On 2025/9/18 16:57, Zhenzhong Duan wrote:
This adds PASID cache sync for RID_PASID, non-RID_PASID isn't supported.
Adds an new entry VTDPASIDCacheEntry in VTDAddressSpace to cache the pasid
entry and track PASID usage and future PASID tagged DMA address translation
support in vIOMMU.
When guest triggers pasid cache invalidation, QEMU will capture it and
update or invalidate pasid cache.
vIOMMU emulator could figure out the reason by fetching latest guest pasid
entry in memory and compare it with cached PASID entry if it's valid.
Signed-off-by: Yi Liu <[email protected]>
Signed-off-by: Zhenzhong Duan <[email protected]>
---
hw/i386/intel_iommu_internal.h | 19 +++-
include/hw/i386/intel_iommu.h | 6 ++
hw/i386/intel_iommu.c | 157 ++++++++++++++++++++++++++++++---
hw/i386/trace-events | 3 +
4 files changed, 173 insertions(+), 12 deletions(-)
diff --git a/hw/i386/intel_iommu_internal.h b/hw/i386/intel_iommu_internal.h
index 9cdc8d5dbb..d400bcee21 100644
--- a/hw/i386/intel_iommu_internal.h
+++ b/hw/i386/intel_iommu_internal.h
@@ -316,6 +316,7 @@ typedef enum VTDFaultReason {
* request while disabled */
VTD_FR_IR_SID_ERR = 0x26, /* Invalid Source-ID */
+ VTD_FR_RTADDR_INV_TTM = 0x31, /* Invalid TTM in RTADDR */
/* PASID directory entry access failure */
VTD_FR_PASID_DIR_ACCESS_ERR = 0x50,
/* The Present(P) field of pasid directory entry is 0 */
@@ -493,6 +494,15 @@ typedef union VTDInvDesc VTDInvDesc;
#define VTD_INV_DESC_PIOTLB_RSVD_VAL0 0xfff000000000f1c0ULL
#define VTD_INV_DESC_PIOTLB_RSVD_VAL1 0xf80ULL
+/* PASID-cache Invalidate Descriptor (pc_inv_dsc) fields */
+#define VTD_INV_DESC_PASIDC_G(x) extract64((x)->val[0], 4, 2)
+#define VTD_INV_DESC_PASIDC_G_DSI 0
+#define VTD_INV_DESC_PASIDC_G_PASID_SI 1
+#define VTD_INV_DESC_PASIDC_G_GLOBAL 3
+#define VTD_INV_DESC_PASIDC_DID(x) extract64((x)->val[0], 16, 16)
+#define VTD_INV_DESC_PASIDC_PASID(x) extract64((x)->val[0], 32, 20)
+#define VTD_INV_DESC_PASIDC_RSVD_VAL0 0xfff000000000f1c0ULL
+
/* Information about page-selective IOTLB invalidate */
struct VTDIOTLBPageInvInfo {
uint16_t domain_id;
@@ -552,6 +562,13 @@ typedef struct VTDRootEntry VTDRootEntry;
#define VTD_SM_CONTEXT_ENTRY_RSVD_VAL0(aw) (0x1e0ULL | ~VTD_HAW_MASK(aw))
#define VTD_SM_CONTEXT_ENTRY_RSVD_VAL1 0xffffffffffe00000ULL
+typedef struct VTDPASIDCacheInfo {
+ uint8_t type;
+ uint16_t did;
+ uint32_t pasid;
+ bool reset;
+} VTDPASIDCacheInfo;
+
/* PASID Table Related Definitions */
#define VTD_PASID_DIR_BASE_ADDR_MASK (~0xfffULL)
#define VTD_PASID_TABLE_BASE_ADDR_MASK (~0xfffULL)
@@ -573,7 +590,7 @@ typedef struct VTDRootEntry VTDRootEntry;
#define VTD_SM_PASID_ENTRY_PT (4ULL << 6)
#define VTD_SM_PASID_ENTRY_AW 7ULL /* Adjusted guest-address-width */
-#define VTD_SM_PASID_ENTRY_DID(val) ((val) & VTD_DOMAIN_ID_MASK)
+#define VTD_SM_PASID_ENTRY_DID(x) extract64((x)->val[1], 0, 16)
#define VTD_SM_PASID_ENTRY_FSPM 3ULL
#define VTD_SM_PASID_ENTRY_FSPTPTR (~0xfffULL)
diff --git a/include/hw/i386/intel_iommu.h b/include/hw/i386/intel_iommu.h
index 3351892da0..ff01e5c82d 100644
--- a/include/hw/i386/intel_iommu.h
+++ b/include/hw/i386/intel_iommu.h
@@ -95,6 +95,11 @@ struct VTDPASIDEntry {
uint64_t val[8];
};
+typedef struct VTDPASIDCacheEntry {
+ struct VTDPASIDEntry pasid_entry;
+ bool valid;
+} VTDPASIDCacheEntry;
+
struct VTDAddressSpace {
PCIBus *bus;
uint8_t devfn;
@@ -107,6 +112,7 @@ struct VTDAddressSpace {
MemoryRegion iommu_ir_fault; /* Interrupt region for catching fault */
IntelIOMMUState *iommu_state;
VTDContextCacheEntry context_cache_entry;
+ VTDPASIDCacheEntry pasid_cache_entry;
QLIST_ENTRY(VTDAddressSpace) next;
/* Superset of notifier flags that this address space has */
IOMMUNotifierFlag notifier_flags;
diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index d37d47115a..24061f6dc6 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -1614,7 +1614,7 @@ static uint16_t vtd_get_domain_id(IntelIOMMUState *s,
if (s->root_scalable) {
vtd_ce_get_pasid_entry(s, ce, &pe, pasid);
- return VTD_SM_PASID_ENTRY_DID(pe.val[1]);
+ return VTD_SM_PASID_ENTRY_DID(&pe);
}
return VTD_CONTEXT_ENTRY_DID(ce->hi);
@@ -3074,6 +3074,144 @@ static bool vtd_process_piotlb_desc(IntelIOMMUState *s,
return true;
}
+static inline int vtd_dev_get_pe_from_pasid(VTDAddressSpace *vtd_as,
+ VTDPASIDEntry *pe)
+{
+ IntelIOMMUState *s = vtd_as->iommu_state;
+ VTDContextEntry ce;
+ int ret;
+
+ if (!s->root_scalable) {
+ return -VTD_FR_RTADDR_INV_TTM;
+ }
+
+ ret = vtd_dev_to_context_entry(s, pci_bus_num(vtd_as->bus), vtd_as->devfn,
+ &ce);
+ if (ret) {
+ return ret;
+ }
+
+ return vtd_ce_get_pasid_entry(s, &ce, pe, vtd_as->pasid);
+}
+
+/*
+ * For each IOMMUFD backed device, update or invalidate pasid cache based on
+ * the value in memory.
