Skip to contentSkip to conversation

IOMMU cleanup

PATCH Superseded Paul Durrant 3 participants 21 messages Sep 7, 2020 – Sep 10, 2020

v5 of 10 Superseded Latest is v10 →

lore
Share & tools
lore
#0 lore raw

v5 cover IOMMU cleanup

From: Paul Durrant <[email protected]>

Paul Durrant (8):
x86/iommu: convert VT-d code to use new page table allocator
iommu: remove unused iommu_ops method and tasklet
iommu: flush I/O TLB if iommu_map() or iommu_unmap() fail
iommu: make map and unmap take a page count, similar to flush remove remaining uses of iommu_legacy_map/unmap common/grant_table: batch flush I/O TLB
iommu: remove the share_p2m operation
iommu: stop calling IOMMU page tables 'p2m tables'

 xen/arch/x86/mm.c                           |  22 +-
 xen/arch/x86/mm/p2m-ept.c                   |  21 +-
 xen/arch/x86/mm/p2m-pt.c                    |  16 +-
 xen/arch/x86/mm/p2m.c                       |  27 ++-
 xen/arch/x86/x86_64/mm.c                    |  20 +-
 xen/common/grant_table.c                    | 210 ++++++++++++------
 xen/common/memory.c                         |   7 +-
 xen/drivers/passthrough/amd/iommu.h         |   2 +-
 xen/drivers/passthrough/amd/iommu_map.c     |   2 +-
 xen/drivers/passthrough/amd/pci_amd_iommu.c |  22 +-
 xen/drivers/passthrough/iommu.c             | 120 +++--------
 xen/drivers/passthrough/vtd/extern.h        |   2 +-
 xen/drivers/passthrough/vtd/iommu.c         | 225 +++++++++-----------
 xen/drivers/passthrough/vtd/x86/vtd.c       |   2 +-
 xen/drivers/passthrough/x86/iommu.c         |   2 +-
 xen/include/xen/iommu.h                     |  36 +---
 16 files changed, 373 insertions(+), 363 deletions(-)

---
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Jan Beulich <[email protected]>
Cc: Julien Grall <[email protected]>
Cc: Jun Nakajima <[email protected]>
Cc: Kevin Tian <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Wei Liu <[email protected]>

Signature · 1 line
2.20.1
#1 lore raw

PATCH v5 1/8 x86/iommu: convert VT-d code to use new page table allocator

From: Paul Durrant <[email protected]>

This patch converts the VT-d code to use the new IOMMU page table allocator function. This allows all the free-ing code to be removed (since it is now handled by the general x86 code) which reduces TLB and cache thrashing as well as shortening the code.

The scope of the mapping_lock in intel_iommu_quarantine_init() has also been increased slightly; it should have always covered accesses to 'arch.vtd.pgd_maddr'.

NOTE: The common IOMMU needs a slight modification to avoid scheduling the
cleanup tasklet if the free_page_table() method is not present (since the tasklet will unconditionally call it).

---
Cc: Kevin Tian <[email protected]>

v2:
- New in v2 (split from "add common page-table allocator")

---
 xen/drivers/passthrough/iommu.c     |   6 +-
 xen/drivers/passthrough/vtd/iommu.c | 101 ++++++++++------------------
 2 files changed, 39 insertions(+), 68 deletions(-)
xen/drivers/passthrough/iommu.c +4 -2
index 1d644844ab..2b1db8022c 100644
@@ -225,8 +225,10 @@ static void iommu_teardown(struct domain *d)
 {
     struct domain_iommu *hd = dom_iommu(d);
 
-    hd->platform_ops->teardown(d);
-    tasklet_schedule(&iommu_pt_cleanup_tasklet);
+    iommu_vcall(hd->platform_ops, teardown, d);
+
+    if ( hd->platform_ops->free_page_table )
+        tasklet_schedule(&iommu_pt_cleanup_tasklet);
 }
 
 void iommu_domain_destroy(struct domain *d)
xen/drivers/passthrough/vtd/iommu.c +35 -66

172 lines not on this page · raw message

Signature · 1 line
2.20.1
#2 lore raw

PATCH v5 2/8 iommu: remove unused iommu_ops method and tasklet

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

The VT-d and AMD IOMMU implementations both use the general x86 IOMMU page table allocator and ARM always shares page tables with CPU. Hence there is no need to retain the free_page_table() method or the tasklet which invokes it.

