File 5d8b7393-AMD-IOMMU-restrict-intremap-table-size... of Package xen

Overview Repositories Revisions Requests Users Attributes Meta

File 5d8b7393-AMD-IOMMU-restrict-intremap-table-sizes.patch of Package xen (Revision d8e25f42ca14b566b71c7b616cfb0f4b)

Currently displaying revision d8e25f42ca14b566b71c7b616cfb0f4b , Show latest

References: bsc#1135799

# Commit 34a9ef62d3d29ddafbec6b3409074dd67eba7109
# Date 2019-09-25 16:02:59 +0200
# Author Jan Beulich <jbeulich@suse.com>
# Committer Jan Beulich <jbeulich@suse.com>
AMD/IOMMU: restrict interrupt remapping table sizes

There's no point setting up tables with more space than a PCI device can
use. For both MSI and MSI-X we can determine how many interrupts could
be set up at most. Tables allocated during ACPI table parsing, however,
will (for now at least) continue to be set up to have maximum size.

Note that until we would want to use sub-page allocations here there's
no point checking whether both MSI and MSI-X are supported by a device -
an order-0 allocation will fit the dual case in any event, no matter
that the MSI-X vector count may be smaller than the MSI one.

On my Rome system this reduces space needed from just over 1k pages to
about 125.

Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>

--- a/xen/drivers/passthrough/amd/iommu_acpi.c
+++ b/xen/drivers/passthrough/amd/iommu_acpi.c
@@ -77,7 +77,7 @@ static void __init add_ivrs_mapping_entr
         {
             if ( !shared_intremap_table )
                 shared_intremap_table = amd_iommu_alloc_intremap_table(
-                    iommu, &shared_intremap_inuse);
+                    iommu, &shared_intremap_inuse, 0);
 
             if ( !shared_intremap_table )
                 panic("No memory for shared IRT\n");
@@ -89,7 +89,7 @@ static void __init add_ivrs_mapping_entr
         {
             ivrs_mappings[alias_id].intremap_table =
                 amd_iommu_alloc_intremap_table(
-                    iommu, &ivrs_mappings[alias_id].intremap_inuse);
+                    iommu, &ivrs_mappings[alias_id].intremap_inuse, 0);
 
             if ( !ivrs_mappings[alias_id].intremap_table )
                 panic("No memory for %04x:%02x:%02x.%u's IRT\n",
--- a/xen/drivers/passthrough/amd/iommu_init.c
+++ b/xen/drivers/passthrough/amd/iommu_init.c
@@ -1290,12 +1290,14 @@ static int __init amd_iommu_setup_device
                 pcidevs_unlock();
             }
 
-            if ( pdev )
+            if ( pdev && (pdev->msix || pdev->msi_maxvec) )
             {
                 ivrs_mappings[bdf].intremap_table =
                     amd_iommu_alloc_intremap_table(
                         ivrs_mappings[bdf].iommu,
-                        &ivrs_mappings[bdf].intremap_inuse);
+                        &ivrs_mappings[bdf].intremap_inuse,
+                        pdev->msix ? pdev->msix->nr_entries
+                                   : pdev->msi_maxvec);
                 if ( !ivrs_mappings[bdf].intremap_table )
                     return -ENOMEM;
 
@@ -1318,11 +1320,8 @@ static int __init amd_iommu_setup_device
             }
 
             amd_iommu_set_intremap_table(
-                dte,
-                ivrs_mappings[bdf].intremap_table
-                ? virt_to_maddr(ivrs_mappings[bdf].intremap_table)
-                : 0,
-                iommu_intremap);
+                dte, ivrs_mappings[bdf].intremap_table,
+                ivrs_mappings[bdf].iommu, iommu_intremap);
         }
     }
 
--- a/xen/drivers/passthrough/amd/iommu_intr.c
+++ b/xen/drivers/passthrough/amd/iommu_intr.c
@@ -69,7 +69,8 @@ union irte_cptr {
     const union irte128 *ptr128;
 } __transparent__;
 
-#define INTREMAP_MAX_ENTRIES (1 << IOMMU_INTREMAP_ORDER)
+#define INTREMAP_MAX_ORDER   0xB
+#define INTREMAP_MAX_ENTRIES (1 << INTREMAP_MAX_ORDER)
 
 struct ioapic_sbdf ioapic_sbdf[MAX_IO_APICS];
 struct hpet_sbdf hpet_sbdf;
@@ -80,17 +81,13 @@ unsigned int nr_ioapic_sbdf;
 
 static void dump_intremap_tables(unsigned char key);
 