+ */
+static void vtd_pasid_cache_sync_locked(gpointer key, gpointer value,
+ gpointer user_data)
+{
+ VTDPASIDCacheInfo *pc_info = user_data;
+ VTDAddressSpace *vtd_as = value;
+ VTDPASIDCacheEntry *pc_entry = &vtd_as->pasid_cache_entry;
+ VTDPASIDEntry pe;
+ uint16_t did;
+
+ /* Ignore emulated device or legacy VFIO backed device */
+ if (!vtd_find_hiod_iommufd(vtd_as)) {
+ return;
+ }
+
+ /* non-RID_PASID isn't supported yet */
+ assert(vtd_as->pasid == PCI_NO_PASID);
+
+ if (vtd_dev_get_pe_from_pasid(vtd_as, &pe)) {
+ /*
+ * No valid pasid entry in guest memory. e.g. pasid entry was modified
+ * to be either all-zero or non-present. Either case means existing
+ * pasid cache should be invalidated.
+ */
+ pc_entry->valid = false;
+ return;
+ }
+
+ /*
+ * VTD_INV_DESC_PASIDC_G_DSI and VTD_INV_DESC_PASIDC_G_PASID_SI require
+ * DID check. If DID doesn't match the value in cache or memory, then
+ * it's not a pasid entry we want to invalidate.
I think comparing DID applies to the case in which pc_entry->valid is
true. If pc_entry->valid is false, this means no cached pc_entry yet. If
pe in guest memory is valid, the pc_entry should be updated/set hence
the bind_pasid operation (added in later patch) would be conducted.
+ */
+ switch (pc_info->type) {
+ case VTD_INV_DESC_PASIDC_G_PASID_SI:
+ case VTD_INV_DESC_PASIDC_G_DSI:
+ if (pc_entry->valid) {
+ did = VTD_SM_PASID_ENTRY_DID(&pc_entry->pasid_entry);
+ if (pc_info->did == did) {
+ break;
+ }
+ }
+ did = VTD_SM_PASID_ENTRY_DID(&pe);
+ if (pc_info->did == did) {
+ break;
+ }
+ return;
+ }
+
+ pc_entry->pasid_entry = pe;
+ pc_entry->valid = true;
+}
+
+static void vtd_pasid_cache_sync(IntelIOMMUState *s, VTDPASIDCacheInfo
*pc_info)
+{
+ if (!s->fsts || !s->root_scalable || !s->dmar_enabled) {
+ return;
+ }
+
+ vtd_iommu_lock(s);
+ g_hash_table_foreach(s->vtd_address_spaces, vtd_pasid_cache_sync_locked,
+ pc_info);
+ vtd_iommu_unlock(s);
+}
+
+static bool vtd_process_pasid_desc(IntelIOMMUState *s,
+ VTDInvDesc *inv_desc)
+{
+ uint16_t did;
+ uint32_t pasid;
+ VTDPASIDCacheInfo pc_info = {};
+ uint64_t mask[4] = {VTD_INV_DESC_PASIDC_RSVD_VAL0, VTD_INV_DESC_ALL_ONE,
+ VTD_INV_DESC_ALL_ONE, VTD_INV_DESC_ALL_ONE};
+
+ if (!vtd_inv_desc_reserved_check(s, inv_desc, mask, true,
+ __func__, "pasid cache inv")) {
+ return false;
+ }
+
+ did = VTD_INV_DESC_PASIDC_DID(inv_desc);
+ pasid = VTD_INV_DESC_PASIDC_PASID(inv_desc);
+ pc_info.type = VTD_INV_DESC_PASIDC_G(inv_desc);
+
+ switch (pc_info.type) {
+ case VTD_INV_DESC_PASIDC_G_DSI:
+ trace_vtd_inv_desc_pasid_cache_dsi(did);
+ pc_info.did = did;
+ break;
+
+ case VTD_INV_DESC_PASIDC_G_PASID_SI:
+ /* PASID selective implies a DID selective */
+ trace_vtd_inv_desc_pasid_cache_psi(did, pasid);
+ /* Currently non-RID_PASID invalidation requests are ignored */
I'm a bit doubting if this is safe given the ATS path (for emulated
device) is merged. ATS path supports non-RID_PASID if emulated device
has PASID cap. The lucky thing is that the ATS path does not have
pasid level cache. So skipping invalidation for non-RID_PASID is not
harmful so far. Just a note to other reviewers although I didn't see a
problem here.
+ if (pasid != RID_PASID) {
+ return true;
+ }
+ pc_info.did = did;
+ pc_info.pasid = pasid;
+ break;
+
Regards,
Yi Liu