---

v2:
- New in v2 (split from "add common page-table allocator")

---
 xen/drivers/passthrough/iommu.c | 25 -------------------------
 xen/include/xen/iommu.h         |  2 --
 2 files changed, 27 deletions(-)
xen/drivers/passthrough/iommu.c +0 -25
index 2b1db8022c..660dc5deb2 100644
@@ -49,10 +49,6 @@ bool_t __read_mostly amd_iommu_perdev_intremap = 1;
 
 DEFINE_PER_CPU(bool_t, iommu_dont_flush_iotlb);
 
-DEFINE_SPINLOCK(iommu_pt_cleanup_lock);
-PAGE_LIST_HEAD(iommu_pt_cleanup_list);
-static struct tasklet iommu_pt_cleanup_tasklet;
-
 static int __init parse_iommu_param(const char *s)
 {
     const char *ss;
@@ -226,9 +222,6 @@ static void iommu_teardown(struct domain *d)
     struct domain_iommu *hd = dom_iommu(d);
 
     iommu_vcall(hd->platform_ops, teardown, d);
-
-    if ( hd->platform_ops->free_page_table )
-        tasklet_schedule(&iommu_pt_cleanup_tasklet);
 }
 
 void iommu_domain_destroy(struct domain *d)
@@ -368,23 +361,6 @@ int iommu_lookup_page(struct domain *d, dfn_t dfn, mfn_t *mfn,
     return iommu_call(hd->platform_ops, lookup_page, d, dfn, mfn, flags);
 }
 
-static void iommu_free_pagetables(void *unused)
-{
-    do {
-        struct page_info *pg;
-
-        spin_lock(&iommu_pt_cleanup_lock);
-        pg = page_list_remove_head(&iommu_pt_cleanup_list);
-        spin_unlock(&iommu_pt_cleanup_lock);
-        if ( !pg )
-            return;
-        iommu_vcall(iommu_get_ops(), free_page_table, pg);
-    } while ( !softirq_pending(smp_processor_id()) );
-
-    tasklet_schedule_on_cpu(&iommu_pt_cleanup_tasklet,
-                            cpumask_cycle(smp_processor_id(), &cpu_online_map));
-}
-
 int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned int page_count,
                       unsigned int flush_flags)
 {
@@ -508,7 +484,6 @@ int __init iommu_setup(void)
 #ifndef iommu_intremap
         printk("Interrupt remapping %sabled\n", iommu_intremap ? "en" : "dis");
 #endif
-        tasklet_init(&iommu_pt_cleanup_tasklet, iommu_free_pagetables, NULL);
     }
 
     return rc;
xen/include/xen/iommu.h +0 -2
index 3272874958..1831dc66b0 100644
@@ -263,8 +263,6 @@ struct iommu_ops {
     int __must_check (*lookup_page)(struct domain *d, dfn_t dfn, mfn_t *mfn,
                                     unsigned int *flags);
 
-    void (*free_page_table)(struct page_info *);
-
 #ifdef CONFIG_X86
     int (*enable_x2apic)(void);
     void (*disable_x2apic)(void);
Signature · 1 line
2.20.1
#3 lore raw

PATCH v5 3/8 iommu: flush I/O TLB if iommu_map() or iommu_unmap() fail

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

This patch adds a full I/O TLB flush to the error paths of iommu_map() and iommu_unmap().

Without this change callers need constructs such as:

rc = iommu_map/unmap(...)
err = iommu_flush(...)
if ( !rc )
rc = err;

With this change, it can be simplified to:

rc = iommu_map/unmap(...)
if ( !rc )
rc = iommu_flush(...)

because, if the map or unmap fails the flush will be unnecessary. This saves a stack variable and generally makes the call sites tidier.