-static unsigned int __init intremap_table_order(const struct amd_iommu *iommu)
-{
-    return iommu->ctrl.ga_en
-           ? get_order_from_bytes(INTREMAP_MAX_ENTRIES * sizeof(union irte128))
-           : get_order_from_bytes(INTREMAP_MAX_ENTRIES * sizeof(union irte32));
-}
+#define intremap_page_order(irt) PFN_ORDER(virt_to_page(irt))
 
 unsigned int amd_iommu_intremap_table_order(
     const void *irt, const struct amd_iommu *iommu)
 {
-    return IOMMU_INTREMAP_ORDER;
+    return intremap_page_order(irt) + PAGE_SHIFT -
+           (iommu->ctrl.ga_en ? 4 : 2);
 }
 
 static unsigned int intremap_table_entries(
@@ -825,7 +822,10 @@ int amd_iommu_free_intremap_table(
 
     if ( *tblp )
     {
-        __free_amd_iommu_tables(*tblp, intremap_table_order(iommu));
+        unsigned int order = intremap_page_order(*tblp);
+
+        intremap_page_order(*tblp) = 0;
+        __free_amd_iommu_tables(*tblp, order);
         *tblp = NULL;
     }
 
@@ -833,15 +833,23 @@ int amd_iommu_free_intremap_table(
 }
 
 void *amd_iommu_alloc_intremap_table(
-    const struct amd_iommu *iommu, unsigned long **inuse_map)
+    const struct amd_iommu *iommu, unsigned long **inuse_map, unsigned int nr)
 {
-    unsigned int order = intremap_table_order(iommu);
-    void *tb = __alloc_amd_iommu_tables(order);
+    unsigned int order;
+    void *tb;
 
+    if ( !nr )
+        nr = INTREMAP_MAX_ENTRIES;
+
+    order = iommu->ctrl.ga_en
+            ? get_order_from_bytes(nr * sizeof(union irte128))
+            : get_order_from_bytes(nr * sizeof(union irte32));
+
+    tb = __alloc_amd_iommu_tables(order);
     if ( tb )
     {
-        unsigned int nr = intremap_table_entries(tb, iommu);
-
+        intremap_page_order(tb) = order;
+        nr = intremap_table_entries(tb, iommu);
         *inuse_map = xzalloc_array(unsigned long, BITS_TO_LONGS(nr));
         if ( *inuse_map )
             memset(tb, 0, PAGE_SIZE << order);
--- a/xen/drivers/passthrough/amd/iommu_map.c
+++ b/xen/drivers/passthrough/amd/iommu_map.c
@@ -208,20 +208,26 @@ void iommu_dte_set_iotlb(uint32_t *dte,
 }
 
 void __init amd_iommu_set_intremap_table(
-    uint32_t *dte, uint64_t intremap_ptr, uint8_t int_valid)
+    uint32_t *dte, const void *ptr, const struct amd_iommu *iommu,
+    bool int_valid)
 {
     uint32_t addr_hi, addr_lo, entry;
 
-    addr_lo = intremap_ptr & DMA_32BIT_MASK;
-    addr_hi = intremap_ptr >> 32;
+    if ( ptr )
+    {
+        addr_lo = virt_to_maddr(ptr) & DMA_32BIT_MASK;
+        addr_hi = virt_to_maddr(ptr) >> 32;
+    }
+    else
+        addr_hi = addr_lo = 0;
 
     entry = dte[5];
     set_field_in_reg_u32(addr_hi, entry,
                          IOMMU_DEV_TABLE_INT_TABLE_PTR_HIGH_MASK,
                          IOMMU_DEV_TABLE_INT_TABLE_PTR_HIGH_SHIFT, &entry);
     /* Fixed and arbitrated interrupts remapepd */
-    set_field_in_reg_u32(intremap_ptr ? IOMMU_DEV_TABLE_INT_CONTROL_TRANSLATED
-                                      : IOMMU_DEV_TABLE_INT_CONTROL_ABORTED,
+    set_field_in_reg_u32(ptr ? IOMMU_DEV_TABLE_INT_CONTROL_TRANSLATED
+                             : IOMMU_DEV_TABLE_INT_CONTROL_ABORTED,
                          entry,
                          IOMMU_DEV_TABLE_INT_CONTROL_MASK,
                          IOMMU_DEV_TABLE_INT_CONTROL_SHIFT, &entry);
@@ -230,8 +236,8 @@ void __init amd_iommu_set_intremap_table
     set_field_in_reg_u32(addr_lo >> 6, 0,
                          IOMMU_DEV_TABLE_INT_TABLE_PTR_LOW_MASK,
                          IOMMU_DEV_TABLE_INT_TABLE_PTR_LOW_SHIFT, &entry);
-    /* 2048 entries */
-    set_field_in_reg_u32(intremap_ptr ? IOMMU_INTREMAP_ORDER : 0, entry,
+    set_field_in_reg_u32(ptr ? amd_iommu_intremap_table_order(ptr, iommu) : 0,
+                         entry,
                          IOMMU_DEV_TABLE_INT_TABLE_LENGTH_MASK,
                          IOMMU_DEV_TABLE_INT_TABLE_LENGTH_SHIFT, &entry);
 
