, vmstate_of, vmstate_struct, vmstate_subsections, vmstate_validate,
zeroable::Zeroable,
@@ -212,13 +214,13 @@ pub struct HPETTimer {
}
impl HPETTimer {
- fn init(&mut self, index: u8, state: &HPETState) {
- *self = HPETTimer {
+ fn new(index: u8, state: *const HPETState) -> HPETTimer {
+ HPETTimer {
index,
// SAFETY: the HPETTimer will only be used after the timer
// is initialized below.
qemu_timer: unsafe { Timer::new() },
- state: NonNull::new((state as *const HPETState).cast_mut()).unwrap(),
+ state: NonNull::new(state.cast_mut()).unwrap(),
config: 0,
cmp: 0,
fsb: 0,
@@ -226,19 +228,15 @@ fn init(&mut self, index: u8, state: &HPETState) {
period: 0,
wrap_flag: 0,
last: 0,
- };
+ }
+ }
+ fn init_timer_with_cell(cell: &BqlRefCell<Self>) {
+ let mut timer = cell.borrow_mut();
// SAFETY: HPETTimer is only used as part of HPETState, which is
// always pinned.
- let qemu_timer = unsafe { Pin::new_unchecked(&mut self.qemu_timer) };
- qemu_timer.init_full(
- None,
- CLOCK_VIRTUAL,
- Timer::NS,
- 0,
- timer_handler,
- &state.timers[self.index as usize],
- )
+ let qemu_timer = unsafe { Pin::new_unchecked(&mut timer.qemu_timer) };
+ qemu_timer.init_full(None, CLOCK_VIRTUAL, Timer::NS, 0, timer_handler, cell);
}
fn get_state(&self) -> &HPETState {
@@ -607,9 +605,18 @@ fn handle_legacy_irq(&self, irq: u32, level: u32) {
}
}
- fn init_timer(&self) {
- for (index, timer) in self.timers.iter().enumerate() {
- timer.borrow_mut().init(index.try_into().unwrap(), self);
+ fn init_timers(this: &mut MaybeUninit<Self>) {
+ let state = this.as_ptr();
+ for index in 0..HPET_MAX_TIMERS {
+ let mut timer = uninit_field_mut!(*this, timers[index]);
+
+ // Initialize in two steps, to avoid calling Timer::init_full on a
+ // temporary that can be moved.
+ let timer = timer.write(BqlRefCell::new(HPETTimer::new(
+ index.try_into().unwrap(),
+ state,
+ )));
+ HPETTimer::init_timer_with_cell(timer);
}
}
@@ -710,6 +717,8 @@ unsafe fn init(&mut self) {
"hpet",
HPET_REG_SPACE_LEN,
);
+
+ Self::init_timers(unsafe { &mut *((self as *mut Self).cast::<MaybeUninit<Self>>()) });
}
fn post_init(&self) {
@@ -731,7 +740,6 @@ fn realize(&self) -> qemu_api::Result<()> {
self.hpet_id.set(HPETFwConfig::assign_hpet_id()?);
- self.init_timer();
// 64-bit General Capabilities and ID Register; LegacyReplacementRoute.
self.capability.set(
HPET_CAP_REV_ID_VALUE << HPET_CAP_REV_ID_SHIFT |
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 07/24] rust: qom: introduce ParentInit
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (5 preceding siblings ...)
2025-06-20 16:40 ` [PULL 06/24] rust: hpet: fully initialize object during instance_init Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 08/24] rust: qom: make ParentInit lifetime-invariant Paolo Bonzini
` (16 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
This is a smart pointer for MaybeUninit; it can be upcasted to the
already-initialized parent classes, or dereferenced to a MaybeUninit
for the class that is being initialized.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/qemu-api/src/qom.rs | 96 +++++++++++++++++++++++++++++++++++++++-
1 file changed, 95 insertions(+), 1 deletion(-)
diff --git a/rust/qemu-api/src/qom.rs b/rust/qemu-api/src/qom.rs
index 14f98fee60a..ef966e570ca 100644
--- a/rust/qemu-api/src/qom.rs
+++ b/rust/qemu-api/src/qom.rs
@@ -95,7 +95,7 @@
use std::{
ffi::{c_void, CStr},
fmt,
- mem::ManuallyDrop,
+ mem::{ManuallyDrop, MaybeUninit},
ops::{Deref, DerefMut},
ptr::NonNull,
};
@@ -206,6 +206,100 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> Result<(), fmt::Error> {
}
}
+/// This struct knows that the superclasses of the object have already been
+/// initialized.
+pub struct ParentInit<'a, T>(&'a mut MaybeUninit<T>);
+
+impl<'a, T> ParentInit<'a, T> {
+ #[inline]
+ pub fn with(obj: &'a mut MaybeUninit<T>, f: impl FnOnce(ParentInit<'a, T>)) {
+ let parent_init = ParentInit(obj);
+ f(parent_init)
+ }
+}
+
+impl<T: ObjectType> ParentInit<'_, T> {
+ /// Return the receiver as a mutable raw pointer to Object.
+ ///
+ /// # Safety
+ ///
+ /// Fields beyond `Object` could be uninitialized and it's your
+ /// responsibility to avoid that they're used when the pointer is
+ /// dereferenced, either directly or through a cast.
+ pub fn as_object_mut_ptr(&self) -> *mut bindings::Object {
+ self.as_object_ptr().cast_mut()
+ }
+
+ /// Return the receiver as a mutable raw pointer to Object.
+ ///
+ /// # Safety
+ ///
+ /// Fields beyond `Object` could be uninitialized and it's your
+ /// responsibility to avoid that they're used when the pointer is
+ /// dereferenced, either directly or through a cast.
+ pub fn as_object_ptr(&self) -> *const bindings::Object {
+ self.0.as_ptr().cast()
+ }
+}
+
+impl<'a, T: ObjectImpl> ParentInit<'a, T> {
+ /// Convert from a derived type to one of its parent types, which
+ /// have already been initialized.
+ ///
+ /// # Safety
+ ///
+ /// Structurally this is always a safe operation; the [`IsA`] trait
+ /// provides static verification trait that `Self` dereferences to `U` or
+ /// a child of `U`, and only parent types of `T` are allowed.
+ ///
+ /// However, while the fields of the resulting reference are initialized,
+ /// calls might use uninitialized fields of the subclass. It is your
+ /// responsibility to avoid this.
+ pub unsafe fn upcast<U: ObjectType>(&self) -> &'a U
+ where
+ T::ParentType: IsA<U>,
+ {
+ // SAFETY: soundness is declared via IsA<U>, which is an unsafe trait;
+ // the parent has been initialized before `instance_init `is called
+ unsafe { &*(self.0.as_ptr().cast::<U>()) }
+ }
+
+ /// Convert from a derived type to one of its parent types, which
+ /// have already been initialized.
+ ///
+ /// # Safety
+ ///
+ /// Structurally this is always a safe operation; the [`IsA`] trait
+ /// provides static verification trait that `Self` dereferences to `U` or
+ /// a child of `U`, and only parent types of `T` are allowed.
+ ///
+ /// However, while the fields of the resulting reference are initialized,
+ /// calls might use uninitialized fields of the subclass. It is your
+ /// responsibility to avoid this.
+ pub unsafe fn upcast_mut<U: ObjectType>(&mut self) -> &'a mut U
+ where
+ T::ParentType: IsA<U>,
+ {
+ // SAFETY: soundness is declared via IsA<U>, which is an unsafe trait;
+ // the parent has been initialized before `instance_init `is called
+ unsafe { &mut *(self.0.as_mut_ptr().cast::<U>()) }
+ }
+}
+
+impl<T> Deref for ParentInit<'_, T> {
+ type Target = MaybeUninit<T>;
+
+ fn deref(&self) -> &Self::Target {
+ self.0
+ }
+}
+
+impl<T> DerefMut for ParentInit<'_, T> {
+ fn deref_mut(&mut self) -> &mut Self::Target {
+ self.0
+ }
+}
+
unsafe extern "C" fn rust_instance_init<T: ObjectImpl>(obj: *mut bindings::Object) {
let mut state = NonNull::new(obj).unwrap().cast::<T>();
// SAFETY: obj is an instance of T, since rust_instance_init<T>
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 08/24] rust: qom: make ParentInit lifetime-invariant
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (6 preceding siblings ...)
2025-06-20 16:40 ` [PULL 07/24] rust: qom: introduce ParentInit Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 09/24] rust: qom: change instance_init to take a ParentInit<> Paolo Bonzini
` (15 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
This is the trick that allows the parent-field initializer to be used
only for the object that it's meant to be initialized. This way,
the owner of a MemoryRegion must be the object that embeds it.
More information is in the comments; it's best explained with a simplified
example.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/qemu-api/src/qom.rs | 89 ++++++++++++++++++++++++++++++++++++++--
1 file changed, 85 insertions(+), 4 deletions(-)
diff --git a/rust/qemu-api/src/qom.rs b/rust/qemu-api/src/qom.rs
index ef966e570ca..04d102591dc 100644
--- a/rust/qemu-api/src/qom.rs
+++ b/rust/qemu-api/src/qom.rs
@@ -95,6 +95,7 @@
use std::{
ffi::{c_void, CStr},
fmt,
+ marker::PhantomData,
mem::{ManuallyDrop, MaybeUninit},
ops::{Deref, DerefMut},
ptr::NonNull,
@@ -208,12 +209,92 @@ fn fmt(&self, f: &mut fmt::Formatter<'_>) -> Result<(), fmt::Error> {
/// This struct knows that the superclasses of the object have already been
/// initialized.
-pub struct ParentInit<'a, T>(&'a mut MaybeUninit<T>);
+///
+/// The declaration of `ParentInit` is.. *"a kind of magic"*. It uses a
+/// technique that is found in several crates, the main ones probably being
+/// `ghost-cell` (in fact it was introduced by the [`GhostCell` paper](https://plv.mpi-sws.org/rustbelt/ghostcell/))
+/// and `generativity`.
+///
+/// The `PhantomData` makes the `ParentInit` type *invariant* with respect to
+/// the lifetime argument `'init`. This, together with the `for<'...>` in
+/// `[ParentInit::with]`, block any attempt of the compiler to be creative when
+/// operating on types of type `ParentInit` and to extend their lifetimes. In
+/// particular, it ensures that the `ParentInit` cannot be made to outlive the
+/// `rust_instance_init()` function that creates it, and therefore that the
+/// `&'init T` reference is valid.
+///
+/// This implementation of the same concept, without the QOM baggage, can help
+/// understanding the effect:
+///
+/// ```
+/// use std::marker::PhantomData;
+///
+/// #[derive(PartialEq, Eq)]
+/// pub struct Jail<'closure, T: Copy>(&'closure T, PhantomData<fn(&'closure ()) -> &'closure ()>);
+///
+/// impl<'closure, T: Copy> Jail<'closure, T> {
+/// fn get(&self) -> T {
+/// *self.0
+/// }
+///
+/// #[inline]
+/// fn with<U>(v: T, f: impl for<'id> FnOnce(Jail<'id, T>) -> U) -> U {
+/// let parent_init = Jail(&v, PhantomData);
+/// f(parent_init)
+/// }
+/// }
+/// ```
+///
+/// It's impossible to escape the `Jail`; `token1` cannot be moved out of the
+/// closure:
+///
+/// ```ignore
+/// let x = 42;
+/// let escape = Jail::with(&x, |token1| {
+/// println!("{}", token1.get());
+/// // fails to compile...
+/// token1
+/// });
+/// // ... so you cannot do this:
+/// println!("{}", escape.get());
+/// ```
+///
+/// Likewise, in the QOM case the `ParentInit` cannot be moved out of
+/// `instance_init()`. Without this trick it would be possible to stash a
+/// `ParentInit` and use it later to access uninitialized memory.
+///
+/// Here is another example, showing how separately-created "identities" stay
+/// isolated:
+///
+/// ```ignore
+/// impl<'closure, T: Copy> Clone for Jail<'closure, T> {
+/// fn clone(&self) -> Jail<'closure, T> {
+/// Jail(self.0, PhantomData)
+/// }
+/// }
+///
+/// fn main() {
+/// Jail::with(42, |token1| {
+/// // this works and returns true: the clone has the same "identity"
+/// println!("{}", token1 == token1.clone());
+/// Jail::with(42, |token2| {
+/// // here the outer token remains accessible...
+/// println!("{}", token1.get());
+/// // ... but the two are separate: this fails to compile:
+/// println!("{}", token1 == token2);
+/// });
+/// });
+/// }
+/// ```
+pub struct ParentInit<'init, T>(
+ &'init mut MaybeUninit<T>,
+ PhantomData<fn(&'init ()) -> &'init ()>,
+);
-impl<'a, T> ParentInit<'a, T> {
+impl<'init, T> ParentInit<'init, T> {
#[inline]
- pub fn with(obj: &'a mut MaybeUninit<T>, f: impl FnOnce(ParentInit<'a, T>)) {
- let parent_init = ParentInit(obj);
+ pub fn with(obj: &'init mut MaybeUninit<T>, f: impl for<'id> FnOnce(ParentInit<'id, T>)) {
+ let parent_init = ParentInit(obj, PhantomData);
f(parent_init)
}
}
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 09/24] rust: qom: change instance_init to take a ParentInit<>
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (7 preceding siblings ...)
2025-06-20 16:40 ` [PULL 08/24] rust: qom: make ParentInit lifetime-invariant Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 10/24] rust: prepare variable definitions for multiple bindgen invocations Paolo Bonzini
` (14 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
This removes undefined behavior associated to writing to uninitialized
fields, and makes it possible to remove "unsafe" from the instance_init
implementation.
However, the init function itself is still unsafe, because it must promise
(as a sort as MaybeUninit::assume_init) that all fields have been
initialized.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/hw/char/pl011/src/device.rs | 34 ++++++++++------------
rust/hw/timer/hpet/src/device.rs | 16 ++++-------
rust/qemu-api/src/memory.rs | 12 ++++----
rust/qemu-api/src/qdev.rs | 49 +++++++++++++++++++-------------
rust/qemu-api/src/qom.rs | 9 ++++--
5 files changed, 63 insertions(+), 57 deletions(-)
diff --git a/rust/hw/char/pl011/src/device.rs b/rust/hw/char/pl011/src/device.rs
index be8387f6f2d..2d416cd9a3c 100644
--- a/rust/hw/char/pl011/src/device.rs
+++ b/rust/hw/char/pl011/src/device.rs
@@ -2,7 +2,7 @@
// Author(s): Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
// SPDX-License-Identifier: GPL-2.0-or-later
-use std::{ffi::CStr, mem::size_of, ptr::addr_of_mut};
+use std::{ffi::CStr, mem::size_of};
use qemu_api::{
chardev::{CharBackend, Chardev, Event},
@@ -11,9 +11,10 @@
memory::{hwaddr, MemoryRegion, MemoryRegionOps, MemoryRegionOpsBuilder},
prelude::*,
qdev::{Clock, ClockEvent, DeviceImpl, DeviceState, Property, ResetType, ResettablePhasesImpl},
- qom::{ObjectImpl, Owned, ParentField},
+ qom::{ObjectImpl, Owned, ParentField, ParentInit},
static_assert,
sysbus::{SysBusDevice, SysBusDeviceImpl},
+ uninit_field_mut,
vmstate::VMStateDescription,
};
@@ -163,7 +164,7 @@ impl PL011Impl for PL011State {
impl ObjectImpl for PL011State {
type ParentType = SysBusDevice;
- const INSTANCE_INIT: Option<unsafe fn(&mut Self)> = Some(Self::init);
+ const INSTANCE_INIT: Option<unsafe fn(ParentInit<Self>)> = Some(Self::init);
const INSTANCE_POST_INIT: Option<fn(&Self)> = Some(Self::post_init);
const CLASS_INIT: fn(&mut Self::Class) = Self::Class::class_init::<Self>;
}
@@ -488,7 +489,7 @@ impl PL011State {
/// `PL011State` type. It must not be called more than once on the same
/// location/instance. All its fields are expected to hold uninitialized
/// values with the sole exception of `parent_obj`.
- unsafe fn init(&mut self) {
+ unsafe fn init(mut this: ParentInit<Self>) {
static PL011_OPS: MemoryRegionOps<PL011State> = MemoryRegionOpsBuilder::<PL011State>::new()
.read(&PL011State::read)
.write(&PL011State::write)
@@ -496,28 +497,23 @@ unsafe fn init(&mut self) {
.impl_sizes(4, 4)
.build();
- // SAFETY:
- //
- // self and self.iomem are guaranteed to be valid at this point since callers
- // must make sure the `self` reference is valid.
+ // SAFETY: this and this.iomem are guaranteed to be valid at this point
MemoryRegion::init_io(
- unsafe { &mut *addr_of_mut!(self.iomem) },
- addr_of_mut!(*self),
+ &mut uninit_field_mut!(*this, iomem),
&PL011_OPS,
"pl011",
0x1000,
);
- self.regs = Default::default();
+ uninit_field_mut!(*this, regs).write(Default::default());
- // SAFETY:
- //
- // self.clock is not initialized at this point; but since `Owned<_>` is
- // not Drop, we can overwrite the undefined value without side effects;
- // it's not sound but, because for all PL011State instances are created
- // by QOM code which calls this function to initialize the fields, at
- // leastno code is able to access an invalid self.clock value.
- self.clock = self.init_clock_in("clk", &Self::clock_update, ClockEvent::ClockUpdate);
+ let clock = DeviceState::init_clock_in(
+ &mut this,
+ "clk",
+ &Self::clock_update,
+ ClockEvent::ClockUpdate,
+ );
+ uninit_field_mut!(*this, clock).write(clock);
}
const fn clock_update(&self, _event: ClockEvent) {
diff --git a/rust/hw/timer/hpet/src/device.rs b/rust/hw/timer/hpet/src/device.rs
index 340ca1d355d..a281927781e 100644
--- a/rust/hw/timer/hpet/src/device.rs
+++ b/rust/hw/timer/hpet/src/device.rs
@@ -21,8 +21,8 @@
hwaddr, MemoryRegion, MemoryRegionOps, MemoryRegionOpsBuilder, MEMTXATTRS_UNSPECIFIED,
},
prelude::*,
- qdev::{DeviceImpl, DeviceMethods, DeviceState, Property, ResetType, ResettablePhasesImpl},
- qom::{ObjectImpl, ObjectType, ParentField},
+ qdev::{DeviceImpl, DeviceState, Property, ResetType, ResettablePhasesImpl},
+ qom::{ObjectImpl, ObjectType, ParentField, ParentInit},
qom_isa,
sysbus::{SysBusDevice, SysBusDeviceImpl},
timer::{Timer, CLOCK_VIRTUAL, NANOSECONDS_PER_SECOND},
@@ -697,7 +697,7 @@ fn set_counter_reg(&self, shift: u32, len: u32, val: u64) {
.set(self.counter.get().deposit(shift, len, val));
}
- unsafe fn init(&mut self) {
+ unsafe fn init(mut this: ParentInit<Self>) {
static HPET_RAM_OPS: MemoryRegionOps<HPETState> =
MemoryRegionOpsBuilder::<HPETState>::new()
.read(&HPETState::read)
@@ -707,18 +707,14 @@ unsafe fn init(&mut self) {
.impl_sizes(4, 8)
.build();
- // SAFETY:
- // self and self.iomem are guaranteed to be valid at this point since callers
- // must make sure the `self` reference is valid.