---
Cc: Jan Beulich <[email protected]>

v5:
- Avoid unnecessary flushing if 'page_order' is 0
- Add missing indirection on 'flush_flags'

v2:
- New in v2

---
 xen/drivers/passthrough/iommu.c | 34 +++++++++++++++++----------------
 1 file changed, 18 insertions(+), 16 deletions(-)
xen/drivers/passthrough/iommu.c +18 -16
index 660dc5deb2..f27facd52d 100644
@@ -274,6 +274,13 @@ int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
         break;
     }
 
+    /*
+     * Something went wrong so, if we were dealing with more than a single
+     * page, flush everything and clear flush flags.
+     */
+    if ( page_order && unlikely(rc) && !iommu_iotlb_flush_all(d, *flush_flags) )
+        *flush_flags = 0;
+
     return rc;
 }
 
@@ -283,14 +290,8 @@ int iommu_legacy_map(struct domain *d, dfn_t dfn, mfn_t mfn,
     unsigned int flush_flags = 0;
     int rc = iommu_map(d, dfn, mfn, page_order, flags, &flush_flags);
 
-    if ( !this_cpu(iommu_dont_flush_iotlb) )
-    {
-        int err = iommu_iotlb_flush(d, dfn, (1u << page_order),
-                                    flush_flags);
-
-        if ( !rc )
-            rc = err;
-    }
+    if ( !this_cpu(iommu_dont_flush_iotlb) && !rc )
+        rc = iommu_iotlb_flush(d, dfn, (1u << page_order), flush_flags);
 
     return rc;
 }
@@ -330,6 +331,13 @@ int iommu_unmap(struct domain *d, dfn_t dfn, unsigned int page_order,
         }
     }
 
+    /*
+     * Something went wrong so, if we were dealing with more than a single
+     * page, flush everything and clear flush flags.
+     */
+    if ( page_order && unlikely(rc) && !iommu_iotlb_flush_all(d, *flush_flags) )
+        *flush_flags = 0;
+
     return rc;
 }
 
@@ -338,14 +346,8 @@ int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned int page_order)
     unsigned int flush_flags = 0;
     int rc = iommu_unmap(d, dfn, page_order, &flush_flags);
 
-    if ( !this_cpu(iommu_dont_flush_iotlb) )
-    {
-        int err = iommu_iotlb_flush(d, dfn, (1u << page_order),
-                                    flush_flags);
-
-        if ( !rc )
-            rc = err;
-    }
+    if ( !this_cpu(iommu_dont_flush_iotlb) && ! rc )
+        rc = iommu_iotlb_flush(d, dfn, (1u << page_order), flush_flags);
 
     return rc;
 }
Signature · 1 line
2.20.1
Reviewed-by on v5 3/8
#4 lore raw

Re: [PATCH v5 3/8] iommu: flush I/O TLB if iommu_map() or iommu_unmap() fail

Quoted 14 lines · Paul Durrant

From: Paul Durrant <[email protected]>

This patch adds a full I/O TLB flush to the error paths of iommu_map() and iommu_unmap().

Without this change callers need constructs such as:

rc = iommu_map/unmap(...)
err = iommu_flush(...)
if ( !rc )
rc = err;

With this change, it can be simplified to:

Quoted 8 lines

rc = iommu_map/unmap(...)
if ( !rc )
rc = iommu_flush(...)

because, if the map or unmap fails the flush will be unnecessary. This saves a stack variable and generally makes the call sites tidier.

with one cosmetic issue taken care of (perhaps while committing):

xen/drivers/passthrough/iommu.c int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned int page_order) · line 349

Quoted 13 lines
xen/drivers/passthrough/iommu.c +1 -8
@@ -338,14 +346,8 @@ int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned int page_order)
     unsigned int flush_flags = 0;
     int rc = iommu_unmap(d, dfn, page_order, &flush_flags);
 
-    if ( !this_cpu(iommu_dont_flush_iotlb) )
-    {
-        int err = iommu_iotlb_flush(d, dfn, (1u << page_order),
-                                    flush_flags);
-
-        if ( !rc )
-            rc = err;
-    }
+    if ( !this_cpu(iommu_dont_flush_iotlb) && ! rc )

There's a stray blank after the latter ! here.

