On 7/17/26 19:30, [email protected] wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s)
to consider:
- [High] The memory layout of `union hv_device_domain_id` is inverted compared
to the expected Hyper-V ABI, resulting in corrupted device domain IDs being
passed to the hypervisor.
--
--- Patch [5]: [PATCH V4 5/9] mshv: Import data structs around device passthru
from hyperv headers ---
commit 02ed2cd292e85f0b6e754b6a29df473f95394315
Author: Mukesh R <[email protected]>
mshv: Import data structs around device passthru from hyperv headers
This commit copies and imports definitions and declarations from Hyper-V
public headers. These additions are related to creating domains in the
hypervisor and attaching devices to them.
diff --git a/include/hyperv/hvhdk_mini.h b/include/hyperv/hvhdk_mini.h
index b4cb2fa26e9bc..90497635c3cbd 100644
--- a/include/hyperv/hvhdk_mini.h
+++ b/include/hyperv/hvhdk_mini.h
[ ... ]
+/* 3 domain types: stage 1, stage 2, and SOC */
+#define HV_DEVICE_DOMAIN_TYPE_S2 0 /* HV_DEVICE_DOMAIN_ID_TYPE_S2 */
+#define HV_DEVICE_DOMAIN_TYPE_S1 1 /* HV_DEVICE_DOMAIN_ID_TYPE_S1 */
+#define HV_DEVICE_DOMAIN_TYPE_SOC 2 /* HV_DEVICE_DOMAIN_ID_TYPE_SOC */
+
+/* ID for stage 2 default domain and NULL domain */
+#define HV_DEVICE_DOMAIN_ID_S2_DEFAULT 0
+#define HV_DEVICE_DOMAIN_ID_S2_NULL 0xFFFFFFFFULL
+
+union hv_device_domain_id {
+ u64 as_uint64;
+ struct {
+ u32 type : 4;
+ u32 reserved : 28;
+ u32 id;
+ };
+} __packed;
[Severity: High]
Is the memory layout of union hv_device_domain_id inverted compared to the
expected Hyper-V ABI?
No.
The macro HV_DEVICE_DOMAIN_ID_S2_NULL is defined as 0xFFFFFFFFULL, which
places 0xFFFFFFFF in the lower 32 bits and 0 in the upper 32 bits. Given
that Stage 2 (S2) domains have a type of 0, this macro implies that the id
should be in the lower 32 bits and the type should be in the upper 32 bits.
HV_DEVICE_DOMAIN_ID_S2_NULL is never assigned to any data struct of
type "union hv_device_domain_id".
However, union hv_device_domain_id defines type as the first bitfield
(occupying the lower 32 bits) and id as the second (occupying the upper 32
bits).
Under this layout, casting 0xFFFFFFFFULL evaluates to type = 15 and id = 0,
instead of the expected type = 0 and id = 0xFFFFFFFF. Could this cause
malformed domain IDs to be sent to the hypervisor?