MemoryRegion::init_io(
- unsafe { &mut *addr_of_mut!(self.iomem) },
- addr_of_mut!(*self),
+ &mut uninit_field_mut!(*this, iomem),
&HPET_RAM_OPS,
"hpet",
HPET_REG_SPACE_LEN,
);
- Self::init_timers(unsafe { &mut *((self as *mut Self).cast::<MaybeUninit<Self>>()) });
+ Self::init_timers(&mut this);
}
fn post_init(&self) {
@@ -900,7 +896,7 @@ unsafe impl ObjectType for HPETState {
impl ObjectImpl for HPETState {
type ParentType = SysBusDevice;
- const INSTANCE_INIT: Option<unsafe fn(&mut Self)> = Some(Self::init);
+ const INSTANCE_INIT: Option<unsafe fn(ParentInit<Self>)> = Some(Self::init);
const INSTANCE_POST_INIT: Option<fn(&Self)> = Some(Self::post_init);
const CLASS_INIT: fn(&mut Self::Class) = Self::Class::class_init::<Self>;
}
diff --git a/rust/qemu-api/src/memory.rs b/rust/qemu-api/src/memory.rs
index 9ef2694bd62..e40fad6cf19 100644
--- a/rust/qemu-api/src/memory.rs
+++ b/rust/qemu-api/src/memory.rs
@@ -16,6 +16,7 @@
callbacks::FnCall,
cell::Opaque,
prelude::*,
+ uninit::MaybeUninitField,
zeroable::Zeroable,
};
@@ -147,7 +148,7 @@ impl MemoryRegion {
#[inline(always)]
unsafe fn do_init_io(
slot: *mut bindings::MemoryRegion,
- owner: *mut Object,
+ owner: *mut bindings::Object,
ops: &'static bindings::MemoryRegionOps,
name: &'static str,
size: u64,
@@ -156,7 +157,7 @@ unsafe fn do_init_io(
let cstr = CString::new(name).unwrap();
memory_region_init_io(
slot,
- owner.cast::<bindings::Object>(),
+ owner,
ops,
owner.cast::<c_void>(),
cstr.as_ptr(),
@@ -166,16 +167,15 @@ unsafe fn do_init_io(
}
pub fn init_io<T: IsA<Object>>(
- &mut self,
- owner: *mut T,
+ this: &mut MaybeUninitField<'_, T, Self>,
ops: &'static MemoryRegionOps<T>,
name: &'static str,
size: u64,
) {
unsafe {
Self::do_init_io(
- self.0.as_mut_ptr(),
- owner.cast::<Object>(),
+ this.as_mut_ptr().cast(),
+ MaybeUninitField::parent_mut(this).cast(),
&ops.0,
name,
size,
diff --git a/rust/qemu-api/src/qdev.rs b/rust/qemu-api/src/qdev.rs
index 0610959f467..36f02fb57db 100644
--- a/rust/qemu-api/src/qdev.rs
+++ b/rust/qemu-api/src/qdev.rs
@@ -19,7 +19,7 @@
error::{Error, Result},
irq::InterruptSource,
prelude::*,
- qom::{ObjectClass, ObjectImpl, Owned},
+ qom::{ObjectClass, ObjectImpl, Owned, ParentInit},
vmstate::VMStateDescription,
};
@@ -247,15 +247,9 @@ unsafe impl ObjectType for DeviceState {
}
qom_isa!(DeviceState: Object);
-/// Trait for methods exposed by the [`DeviceState`] class. The methods can be
-/// called on all objects that have the trait `IsA<DeviceState>`.
-///
-/// The trait should only be used through the blanket implementation,
-/// which guarantees safety via `IsA`.
-pub trait DeviceMethods: ObjectDeref
-where
- Self::Target: IsA<DeviceState>,
-{
+/// Initialization methods take a [`ParentInit`] and can be called as
+/// associated functions.
+impl DeviceState {
/// Add an input clock named `name`. Invoke the callback with
/// `self` as the first parameter for the events that are requested.
///
@@ -266,12 +260,15 @@ pub trait DeviceMethods: ObjectDeref
/// which Rust code has a reference to a child object) it would be
/// possible for this function to return a `&Clock` too.
#[inline]
- fn init_clock_in<F: for<'a> FnCall<(&'a Self::Target, ClockEvent)>>(
- &self,
+ pub fn init_clock_in<T: DeviceImpl, F: for<'a> FnCall<(&'a T, ClockEvent)>>(
+ this: &mut ParentInit<T>,
name: &str,
_cb: &F,
events: ClockEvent,
- ) -> Owned<Clock> {
+ ) -> Owned<Clock>
+ where
+ T::ParentType: IsA<DeviceState>,
+ {
fn do_init_clock_in(
dev: &DeviceState,
name: &str,
@@ -287,10 +284,10 @@ fn do_init_clock_in(
unsafe {
let cstr = CString::new(name).unwrap();
let clk = bindings::qdev_init_clock_in(
- dev.as_mut_ptr(),
+ dev.0.as_mut_ptr(),
cstr.as_ptr(),
cb,
- dev.as_void_ptr(),
+ dev.0.as_void_ptr(),
events.0,
);
@@ -307,12 +304,12 @@ fn do_init_clock_in(
// SAFETY: the opaque is "this", which is indeed a pointer to T
F::call((unsafe { &*(opaque.cast::<T>()) }, event))
}
- Some(rust_clock_cb::<Self::Target, F>)
+ Some(rust_clock_cb::<T, F>)
} else {
None
};
- do_init_clock_in(self.upcast(), name, cb, events)
+ do_init_clock_in(unsafe { this.upcast_mut() }, name, cb, events)
}
/// Add an output clock named `name`.
@@ -324,16 +321,30 @@ fn do_init_clock_in(
/// which Rust code has a reference to a child object) it would be
/// possible for this function to return a `&Clock` too.
#[inline]
- fn init_clock_out(&self, name: &str) -> Owned<Clock> {
+ pub fn init_clock_out<T: DeviceImpl>(this: &mut ParentInit<T>, name: &str) -> Owned<Clock>
+ where
+ T::ParentType: IsA<DeviceState>,
+ {
unsafe {
let cstr = CString::new(name).unwrap();
- let clk = bindings::qdev_init_clock_out(self.upcast().as_mut_ptr(), cstr.as_ptr());
+ let dev: &mut DeviceState = this.upcast_mut();
+ let clk = bindings::qdev_init_clock_out(dev.0.as_mut_ptr(), cstr.as_ptr());
let clk: &Clock = Clock::from_raw(clk);
Owned::from(clk)
}
}
+}
+/// Trait for methods exposed by the [`DeviceState`] class. The methods can be
+/// called on all objects that have the trait `IsA<DeviceState>`.
+///
+/// The trait should only be used through the blanket implementation,
+/// which guarantees safety via `IsA`.
+pub trait DeviceMethods: ObjectDeref
+where
+ Self::Target: IsA<DeviceState>,
+{
fn prop_set_chr(&self, propname: &str, chr: &Owned<Chardev>) {
assert!(bql_locked());
let c_propname = CString::new(propname).unwrap();
diff --git a/rust/qemu-api/src/qom.rs b/rust/qemu-api/src/qom.rs
index 04d102591dc..e20ee014cb1 100644
--- a/rust/qemu-api/src/qom.rs
+++ b/rust/qemu-api/src/qom.rs
@@ -382,12 +382,15 @@ fn deref_mut(&mut self) -> &mut Self::Target {
}
unsafe extern "C" fn rust_instance_init<T: ObjectImpl>(obj: *mut bindings::Object) {
- let mut state = NonNull::new(obj).unwrap().cast::<T>();
+ let mut state = NonNull::new(obj).unwrap().cast::<MaybeUninit<T>>();
+
// SAFETY: obj is an instance of T, since rust_instance_init<T>
// is called from QOM core as the instance_init function
// for class T
unsafe {
- T::INSTANCE_INIT.unwrap()(state.as_mut());
+ ParentInit::with(state.as_mut(), |parent_init| {
+ T::INSTANCE_INIT.unwrap()(parent_init);
+ });
}
}
@@ -654,7 +657,7 @@ pub trait ObjectImpl: ObjectType + IsA<Object> {
///
/// FIXME: The argument is not really a valid reference. `&mut
/// MaybeUninit<Self>` would be a better description.
- const INSTANCE_INIT: Option<unsafe fn(&mut Self)> = None;
+ const INSTANCE_INIT: Option<unsafe fn(ParentInit<Self>)> = None;
/// Function that is called to finish initialization of an object, once
/// `INSTANCE_INIT` functions have been called.
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 10/24] rust: prepare variable definitions for multiple bindgen invocations
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (8 preceding siblings ...)
2025-06-20 16:40 ` [PULL 09/24] rust: qom: change instance_init to take a ParentInit<> Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 11/24] rust: move rust.bindgen to qemu-api crate Paolo Bonzini
` (13 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
When splitting the QEMU Rust bindings into multiple crates, the
bindgen-generated structs also have to be split so that it's
possible to add "impl" blocks (e.g. for Sync/Send or Default,
or even for utility methods in cases such as VMStateFlags).
Tweak various variable definitions in meson.build, to avoid naming
conflicts once there will be multiple bindgen invocations.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
meson.build | 21 ++++++++++++---------
rust/meson.build | 2 +-
rust/qemu-api/meson.build | 2 +-
3 files changed, 14 insertions(+), 11 deletions(-)
diff --git a/meson.build b/meson.build
index ed60be2a2d0..19ffa9cb34c 100644
--- a/meson.build
+++ b/meson.build
@@ -4200,10 +4200,11 @@ foreach target_base_arch, config_base_arch : config_base_arch_mak
endforeach
if have_rust
+ bindings_incdir = include_directories('.', 'include')
# We would like to use --generate-cstr, but it is only available
# starting with bindgen 0.66.0. The oldest supported versions
# is 0.60.x (Debian 12 has 0.60.1) which introduces --allowlist-file.
- bindgen_args = [
+ bindgen_args_common = [
'--disable-header-comment',
'--raw-line', '// @generated',
'--ctypes-prefix', 'std::os::raw',
@@ -4219,20 +4220,22 @@ if have_rust
]
if not rustfmt.found()
if bindgen.version().version_compare('<0.65.0')
- bindgen_args += ['--no-rustfmt-bindings']
+ bindgen_args_common += ['--no-rustfmt-bindings']
else
- bindgen_args += ['--formatter', 'none']
+ bindgen_args_common += ['--formatter', 'none']
endif
endif
if bindgen.version().version_compare('>=0.66.0')
- bindgen_args += ['--rust-target', '1.59']
+ bindgen_args_common += ['--rust-target', '1.59']
endif
if bindgen.version().version_compare('<0.61.0')
# default in 0.61+
- bindgen_args += ['--size_t-is-usize']
+ bindgen_args_common += ['--size_t-is-usize']
else
- bindgen_args += ['--merge-extern-blocks']
+ bindgen_args_common += ['--merge-extern-blocks']
endif
+
+ bindgen_args = []
c_enums = [
'DeviceCategory',
'GpioPolarity',
@@ -4264,13 +4267,13 @@ if have_rust
# this case you must pass the path to `clang` and `libclang` to your build
# command invocation using the environment variables CLANG_PATH and
# LIBCLANG_PATH
- bindings_rs = rust.bindgen(
+ _qemu_api_bindings_inc_rs = rust.bindgen(
input: 'rust/wrapper.h',
dependencies: common_ss.all_dependencies(),
output: 'bindings.inc.rs',
- include_directories: include_directories('.', 'include'),
+ include_directories: bindings_incdir,
bindgen_version: ['>=0.60.0'],
- args: bindgen_args,
+ args: bindgen_args_common + bindgen_args,
)
subdir('rust')
endif
diff --git a/rust/meson.build b/rust/meson.build
index 99ae7956cd0..e9f0879e296 100644
--- a/rust/meson.build
+++ b/rust/meson.build
@@ -33,5 +33,5 @@ if cargo.found()
command: [config_host['MESON'], 'devenv',
'--workdir', '@CURRENT_SOURCE_DIR@',
cargo, 'fmt'],
- depends: bindings_rs)
+ depends: _qemu_api_bindings_inc_rs)
endif
diff --git a/rust/qemu-api/meson.build b/rust/qemu-api/meson.build
index 33653b4a28e..64c04dfd74b 100644
--- a/rust/qemu-api/meson.build
+++ b/rust/qemu-api/meson.build
@@ -32,7 +32,7 @@ _qemu_api_rs = static_library(
'src/vmstate.rs',
'src/zeroable.rs',
],
- {'.' : bindings_rs},
+ {'.' : _qemu_api_bindings_inc_rs},
),
override_options: ['rust_std=2021', 'build.rust_std=2021'],
rust_abi: 'rust',
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 11/24] rust: move rust.bindgen to qemu-api crate
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (9 preceding siblings ...)
2025-06-20 16:40 ` [PULL 10/24] rust: prepare variable definitions for multiple bindgen invocations Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 12/24] rust/qemu-api: Add initial logging support based on C API Paolo Bonzini
` (12 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu
Once qemu-api is split in multiple crates, each of them will have
its own invocation of bindgen. There cannot be only one, because
there are occasional "impl" blocks for the bindgen-generated
structs (e.g. VMStateFlags or QOM classes) that have to
reside in the same crate as the bindgen-generated code.
For now, prepare for this new organization by invoking bindgen
within the qemu-api crate's build definitions; it's also a
much better place to list enums that need specific treatment
from bindgen.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
meson.build | 41 -----------------------------------
rust/{ => qemu-api}/wrapper.h | 0
rust/meson.build | 4 +++-
rust/qemu-api/build.rs | 2 +-
rust/qemu-api/meson.build | 41 +++++++++++++++++++++++++++++++++++
5 files changed, 45 insertions(+), 43 deletions(-)
rename rust/{ => qemu-api}/wrapper.h (100%)
diff --git a/meson.build b/meson.build
index 19ffa9cb34c..4676908dbb2 100644
--- a/meson.build
+++ b/meson.build
@@ -4234,47 +4234,6 @@ if have_rust
else
bindgen_args_common += ['--merge-extern-blocks']
endif
-
- bindgen_args = []
- c_enums = [
- 'DeviceCategory',
- 'GpioPolarity',
- 'MachineInitPhase',
- 'MemoryDeviceInfoKind',
- 'MigrationPolicy',
- 'MigrationPriority',
- 'QEMUChrEvent',
- 'QEMUClockType',
- 'ResetType',
- 'device_endian',
- 'module_init_type',
- ]
- foreach enum : c_enums
- bindgen_args += ['--rustified-enum', enum]
- endforeach
- c_bitfields = [
- 'ClockEvent',
- 'VMStateFlags',
- ]
- foreach enum : c_bitfields
- bindgen_args += ['--bitfield-enum', enum]
- endforeach
-
- # TODO: Remove this comment when the clang/libclang mismatch issue is solved.
- #
- # Rust bindings generation with `bindgen` might fail in some cases where the
- # detected `libclang` does not match the expected `clang` version/target. In
- # this case you must pass the path to `clang` and `libclang` to your build
- # command invocation using the environment variables CLANG_PATH and
- # LIBCLANG_PATH
- _qemu_api_bindings_inc_rs = rust.bindgen(
- input: 'rust/wrapper.h',
- dependencies: common_ss.all_dependencies(),
- output: 'bindings.inc.rs',
- include_directories: bindings_incdir,
- bindgen_version: ['>=0.60.0'],
- args: bindgen_args_common + bindgen_args,
- )
subdir('rust')
endif
diff --git a/rust/wrapper.h b/rust/qemu-api/wrapper.h
similarity index 100%
rename from rust/wrapper.h
rename to rust/qemu-api/wrapper.h
diff --git a/rust/meson.build b/rust/meson.build
index e9f0879e296..331f11b7e72 100644
--- a/rust/meson.build
+++ b/rust/meson.build
@@ -20,6 +20,8 @@ proc_macro2_rs_native = dependency('proc-macro2-1-rs', native: true)
qemuutil_rs = qemuutil.partial_dependency(link_args: true, links: true)
+genrs = []
+
subdir('qemu-api-macros')
subdir('bits')
subdir('qemu-api')
@@ -33,5 +35,5 @@ if cargo.found()
command: [config_host['MESON'], 'devenv',
'--workdir', '@CURRENT_SOURCE_DIR@',
cargo, 'fmt'],
- depends: _qemu_api_bindings_inc_rs)
+ depends: genrs)
endif
diff --git a/rust/qemu-api/build.rs b/rust/qemu-api/build.rs
index 1e720641d2b..7849486c1ba 100644
--- a/rust/qemu-api/build.rs
+++ b/rust/qemu-api/build.rs
@@ -14,7 +14,7 @@ fn main() -> Result<()> {
let path = env::var("MESON_BUILD_ROOT")
.unwrap_or_else(|_| format!("{}/src", env!("CARGO_MANIFEST_DIR")));
- let file = format!("{path}/bindings.inc.rs");
+ let file = format!("{path}/rust/qemu-api/bindings.inc.rs");
let file = Path::new(&file);
if !Path::new(&file).exists() {
panic!(concat!(
diff --git a/rust/qemu-api/meson.build b/rust/qemu-api/meson.build
index 64c04dfd74b..5b8c7e5e8d5 100644
--- a/rust/qemu-api/meson.build
+++ b/rust/qemu-api/meson.build
@@ -7,6 +7,47 @@ if get_option('debug_mutex')
_qemu_api_cfg += ['--cfg', 'feature="debug_cell"']
endif
+c_enums = [
+ 'DeviceCategory',
+ 'GpioPolarity',
+ 'MachineInitPhase',
+ 'MemoryDeviceInfoKind',
+ 'MigrationPolicy',
+ 'MigrationPriority',
+ 'QEMUChrEvent',
+ 'QEMUClockType',
+ 'ResetType',
+ 'device_endian',
+ 'module_init_type',
+]
+_qemu_api_bindgen_args = []
+foreach enum : c_enums
+ _qemu_api_bindgen_args += ['--rustified-enum', enum]
+endforeach
+c_bitfields = [
+ 'ClockEvent',
+ 'VMStateFlags',
+]
+foreach enum : c_bitfields
+ _qemu_api_bindgen_args += ['--bitfield-enum', enum]
+endforeach
+
+# TODO: Remove this comment when the clang/libclang mismatch issue is solved.
+#
+# Rust bindings generation with `bindgen` might fail in some cases where the
+# detected `libclang` does not match the expected `clang` version/target. In
+# this case you must pass the path to `clang` and `libclang` to your build
+# command invocation using the environment variables CLANG_PATH and
+# LIBCLANG_PATH
+_qemu_api_bindings_inc_rs = rust.bindgen(
+ input: 'wrapper.h',
+ dependencies: common_ss.all_dependencies(),
+ output: 'bindings.inc.rs',
+ include_directories: bindings_incdir,
+ bindgen_version: ['>=0.60.0'],
+ args: bindgen_args_common + _qemu_api_bindgen_args,
+ )
+
_qemu_api_rs = static_library(
'qemu_api',
structured_sources(
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 12/24] rust/qemu-api: Add initial logging support based on C API
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (10 preceding siblings ...)
2025-06-20 16:40 ` [PULL 11/24] rust: move rust.bindgen to qemu-api crate Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 13/24] rust: pl011: Implement logging Paolo Bonzini
` (11 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Bernhard Beschow
From: Bernhard Beschow <shentey@gmail.com>
A log_mask_ln!() macro is provided which expects similar arguments as the
C version. However, the formatting works as one would expect from Rust.
To maximize code reuse the macro is just a thin wrapper around
qemu_log(). Also, just the bare minimum of logging masks is provided
which should suffice for the current use case of Rust in QEMU.
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250615112037.11992-2-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
docs/devel/rust.rst | 1 +
rust/qemu-api/wrapper.h | 2 +
rust/qemu-api/meson.build | 1 +
rust/qemu-api/src/lib.rs | 1 +
rust/qemu-api/src/log.rs | 73 ++++++++++++++++++++++++++++++++++++
rust/qemu-api/src/prelude.rs | 2 +
6 files changed, 80 insertions(+)
create mode 100644 rust/qemu-api/src/log.rs
diff --git a/docs/devel/rust.rst b/docs/devel/rust.rst
index 47e9677fcb0..dc8c44109e1 100644
--- a/docs/devel/rust.rst
+++ b/docs/devel/rust.rst
@@ -162,6 +162,7 @@ module status
``errno`` complete
``error`` stable
``irq`` complete
+``log`` proof of concept
``memory`` stable
``module`` complete
``qdev`` stable
diff --git a/rust/qemu-api/wrapper.h b/rust/qemu-api/wrapper.h
index 6060d3ba1ab..15a1b19847f 100644
--- a/rust/qemu-api/wrapper.h
+++ b/rust/qemu-api/wrapper.h
@@ -48,6 +48,8 @@ typedef enum memory_order {
#endif /* __CLANG_STDATOMIC_H */
#include "qemu/osdep.h"
+#include "qemu/log.h"
+#include "qemu/log-for-trace.h"
#include "qemu/module.h"
#include "qemu-io.h"
#include "system/system.h"
diff --git a/rust/qemu-api/meson.build b/rust/qemu-api/meson.build
index 5b8c7e5e8d5..a090297c458 100644
--- a/rust/qemu-api/meson.build
+++ b/rust/qemu-api/meson.build
@@ -62,6 +62,7 @@ _qemu_api_rs = static_library(
'src/errno.rs',
'src/error.rs',
'src/irq.rs',
+ 'src/log.rs',
'src/memory.rs',
'src/module.rs',
'src/prelude.rs',
diff --git a/rust/qemu-api/src/lib.rs b/rust/qemu-api/src/lib.rs
index c78198f0f41..86dcd8ef17a 100644
--- a/rust/qemu-api/src/lib.rs
+++ b/rust/qemu-api/src/lib.rs
@@ -21,6 +21,7 @@
pub mod errno;
pub mod error;
pub mod irq;
+pub mod log;
pub mod memory;
pub mod module;
pub mod qdev;
diff --git a/rust/qemu-api/src/log.rs b/rust/qemu-api/src/log.rs
new file mode 100644
index 00000000000..d6c3d6c1b63
--- /dev/null
+++ b/rust/qemu-api/src/log.rs
@@ -0,0 +1,73 @@
+// Copyright 2025 Bernhard Beschow <shentey@gmail.com>
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+//! Bindings for QEMU's logging infrastructure
+
+#[repr(u32)]
+/// Represents specific error categories within QEMU's logging system.