Jan

#5 lore raw

PATCH v5 4/8 iommu: make map and unmap take a page count, similar to flush

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

At the moment iommu_map() and iommu_unmap() take a page order rather than a count, whereas iommu_iotlb_flush() takes a page count rather than an order. This patch makes them consistent with each other, opting for a page count since CPU page orders are not necessarily the same as those of an IOMMU.

NOTE: The 'page_count' parameter is also made an unsigned long in all the
aforementioned functions.

---
Cc: Jan Beulich <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Julien Grall <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Jun Nakajima <[email protected]>
Cc: Kevin Tian <[email protected]>

v5:
- Re-worked to just use a page count, rather than both a count and an order

v2:
- New in v2

---
 xen/arch/x86/mm.c                       |  6 +++--
 xen/arch/x86/mm/p2m-ept.c               |  6 ++---
 xen/arch/x86/mm/p2m-pt.c                |  4 +--
 xen/arch/x86/mm/p2m.c                   |  5 ++--
 xen/arch/x86/x86_64/mm.c                |  4 +--
 xen/common/grant_table.c                |  6 ++---
 xen/drivers/passthrough/amd/iommu.h     |  2 +-
 xen/drivers/passthrough/amd/iommu_map.c |  2 +-
 xen/drivers/passthrough/iommu.c         | 35 +++++++++++--------------
 xen/drivers/passthrough/vtd/iommu.c     |  4 +--
 xen/drivers/passthrough/x86/iommu.c     |  2 +-
 xen/include/xen/iommu.h                 | 12 ++++-----
 12 files changed, 44 insertions(+), 44 deletions(-)
xen/arch/x86/mm.c +4 -2
index 35ec0e11f6..e133ccf0d8 100644
@@ -2966,9 +2966,11 @@ static int _get_page_type(struct page_info *page, unsigned long type,
             mfn_t mfn = page_to_mfn(page);
 
             if ( (x & PGT_type_mask) == PGT_writable_page )
-                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)), PAGE_ORDER_4K);
+                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)),
+                                        1ul << PAGE_ORDER_4K);
             else
-                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn, PAGE_ORDER_4K,
+                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn,
+                                      1ul << PAGE_ORDER_4K,
                                       IOMMUF_readable | IOMMUF_writable);
 
             if ( unlikely(rc) )
xen/arch/x86/mm/p2m-ept.c +3 -3
index b8154a7ecc..12cf38f6eb 100644
@@ -843,14 +843,14 @@ out:
          need_modify_vtd_table )
     {
         if ( iommu_use_hap_pt(d) )
-            rc = iommu_iotlb_flush(d, _dfn(gfn), (1u << order),
+            rc = iommu_iotlb_flush(d, _dfn(gfn), 1ul << order,
                                    (iommu_flags ? IOMMU_FLUSHF_added : 0) |
                                    (vtd_pte_present ? IOMMU_FLUSHF_modified
                                                     : 0));
         else if ( need_iommu_pt_sync(d) )
             rc = iommu_flags ?
-                iommu_legacy_map(d, _dfn(gfn), mfn, order, iommu_flags) :
-                iommu_legacy_unmap(d, _dfn(gfn), order);
+                iommu_legacy_map(d, _dfn(gfn), mfn, 1ul << order, iommu_flags) :
+                iommu_legacy_unmap(d, _dfn(gfn), 1ul << order);
     }
 