--- a/xen/drivers/passthrough/amd/pci_amd_iommu.c
+++ b/xen/drivers/passthrough/amd/pci_amd_iommu.c
@@ -512,18 +512,22 @@ static int amd_iommu_add_device(u8 devfn
     {
         unsigned long flags;
 
-        ivrs_mappings[bdf].intremap_table =
-            amd_iommu_alloc_intremap_table(
-                iommu, &ivrs_mappings[bdf].intremap_inuse);
-        if ( !ivrs_mappings[bdf].intremap_table )
-            return -ENOMEM;
+        if ( pdev->msix || pdev->msi_maxvec )
+        {
+            ivrs_mappings[bdf].intremap_table =
+                amd_iommu_alloc_intremap_table(
+                    iommu, &ivrs_mappings[bdf].intremap_inuse,
+                    pdev->msix ? pdev->msix->nr_entries
+                               : pdev->msi_maxvec);
+            if ( !ivrs_mappings[bdf].intremap_table )
+                return -ENOMEM;
+        }
 
         spin_lock_irqsave(&iommu->lock, flags);
 
         amd_iommu_set_intremap_table(
             iommu->dev_table.buffer + (bdf * IOMMU_DEV_TABLE_ENTRY_SIZE),
-            virt_to_maddr(ivrs_mappings[bdf].intremap_table),
-            iommu_intremap);
+            ivrs_mappings[bdf].intremap_table, iommu, iommu_intremap);
 
         amd_iommu_flush_device(iommu, bdf);
 
--- a/xen/include/asm-x86/hvm/svm/amd-iommu-defs.h
+++ b/xen/include/asm-x86/hvm/svm/amd-iommu-defs.h
@@ -107,9 +107,6 @@
 #define IOMMU_DEV_TABLE_INT_CONTROL_FORWARDED	0x1
 #define IOMMU_DEV_TABLE_INT_CONTROL_TRANSLATED	0x2
 
-/* For now, we always allocate the maximum: 2048 entries. */
-#define IOMMU_INTREMAP_ORDER			0xB
-
 /* DeviceTable Entry[31:0] */
 #define IOMMU_DEV_TABLE_VALID_MASK			0x00000001
 #define IOMMU_DEV_TABLE_VALID_SHIFT			0
--- a/xen/include/asm-x86/hvm/svm/amd-iommu-proto.h
+++ b/xen/include/asm-x86/hvm/svm/amd-iommu-proto.h
@@ -76,7 +76,7 @@ void amd_iommu_share_p2m(struct domain *
 /* device table functions */
 int get_dma_requestor_id(u16 seg, u16 bdf);
 void amd_iommu_set_intremap_table(
-    u32 *dte, u64 intremap_ptr, u8 int_valid);
+    uint32_t *dte, const void *ptr, const struct amd_iommu *iommu, bool valid);
 void amd_iommu_set_root_page_table(
     u32 *dte, u64 root_ptr, u16 domain_id, u8 paging_mode, u8 valid);
 void iommu_dte_set_iotlb(u32 *dte, u8 i);
@@ -101,7 +101,7 @@ struct amd_iommu *find_iommu_for_device(
 bool iov_supports_xt(void);
 int amd_iommu_setup_ioapic_remapping(void);
 void *amd_iommu_alloc_intremap_table(
-    const struct amd_iommu *, unsigned long **);
+    const struct amd_iommu *, unsigned long **, unsigned int nr);
 int amd_iommu_free_intremap_table(
     const struct amd_iommu *, struct ivrs_mappings *, uint16_t);
 unsigned int amd_iommu_intremap_table_order(

Places

File 5d8b7393-AMD-IOMMU-restrict-intremap-table-sizes.patch of Package xen (Revision d8e25f42ca14b566b71c7b616cfb0f4b)

Places