+///
+/// The `Log` enum provides a Rust abstraction for logging errors, corresponding
+/// to a subset of the error categories defined in the C implementation.
+pub enum Log {
+ /// Log invalid access caused by the guest.
+ /// Corresponds to `LOG_GUEST_ERROR` in the C implementation.
+ GuestError = crate::bindings::LOG_GUEST_ERROR,
+
+ /// Log guest access of unimplemented functionality.
+ /// Corresponds to `LOG_UNIMP` in the C implementation.
+ Unimp = crate::bindings::LOG_UNIMP,
+}
+
+/// A macro to log messages conditionally based on a provided mask.
+///
+/// The `log_mask_ln` macro checks whether the given mask matches the current
+/// log level and, if so, formats and logs the message. It is the Rust
+/// counterpart of the `qemu_log_mask()` macro in the C implementation.
+///
+/// # Parameters
+///
+/// - `$mask`: A log level mask. This should be a variant of the `Log` enum.
+/// - `$fmt`: A format string following the syntax and rules of the `format!`
+/// macro. It specifies the structure of the log message.
+/// - `$args`: Optional arguments to be interpolated into the format string.
+///
+/// # Example
+///
+/// ```
+/// use qemu_api::{log::Log, log_mask_ln};
+///
+/// let error_address = 0xbad;
+/// log_mask_ln!(Log::GuestError, "Address 0x{error_address:x} out of range");
+/// ```
+///
+/// It is also possible to use printf-style formatting, as well as having a
+/// trailing `,`:
+///
+/// ```
+/// use qemu_api::{log::Log, log_mask_ln};
+///
+/// let error_address = 0xbad;
+/// log_mask_ln!(
+/// Log::GuestError,
+/// "Address 0x{:x} out of range",
+/// error_address,
+/// );
+/// ```
+#[macro_export]
+macro_rules! log_mask_ln {
+ ($mask:expr, $fmt:tt $($args:tt)*) => {{
+ // Type assertion to enforce type `Log` for $mask
+ let _: Log = $mask;
+
+ if unsafe {
+ (::qemu_api::bindings::qemu_loglevel & ($mask as std::os::raw::c_int)) != 0
+ } {
+ let formatted_string = format!("{}\n", format_args!($fmt $($args)*));
+ let c_string = std::ffi::CString::new(formatted_string).unwrap();
+
+ unsafe {
+ ::qemu_api::bindings::qemu_log(c_string.as_ptr());
+ }
+ }
+ }};
+}
diff --git a/rust/qemu-api/src/prelude.rs b/rust/qemu-api/src/prelude.rs
index 43bfcd5fcab..8f9e23ee2c5 100644
--- a/rust/qemu-api/src/prelude.rs
+++ b/rust/qemu-api/src/prelude.rs
@@ -11,6 +11,8 @@
pub use crate::errno;
+pub use crate::log_mask_ln;
+
pub use crate::qdev::DeviceMethods;
pub use crate::qom::InterfaceType;
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 13/24] rust: pl011: Implement logging
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (11 preceding siblings ...)
2025-06-20 16:40 ` [PULL 12/24] rust/qemu-api: Add initial logging support based on C API Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 14/24] rust: pl011: Add missing logging to match C version Paolo Bonzini
` (10 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Bernhard Beschow
From: Bernhard Beschow <shentey@gmail.com>
Now that there is logging support in Rust for QEMU, use it in the pl011
device.
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250615112037.11992-3-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/hw/char/pl011/src/device.rs | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/rust/hw/char/pl011/src/device.rs b/rust/hw/char/pl011/src/device.rs
index 2d416cd9a3c..92dc295540e 100644
--- a/rust/hw/char/pl011/src/device.rs
+++ b/rust/hw/char/pl011/src/device.rs
@@ -8,6 +8,8 @@
chardev::{CharBackend, Chardev, Event},
impl_vmstate_forward,
irq::{IRQState, InterruptSource},
+ log::Log,
+ log_mask_ln,
memory::{hwaddr, MemoryRegion, MemoryRegionOps, MemoryRegionOpsBuilder},
prelude::*,
qdev::{Clock, ClockEvent, DeviceImpl, DeviceState, Property, ResetType, ResettablePhasesImpl},
@@ -276,8 +278,7 @@ pub(self) fn write(
DMACR => {
self.dmacr = value;
if value & 3 > 0 {
- // qemu_log_mask(LOG_UNIMP, "pl011: DMA not implemented\n");
- eprintln!("pl011: DMA not implemented");
+ log_mask_ln!(Log::Unimp, "pl011: DMA not implemented");
}
}
}
@@ -534,7 +535,7 @@ fn read(&self, offset: hwaddr, _size: u32) -> u64 {
u64::from(device_id[(offset - 0xfe0) >> 2])
}
Err(_) => {
- // qemu_log_mask(LOG_GUEST_ERROR, "pl011_read: Bad offset 0x%x\n", (int)offset);
+ log_mask_ln!(Log::GuestError, "PL011State::read: Bad offset {offset}");
0
}
Ok(field) => {
@@ -566,7 +567,10 @@ fn write(&self, offset: hwaddr, value: u64, _size: u32) {
.borrow_mut()
.write(field, value as u32, &self.char_backend);
} else {
- eprintln!("write bad offset {offset} value {value}");
+ log_mask_ln!(
+ Log::GuestError,
+ "PL011State::write: Bad offset {offset} value {value}"
+ );
}
if update_irq {
self.update();
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 14/24] rust: pl011: Add missing logging to match C version
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (12 preceding siblings ...)
2025-06-20 16:40 ` [PULL 13/24] rust: pl011: Implement logging Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 15/24] rust: hpet: fix new warning Paolo Bonzini
` (9 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Bernhard Beschow
From: Bernhard Beschow <shentey@gmail.com>
Co-developed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Link: https://lore.kernel.org/r/20250615112037.11992-4-shentey@gmail.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/hw/char/pl011/src/device.rs | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/rust/hw/char/pl011/src/device.rs b/rust/hw/char/pl011/src/device.rs
index 92dc295540e..5b53f2649f1 100644
--- a/rust/hw/char/pl011/src/device.rs
+++ b/rust/hw/char/pl011/src/device.rs
@@ -305,6 +305,12 @@ fn read_data_register(&mut self, update: &mut bool) -> u32 {
}
fn write_data_register(&mut self, value: u32) -> bool {
+ if !self.control.enable_uart() {
+ log_mask_ln!(Log::GuestError, "PL011 data written to disabled UART");
+ }
+ if !self.control.enable_transmit() {
+ log_mask_ln!(Log::GuestError, "PL011 data written to disabled TX UART");
+ }
// interrupts always checked
let _ = self.loopback_tx(value.into());
self.int_level |= Interrupt::TX;
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 15/24] rust: hpet: fix new warning
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (13 preceding siblings ...)
2025-06-20 16:40 ` [PULL 14/24] rust: pl011: Add missing logging to match C version Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 16/24] i386/cpu: Move adjustment of CPUID_EXT_PDCM before feature_dependencies[] check Paolo Bonzini
` (8 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Zhao Liu, Stefan Hajnoczi
Nightly rustc complains that HPETAddrDecode has a lifetime but it is not
clearly noted that it comes from &self. Apply the compiler's suggestion
to shut it up.
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
rust/hw/timer/hpet/src/device.rs | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/rust/hw/timer/hpet/src/device.rs b/rust/hw/timer/hpet/src/device.rs
index a281927781e..acf7251029e 100644
--- a/rust/hw/timer/hpet/src/device.rs
+++ b/rust/hw/timer/hpet/src/device.rs
@@ -771,7 +771,7 @@ fn reset_hold(&self, _type: ResetType) {
self.rtc_irq_level.set(0);
}
- fn decode(&self, mut addr: hwaddr, size: u32) -> HPETAddrDecode {
+ fn decode(&self, mut addr: hwaddr, size: u32) -> HPETAddrDecode<'_> {
let shift = ((addr & 4) * 8) as u32;
let len = std::cmp::min(size * 8, 64 - shift);
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 16/24] i386/cpu: Move adjustment of CPUID_EXT_PDCM before feature_dependencies[] check
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (14 preceding siblings ...)
2025-06-20 16:40 ` [PULL 15/24] rust: hpet: fix new warning Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 17/24] i386/cpu: Warn about why CPUID_EXT_PDCM is not available Paolo Bonzini
` (7 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li, Zhao Liu
From: Xiaoyao Li <xiaoyao.li@intel.com>
There is one entry relates to CPUID_EXT_PDCM in feature_dependencies[].
So it needs to get correct value of CPUID_EXT_PDCM before using
feature_dependencies[] to apply dependencies.
Besides, it also ensures CPUID_EXT_PDCM value is tracked in
env->features[FEAT_1_ECX].
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Link: https://lore.kernel.org/r/20250304052450.465445-2-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/cpu.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 40aefb38f6d..29bce67c3a1 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -7349,9 +7349,6 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
if (threads_per_pkg > 1) {
*ebx |= threads_per_pkg << 16;
}
- if (!cpu->enable_pmu) {
- *ecx &= ~CPUID_EXT_PDCM;
- }
break;
case 2:
/* cache info: needed for Pentium Pro compatibility */
@@ -8341,6 +8338,10 @@ void x86_cpu_expand_features(X86CPU *cpu, Error **errp)
}
}
+ if (!cpu->enable_pmu) {
+ env->features[FEAT_1_ECX] &= ~CPUID_EXT_PDCM;
+ }
+
for (i = 0; i < ARRAY_SIZE(feature_dependencies); i++) {
FeatureDep *d = &feature_dependencies[i];
if (!(env->features[d->from.index] & d->from.mask)) {
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 17/24] i386/cpu: Warn about why CPUID_EXT_PDCM is not available
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (15 preceding siblings ...)
2025-06-20 16:40 ` [PULL 16/24] i386/cpu: Move adjustment of CPUID_EXT_PDCM before feature_dependencies[] check Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 18/24] i386/tdx: Error and exit when named cpu model is requested Paolo Bonzini
` (6 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li, Zhao Liu
From: Xiaoyao Li <xiaoyao.li@intel.com>
When user requests PDCM explicitly via "+pdcm" without PMU enabled, emit
a warning to inform the user.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Link: https://lore.kernel.org/r/20250304052450.465445-3-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/cpu.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 29bce67c3a1..0d35e95430f 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -8339,6 +8339,9 @@ void x86_cpu_expand_features(X86CPU *cpu, Error **errp)
}
if (!cpu->enable_pmu) {
+ mark_unavailable_features(cpu, FEAT_1_ECX,
+ env->user_features[FEAT_1_ECX] & CPUID_EXT_PDCM,
+ "This feature is not available due to PMU being disabled");
env->features[FEAT_1_ECX] &= ~CPUID_EXT_PDCM;
}
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 18/24] i386/tdx: Error and exit when named cpu model is requested
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (16 preceding siblings ...)
2025-06-20 16:40 ` [PULL 17/24] i386/cpu: Warn about why CPUID_EXT_PDCM is not available Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 19/24] i386/cpu: Rename enable_cpuid_0x1f to force_cpuid_0x1f Paolo Bonzini
` (5 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li
From: Xiaoyao Li <xiaoyao.li@intel.com>
Currently, it gets below error when requesting any named cpu model with
"-cpu" to boot a TDX VM:
qemu-system-x86_64: KVM_TDX_INIT_VM failed: Invalid argument
It misleads people to think it's the bug of KVM or QEMU. It is just that
current QEMU doesn't support named cpu model for TDX.
To support named cpu models for TDX guest, there are opens to be
finalized and needs a mount of additional work.
For now, explicitly check the case when named cpu model is requested.
Error report a hint and exit.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Link: https://lore.kernel.org/r/20250612133801.2238342-1-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/kvm/tdx.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/target/i386/kvm/tdx.c b/target/i386/kvm/tdx.c
index 820ca3614e2..2b52de9d71b 100644
--- a/target/i386/kvm/tdx.c
+++ b/target/i386/kvm/tdx.c
@@ -739,8 +739,14 @@ static int tdx_kvm_type(X86ConfidentialGuest *cg)
static void tdx_cpu_instance_init(X86ConfidentialGuest *cg, CPUState *cpu)
{
+ X86CPUClass *xcc = X86_CPU_GET_CLASS(cpu);
X86CPU *x86cpu = X86_CPU(cpu);
+ if (xcc->model) {
+ error_report("Named cpu model is not supported for TDX yet!");
+ exit(1);
+ }
+
object_property_set_bool(OBJECT(cpu), "pmu", false, &error_abort);
/* invtsc is fixed1 for TD guest */
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 19/24] i386/cpu: Rename enable_cpuid_0x1f to force_cpuid_0x1f
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (17 preceding siblings ...)
2025-06-20 16:40 ` [PULL 18/24] i386/tdx: Error and exit when named cpu model is requested Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 20/24] i386/tdx: Fix the typo of the comment of struct TdxGuest Paolo Bonzini
` (4 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li, Igor Mammedov, Daniel P. Berrangé
From: Xiaoyao Li <xiaoyao.li@intel.com>
The name of "enable_cpuid_0x1f" isn't right to its behavior because the
leaf 0x1f can be enabled even when "enable_cpuid_0x1f" is false.
Rename it to "force_cpuid_0x1f" to better reflect its behavior.
Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-2-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/cpu.h | 4 ++--
target/i386/kvm/tdx.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index 5910dcf74d4..51e10139dfd 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -2269,7 +2269,7 @@ struct ArchCPU {
bool enable_cpuid_0xb;
/* Force to enable cpuid 0x1f */
- bool enable_cpuid_0x1f;
+ bool force_cpuid_0x1f;
/* Enable auto level-increase for all CPUID leaves */
bool full_cpuid_auto_level;
@@ -2539,7 +2539,7 @@ void mark_forced_on_features(X86CPU *cpu, FeatureWord w, uint64_t mask,
static inline bool x86_has_cpuid_0x1f(X86CPU *cpu)
{
- return cpu->enable_cpuid_0x1f ||
+ return cpu->force_cpuid_0x1f ||
x86_has_extended_topo(cpu->env.avail_cpu_topo);
}
diff --git a/target/i386/kvm/tdx.c b/target/i386/kvm/tdx.c
index 2b52de9d71b..acbe7497543 100644
--- a/target/i386/kvm/tdx.c
+++ b/target/i386/kvm/tdx.c
@@ -752,7 +752,7 @@ static void tdx_cpu_instance_init(X86ConfidentialGuest *cg, CPUState *cpu)
/* invtsc is fixed1 for TD guest */
object_property_set_bool(OBJECT(cpu), "invtsc", true, &error_abort);
- x86cpu->enable_cpuid_0x1f = true;
+ x86cpu->force_cpuid_0x1f = true;
}
static uint32_t tdx_adjust_cpuid_features(X86ConfidentialGuest *cg,
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 20/24] i386/tdx: Fix the typo of the comment of struct TdxGuest
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (18 preceding siblings ...)
2025-06-20 16:40 ` [PULL 19/24] i386/cpu: Rename enable_cpuid_0x1f to force_cpuid_0x1f Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 21/24] i386/tdx: Clarify the error message of mrconfigid/mrowner/mrownerconfig Paolo Bonzini
` (3 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li, Daniel P. Berrangé, Igor Mammedov
From: Xiaoyao Li <xiaoyao.li@intel.com>
Change sha348 to sha384.
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-3-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/kvm/tdx.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/target/i386/kvm/tdx.h b/target/i386/kvm/tdx.h
index 04b5afe199f..8dd66e90149 100644
--- a/target/i386/kvm/tdx.h
+++ b/target/i386/kvm/tdx.h
@@ -40,9 +40,9 @@ typedef struct TdxGuest {
bool initialized;
uint64_t attributes; /* TD attributes */
uint64_t xfam;
- char *mrconfigid; /* base64 encoded sha348 digest */
- char *mrowner; /* base64 encoded sha348 digest */
- char *mrownerconfig; /* base64 encoded sha348 digest */
+ char *mrconfigid; /* base64 encoded sha384 digest */
+ char *mrowner; /* base64 encoded sha384 digest */
+ char *mrownerconfig; /* base64 encoded sha384 digest */
MemoryRegion *tdvf_mr;
TdxFirmware tdvf;
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 21/24] i386/tdx: Clarify the error message of mrconfigid/mrowner/mrownerconfig
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (19 preceding siblings ...)
2025-06-20 16:40 ` [PULL 20/24] i386/tdx: Fix the typo of the comment of struct TdxGuest Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 22/24] update Linux headers to v6.16-rc3 Paolo Bonzini
` (2 subsequent siblings)
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Xiaoyao Li, Daniel P. Berrangé, Igor Mammedov
From: Xiaoyao Li <xiaoyao.li@intel.com>
The error message is misleading - we successfully decoded the data,
the decoded data was simply with the wrong length.
Change the error message to show it is an length check failure with both
the received and expected values.
Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Link: https://lore.kernel.org/r/20250603050305.1704586-4-xiaoyao.li@intel.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/kvm/tdx.c | 12 +++++++++---
1 file changed, 9 insertions(+), 3 deletions(-)
diff --git a/target/i386/kvm/tdx.c b/target/i386/kvm/tdx.c
index acbe7497543..2284167141a 100644
--- a/target/i386/kvm/tdx.c
+++ b/target/i386/kvm/tdx.c
@@ -1032,7 +1032,9 @@ int tdx_pre_create_vcpu(CPUState *cpu, Error **errp)
return -1;
}
if (data_len != QCRYPTO_HASH_DIGEST_LEN_SHA384) {
- error_setg(errp, "TDX: failed to decode mrconfigid");
+ error_setg(errp, "TDX 'mrconfigid' sha384 digest was %ld bytes, "
+ "expected %d bytes", data_len,
+ QCRYPTO_HASH_DIGEST_LEN_SHA384);
return -1;
}
memcpy(init_vm->mrconfigid, data, data_len);
@@ -1045,7 +1047,9 @@ int tdx_pre_create_vcpu(CPUState *cpu, Error **errp)
return -1;
}
if (data_len != QCRYPTO_HASH_DIGEST_LEN_SHA384) {
- error_setg(errp, "TDX: failed to decode mrowner");
+ error_setg(errp, "TDX 'mrowner' sha384 digest was %ld bytes, "
+ "expected %d bytes", data_len,
+ QCRYPTO_HASH_DIGEST_LEN_SHA384);
return -1;
}
memcpy(init_vm->mrowner, data, data_len);
@@ -1058,7 +1062,9 @@ int tdx_pre_create_vcpu(CPUState *cpu, Error **errp)
return -1;
}
if (data_len != QCRYPTO_HASH_DIGEST_LEN_SHA384) {
- error_setg(errp, "TDX: failed to decode mrownerconfig");
+ error_setg(errp, "TDX 'mrownerconfig' sha384 digest was %ld bytes, "
+ "expected %d bytes", data_len,
+ QCRYPTO_HASH_DIGEST_LEN_SHA384);
return -1;
}
memcpy(init_vm->mrownerconfig, data, data_len);
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 22/24] update Linux headers to v6.16-rc3
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (20 preceding siblings ...)
2025-06-20 16:40 ` [PULL 21/24] i386/tdx: Clarify the error message of mrconfigid/mrowner/mrownerconfig Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 23/24] i386/tdx: handle TDG.VP.VMCALL<GetTdVmCallInfo> Paolo Bonzini
2025-06-20 16:40 ` [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote> Paolo Bonzini
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
include/standard-headers/asm-x86/setup_data.h | 13 +-
include/standard-headers/drm/drm_fourcc.h | 45 +++++++
include/standard-headers/linux/ethtool.h | 124 +++++++++---------
include/standard-headers/linux/fuse.h | 6 +-
.../linux/input-event-codes.h | 3 +-
include/standard-headers/linux/pci_regs.h | 12 +-
include/standard-headers/linux/virtio_gpu.h | 3 +-
include/standard-headers/linux/virtio_pci.h | 1 +
| 9 +-
| 1 +
| 4 +-
| 25 ++++
| 4 +-
13 files changed, 177 insertions(+), 73 deletions(-)
diff --git a/include/standard-headers/asm-x86/setup_data.h b/include/standard-headers/asm-x86/setup_data.h
index a483d72f428..2e446c1d858 100644
--- a/include/standard-headers/asm-x86/setup_data.h
+++ b/include/standard-headers/asm-x86/setup_data.h
@@ -13,7 +13,8 @@
#define SETUP_CC_BLOB 7
#define SETUP_IMA 8
#define SETUP_RNG_SEED 9
-#define SETUP_ENUM_MAX SETUP_RNG_SEED
+#define SETUP_KEXEC_KHO 10
+#define SETUP_ENUM_MAX SETUP_KEXEC_KHO
#define SETUP_INDIRECT (1<<31)
#define SETUP_TYPE_MAX (SETUP_ENUM_MAX | SETUP_INDIRECT)
@@ -78,6 +79,16 @@ struct ima_setup_data {
uint64_t size;
} QEMU_PACKED;
+/*
+ * Locations of kexec handover metadata
+ */
+struct kho_data {
+ uint64_t fdt_addr;
+ uint64_t fdt_size;
+ uint64_t scratch_addr;
+ uint64_t scratch_size;
+} QEMU_PACKED;
+
#endif /* __ASSEMBLER__ */
#endif /* _ASM_X86_SETUP_DATA_H */
diff --git a/include/standard-headers/drm/drm_fourcc.h b/include/standard-headers/drm/drm_fourcc.h
index a8b759dcbc8..c8309d378bf 100644
--- a/include/standard-headers/drm/drm_fourcc.h
+++ b/include/standard-headers/drm/drm_fourcc.h
@@ -421,6 +421,7 @@ extern "C" {
#define DRM_FORMAT_MOD_VENDOR_ALLWINNER 0x09
#define DRM_FORMAT_MOD_VENDOR_AMLOGIC 0x0a
#define DRM_FORMAT_MOD_VENDOR_MTK 0x0b
+#define DRM_FORMAT_MOD_VENDOR_APPLE 0x0c
/* add more to the end as needed */
@@ -1493,6 +1494,50 @@ drm_fourcc_canonicalize_nvidia_format_mod(uint64_t modifier)
/* alias for the most common tiling format */
#define DRM_FORMAT_MOD_MTK_16L_32S_TILE DRM_FORMAT_MOD_MTK(MTK_FMT_MOD_TILE_16L32S)
+/*
+ * Apple GPU-tiled layouts.
+ *
+ * Apple GPUs support nonlinear tilings with optional lossless compression.
+ *
+ * GPU-tiled images are divided into 16KiB tiles:
+ *
+ * Bytes per pixel Tile size
+ * --------------- ---------
+ * 1 128x128
+ * 2 128x64
+ * 4 64x64
+ * 8 64x32
+ * 16 32x32
+ *
+ * Tiles are raster-order. Pixels within a tile are interleaved (Morton order).
+ *
+ * Compressed images pad the body to 128-bytes and are immediately followed by a
+ * metadata section. The metadata section rounds the image dimensions to
+ * powers-of-two and contains 8 bytes for each 16x16 compression subtile.
+ * Subtiles are interleaved (Morton order).
+ *
+ * All images are 128-byte aligned.
+ *
+ * These layouts fundamentally do not have meaningful strides. No matter how we
+ * specify strides for these layouts, userspace unaware of Apple image layouts
+ * will be unable to use correctly the specified stride for any purpose.
+ * Userspace aware of the image layouts do not use strides. The most "correct"
+ * convention would be setting the image stride to 0. Unfortunately, some
+ * software assumes the stride is at least (width * bytes per pixel). We
+ * therefore require that stride equals (width * bytes per pixel). Since the
+ * stride is arbitrary here, we pick the simplest convention.
+ *
+ * Although containing two sections, compressed image layouts are treated in
+ * software as a single plane. This is modelled after AFBC, a similar
+ * scheme. Attempting to separate the sections to be "explicit" in DRM would
+ * only generate more confusion, as software does not treat the image this way.
+ *
+ * For detailed information on the hardware image layouts, see
+ * https://docs.mesa3d.org/drivers/asahi.html#image-layouts
+ */
+#define DRM_FORMAT_MOD_APPLE_GPU_TILED fourcc_mod_code(APPLE, 1)
+#define DRM_FORMAT_MOD_APPLE_GPU_TILED_COMPRESSED fourcc_mod_code(APPLE, 2)
+
/*
* AMD modifiers
*
diff --git a/include/standard-headers/linux/ethtool.h b/include/standard-headers/linux/ethtool.h
index 5d1ad5fdeab..cef0d207a62 100644
--- a/include/standard-headers/linux/ethtool.h
+++ b/include/standard-headers/linux/ethtool.h
@@ -2295,71 +2295,75 @@ static inline int ethtool_validate_duplex(uint8_t duplex)
#define RXH_XFRM_SYM_OR_XOR (1 << 1)
#define RXH_XFRM_NO_CHANGE 0xff
-/* L2-L4 network traffic flow types */
-#define TCP_V4_FLOW 0x01 /* hash or spec (tcp_ip4_spec) */
-#define UDP_V4_FLOW 0x02 /* hash or spec (udp_ip4_spec) */
-#define SCTP_V4_FLOW 0x03 /* hash or spec (sctp_ip4_spec) */
-#define AH_ESP_V4_FLOW 0x04 /* hash only */
-#define TCP_V6_FLOW 0x05 /* hash or spec (tcp_ip6_spec; nfc only) */
-#define UDP_V6_FLOW 0x06 /* hash or spec (udp_ip6_spec; nfc only) */
-#define SCTP_V6_FLOW 0x07 /* hash or spec (sctp_ip6_spec; nfc only) */
-#define AH_ESP_V6_FLOW 0x08 /* hash only */
-#define AH_V4_FLOW 0x09 /* hash or spec (ah_ip4_spec) */
-#define ESP_V4_FLOW 0x0a /* hash or spec (esp_ip4_spec) */
-#define AH_V6_FLOW 0x0b /* hash or spec (ah_ip6_spec; nfc only) */
-#define ESP_V6_FLOW 0x0c /* hash or spec (esp_ip6_spec; nfc only) */
-#define IPV4_USER_FLOW 0x0d /* spec only (usr_ip4_spec) */
-#define IP_USER_FLOW IPV4_USER_FLOW
-#define IPV6_USER_FLOW 0x0e /* spec only (usr_ip6_spec; nfc only) */
-#define IPV4_FLOW 0x10 /* hash only */
-#define IPV6_FLOW 0x11 /* hash only */
-#define ETHER_FLOW 0x12 /* spec only (ether_spec) */
+enum {
+ /* L2-L4 network traffic flow types */
+ TCP_V4_FLOW = 0x01, /* hash or spec (tcp_ip4_spec) */
+ UDP_V4_FLOW = 0x02, /* hash or spec (udp_ip4_spec) */
+ SCTP_V4_FLOW = 0x03, /* hash or spec (sctp_ip4_spec) */
+ AH_ESP_V4_FLOW = 0x04, /* hash only */
+ TCP_V6_FLOW = 0x05, /* hash or spec (tcp_ip6_spec; nfc only) */
+ UDP_V6_FLOW = 0x06, /* hash or spec (udp_ip6_spec; nfc only) */
+ SCTP_V6_FLOW = 0x07, /* hash or spec (sctp_ip6_spec; nfc only) */
+ AH_ESP_V6_FLOW = 0x08, /* hash only */
+ AH_V4_FLOW = 0x09, /* hash or spec (ah_ip4_spec) */
+ ESP_V4_FLOW = 0x0a, /* hash or spec (esp_ip4_spec) */
+ AH_V6_FLOW = 0x0b, /* hash or spec (ah_ip6_spec; nfc only) */
+ ESP_V6_FLOW = 0x0c, /* hash or spec (esp_ip6_spec; nfc only) */
+ IPV4_USER_FLOW = 0x0d, /* spec only (usr_ip4_spec) */
+ IP_USER_FLOW = IPV4_USER_FLOW,
+ IPV6_USER_FLOW = 0x0e, /* spec only (usr_ip6_spec; nfc only) */
+ IPV4_FLOW = 0x10, /* hash only */
+ IPV6_FLOW = 0x11, /* hash only */
+ ETHER_FLOW = 0x12, /* spec only (ether_spec) */
-/* Used for GTP-U IPv4 and IPv6.
- * The format of GTP packets only includes
- * elements such as TEID and GTP version.
- * It is primarily intended for data communication of the UE.
- */
-#define GTPU_V4_FLOW 0x13 /* hash only */
-#define GTPU_V6_FLOW 0x14 /* hash only */
+ /* Used for GTP-U IPv4 and IPv6.
+ * The format of GTP packets only includes
+ * elements such as TEID and GTP version.
+ * It is primarily intended for data communication of the UE.
+ */
+ GTPU_V4_FLOW = 0x13, /* hash only */
+ GTPU_V6_FLOW = 0x14, /* hash only */
-/* Use for GTP-C IPv4 and v6.
- * The format of these GTP packets does not include TEID.
- * Primarily expected to be used for communication
- * to create sessions for UE data communication,
- * commonly referred to as CSR (Create Session Request).
- */
-#define GTPC_V4_FLOW 0x15 /* hash only */
-#define GTPC_V6_FLOW 0x16 /* hash only */
+ /* Use for GTP-C IPv4 and v6.
+ * The format of these GTP packets does not include TEID.
+ * Primarily expected to be used for communication
+ * to create sessions for UE data communication,
+ * commonly referred to as CSR (Create Session Request).
+ */
+ GTPC_V4_FLOW = 0x15, /* hash only */
+ GTPC_V6_FLOW = 0x16, /* hash only */
-/* Use for GTP-C IPv4 and v6.
- * Unlike GTPC_V4_FLOW, the format of these GTP packets includes TEID.
- * After session creation, it becomes this packet.
- * This is mainly used for requests to realize UE handover.
- */
-#define GTPC_TEID_V4_FLOW 0x17 /* hash only */
-#define GTPC_TEID_V6_FLOW 0x18 /* hash only */
+ /* Use for GTP-C IPv4 and v6.
+ * Unlike GTPC_V4_FLOW, the format of these GTP packets includes TEID.
+ * After session creation, it becomes this packet.
+ * This is mainly used for requests to realize UE handover.
+ */
+ GTPC_TEID_V4_FLOW = 0x17, /* hash only */
+ GTPC_TEID_V6_FLOW = 0x18, /* hash only */
-/* Use for GTP-U and extended headers for the PSC (PDU Session Container).
- * The format of these GTP packets includes TEID and QFI.
- * In 5G communication using UPF (User Plane Function),
- * data communication with this extended header is performed.
- */
-#define GTPU_EH_V4_FLOW 0x19 /* hash only */
-#define GTPU_EH_V6_FLOW 0x1a /* hash only */
+ /* Use for GTP-U and extended headers for the PSC (PDU Session Container).
+ * The format of these GTP packets includes TEID and QFI.
+ * In 5G communication using UPF (User Plane Function),
+ * data communication with this extended header is performed.
+ */
+ GTPU_EH_V4_FLOW = 0x19, /* hash only */
+ GTPU_EH_V6_FLOW = 0x1a, /* hash only */
-/* Use for GTP-U IPv4 and v6 PSC (PDU Session Container) extended headers.
- * This differs from GTPU_EH_V(4|6)_FLOW in that it is distinguished by
- * UL/DL included in the PSC.
- * There are differences in the data included based on Downlink/Uplink,
- * and can be used to distinguish packets.
- * The functions described so far are useful when you want to
- * handle communication from the mobile network in UPF, PGW, etc.
- */
-#define GTPU_UL_V4_FLOW 0x1b /* hash only */
-#define GTPU_UL_V6_FLOW 0x1c /* hash only */
-#define GTPU_DL_V4_FLOW 0x1d /* hash only */
-#define GTPU_DL_V6_FLOW 0x1e /* hash only */
+ /* Use for GTP-U IPv4 and v6 PSC (PDU Session Container) extended headers.
+ * This differs from GTPU_EH_V(4|6)_FLOW in that it is distinguished by
+ * UL/DL included in the PSC.
+ * There are differences in the data included based on Downlink/Uplink,
+ * and can be used to distinguish packets.
+ * The functions described so far are useful when you want to
+ * handle communication from the mobile network in UPF, PGW, etc.
+ */
+ GTPU_UL_V4_FLOW = 0x1b, /* hash only */
+ GTPU_UL_V6_FLOW = 0x1c, /* hash only */
+ GTPU_DL_V4_FLOW = 0x1d, /* hash only */
+ GTPU_DL_V6_FLOW = 0x1e, /* hash only */
+
+ __FLOW_TYPE_COUNT,
+};
/* Flag to enable additional fields in struct ethtool_rx_flow_spec */
#define FLOW_EXT 0x80000000
diff --git a/include/standard-headers/linux/fuse.h b/include/standard-headers/linux/fuse.h
index a2b5815d890..d8b2fd67e16 100644
--- a/include/standard-headers/linux/fuse.h
+++ b/include/standard-headers/linux/fuse.h
@@ -232,6 +232,9 @@
*
* 7.43
* - add FUSE_REQUEST_TIMEOUT
+ *
+ * 7.44
+ * - add FUSE_NOTIFY_INC_EPOCH
*/
#ifndef _LINUX_FUSE_H
@@ -263,7 +266,7 @@
#define FUSE_KERNEL_VERSION 7
/** Minor version number of this interface */
-#define FUSE_KERNEL_MINOR_VERSION 43
+#define FUSE_KERNEL_MINOR_VERSION 44
/** The node ID of the root inode */
#define FUSE_ROOT_ID 1
@@ -667,6 +670,7 @@ enum fuse_notify_code {
FUSE_NOTIFY_RETRIEVE = 5,
FUSE_NOTIFY_DELETE = 6,
FUSE_NOTIFY_RESEND = 7,
+ FUSE_NOTIFY_INC_EPOCH = 8,
FUSE_NOTIFY_CODE_MAX,
};
diff --git a/include/standard-headers/linux/input-event-codes.h b/include/standard-headers/linux/input-event-codes.h
index 09ba0ad8783..a82ff795e06 100644
--- a/include/standard-headers/linux/input-event-codes.h
+++ b/include/standard-headers/linux/input-event-codes.h
@@ -925,7 +925,8 @@
#define SW_MUTE_DEVICE 0x0e /* set = device disabled */
#define SW_PEN_INSERTED 0x0f /* set = pen inserted */
#define SW_MACHINE_COVER 0x10 /* set = cover closed */
-#define SW_MAX_ 0x10
+#define SW_USB_INSERT 0x11 /* set = USB audio device connected */
+#define SW_MAX_ 0x11
#define SW_CNT (SW_MAX_+1)
/*
diff --git a/include/standard-headers/linux/pci_regs.h b/include/standard-headers/linux/pci_regs.h
index ba326710f9c..a3a3e942ded 100644
--- a/include/standard-headers/linux/pci_regs.h
+++ b/include/standard-headers/linux/pci_regs.h
@@ -750,7 +750,8 @@
#define PCI_EXT_CAP_ID_NPEM 0x29 /* Native PCIe Enclosure Management */
#define PCI_EXT_CAP_ID_PL_32GT 0x2A /* Physical Layer 32.0 GT/s */
#define PCI_EXT_CAP_ID_DOE 0x2E /* Data Object Exchange */
-#define PCI_EXT_CAP_ID_MAX PCI_EXT_CAP_ID_DOE
+#define PCI_EXT_CAP_ID_PL_64GT 0x31 /* Physical Layer 64.0 GT/s */
+#define PCI_EXT_CAP_ID_MAX PCI_EXT_CAP_ID_PL_64GT
#define PCI_EXT_CAP_DSN_SIZEOF 12
#define PCI_EXT_CAP_MCAST_ENDPOINT_SIZEOF 40
@@ -1144,12 +1145,21 @@
#define PCI_DLF_CAP 0x04 /* Capabilities Register */
#define PCI_DLF_EXCHANGE_ENABLE 0x80000000 /* Data Link Feature Exchange Enable */
+/* Secondary PCIe Capability 8.0 GT/s */
+#define PCI_SECPCI_LE_CTRL 0x0c /* Lane Equalization Control Register */
+
/* Physical Layer 16.0 GT/s */
#define PCI_PL_16GT_LE_CTRL 0x20 /* Lane Equalization Control Register */
#define PCI_PL_16GT_LE_CTRL_DSP_TX_PRESET_MASK 0x0000000F
#define PCI_PL_16GT_LE_CTRL_USP_TX_PRESET_MASK 0x000000F0
#define PCI_PL_16GT_LE_CTRL_USP_TX_PRESET_SHIFT 4
+/* Physical Layer 32.0 GT/s */
+#define PCI_PL_32GT_LE_CTRL 0x20 /* Lane Equalization Control Register */
+
+/* Physical Layer 64.0 GT/s */
+#define PCI_PL_64GT_LE_CTRL 0x20 /* Lane Equalization Control Register */
+
/* Native PCIe Enclosure Management */
#define PCI_NPEM_CAP 0x04 /* NPEM capability register */
#define PCI_NPEM_CAP_CAPABLE 0x00000001 /* NPEM Capable */
diff --git a/include/standard-headers/linux/virtio_gpu.h b/include/standard-headers/linux/virtio_gpu.h
index 6459fdb9fb5..00cd3f04af9 100644
--- a/include/standard-headers/linux/virtio_gpu.h
+++ b/include/standard-headers/linux/virtio_gpu.h
@@ -309,8 +309,9 @@ struct virtio_gpu_cmd_submit {
#define VIRTIO_GPU_CAPSET_VIRGL 1
#define VIRTIO_GPU_CAPSET_VIRGL2 2
-/* 3 is reserved for gfxstream */
+#define VIRTIO_GPU_CAPSET_GFXSTREAM_VULKAN 3
#define VIRTIO_GPU_CAPSET_VENUS 4
+#define VIRTIO_GPU_CAPSET_CROSS_DOMAIN 5
#define VIRTIO_GPU_CAPSET_DRM 6
/* VIRTIO_GPU_CMD_GET_CAPSET_INFO */
diff --git a/include/standard-headers/linux/virtio_pci.h b/include/standard-headers/linux/virtio_pci.h
index 91fec6f5029..09e964e6eee 100644
--- a/include/standard-headers/linux/virtio_pci.h
+++ b/include/standard-headers/linux/virtio_pci.h
@@ -246,6 +246,7 @@ struct virtio_pci_cfg_cap {
#define VIRTIO_ADMIN_CMD_LIST_USE 0x1
/* Admin command group type. */
+#define VIRTIO_ADMIN_GROUP_TYPE_SELF 0x0
#define VIRTIO_ADMIN_GROUP_TYPE_SRIOV 0x1
/* Transitional device admin command. */
--git a/linux-headers/asm-arm64/kvm.h b/linux-headers/asm-arm64/kvm.h
index 4e6aff08df8..f4d9baafa1c 100644
--- a/linux-headers/asm-arm64/kvm.h
+++ b/linux-headers/asm-arm64/kvm.h
@@ -419,10 +419,11 @@ enum {
/* Device Control API on vcpu fd */
#define KVM_ARM_VCPU_PMU_V3_CTRL 0
-#define KVM_ARM_VCPU_PMU_V3_IRQ 0
-#define KVM_ARM_VCPU_PMU_V3_INIT 1
-#define KVM_ARM_VCPU_PMU_V3_FILTER 2
-#define KVM_ARM_VCPU_PMU_V3_SET_PMU 3
+#define KVM_ARM_VCPU_PMU_V3_IRQ 0
+#define KVM_ARM_VCPU_PMU_V3_INIT 1
+#define KVM_ARM_VCPU_PMU_V3_FILTER 2
+#define KVM_ARM_VCPU_PMU_V3_SET_PMU 3
+#define KVM_ARM_VCPU_PMU_V3_SET_NR_COUNTERS 4
#define KVM_ARM_VCPU_TIMER_CTRL 1
#define KVM_ARM_VCPU_TIMER_IRQ_VTIMER 0
#define KVM_ARM_VCPU_TIMER_IRQ_PTIMER 1
--git a/linux-headers/asm-x86/kvm.h b/linux-headers/asm-x86/kvm.h
index 7fb57ccb2a7..cd275ae76d2 100644
--- a/linux-headers/asm-x86/kvm.h
+++ b/linux-headers/asm-x86/kvm.h
@@ -843,6 +843,7 @@ struct kvm_sev_snp_launch_start {
};
/* Kept in sync with firmware values for simplicity. */
+#define KVM_SEV_PAGE_TYPE_INVALID 0x0
#define KVM_SEV_SNP_PAGE_TYPE_NORMAL 0x1
#define KVM_SEV_SNP_PAGE_TYPE_ZERO 0x3
#define KVM_SEV_SNP_PAGE_TYPE_UNMEASURED 0x4
--git a/linux-headers/linux/bits.h b/linux-headers/linux/bits.h
index 58596d18f43..9243f389751 100644
--- a/linux-headers/linux/bits.h
+++ b/linux-headers/linux/bits.h
@@ -4,9 +4,9 @@
#ifndef _LINUX_BITS_H
#define _LINUX_BITS_H
-#define __GENMASK(h, l) (((~_UL(0)) << (l)) & (~_UL(0) >> (BITS_PER_LONG - 1 - (h))))
+#define __GENMASK(h, l) (((~_UL(0)) << (l)) & (~_UL(0) >> (__BITS_PER_LONG - 1 - (h))))
-#define __GENMASK_ULL(h, l) (((~_ULL(0)) << (l)) & (~_ULL(0) >> (BITS_PER_LONG_LONG - 1 - (h))))
+#define __GENMASK_ULL(h, l) (((~_ULL(0)) << (l)) & (~_ULL(0) >> (__BITS_PER_LONG_LONG - 1 - (h))))
#define __GENMASK_U128(h, l) \
((_BIT128((h)) << 1) - (_BIT128(l)))
--git a/linux-headers/linux/kvm.h b/linux-headers/linux/kvm.h
index 99cc82a275c..0690743944b 100644
--- a/linux-headers/linux/kvm.h
+++ b/linux-headers/linux/kvm.h
@@ -178,6 +178,7 @@ struct kvm_xen_exit {
#define KVM_EXIT_NOTIFY 37
#define KVM_EXIT_LOONGARCH_IOCSR 38
#define KVM_EXIT_MEMORY_FAULT 39
+#define KVM_EXIT_TDX 40
/* For KVM_EXIT_INTERNAL_ERROR */
/* Emulate instruction failed. */
@@ -439,6 +440,27 @@ struct kvm_run {
__u64 gpa;
__u64 size;
} memory_fault;
+ /* KVM_EXIT_TDX */
+ struct {
+ __u64 flags;
+ __u64 nr;
+ union {
+ struct {
+ __u64 ret;
+ __u64 data[5];
+ } unknown;
+ struct {
+ __u64 ret;
+ __u64 gpa;
+ __u64 size;
+ } get_quote;
+ struct {
+ __u64 ret;
+ __u64 leaf;
+ __u64 r11, r12, r13, r14;
+ } get_tdvmcall_info;
+ };
+ } tdx;
/* Fix the size of the union. */
char padding[256];
};
@@ -923,6 +945,9 @@ struct kvm_enable_cap {
#define KVM_CAP_X86_APIC_BUS_CYCLES_NS 237
#define KVM_CAP_X86_GUEST_MODE 238
#define KVM_CAP_ARM_WRITABLE_IMP_ID_REGS 239
+#define KVM_CAP_ARM_EL2 240
+#define KVM_CAP_ARM_EL2_E2H0 241
+#define KVM_CAP_RISCV_MP_STATE_RESET 242
struct kvm_irq_routing_irqchip {
__u32 irqchip;
--git a/linux-headers/linux/vhost.h b/linux-headers/linux/vhost.h
index b95dd84eef2..d4b3e2ae131 100644
--- a/linux-headers/linux/vhost.h
+++ b/linux-headers/linux/vhost.h
@@ -28,10 +28,10 @@
/* Set current process as the (exclusive) owner of this file descriptor. This
* must be called before any other vhost command. Further calls to
- * VHOST_OWNER_SET fail until VHOST_OWNER_RESET is called. */
+ * VHOST_SET_OWNER fail until VHOST_RESET_OWNER is called. */
#define VHOST_SET_OWNER _IO(VHOST_VIRTIO, 0x01)
/* Give up ownership, and reset the device to default values.
- * Allows subsequent call to VHOST_OWNER_SET to succeed. */
+ * Allows subsequent call to VHOST_SET_OWNER to succeed. */
#define VHOST_RESET_OWNER _IO(VHOST_VIRTIO, 0x02)
/* Set up/modify memory layout */
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 23/24] i386/tdx: handle TDG.VP.VMCALL<GetTdVmCallInfo>
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (21 preceding siblings ...)
2025-06-20 16:40 ` [PULL 22/24] update Linux headers to v6.16-rc3 Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-20 16:40 ` [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote> Paolo Bonzini
23 siblings, 0 replies; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Binbin Wu
From: Binbin Wu <binbin.wu@linux.intel.com>
Signed-off-by: Binbin Wu <binbin.wu@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
target/i386/kvm/tdx.h | 9 +++++++++
target/i386/kvm/kvm.c | 12 ++++++++++++
target/i386/kvm/tdx-stub.c | 4 ++++
target/i386/kvm/tdx.c | 12 ++++++++++++
4 files changed, 37 insertions(+)
diff --git a/target/i386/kvm/tdx.h b/target/i386/kvm/tdx.h
index 8dd66e90149..0dd41d5811a 100644
--- a/target/i386/kvm/tdx.h
+++ b/target/i386/kvm/tdx.h
@@ -21,6 +21,14 @@ typedef struct TdxGuestClass {
/* TDX requires bus frequency 25MHz */
#define TDX_APIC_BUS_CYCLES_NS 40
+#define TDVMCALL_GET_TD_VM_CALL_INFO 0x10000
+
+#define TDG_VP_VMCALL_SUCCESS 0x0000000000000000ULL
+#define TDG_VP_VMCALL_RETRY 0x0000000000000001ULL
+#define TDG_VP_VMCALL_INVALID_OPERAND 0x8000000000000000ULL
+#define TDG_VP_VMCALL_GPA_INUSE 0x8000000000000001ULL
+#define TDG_VP_VMCALL_ALIGN_ERROR 0x8000000000000002ULL
+
enum TdxRamType {
TDX_RAM_UNACCEPTED,
TDX_RAM_ADDED,
@@ -61,5 +69,6 @@ int tdx_pre_create_vcpu(CPUState *cpu, Error **errp);
void tdx_set_tdvf_region(MemoryRegion *tdvf_mr);
int tdx_parse_tdvf(void *flash_ptr, int size);
int tdx_handle_report_fatal_error(X86CPU *cpu, struct kvm_run *run);
+void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run);
#endif /* QEMU_I386_TDX_H */
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 56a6b9b6381..8ef29fc1fb1 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -6170,6 +6170,18 @@ int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
break;
}
break;
+ case KVM_EXIT_TDX:
+ /*
+ * run->tdx is already set up for the case where userspace
+ * does not handle the TDVMCALL.
+ */
+ switch (run->tdx.nr) {
+ case TDVMCALL_GET_TD_VM_CALL_INFO:
+ tdx_handle_get_tdvmcall_info(cpu, run);
+ break;
+ }
+ ret = 0;
+ break;
default:
fprintf(stderr, "KVM: unknown exit reason %d\n", run->exit_reason);
ret = -1;
diff --git a/target/i386/kvm/tdx-stub.c b/target/i386/kvm/tdx-stub.c
index 720a4ff046e..62a12a06775 100644
--- a/target/i386/kvm/tdx-stub.c
+++ b/target/i386/kvm/tdx-stub.c
@@ -18,3 +18,7 @@ int tdx_handle_report_fatal_error(X86CPU *cpu, struct kvm_run *run)
{
return -EINVAL;
}
+
+void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run)
+{
+}
diff --git a/target/i386/kvm/tdx.c b/target/i386/kvm/tdx.c
index 2284167141a..ef10a193474 100644
--- a/target/i386/kvm/tdx.c
+++ b/target/i386/kvm/tdx.c
@@ -1120,6 +1120,18 @@ int tdx_parse_tdvf(void *flash_ptr, int size)
return tdvf_parse_metadata(&tdx_guest->tdvf, flash_ptr, size);
}
+void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run)
+{
+ if (run->tdx.get_tdvmcall_info.leaf != 1) {
+ return;
+ }
+
+ run->tdx.get_tdvmcall_info.r11 = 0;
+ run->tdx.get_tdvmcall_info.r12 = 0;
+ run->tdx.get_tdvmcall_info.r13 = 0;
+ run->tdx.get_tdvmcall_info.r14 = 0;
+}
+
static void tdx_panicked_on_fatal_error(X86CPU *cpu, uint64_t error_code,
char *message, uint64_t gpa)
{
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-20 16:40 [PULL 00/24] i386 (TCG, TDX), Rust changes for 2025-06-20 Paolo Bonzini
` (22 preceding siblings ...)
2025-06-20 16:40 ` [PULL 23/24] i386/tdx: handle TDG.VP.VMCALL<GetTdVmCallInfo> Paolo Bonzini
@ 2025-06-20 16:40 ` Paolo Bonzini
2025-06-23 6:43 ` Cédric Le Goater
23 siblings, 1 reply; 36+ messages in thread
From: Paolo Bonzini @ 2025-06-20 16:40 UTC (permalink / raw)
To: qemu-devel; +Cc: Isaku Yamahata, Chenyi Qiang, Xiaoyao Li
From: Isaku Yamahata <isaku.yamahata@intel.com>
Add property "quote-generation-socket" to tdx-guest, which is a property
of type SocketAddress to specify Quote Generation Service(QGS).
On request of GetQuote, it connects to the QGS socket, read request
data from shared guest memory, send the request data to the QGS,
and store the response into shared guest memory, at last notify
TD guest by interrupt.
command line example:
qemu-system-x86_64 \
-object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
-machine confidential-guest-support=tdx0
Note, above example uses the unix socket. It can be other types, like vsock,
which depends on the implementation of QGS.
To avoid no response from QGS server, setup a timer for the transaction.
If timeout, make it an error and interrupt guest. Define the threshold of
time to 30s at present, maybe change to other value if not appropriate.
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
qapi/qom.json | 8 +-
target/i386/kvm/tdx-quote-generator.h | 82 +++++++
target/i386/kvm/tdx.h | 10 +
target/i386/kvm/kvm.c | 3 +
target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
target/i386/kvm/tdx-stub.c | 4 +
target/i386/kvm/tdx.c | 176 ++++++++++++++-
target/i386/kvm/meson.build | 2 +-
8 files changed, 582 insertions(+), 3 deletions(-)
create mode 100644 target/i386/kvm/tdx-quote-generator.h
create mode 100644 target/i386/kvm/tdx-quote-generator.c
diff --git a/qapi/qom.json b/qapi/qom.json
index 3e8debf78c2..b133b064471 100644
--- a/qapi/qom.json
+++ b/qapi/qom.json
@@ -1071,6 +1071,11 @@
# e.g., specific to the workload rather than the run-time or OS
# (base64 encoded SHA384 digest). Defaults to all zeros.
#
+# @quote-generation-socket: socket address for Quote Generation
+# Service (QGS). QGS is a daemon running on the host. Without
+# it, the guest will not be able to get a TD quote for
+# attestation.
+#
# Since: 10.1
##
{ 'struct': 'TdxGuestProperties',
@@ -1078,7 +1083,8 @@
'*sept-ve-disable': 'bool',
'*mrconfigid': 'str',
'*mrowner': 'str',
- '*mrownerconfig': 'str' } }
+ '*mrownerconfig': 'str',
+ '*quote-generation-socket': 'SocketAddress' } }
##
# @ThreadContextProperties:
diff --git a/target/i386/kvm/tdx-quote-generator.h b/target/i386/kvm/tdx-quote-generator.h
new file mode 100644
index 00000000000..3bd9b8ef331
--- /dev/null
+++ b/target/i386/kvm/tdx-quote-generator.h
@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+
+#ifndef QEMU_I386_TDX_QUOTE_GENERATOR_H
+#define QEMU_I386_TDX_QUOTE_GENERATOR_H
+
+#include "qom/object_interfaces.h"
+#include "io/channel-socket.h"
+#include "exec/hwaddr.h"
+
+#define TDX_GET_QUOTE_STRUCTURE_VERSION 1ULL
+
+#define TDX_VP_GET_QUOTE_SUCCESS 0ULL
+#define TDX_VP_GET_QUOTE_IN_FLIGHT (-1ULL)
+#define TDX_VP_GET_QUOTE_ERROR 0x8000000000000000ULL
+#define TDX_VP_GET_QUOTE_QGS_UNAVAILABLE 0x8000000000000001ULL
+
+/* Limit to avoid resource starvation. */
+#define TDX_GET_QUOTE_MAX_BUF_LEN (128 * 1024)
+#define TDX_MAX_GET_QUOTE_REQUEST 16
+
+#define TDX_GET_QUOTE_HDR_SIZE 24
+
+/* Format of pages shared with guest. */
+struct tdx_get_quote_header {
+ /* Format version: must be 1 in little endian. */
+ uint64_t structure_version;
+
+ /*
+ * GetQuote status code in little endian:
+ * Guest must set error_code to 0 to avoid information leak.
+ * Qemu sets this before interrupting guest.
+ */
+ uint64_t error_code;
+
+ /*
+ * in-message size in little endian: The message will follow this header.
+ * The in-message will be send to QGS.
+ */
+ uint32_t in_len;
+
+ /*
+ * out-message size in little endian:
+ * On request, out_len must be zero to avoid information leak.
+ * On return, message size from QGS. Qemu overwrites this field.
+ * The message will follows this header. The in-message is overwritten.
+ */
+ uint32_t out_len;
+
+ /*
+ * Message buffer follows.
+ * Guest sets message that will be send to QGS. If out_len > in_len, guest
+ * should zero remaining buffer to avoid information leak.
+ * Qemu overwrites this buffer with a message returned from QGS.
+ */
+};
+
+typedef struct TdxGenerateQuoteTask {
+ hwaddr buf_gpa;
+ hwaddr payload_gpa;
+ uint64_t payload_len;
+
+ char *send_data;
+ uint64_t send_data_size;
+ uint64_t send_data_sent;
+
+ char *receive_buf;
+ uint64_t receive_buf_received;
+
+ uint64_t status_code;
+ struct tdx_get_quote_header hdr;
+
+ QIOChannelSocket *sioc;
+ guint watch;
+ QEMUTimer timer;
+
+ void (*completion)(struct TdxGenerateQuoteTask *task);
+ void *opaque;
+} TdxGenerateQuoteTask;
+
+void tdx_generate_quote(TdxGenerateQuoteTask *task, SocketAddress *qg_sock_addr);
+
+#endif /* QEMU_I386_TDX_QUOTE_GENERATOR_H */
diff --git a/target/i386/kvm/tdx.h b/target/i386/kvm/tdx.h
index 0dd41d5811a..35a09c19c52 100644
--- a/target/i386/kvm/tdx.h
+++ b/target/i386/kvm/tdx.h
@@ -11,6 +11,8 @@
#include "cpu.h"
#include "hw/i386/tdvf.h"
+#include "tdx-quote-generator.h"
+
#define TYPE_TDX_GUEST "tdx-guest"
#define TDX_GUEST(obj) OBJECT_CHECK(TdxGuest, (obj), TYPE_TDX_GUEST)
@@ -22,6 +24,7 @@ typedef struct TdxGuestClass {
#define TDX_APIC_BUS_CYCLES_NS 40
#define TDVMCALL_GET_TD_VM_CALL_INFO 0x10000
+#define TDVMCALL_GET_QUOTE 0x10002
#define TDG_VP_VMCALL_SUCCESS 0x0000000000000000ULL
#define TDG_VP_VMCALL_RETRY 0x0000000000000001ULL
@@ -29,6 +32,8 @@ typedef struct TdxGuestClass {
#define TDG_VP_VMCALL_GPA_INUSE 0x8000000000000001ULL
#define TDG_VP_VMCALL_ALIGN_ERROR 0x8000000000000002ULL
+#define TDG_VP_VMCALL_SUBFUNC_GET_QUOTE 0x0000000000000001ULL
+
enum TdxRamType {
TDX_RAM_UNACCEPTED,
TDX_RAM_ADDED,
@@ -57,6 +62,10 @@ typedef struct TdxGuest {
uint32_t nr_ram_entries;
TdxRamEntry *ram_entries;
+
+ /* GetQuote */
+ SocketAddress *qg_sock_addr;
+ int num;
} TdxGuest;
#ifdef CONFIG_TDX
@@ -69,6 +78,7 @@ int tdx_pre_create_vcpu(CPUState *cpu, Error **errp);
void tdx_set_tdvf_region(MemoryRegion *tdvf_mr);
int tdx_parse_tdvf(void *flash_ptr, int size);
int tdx_handle_report_fatal_error(X86CPU *cpu, struct kvm_run *run);
+void tdx_handle_get_quote(X86CPU *cpu, struct kvm_run *run);
void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run);
#endif /* QEMU_I386_TDX_H */
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 8ef29fc1fb1..234878c613f 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -6176,6 +6176,9 @@ int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
* does not handle the TDVMCALL.
*/
switch (run->tdx.nr) {
+ case TDVMCALL_GET_QUOTE:
+ tdx_handle_get_quote(cpu, run);
+ break;
case TDVMCALL_GET_TD_VM_CALL_INFO:
tdx_handle_get_tdvmcall_info(cpu, run);
break;
diff --git a/target/i386/kvm/tdx-quote-generator.c b/target/i386/kvm/tdx-quote-generator.c
new file mode 100644
index 00000000000..f59715f6175
--- /dev/null
+++ b/target/i386/kvm/tdx-quote-generator.c
@@ -0,0 +1,300 @@
+/*
+ * QEMU TDX Quote Generation Support
+ *
+ * Copyright (c) 2025 Intel Corporation
+ *
+ * Author:
+ * Xiaoyao Li <xiaoyao.li@intel.com>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "qapi/error.h"
+#include "qapi/qapi-visit-sockets.h"
+
+#include "tdx-quote-generator.h"
+
+#define QGS_MSG_LIB_MAJOR_VER 1
+#define QGS_MSG_LIB_MINOR_VER 1
+
+typedef enum _qgs_msg_type_t {
+ GET_QUOTE_REQ = 0,
+ GET_QUOTE_RESP = 1,
+ GET_COLLATERAL_REQ = 2,
+ GET_COLLATERAL_RESP = 3,
+ GET_PLATFORM_INFO_REQ = 4,
+ GET_PLATFORM_INFO_RESP = 5,
+ QGS_MSG_TYPE_MAX
+} qgs_msg_type_t;
+
+typedef struct _qgs_msg_header_t {
+ uint16_t major_version;
+ uint16_t minor_version;
+ uint32_t type;
+ uint32_t size; // size of the whole message, include this header, in byte
+ uint32_t error_code; // used in response only
+} qgs_msg_header_t;
+
+typedef struct _qgs_msg_get_quote_req_t {
+ qgs_msg_header_t header; // header.type = GET_QUOTE_REQ
+ uint32_t report_size; // cannot be 0
+ uint32_t id_list_size; // length of id_list, in byte, can be 0
+} qgs_msg_get_quote_req_t;
+
+typedef struct _qgs_msg_get_quote_resp_s {
+ qgs_msg_header_t header; // header.type = GET_QUOTE_RESP
+ uint32_t selected_id_size; // can be 0 in case only one id is sent in request
+ uint32_t quote_size; // length of quote_data, in byte
+ uint8_t id_quote[]; // selected id followed by quote
+} qgs_msg_get_quote_resp_t;
+
+#define HEADER_SIZE 4
+
+static uint32_t decode_header(const char *buf, size_t len) {
+ if (len < HEADER_SIZE) {
+ return 0;
+ }
+ uint32_t msg_size = 0;
+ for (uint32_t i = 0; i < HEADER_SIZE; ++i) {
+ msg_size = msg_size * 256 + (buf[i] & 0xFF);
+ }
+ return msg_size;
+}
+
+static void encode_header(char *buf, size_t len, uint32_t size) {
+ assert(len >= HEADER_SIZE);
+ buf[0] = ((size >> 24) & 0xFF);
+ buf[1] = ((size >> 16) & 0xFF);
+ buf[2] = ((size >> 8) & 0xFF);
+ buf[3] = (size & 0xFF);
+}
+
+static void tdx_generate_quote_cleanup(TdxGenerateQuoteTask *task)
+{
+ timer_del(&task->timer);
+
+ g_source_remove(task->watch);
+ qio_channel_close(QIO_CHANNEL(task->sioc), NULL);
+ object_unref(OBJECT(task->sioc));
+
+ task->completion(task);
+}
+
+static gboolean tdx_get_quote_read(QIOChannel *ioc, GIOCondition condition,
+ gpointer opaque)
+{
+ TdxGenerateQuoteTask *task = opaque;
+ Error *err = NULL;
+ int ret;
+
+ ret = qio_channel_read(ioc, task->receive_buf + task->receive_buf_received,
+ task->payload_len - task->receive_buf_received, &err);
+ if (ret < 0) {
+ if (ret == QIO_CHANNEL_ERR_BLOCK) {
+ return G_SOURCE_CONTINUE;
+ } else {
+ error_report_err(err);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ }
+
+ if (ret == 0) {
+ error_report("End of file before reply received");
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+
+ task->receive_buf_received += ret;
+ if (task->receive_buf_received >= HEADER_SIZE) {
+ uint32_t len = decode_header(task->receive_buf,
+ task->receive_buf_received);
+ if (len == 0 ||
+ len > (task->payload_len - HEADER_SIZE)) {
+ error_report("Message len %u must be non-zero & less than %zu",
+ len, (task->payload_len - HEADER_SIZE));
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+
+ /* Now we know the size, shrink to fit */
+ task->payload_len = HEADER_SIZE + len;
+ task->receive_buf = g_renew(char,
+ task->receive_buf,
+ task->payload_len);
+ }
+
+ if (task->receive_buf_received >= (sizeof(qgs_msg_header_t) + HEADER_SIZE)) {
+ qgs_msg_header_t *hdr = (qgs_msg_header_t *)(task->receive_buf + HEADER_SIZE);
+ if (hdr->major_version != QGS_MSG_LIB_MAJOR_VER ||
+ hdr->minor_version != QGS_MSG_LIB_MINOR_VER) {
+ error_report("Invalid QGS message header version %d.%d",
+ hdr->major_version,
+ hdr->minor_version);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ if (hdr->type != GET_QUOTE_RESP) {
+ error_report("Invalid QGS message type %d",
+ hdr->type);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ if (hdr->size > (task->payload_len - HEADER_SIZE)) {
+ error_report("QGS message size %d exceeds payload capacity %zu",
+ hdr->size, task->payload_len);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ if (hdr->error_code != 0) {
+ error_report("QGS message error code %d",
+ hdr->error_code);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ }
+ if (task->receive_buf_received >= (sizeof(qgs_msg_get_quote_resp_t) + HEADER_SIZE)) {
+ qgs_msg_get_quote_resp_t *msg = (qgs_msg_get_quote_resp_t *)(task->receive_buf + HEADER_SIZE);
+ if (msg->selected_id_size != 0) {
+ error_report("QGS message selected ID was %d not 0",
+ msg->selected_id_size);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+
+ if ((task->payload_len - HEADER_SIZE - sizeof(qgs_msg_get_quote_resp_t)) !=
+ msg->quote_size) {
+ error_report("QGS quote size %d should be %zu",
+ msg->quote_size,
+ (task->payload_len - sizeof(qgs_msg_get_quote_resp_t)));
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ goto end;
+ }
+ }
+
+ if (task->receive_buf_received == task->payload_len) {
+ size_t strip = HEADER_SIZE + sizeof(qgs_msg_get_quote_resp_t);
+ memmove(task->receive_buf,
+ task->receive_buf + strip,
+ task->receive_buf_received - strip);
+ task->receive_buf_received -= strip;
+ task->status_code = TDX_VP_GET_QUOTE_SUCCESS;
+ goto end;
+ }
+
+ return G_SOURCE_CONTINUE;
+
+end:
+ tdx_generate_quote_cleanup(task);
+ return G_SOURCE_REMOVE;
+}
+
+static gboolean tdx_send_report(QIOChannel *ioc, GIOCondition condition,
+ gpointer opaque)
+{
+ TdxGenerateQuoteTask *task = opaque;
+ Error *err = NULL;
+ int ret;
+
+ ret = qio_channel_write(ioc, task->send_data + task->send_data_sent,
+ task->send_data_size - task->send_data_sent, &err);
+ if (ret < 0) {
+ if (ret == QIO_CHANNEL_ERR_BLOCK) {
+ ret = 0;
+ } else {
+ error_report_err(err);
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ tdx_generate_quote_cleanup(task);
+ goto end;
+ }
+ }
+ task->send_data_sent += ret;
+
+ if (task->send_data_sent == task->send_data_size) {
+ task->watch = qio_channel_add_watch(QIO_CHANNEL(task->sioc), G_IO_IN,
+ tdx_get_quote_read, task, NULL);
+ goto end;
+ }
+
+ return G_SOURCE_CONTINUE;
+
+end:
+ return G_SOURCE_REMOVE;
+}
+
+static void tdx_quote_generator_connected(QIOTask *qio_task, gpointer opaque)
+{
+ TdxGenerateQuoteTask *task = opaque;
+ Error *err = NULL;
+ int ret;
+
+ ret = qio_task_propagate_error(qio_task, &err);
+ if (ret) {
+ error_report_err(err);
+ task->status_code = TDX_VP_GET_QUOTE_QGS_UNAVAILABLE;
+ tdx_generate_quote_cleanup(task);
+ return;
+ }
+
+ task->watch = qio_channel_add_watch(QIO_CHANNEL(task->sioc), G_IO_OUT,
+ tdx_send_report, task, NULL);
+}
+
+#define TRANSACTION_TIMEOUT 30000
+
+static void getquote_expired(void *opaque)
+{
+ TdxGenerateQuoteTask *task = opaque;
+
+ task->status_code = TDX_VP_GET_QUOTE_ERROR;
+ tdx_generate_quote_cleanup(task);
+}
+
+static void setup_get_quote_timer(TdxGenerateQuoteTask *task)
+{
+ int64_t time;
+
+ timer_init_ms(&task->timer, QEMU_CLOCK_VIRTUAL, getquote_expired, task);
+ time = qemu_clock_get_ms(QEMU_CLOCK_VIRTUAL);
+ timer_mod(&task->timer, time + TRANSACTION_TIMEOUT);
+}
+
+void tdx_generate_quote(TdxGenerateQuoteTask *task,
+ SocketAddress *qg_sock_addr)
+{
+ QIOChannelSocket *sioc;
+ qgs_msg_get_quote_req_t msg;
+
+ /* Prepare a QGS message prelude */
+ msg.header.major_version = QGS_MSG_LIB_MAJOR_VER;
+ msg.header.minor_version = QGS_MSG_LIB_MINOR_VER;
+ msg.header.type = GET_QUOTE_REQ;
+ msg.header.size = sizeof(msg) + task->send_data_size;
+ msg.header.error_code = 0;
+ msg.report_size = task->send_data_size;
+ msg.id_list_size = 0;
+
+ /* Make room to add the QGS message prelude */
+ task->send_data = g_renew(char,
+ task->send_data,
+ task->send_data_size + sizeof(msg) + HEADER_SIZE);
+ memmove(task->send_data + sizeof(msg) + HEADER_SIZE,
+ task->send_data,
+ task->send_data_size);
+ memcpy(task->send_data + HEADER_SIZE,
+ &msg,
+ sizeof(msg));
+ encode_header(task->send_data, HEADER_SIZE, task->send_data_size + sizeof(msg));
+ task->send_data_size += sizeof(msg) + HEADER_SIZE;
+
+ sioc = qio_channel_socket_new();
+ task->sioc = sioc;
+
+ setup_get_quote_timer(task);
+
+ qio_channel_socket_connect_async(sioc, qg_sock_addr,
+ tdx_quote_generator_connected, task,
+ NULL, NULL);
+}
diff --git a/target/i386/kvm/tdx-stub.c b/target/i386/kvm/tdx-stub.c
index 62a12a06775..76fee49eff0 100644
--- a/target/i386/kvm/tdx-stub.c
+++ b/target/i386/kvm/tdx-stub.c
@@ -19,6 +19,10 @@ int tdx_handle_report_fatal_error(X86CPU *cpu, struct kvm_run *run)
return -EINVAL;
}
+void tdx_handle_get_quote(X86CPU *cpu, struct kvm_run *run)
+{
+}
+
void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run)
{
}
diff --git a/target/i386/kvm/tdx.c b/target/i386/kvm/tdx.c
index ef10a193474..e809e4b2dfa 100644
--- a/target/i386/kvm/tdx.c
+++ b/target/i386/kvm/tdx.c
@@ -14,12 +14,14 @@
#include "qemu/base64.h"
#include "qemu/mmap-alloc.h"
#include "qapi/error.h"
+#include "qapi/qapi-visit-sockets.h"
#include "qom/object_interfaces.h"
#include "crypto/hash.h"
#include "system/kvm_int.h"
#include "system/runstate.h"
#include "system/system.h"
#include "system/ramblock.h"
+#include "system/address-spaces.h"
#include <linux/kvm_para.h>
@@ -32,6 +34,7 @@
#include "hw/i386/tdvf-hob.h"
#include "kvm_i386.h"
#include "tdx.h"
+#include "tdx-quote-generator.h"
#include "standard-headers/asm-x86/kvm_para.h"
@@ -1120,13 +1123,146 @@ int tdx_parse_tdvf(void *flash_ptr, int size)
return tdvf_parse_metadata(&tdx_guest->tdvf, flash_ptr, size);
}
+static void tdx_get_quote_completion(TdxGenerateQuoteTask *task)
+{
+ TdxGuest *tdx = task->opaque;
+ int ret;
+
+ /* Maintain the number of in-flight requests. */
+ qemu_mutex_lock(&tdx->lock);
+ tdx->num--;
+ qemu_mutex_unlock(&tdx->lock);
+
+ if (task->status_code == TDX_VP_GET_QUOTE_SUCCESS) {
+ ret = address_space_write(&address_space_memory, task->payload_gpa,
+ MEMTXATTRS_UNSPECIFIED, task->receive_buf,
+ task->receive_buf_received);
+ if (ret != MEMTX_OK) {
+ error_report("TDX: get-quote: failed to write quote data.");
+ } else {
+ task->hdr.out_len = cpu_to_le64(task->receive_buf_received);
+ }
+ }
+ task->hdr.error_code = cpu_to_le64(task->status_code);
+
+ /* Publish the response contents before marking this request completed. */
+ smp_wmb();
+ ret = address_space_write(&address_space_memory, task->buf_gpa,
+ MEMTXATTRS_UNSPECIFIED, &task->hdr,
+ TDX_GET_QUOTE_HDR_SIZE);
+ if (ret != MEMTX_OK) {
+ error_report("TDX: get-quote: failed to update GetQuote header.");
+ }
+
+ g_free(task->send_data);
+ g_free(task->receive_buf);
+ g_free(task);
+ object_unref(tdx);
+}
+
+void tdx_handle_get_quote(X86CPU *cpu, struct kvm_run *run)
+{
+ TdxGenerateQuoteTask *task;
+ struct tdx_get_quote_header hdr;
+ hwaddr buf_gpa = run->tdx.get_quote.gpa;
+ uint64_t buf_len = run->tdx.get_quote.size;
+
+ QEMU_BUILD_BUG_ON(sizeof(struct tdx_get_quote_header) != TDX_GET_QUOTE_HDR_SIZE);
+
+ run->tdx.get_quote.ret = TDG_VP_VMCALL_INVALID_OPERAND;
+
+ if (buf_len == 0) {
+ return;
+ }
+
+ if (!QEMU_IS_ALIGNED(buf_gpa, 4096) || !QEMU_IS_ALIGNED(buf_len, 4096)) {
+ run->tdx.get_quote.ret = TDG_VP_VMCALL_ALIGN_ERROR;
+ return;
+ }
+
+ if (address_space_read(&address_space_memory, buf_gpa, MEMTXATTRS_UNSPECIFIED,
+ &hdr, TDX_GET_QUOTE_HDR_SIZE) != MEMTX_OK) {
+ error_report("TDX: get-quote: failed to read GetQuote header.");
+ return;
+ }
+
+ if (le64_to_cpu(hdr.structure_version) != TDX_GET_QUOTE_STRUCTURE_VERSION) {
+ return;
+ }
+
+ /* Only safe-guard check to avoid too large buffer size. */
+ if (buf_len > TDX_GET_QUOTE_MAX_BUF_LEN ||
+ le32_to_cpu(hdr.in_len) > buf_len - TDX_GET_QUOTE_HDR_SIZE) {
+ return;
+ }
+
+ if (!tdx_guest->qg_sock_addr) {
+ hdr.error_code = cpu_to_le64(TDX_VP_GET_QUOTE_QGS_UNAVAILABLE);
+ if (address_space_write(&address_space_memory, buf_gpa,
+ MEMTXATTRS_UNSPECIFIED,
+ &hdr, TDX_GET_QUOTE_HDR_SIZE) != MEMTX_OK) {
+ error_report("TDX: failed to update GetQuote header.");
+ return;
+ }
+ run->tdx.get_quote.ret = TDG_VP_VMCALL_SUCCESS;
+ return;
+ }
+
+ qemu_mutex_lock(&tdx_guest->lock);
+ if (tdx_guest->num >= TDX_MAX_GET_QUOTE_REQUEST) {
+ qemu_mutex_unlock(&tdx_guest->lock);
+ run->tdx.get_quote.ret = TDG_VP_VMCALL_RETRY;
+ return;
+ }
+ tdx_guest->num++;
+ qemu_mutex_unlock(&tdx_guest->lock);
+
+ task = g_new(TdxGenerateQuoteTask, 1);
+ task->buf_gpa = buf_gpa;
+ task->payload_gpa = buf_gpa + TDX_GET_QUOTE_HDR_SIZE;
+ task->payload_len = buf_len - TDX_GET_QUOTE_HDR_SIZE;
+ task->hdr = hdr;
+ task->completion = tdx_get_quote_completion;
+
+ task->send_data_size = le32_to_cpu(hdr.in_len);
+ task->send_data = g_malloc(task->send_data_size);
+ task->send_data_sent = 0;
+
+ if (address_space_read(&address_space_memory, task->payload_gpa,
+ MEMTXATTRS_UNSPECIFIED, task->send_data,
+ task->send_data_size) != MEMTX_OK) {
+ goto out_free;
+ }
+
+ /* Mark the buffer in-flight. */
+ hdr.error_code = cpu_to_le64(TDX_VP_GET_QUOTE_IN_FLIGHT);
+ if (address_space_write(&address_space_memory, buf_gpa,
+ MEMTXATTRS_UNSPECIFIED,
+ &hdr, TDX_GET_QUOTE_HDR_SIZE) != MEMTX_OK) {
+ goto out_free;
+ }
+
+ task->receive_buf = g_malloc0(task->payload_len);
+ task->receive_buf_received = 0;
+ task->opaque = tdx_guest;
+
+ object_ref(tdx_guest);
+ tdx_generate_quote(task, tdx_guest->qg_sock_addr);
+ run->tdx.get_quote.ret = TDG_VP_VMCALL_SUCCESS;
+ return;
+
+out_free:
+ g_free(task->send_data);
+ g_free(task);
+}
+
void tdx_handle_get_tdvmcall_info(X86CPU *cpu, struct kvm_run *run)
{
if (run->tdx.get_tdvmcall_info.leaf != 1) {
return;
}
- run->tdx.get_tdvmcall_info.r11 = 0;
+ run->tdx.get_tdvmcall_info.r11 = TDG_VP_VMCALL_SUBFUNC_GET_QUOTE;
run->tdx.get_tdvmcall_info.r12 = 0;
run->tdx.get_tdvmcall_info.r13 = 0;
run->tdx.get_tdvmcall_info.r14 = 0;
@@ -1263,6 +1399,37 @@ static void tdx_guest_set_mrownerconfig(Object *obj, const char *value, Error **
tdx->mrownerconfig = g_strdup(value);
}
+static void tdx_guest_get_qgs(Object *obj, Visitor *v,
+ const char *name, void *opaque,
+ Error **errp)
+{
+ TdxGuest *tdx = TDX_GUEST(obj);
+
+ if (!tdx->qg_sock_addr) {
+ error_setg(errp, "quote-generation-socket is not set");
+ return;
+ }
+ visit_type_SocketAddress(v, name, &tdx->qg_sock_addr, errp);
+}
+
+static void tdx_guest_set_qgs(Object *obj, Visitor *v,
+ const char *name, void *opaque,
+ Error **errp)
+{
+ TdxGuest *tdx = TDX_GUEST(obj);
+ SocketAddress *sock = NULL;
+
+ if (!visit_type_SocketAddress(v, name, &sock, errp)) {
+ return;
+ }
+
+ if (tdx->qg_sock_addr) {
+ qapi_free_SocketAddress(tdx->qg_sock_addr);
+ }
+
+ tdx->qg_sock_addr = sock;
+}
+
/* tdx guest */
OBJECT_DEFINE_TYPE_WITH_INTERFACES(TdxGuest,
tdx_guest,
@@ -1294,6 +1461,13 @@ static void tdx_guest_init(Object *obj)
object_property_add_str(obj, "mrownerconfig",
tdx_guest_get_mrownerconfig,
tdx_guest_set_mrownerconfig);
+
+ object_property_add(obj, "quote-generation-socket", "SocketAddress",
+ tdx_guest_get_qgs,
+ tdx_guest_set_qgs,
+ NULL, NULL);
+
+ qemu_mutex_init(&tdx->lock);
}
static void tdx_guest_finalize(Object *obj)
diff --git a/target/i386/kvm/meson.build b/target/i386/kvm/meson.build
index 3f44cdedb75..2675bf89027 100644
--- a/target/i386/kvm/meson.build
+++ b/target/i386/kvm/meson.build
@@ -8,7 +8,7 @@ i386_kvm_ss.add(files(
i386_kvm_ss.add(when: 'CONFIG_XEN_EMU', if_true: files('xen-emu.c'))
-i386_kvm_ss.add(when: 'CONFIG_TDX', if_true: files('tdx.c'), if_false: files('tdx-stub.c'))
+i386_kvm_ss.add(when: 'CONFIG_TDX', if_true: files('tdx.c', 'tdx-quote-generator.c'), if_false: files('tdx-stub.c'))
i386_system_ss.add(when: 'CONFIG_HYPERV', if_true: files('hyperv.c'), if_false: files('hyperv-stub.c'))
--
2.49.0
^ permalink raw reply related [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-20 16:40 ` [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote> Paolo Bonzini
@ 2025-06-23 6:43 ` Cédric Le Goater
2025-06-23 7:03 ` Xiaoyao Li
0 siblings, 1 reply; 36+ messages in thread
From: Cédric Le Goater @ 2025-06-23 6:43 UTC (permalink / raw)
To: Paolo Bonzini, qemu-devel; +Cc: Isaku Yamahata, Chenyi Qiang, Xiaoyao Li
Hello,
On 6/20/25 18:40, Paolo Bonzini wrote:
> From: Isaku Yamahata <isaku.yamahata@intel.com>
>
> Add property "quote-generation-socket" to tdx-guest, which is a property
> of type SocketAddress to specify Quote Generation Service(QGS).
>
> On request of GetQuote, it connects to the QGS socket, read request
> data from shared guest memory, send the request data to the QGS,
> and store the response into shared guest memory, at last notify
> TD guest by interrupt.
>
> command line example:
> qemu-system-x86_64 \
> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> -machine confidential-guest-support=tdx0
>
> Note, above example uses the unix socket. It can be other types, like vsock,
> which depends on the implementation of QGS.
>
> To avoid no response from QGS server, setup a timer for the transaction.
> If timeout, make it an error and interrupt guest. Define the threshold of
> time to 30s at present, maybe change to other value if not appropriate.
>
> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> qapi/qom.json | 8 +-
> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> target/i386/kvm/tdx.h | 10 +
> target/i386/kvm/kvm.c | 3 +
> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> target/i386/kvm/tdx-stub.c | 4 +
> target/i386/kvm/tdx.c | 176 ++++++++++++++-
> target/i386/kvm/meson.build | 2 +-
> 8 files changed, 582 insertions(+), 3 deletions(-)
> create mode 100644 target/i386/kvm/tdx-quote-generator.h
> create mode 100644 target/i386/kvm/tdx-quote-generator.c
These changes broke the build on 32-bit host.
Could you please send a patch to avoid compiling TDX in such environment ?
Thanks,
C.
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 6:43 ` Cédric Le Goater
@ 2025-06-23 7:03 ` Xiaoyao Li
2025-06-23 7:44 ` Cédric Le Goater
2025-06-23 8:02 ` Daniel P. Berrangé
0 siblings, 2 replies; 36+ messages in thread
From: Xiaoyao Li @ 2025-06-23 7:03 UTC (permalink / raw)
To: Cédric Le Goater, Paolo Bonzini, qemu-devel
Cc: Isaku Yamahata, Chenyi Qiang
On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> Hello,
>
> On 6/20/25 18:40, Paolo Bonzini wrote:
>> From: Isaku Yamahata <isaku.yamahata@intel.com>
>>
>> Add property "quote-generation-socket" to tdx-guest, which is a property
>> of type SocketAddress to specify Quote Generation Service(QGS).
>>
>> On request of GetQuote, it connects to the QGS socket, read request
>> data from shared guest memory, send the request data to the QGS,
>> and store the response into shared guest memory, at last notify
>> TD guest by interrupt.
>>
>> command line example:
>> qemu-system-x86_64 \
>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
>> socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
>> -machine confidential-guest-support=tdx0
>>
>> Note, above example uses the unix socket. It can be other types, like
>> vsock,
>> which depends on the implementation of QGS.
>>
>> To avoid no response from QGS server, setup a timer for the transaction.
>> If timeout, make it an error and interrupt guest. Define the threshold of
>> time to 30s at present, maybe change to other value if not appropriate.
>>
>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
>> qapi/qom.json | 8 +-
>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
>> target/i386/kvm/tdx.h | 10 +
>> target/i386/kvm/kvm.c | 3 +
>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
>> target/i386/kvm/tdx-stub.c | 4 +
>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
>> target/i386/kvm/meson.build | 2 +-
>> 8 files changed, 582 insertions(+), 3 deletions(-)
>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
>
> These changes broke the build on 32-bit host.
>
> Could you please send a patch to avoid compiling TDX in such environment ?
Paolo is on vacation.
I would like to help, but I don't have 32-bit host environment on hand.
Do you know how to set up such environment quickly? (I tried to set up
within a 32-bit VM but the 32-bit OS is too old and I didn't get it work
to install the required package for building QEMU)
> Thanks,
>
> C.
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 7:03 ` Xiaoyao Li
@ 2025-06-23 7:44 ` Cédric Le Goater
2025-06-23 8:02 ` Daniel P. Berrangé
1 sibling, 0 replies; 36+ messages in thread
From: Cédric Le Goater @ 2025-06-23 7:44 UTC (permalink / raw)
To: Xiaoyao Li, Paolo Bonzini, qemu-devel; +Cc: Isaku Yamahata, Chenyi Qiang
On 6/23/25 09:03, Xiaoyao Li wrote:
> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
>> Hello,
>>
>> On 6/20/25 18:40, Paolo Bonzini wrote:
>>> From: Isaku Yamahata <isaku.yamahata@intel.com>
>>>
>>> Add property "quote-generation-socket" to tdx-guest, which is a property
>>> of type SocketAddress to specify Quote Generation Service(QGS).
>>>
>>> On request of GetQuote, it connects to the QGS socket, read request
>>> data from shared guest memory, send the request data to the QGS,
>>> and store the response into shared guest memory, at last notify
>>> TD guest by interrupt.
>>>
>>> command line example:
>>> qemu-system-x86_64 \
>>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation- socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
>>> -machine confidential-guest-support=tdx0
>>>
>>> Note, above example uses the unix socket. It can be other types, like vsock,
>>> which depends on the implementation of QGS.
>>>
>>> To avoid no response from QGS server, setup a timer for the transaction.
>>> If timeout, make it an error and interrupt guest. Define the threshold of
>>> time to 30s at present, maybe change to other value if not appropriate.
>>>
>>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
>>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>> ---
>>> qapi/qom.json | 8 +-
>>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
>>> target/i386/kvm/tdx.h | 10 +
>>> target/i386/kvm/kvm.c | 3 +
>>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
>>> target/i386/kvm/tdx-stub.c | 4 +
>>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
>>> target/i386/kvm/meson.build | 2 +-
>>> 8 files changed, 582 insertions(+), 3 deletions(-)
>>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
>>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
>>
>> These changes broke the build on 32-bit host.
>>
>> Could you please send a patch to avoid compiling TDX in such environment ?
>
> Paolo is on vacation.
> > I would like to help, but I don't have 32-bit host environment on hand. Do you know how to set up such environment quickly? (I tried to set up within a 32-bit VM but the 32-bit OS is too old and I didn't get it work to install the required package for building QEMU)
debian should work fine :
$ uname -a
Linux vm15 6.10.12-686-pae #1 SMP PREEMPT_DYNAMIC Debian 6.10.12-1 (2024-10-01) i686 GNU/Linux
$ cat /etc/os-release
PRETTY_NAME="Debian GNU/Linux trixie/sid"
NAME="Debian GNU/Linux"
...
Thanks,
C.
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 7:03 ` Xiaoyao Li
2025-06-23 7:44 ` Cédric Le Goater
@ 2025-06-23 8:02 ` Daniel P. Berrangé
2025-06-23 13:04 ` Stefan Hajnoczi
1 sibling, 1 reply; 36+ messages in thread
From: Daniel P. Berrangé @ 2025-06-23 8:02 UTC (permalink / raw)
To: Xiaoyao Li
Cc: Cédric Le Goater, Paolo Bonzini, qemu-devel, Isaku Yamahata,
Chenyi Qiang
On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> > Hello,
> >
> > On 6/20/25 18:40, Paolo Bonzini wrote:
> > > From: Isaku Yamahata <isaku.yamahata@intel.com>
> > >
> > > Add property "quote-generation-socket" to tdx-guest, which is a property
> > > of type SocketAddress to specify Quote Generation Service(QGS).
> > >
> > > On request of GetQuote, it connects to the QGS socket, read request
> > > data from shared guest memory, send the request data to the QGS,
> > > and store the response into shared guest memory, at last notify
> > > TD guest by interrupt.
> > >
> > > command line example:
> > > qemu-system-x86_64 \
> > > -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> > > socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> > > -machine confidential-guest-support=tdx0
> > >
> > > Note, above example uses the unix socket. It can be other types,
> > > like vsock,
> > > which depends on the implementation of QGS.
> > >
> > > To avoid no response from QGS server, setup a timer for the transaction.
> > > If timeout, make it an error and interrupt guest. Define the threshold of
> > > time to 30s at present, maybe change to other value if not appropriate.
> > >
> > > Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > > Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > > ---
> > > qapi/qom.json | 8 +-
> > > target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> > > target/i386/kvm/tdx.h | 10 +
> > > target/i386/kvm/kvm.c | 3 +
> > > target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> > > target/i386/kvm/tdx-stub.c | 4 +
> > > target/i386/kvm/tdx.c | 176 ++++++++++++++-
> > > target/i386/kvm/meson.build | 2 +-
> > > 8 files changed, 582 insertions(+), 3 deletions(-)
> > > create mode 100644 target/i386/kvm/tdx-quote-generator.h
> > > create mode 100644 target/i386/kvm/tdx-quote-generator.c
> >
> > These changes broke the build on 32-bit host.
> >
> > Could you please send a patch to avoid compiling TDX in such environment ?
>
> Paolo is on vacation.
>
> I would like to help, but I don't have 32-bit host environment on hand. Do
> you know how to set up such environment quickly? (I tried to set up within a
> 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> the required package for building QEMU)
You should be able to use QEMU's docker containers to get yourself a
Debian i386 container, on a x86_64 host.
With regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 8:02 ` Daniel P. Berrangé
@ 2025-06-23 13:04 ` Stefan Hajnoczi
2025-06-23 13:08 ` Daniel P. Berrangé
0 siblings, 1 reply; 36+ messages in thread
From: Stefan Hajnoczi @ 2025-06-23 13:04 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Xiaoyao Li, Cédric Le Goater, Paolo Bonzini, qemu-devel,
Isaku Yamahata, Chenyi Qiang
On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>
> On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> > On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> > > Hello,
> > >
> > > On 6/20/25 18:40, Paolo Bonzini wrote:
> > > > From: Isaku Yamahata <isaku.yamahata@intel.com>
> > > >
> > > > Add property "quote-generation-socket" to tdx-guest, which is a property
> > > > of type SocketAddress to specify Quote Generation Service(QGS).
> > > >
> > > > On request of GetQuote, it connects to the QGS socket, read request
> > > > data from shared guest memory, send the request data to the QGS,
> > > > and store the response into shared guest memory, at last notify
> > > > TD guest by interrupt.
> > > >
> > > > command line example:
> > > > qemu-system-x86_64 \
> > > > -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> > > > socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> > > > -machine confidential-guest-support=tdx0
> > > >
> > > > Note, above example uses the unix socket. It can be other types,
> > > > like vsock,
> > > > which depends on the implementation of QGS.
> > > >
> > > > To avoid no response from QGS server, setup a timer for the transaction.
> > > > If timeout, make it an error and interrupt guest. Define the threshold of
> > > > time to 30s at present, maybe change to other value if not appropriate.
> > > >
> > > > Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > > > Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > > > ---
> > > > qapi/qom.json | 8 +-
> > > > target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> > > > target/i386/kvm/tdx.h | 10 +
> > > > target/i386/kvm/kvm.c | 3 +
> > > > target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> > > > target/i386/kvm/tdx-stub.c | 4 +
> > > > target/i386/kvm/tdx.c | 176 ++++++++++++++-
> > > > target/i386/kvm/meson.build | 2 +-
> > > > 8 files changed, 582 insertions(+), 3 deletions(-)
> > > > create mode 100644 target/i386/kvm/tdx-quote-generator.h
> > > > create mode 100644 target/i386/kvm/tdx-quote-generator.c
> > >
> > > These changes broke the build on 32-bit host.
> > >
> > > Could you please send a patch to avoid compiling TDX in such environment ?
> >
> > Paolo is on vacation.
> >
> > I would like to help, but I don't have 32-bit host environment on hand. Do
> > you know how to set up such environment quickly? (I tried to set up within a
> > 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> > the required package for building QEMU)
>
> You should be able to use QEMU's docker containers to get yourself a
> Debian i386 container, on a x86_64 host.
The cross-i686-system (Debian) build CI job succeeded:
https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
I wonder why the CI didn't catch the issue?
Stefan
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 13:04 ` Stefan Hajnoczi
@ 2025-06-23 13:08 ` Daniel P. Berrangé
2025-06-23 18:57 ` Stefan Hajnoczi
0 siblings, 1 reply; 36+ messages in thread
From: Daniel P. Berrangé @ 2025-06-23 13:08 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: Xiaoyao Li, Cédric Le Goater, Paolo Bonzini, qemu-devel,
Isaku Yamahata, Chenyi Qiang
On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
> On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> >
> > On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> > > On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> > > > Hello,
> > > >
> > > > On 6/20/25 18:40, Paolo Bonzini wrote:
> > > > > From: Isaku Yamahata <isaku.yamahata@intel.com>
> > > > >
> > > > > Add property "quote-generation-socket" to tdx-guest, which is a property
> > > > > of type SocketAddress to specify Quote Generation Service(QGS).
> > > > >
> > > > > On request of GetQuote, it connects to the QGS socket, read request
> > > > > data from shared guest memory, send the request data to the QGS,
> > > > > and store the response into shared guest memory, at last notify
> > > > > TD guest by interrupt.
> > > > >
> > > > > command line example:
> > > > > qemu-system-x86_64 \
> > > > > -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> > > > > socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> > > > > -machine confidential-guest-support=tdx0
> > > > >
> > > > > Note, above example uses the unix socket. It can be other types,
> > > > > like vsock,
> > > > > which depends on the implementation of QGS.
> > > > >
> > > > > To avoid no response from QGS server, setup a timer for the transaction.
> > > > > If timeout, make it an error and interrupt guest. Define the threshold of
> > > > > time to 30s at present, maybe change to other value if not appropriate.
> > > > >
> > > > > Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > > > > Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > > Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > > Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > > > > ---
> > > > > qapi/qom.json | 8 +-
> > > > > target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> > > > > target/i386/kvm/tdx.h | 10 +
> > > > > target/i386/kvm/kvm.c | 3 +
> > > > > target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> > > > > target/i386/kvm/tdx-stub.c | 4 +
> > > > > target/i386/kvm/tdx.c | 176 ++++++++++++++-
> > > > > target/i386/kvm/meson.build | 2 +-
> > > > > 8 files changed, 582 insertions(+), 3 deletions(-)
> > > > > create mode 100644 target/i386/kvm/tdx-quote-generator.h
> > > > > create mode 100644 target/i386/kvm/tdx-quote-generator.c
> > > >
> > > > These changes broke the build on 32-bit host.
> > > >
> > > > Could you please send a patch to avoid compiling TDX in such environment ?
> > >
> > > Paolo is on vacation.
> > >
> > > I would like to help, but I don't have 32-bit host environment on hand. Do
> > > you know how to set up such environment quickly? (I tried to set up within a
> > > 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> > > the required package for building QEMU)
> >
> > You should be able to use QEMU's docker containers to get yourself a
> > Debian i386 container, on a x86_64 host.
>
> The cross-i686-system (Debian) build CI job succeeded:
> https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
>
> I wonder why the CI didn't catch the issue?
It didn't build the x86_64 target:
--target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
so in turn didn't build any TDX code
With regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 13:08 ` Daniel P. Berrangé
@ 2025-06-23 18:57 ` Stefan Hajnoczi
2025-06-23 20:37 ` Cédric Le Goater
0 siblings, 1 reply; 36+ messages in thread
From: Stefan Hajnoczi @ 2025-06-23 18:57 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Xiaoyao Li, Cédric Le Goater, Paolo Bonzini, qemu-devel,
Isaku Yamahata, Chenyi Qiang
On Mon, Jun 23, 2025 at 9:09 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>
> On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
> > On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> > >
> > > On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> > > > On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> > > > > Hello,
> > > > >
> > > > > On 6/20/25 18:40, Paolo Bonzini wrote:
> > > > > > From: Isaku Yamahata <isaku.yamahata@intel.com>
> > > > > >
> > > > > > Add property "quote-generation-socket" to tdx-guest, which is a property
> > > > > > of type SocketAddress to specify Quote Generation Service(QGS).
> > > > > >
> > > > > > On request of GetQuote, it connects to the QGS socket, read request
> > > > > > data from shared guest memory, send the request data to the QGS,
> > > > > > and store the response into shared guest memory, at last notify
> > > > > > TD guest by interrupt.
> > > > > >
> > > > > > command line example:
> > > > > > qemu-system-x86_64 \
> > > > > > -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> > > > > > socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> > > > > > -machine confidential-guest-support=tdx0
> > > > > >
> > > > > > Note, above example uses the unix socket. It can be other types,
> > > > > > like vsock,
> > > > > > which depends on the implementation of QGS.
> > > > > >
> > > > > > To avoid no response from QGS server, setup a timer for the transaction.
> > > > > > If timeout, make it an error and interrupt guest. Define the threshold of
> > > > > > time to 30s at present, maybe change to other value if not appropriate.
> > > > > >
> > > > > > Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > > > > > Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > > > Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > > > > > Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > > Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > > Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > > > > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > > > > > ---
> > > > > > qapi/qom.json | 8 +-
> > > > > > target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> > > > > > target/i386/kvm/tdx.h | 10 +
> > > > > > target/i386/kvm/kvm.c | 3 +
> > > > > > target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> > > > > > target/i386/kvm/tdx-stub.c | 4 +
> > > > > > target/i386/kvm/tdx.c | 176 ++++++++++++++-
> > > > > > target/i386/kvm/meson.build | 2 +-
> > > > > > 8 files changed, 582 insertions(+), 3 deletions(-)
> > > > > > create mode 100644 target/i386/kvm/tdx-quote-generator.h
> > > > > > create mode 100644 target/i386/kvm/tdx-quote-generator.c
> > > > >
> > > > > These changes broke the build on 32-bit host.
> > > > >
> > > > > Could you please send a patch to avoid compiling TDX in such environment ?
> > > >
> > > > Paolo is on vacation.
> > > >
> > > > I would like to help, but I don't have 32-bit host environment on hand. Do
> > > > you know how to set up such environment quickly? (I tried to set up within a
> > > > 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> > > > the required package for building QEMU)
> > >
> > > You should be able to use QEMU's docker containers to get yourself a
> > > Debian i386 container, on a x86_64 host.
> >
> > The cross-i686-system (Debian) build CI job succeeded:
> > https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
> >
> > I wonder why the CI didn't catch the issue?
>
> It didn't build the x86_64 target:
>
> --target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
>
> so in turn didn't build any TDX code
Here are the targets that were built by the CI job:
target list : avr-softmmu m68k-softmmu microblazeel-softmmu
or1k-softmmu rx-softmmu sh4eb-softmmu tricore-softmmu xtensaeb-softmmu
64-bit targets are not supported on 32-bit hosts since commit
acce728cbc6c ("meson: Disallow 64-bit on 32-bit emulation"). I don't
think the x86_64 target can be built on 32-bit hosts.
But notice that i386-softmmu is missing from the target list. That
could be why the CI job succeeded.
Cédric: What were your ./configure options?
Stefan
>
> With regards,
> Daniel
> --
> |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org -o- https://fstop138.berrange.com :|
> |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
>
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 18:57 ` Stefan Hajnoczi
@ 2025-06-23 20:37 ` Cédric Le Goater
2025-06-24 15:03 ` Stefan Hajnoczi
0 siblings, 1 reply; 36+ messages in thread
From: Cédric Le Goater @ 2025-06-23 20:37 UTC (permalink / raw)
To: Stefan Hajnoczi, Daniel P. Berrangé
Cc: Xiaoyao Li, Paolo Bonzini, qemu-devel, Isaku Yamahata,
Chenyi Qiang
On 6/23/25 20:57, Stefan Hajnoczi wrote:
> On Mon, Jun 23, 2025 at 9:09 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>>
>> On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
>>> On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>>>>
>>>> On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
>>>>> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
>>>>>> Hello,
>>>>>>
>>>>>> On 6/20/25 18:40, Paolo Bonzini wrote:
>>>>>>> From: Isaku Yamahata <isaku.yamahata@intel.com>
>>>>>>>
>>>>>>> Add property "quote-generation-socket" to tdx-guest, which is a property
>>>>>>> of type SocketAddress to specify Quote Generation Service(QGS).
>>>>>>>
>>>>>>> On request of GetQuote, it connects to the QGS socket, read request
>>>>>>> data from shared guest memory, send the request data to the QGS,
>>>>>>> and store the response into shared guest memory, at last notify
>>>>>>> TD guest by interrupt.
>>>>>>>
>>>>>>> command line example:
>>>>>>> qemu-system-x86_64 \
>>>>>>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
>>>>>>> socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
>>>>>>> -machine confidential-guest-support=tdx0
>>>>>>>
>>>>>>> Note, above example uses the unix socket. It can be other types,
>>>>>>> like vsock,
>>>>>>> which depends on the implementation of QGS.
>>>>>>>
>>>>>>> To avoid no response from QGS server, setup a timer for the transaction.
>>>>>>> If timeout, make it an error and interrupt guest. Define the threshold of
>>>>>>> time to 30s at present, maybe change to other value if not appropriate.
>>>>>>>
>>>>>>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
>>>>>>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>>>>>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>>>>>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>>>>>> ---
>>>>>>> qapi/qom.json | 8 +-
>>>>>>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
>>>>>>> target/i386/kvm/tdx.h | 10 +
>>>>>>> target/i386/kvm/kvm.c | 3 +
>>>>>>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
>>>>>>> target/i386/kvm/tdx-stub.c | 4 +
>>>>>>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
>>>>>>> target/i386/kvm/meson.build | 2 +-
>>>>>>> 8 files changed, 582 insertions(+), 3 deletions(-)
>>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
>>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
>>>>>>
>>>>>> These changes broke the build on 32-bit host.
>>>>>>
>>>>>> Could you please send a patch to avoid compiling TDX in such environment ?
>>>>>
>>>>> Paolo is on vacation.
>>>>>
>>>>> I would like to help, but I don't have 32-bit host environment on hand. Do
>>>>> you know how to set up such environment quickly? (I tried to set up within a
>>>>> 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
>>>>> the required package for building QEMU)
>>>>
>>>> You should be able to use QEMU's docker containers to get yourself a
>>>> Debian i386 container, on a x86_64 host.
>>>
>>> The cross-i686-system (Debian) build CI job succeeded:
>>> https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
>>>
>>> I wonder why the CI didn't catch the issue?
>>
>> It didn't build the x86_64 target:
>>
>> --target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
>>
>> so in turn didn't build any TDX code
>
> Here are the targets that were built by the CI job:
>
> target list : avr-softmmu m68k-softmmu microblazeel-softmmu
> or1k-softmmu rx-softmmu sh4eb-softmmu tricore-softmmu xtensaeb-softmmu
>
> 64-bit targets are not supported on 32-bit hosts since commit
> acce728cbc6c ("meson: Disallow 64-bit on 32-bit emulation"). I don't
> think the x86_64 target can be built on 32-bit hosts.
>
> But notice that i386-softmmu is missing from the target list. That
> could be why the CI job succeeded.
>
> Cédric: What were your ./configure options?
Just :
--target-list=arm-softmmu,ppc-softmmu,i386-softmmu
C.
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-23 20:37 ` Cédric Le Goater
@ 2025-06-24 15:03 ` Stefan Hajnoczi
2025-06-24 15:08 ` Daniel P. Berrangé
0 siblings, 1 reply; 36+ messages in thread
From: Stefan Hajnoczi @ 2025-06-24 15:03 UTC (permalink / raw)
To: Cédric Le Goater
Cc: Daniel P. Berrangé, Xiaoyao Li, Paolo Bonzini, qemu-devel,
Isaku Yamahata, Chenyi Qiang
On Mon, Jun 23, 2025 at 4:37 PM Cédric Le Goater <clg@kaod.org> wrote:
>
> On 6/23/25 20:57, Stefan Hajnoczi wrote:
> > On Mon, Jun 23, 2025 at 9:09 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> >>
> >> On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
> >>> On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> >>>>
> >>>> On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> >>>>> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> >>>>>> Hello,
> >>>>>>
> >>>>>> On 6/20/25 18:40, Paolo Bonzini wrote:
> >>>>>>> From: Isaku Yamahata <isaku.yamahata@intel.com>
> >>>>>>>
> >>>>>>> Add property "quote-generation-socket" to tdx-guest, which is a property
> >>>>>>> of type SocketAddress to specify Quote Generation Service(QGS).
> >>>>>>>
> >>>>>>> On request of GetQuote, it connects to the QGS socket, read request
> >>>>>>> data from shared guest memory, send the request data to the QGS,
> >>>>>>> and store the response into shared guest memory, at last notify
> >>>>>>> TD guest by interrupt.
> >>>>>>>
> >>>>>>> command line example:
> >>>>>>> qemu-system-x86_64 \
> >>>>>>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> >>>>>>> socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> >>>>>>> -machine confidential-guest-support=tdx0
> >>>>>>>
> >>>>>>> Note, above example uses the unix socket. It can be other types,
> >>>>>>> like vsock,
> >>>>>>> which depends on the implementation of QGS.
> >>>>>>>
> >>>>>>> To avoid no response from QGS server, setup a timer for the transaction.
> >>>>>>> If timeout, make it an error and interrupt guest. Define the threshold of
> >>>>>>> time to 30s at present, maybe change to other value if not appropriate.
> >>>>>>>
> >>>>>>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> >>>>>>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> >>>>>>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> >>>>>>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> >>>>>>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> >>>>>>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> >>>>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> >>>>>>> ---
> >>>>>>> qapi/qom.json | 8 +-
> >>>>>>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> >>>>>>> target/i386/kvm/tdx.h | 10 +
> >>>>>>> target/i386/kvm/kvm.c | 3 +
> >>>>>>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> >>>>>>> target/i386/kvm/tdx-stub.c | 4 +
> >>>>>>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
> >>>>>>> target/i386/kvm/meson.build | 2 +-
> >>>>>>> 8 files changed, 582 insertions(+), 3 deletions(-)
> >>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
> >>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
> >>>>>>
> >>>>>> These changes broke the build on 32-bit host.
> >>>>>>
> >>>>>> Could you please send a patch to avoid compiling TDX in such environment ?
> >>>>>
> >>>>> Paolo is on vacation.
> >>>>>
> >>>>> I would like to help, but I don't have 32-bit host environment on hand. Do
> >>>>> you know how to set up such environment quickly? (I tried to set up within a
> >>>>> 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> >>>>> the required package for building QEMU)
> >>>>
> >>>> You should be able to use QEMU's docker containers to get yourself a
> >>>> Debian i386 container, on a x86_64 host.
> >>>
> >>> The cross-i686-system (Debian) build CI job succeeded:
> >>> https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
> >>>
> >>> I wonder why the CI didn't catch the issue?
> >>
> >> It didn't build the x86_64 target:
> >>
> >> --target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
> >>
> >> so in turn didn't build any TDX code
> >
> > Here are the targets that were built by the CI job:
> >
> > target list : avr-softmmu m68k-softmmu microblazeel-softmmu
> > or1k-softmmu rx-softmmu sh4eb-softmmu tricore-softmmu xtensaeb-softmmu
> >
> > 64-bit targets are not supported on 32-bit hosts since commit
> > acce728cbc6c ("meson: Disallow 64-bit on 32-bit emulation"). I don't
> > think the x86_64 target can be built on 32-bit hosts.
> >
> > But notice that i386-softmmu is missing from the target list. That
> > could be why the CI job succeeded.
> >
> > Cédric: What were your ./configure options?
>
> Just :
>
> --target-list=arm-softmmu,ppc-softmmu,i386-softmmu
cross-i686-tci builds i386-softmmu on debian-i686-cross with
--disable-kvm, but the TDX code isn't compiled:
https://gitlab.com/qemu-project/qemu/-/jobs/10440424393
A new CI job could be added to cover TDX on 32-bit hosts if you think
it's worthwhile. Maybe it's a corner case and not worth spending CI
cycles on?
Stefan
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-24 15:03 ` Stefan Hajnoczi
@ 2025-06-24 15:08 ` Daniel P. Berrangé
2025-06-24 15:22 ` Cédric Le Goater
0 siblings, 1 reply; 36+ messages in thread
From: Daniel P. Berrangé @ 2025-06-24 15:08 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: Cédric Le Goater, Xiaoyao Li, Paolo Bonzini, qemu-devel,
Isaku Yamahata, Chenyi Qiang
On Tue, Jun 24, 2025 at 11:03:19AM -0400, Stefan Hajnoczi wrote:
> On Mon, Jun 23, 2025 at 4:37 PM Cédric Le Goater <clg@kaod.org> wrote:
> >
> > On 6/23/25 20:57, Stefan Hajnoczi wrote:
> > > On Mon, Jun 23, 2025 at 9:09 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> > >>
> > >> On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
> > >>> On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
> > >>>>
> > >>>> On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
> > >>>>> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
> > >>>>>> Hello,
> > >>>>>>
> > >>>>>> On 6/20/25 18:40, Paolo Bonzini wrote:
> > >>>>>>> From: Isaku Yamahata <isaku.yamahata@intel.com>
> > >>>>>>>
> > >>>>>>> Add property "quote-generation-socket" to tdx-guest, which is a property
> > >>>>>>> of type SocketAddress to specify Quote Generation Service(QGS).
> > >>>>>>>
> > >>>>>>> On request of GetQuote, it connects to the QGS socket, read request
> > >>>>>>> data from shared guest memory, send the request data to the QGS,
> > >>>>>>> and store the response into shared guest memory, at last notify
> > >>>>>>> TD guest by interrupt.
> > >>>>>>>
> > >>>>>>> command line example:
> > >>>>>>> qemu-system-x86_64 \
> > >>>>>>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
> > >>>>>>> socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
> > >>>>>>> -machine confidential-guest-support=tdx0
> > >>>>>>>
> > >>>>>>> Note, above example uses the unix socket. It can be other types,
> > >>>>>>> like vsock,
> > >>>>>>> which depends on the implementation of QGS.
> > >>>>>>>
> > >>>>>>> To avoid no response from QGS server, setup a timer for the transaction.
> > >>>>>>> If timeout, make it an error and interrupt guest. Define the threshold of
> > >>>>>>> time to 30s at present, maybe change to other value if not appropriate.
> > >>>>>>>
> > >>>>>>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> > >>>>>>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > >>>>>>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
> > >>>>>>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > >>>>>>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > >>>>>>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
> > >>>>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> > >>>>>>> ---
> > >>>>>>> qapi/qom.json | 8 +-
> > >>>>>>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
> > >>>>>>> target/i386/kvm/tdx.h | 10 +
> > >>>>>>> target/i386/kvm/kvm.c | 3 +
> > >>>>>>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
> > >>>>>>> target/i386/kvm/tdx-stub.c | 4 +
> > >>>>>>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
> > >>>>>>> target/i386/kvm/meson.build | 2 +-
> > >>>>>>> 8 files changed, 582 insertions(+), 3 deletions(-)
> > >>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
> > >>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
> > >>>>>>
> > >>>>>> These changes broke the build on 32-bit host.
> > >>>>>>
> > >>>>>> Could you please send a patch to avoid compiling TDX in such environment ?
> > >>>>>
> > >>>>> Paolo is on vacation.
> > >>>>>
> > >>>>> I would like to help, but I don't have 32-bit host environment on hand. Do
> > >>>>> you know how to set up such environment quickly? (I tried to set up within a
> > >>>>> 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
> > >>>>> the required package for building QEMU)
> > >>>>
> > >>>> You should be able to use QEMU's docker containers to get yourself a
> > >>>> Debian i386 container, on a x86_64 host.
> > >>>
> > >>> The cross-i686-system (Debian) build CI job succeeded:
> > >>> https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
> > >>>
> > >>> I wonder why the CI didn't catch the issue?
> > >>
> > >> It didn't build the x86_64 target:
> > >>
> > >> --target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
> > >>
> > >> so in turn didn't build any TDX code
> > >
> > > Here are the targets that were built by the CI job:
> > >
> > > target list : avr-softmmu m68k-softmmu microblazeel-softmmu
> > > or1k-softmmu rx-softmmu sh4eb-softmmu tricore-softmmu xtensaeb-softmmu
> > >
> > > 64-bit targets are not supported on 32-bit hosts since commit
> > > acce728cbc6c ("meson: Disallow 64-bit on 32-bit emulation"). I don't
> > > think the x86_64 target can be built on 32-bit hosts.
> > >
> > > But notice that i386-softmmu is missing from the target list. That
> > > could be why the CI job succeeded.
> > >
> > > Cédric: What were your ./configure options?
> >
> > Just :
> >
> > --target-list=arm-softmmu,ppc-softmmu,i386-softmmu
>
> cross-i686-tci builds i386-softmmu on debian-i686-cross with
> --disable-kvm, but the TDX code isn't compiled:
> https://gitlab.com/qemu-project/qemu/-/jobs/10440424393
>
> A new CI job could be added to cover TDX on 32-bit hosts if you think
> it's worthwhile. Maybe it's a corner case and not worth spending CI
> cycles on?
We should not set any expectation of support for TDX on 32-bit host
or guest emulator. IMHO the only right answer here is to disable
build of TDX for anything other than x86_64-softmmu w/ KVM.
With regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 36+ messages in thread* Re: [PULL 24/24] i386/tdx: handle TDG.VP.VMCALL<GetQuote>
2025-06-24 15:08 ` Daniel P. Berrangé
@ 2025-06-24 15:22 ` Cédric Le Goater
0 siblings, 0 replies; 36+ messages in thread
From: Cédric Le Goater @ 2025-06-24 15:22 UTC (permalink / raw)
To: Daniel P. Berrangé, Stefan Hajnoczi
Cc: Xiaoyao Li, Paolo Bonzini, qemu-devel, Isaku Yamahata,
Chenyi Qiang
On 6/24/25 17:08, Daniel P. Berrangé wrote:
> On Tue, Jun 24, 2025 at 11:03:19AM -0400, Stefan Hajnoczi wrote:
>> On Mon, Jun 23, 2025 at 4:37 PM Cédric Le Goater <clg@kaod.org> wrote:
>>>
>>> On 6/23/25 20:57, Stefan Hajnoczi wrote:
>>>> On Mon, Jun 23, 2025 at 9:09 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>>>>>
>>>>> On Mon, Jun 23, 2025 at 09:04:33AM -0400, Stefan Hajnoczi wrote:
>>>>>> On Mon, Jun 23, 2025 at 4:04 AM Daniel P. Berrangé <berrange@redhat.com> wrote:
>>>>>>>
>>>>>>> On Mon, Jun 23, 2025 at 03:03:19PM +0800, Xiaoyao Li wrote:
>>>>>>>> On 6/23/2025 2:43 PM, Cédric Le Goater wrote:
>>>>>>>>> Hello,
>>>>>>>>>
>>>>>>>>> On 6/20/25 18:40, Paolo Bonzini wrote:
>>>>>>>>>> From: Isaku Yamahata <isaku.yamahata@intel.com>
>>>>>>>>>>
>>>>>>>>>> Add property "quote-generation-socket" to tdx-guest, which is a property
>>>>>>>>>> of type SocketAddress to specify Quote Generation Service(QGS).
>>>>>>>>>>
>>>>>>>>>> On request of GetQuote, it connects to the QGS socket, read request
>>>>>>>>>> data from shared guest memory, send the request data to the QGS,
>>>>>>>>>> and store the response into shared guest memory, at last notify
>>>>>>>>>> TD guest by interrupt.
>>>>>>>>>>
>>>>>>>>>> command line example:
>>>>>>>>>> qemu-system-x86_64 \
>>>>>>>>>> -object '{"qom-type":"tdx-guest","id":"tdx0","quote-generation-
>>>>>>>>>> socket":{"type":"unix", "path":"/var/run/tdx-qgs/qgs.socket"}}' \
>>>>>>>>>> -machine confidential-guest-support=tdx0
>>>>>>>>>>
>>>>>>>>>> Note, above example uses the unix socket. It can be other types,
>>>>>>>>>> like vsock,
>>>>>>>>>> which depends on the implementation of QGS.
>>>>>>>>>>
>>>>>>>>>> To avoid no response from QGS server, setup a timer for the transaction.
>>>>>>>>>> If timeout, make it an error and interrupt guest. Define the threshold of
>>>>>>>>>> time to 30s at present, maybe change to other value if not appropriate.
>>>>>>>>>>
>>>>>>>>>> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
>>>>>>>>>> Co-developed-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>>>>>>>>> Signed-off-by: Chenyi Qiang <chenyi.qiang@intel.com>
>>>>>>>>>> Co-developed-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>>>>> Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>>>>> Tested-by: Xiaoyao Li <xiaoyao.li@intel.com>
>>>>>>>>>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>>>>>>>>>> ---
>>>>>>>>>> qapi/qom.json | 8 +-
>>>>>>>>>> target/i386/kvm/tdx-quote-generator.h | 82 +++++++
>>>>>>>>>> target/i386/kvm/tdx.h | 10 +
>>>>>>>>>> target/i386/kvm/kvm.c | 3 +
>>>>>>>>>> target/i386/kvm/tdx-quote-generator.c | 300 ++++++++++++++++++++++++++
>>>>>>>>>> target/i386/kvm/tdx-stub.c | 4 +
>>>>>>>>>> target/i386/kvm/tdx.c | 176 ++++++++++++++-
>>>>>>>>>> target/i386/kvm/meson.build | 2 +-
>>>>>>>>>> 8 files changed, 582 insertions(+), 3 deletions(-)
>>>>>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.h
>>>>>>>>>> create mode 100644 target/i386/kvm/tdx-quote-generator.c
>>>>>>>>>
>>>>>>>>> These changes broke the build on 32-bit host.
>>>>>>>>>
>>>>>>>>> Could you please send a patch to avoid compiling TDX in such environment ?
>>>>>>>>
>>>>>>>> Paolo is on vacation.
>>>>>>>>
>>>>>>>> I would like to help, but I don't have 32-bit host environment on hand. Do
>>>>>>>> you know how to set up such environment quickly? (I tried to set up within a
>>>>>>>> 32-bit VM but the 32-bit OS is too old and I didn't get it work to install
>>>>>>>> the required package for building QEMU)
>>>>>>>
>>>>>>> You should be able to use QEMU's docker containers to get yourself a
>>>>>>> Debian i386 container, on a x86_64 host.
>>>>>>
>>>>>> The cross-i686-system (Debian) build CI job succeeded:
>>>>>> https://gitlab.com/qemu-project/qemu/-/jobs/10423776600
>>>>>>
>>>>>> I wonder why the CI didn't catch the issue?
>>>>>
>>>>> It didn't build the x86_64 target:
>>>>>
>>>>> --target-list-exclude="arm-softmmu i386-softmmu microblaze-softmmu mips-softmmu mipsel-softmmu mips64-softmmu ppc-softmmu riscv32-softmmu sh4-softmmu sparc-softmmu xtensa-softmmu $CROSS_SKIP_TARGETS"
>>>>>
>>>>> so in turn didn't build any TDX code
>>>>
>>>> Here are the targets that were built by the CI job:
>>>>
>>>> target list : avr-softmmu m68k-softmmu microblazeel-softmmu
>>>> or1k-softmmu rx-softmmu sh4eb-softmmu tricore-softmmu xtensaeb-softmmu
>>>>
>>>> 64-bit targets are not supported on 32-bit hosts since commit
>>>> acce728cbc6c ("meson: Disallow 64-bit on 32-bit emulation"). I don't
>>>> think the x86_64 target can be built on 32-bit hosts.
>>>>
>>>> But notice that i386-softmmu is missing from the target list. That
>>>> could be why the CI job succeeded.
>>>>
>>>> Cédric: What were your ./configure options?
>>>
>>> Just :
>>>
>>> --target-list=arm-softmmu,ppc-softmmu,i386-softmmu
>>
>> cross-i686-tci builds i386-softmmu on debian-i686-cross with
>> --disable-kvm, but the TDX code isn't compiled:
>> https://gitlab.com/qemu-project/qemu/-/jobs/10440424393
>>
>> A new CI job could be added to cover TDX on 32-bit hosts if you think
>> it's worthwhile. Maybe it's a corner case and not worth spending CI
>> cycles on?
>
> We should not set any expectation of support for TDX on 32-bit host
> or guest emulator. IMHO the only right answer here is to disable
> build of TDX for anything other than x86_64-softmmu w/ KVM.
yes. This is what I was reminding the developers.
Thanks,
C.
^ permalink raw reply [flat|nested] 36+ messages in thread