     unmap_domain_page(table);
xen/arch/x86/mm/p2m-pt.c +2 -2
index badb26bc34..3af51be78e 100644
@@ -679,9 +679,9 @@ p2m_pt_set_entry(struct p2m_domain *p2m, gfn_t gfn_, mfn_t mfn,
     if ( need_iommu_pt_sync(p2m->domain) &&
          (iommu_old_flags != iommu_pte_flags || old_mfn != mfn_x(mfn)) )
         rc = iommu_pte_flags
-             ? iommu_legacy_map(d, _dfn(gfn), mfn, page_order,
+             ? iommu_legacy_map(d, _dfn(gfn), mfn, 1ul << page_order,
                                 iommu_pte_flags)
-             : iommu_legacy_unmap(d, _dfn(gfn), page_order);
+             : iommu_legacy_unmap(d, _dfn(gfn), 1ul << page_order);
 
     /*
      * Free old intermediate tables if necessary.  This has to be the
xen/arch/x86/mm/p2m.c +3 -2
index db7bde0230..928344be30 100644
@@ -1352,7 +1352,8 @@ int set_identity_p2m_entry(struct domain *d, unsigned long gfn_l,
     {
         if ( !is_iommu_enabled(d) )
             return 0;
-        return iommu_legacy_map(d, _dfn(gfn_l), _mfn(gfn_l), PAGE_ORDER_4K,
+        return iommu_legacy_map(d, _dfn(gfn_l), _mfn(gfn_l),
+                                1ul << PAGE_ORDER_4K,
                                 IOMMUF_readable | IOMMUF_writable);
     }
 
@@ -1443,7 +1444,7 @@ int clear_identity_p2m_entry(struct domain *d, unsigned long gfn_l)
     {
         if ( !is_iommu_enabled(d) )
             return 0;
-        return iommu_legacy_unmap(d, _dfn(gfn_l), PAGE_ORDER_4K);
+        return iommu_legacy_unmap(d, _dfn(gfn_l), 1ul << PAGE_ORDER_4K);
     }
 
     gfn_lock(p2m, gfn, 0);
xen/arch/x86/x86_64/mm.c +2 -2
index 28bc98f8f2..6f2c43181a 100644
@@ -1480,7 +1480,7 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
     {
         for ( i = spfn; i < epfn; i++ )
             if ( iommu_legacy_map(hardware_domain, _dfn(i), _mfn(i),
-                                  PAGE_ORDER_4K,
+                                  1ul << PAGE_ORDER_4K,
                                   IOMMUF_readable | IOMMUF_writable) )
                 break;
         if ( i != epfn )
@@ -1488,7 +1488,7 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
             while (i-- > old_max)
                 /* If statement to satisfy __must_check. */
                 if ( iommu_legacy_unmap(hardware_domain, _dfn(i),
-                                        PAGE_ORDER_4K) )
+                                        1ul << PAGE_ORDER_4K) )
                     continue;
 
             goto destroy_m2p;
xen/common/grant_table.c +3 -3
index 9f0cae52c0..c5a1ef5fc3 100644
@@ -1225,7 +1225,7 @@ map_grant_ref(
             kind = IOMMUF_readable;
         else
             kind = 0;
-        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 0, kind) )
+        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 1ul, kind) )
         {
             double_gt_unlock(lgt, rgt);
             rc = GNTST_general_error;
@@ -1479,9 +1479,9 @@ unmap_common(
 
         kind = mapkind(lgt, rd, op->mfn);
         if ( !kind )
-            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 0);
+            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 1ul);
         else if ( !(kind & MAPKIND_WRITE) )
-            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 0,
+            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 1ul,
                                    IOMMUF_readable);
 
         double_gt_unlock(lgt, rgt);
xen/drivers/passthrough/amd/iommu.h +1 -1
index e2d174f3b4..f1f0415469 100644
@@ -231,7 +231,7 @@ int amd_iommu_reserve_domain_unity_map(struct domain *domain,
                                        paddr_t phys_addr, unsigned long size,
                                        int iw, int ir);
 int __must_check amd_iommu_flush_iotlb_pages(struct domain *d, dfn_t dfn,
-                                             unsigned int page_count,
+                                             unsigned long page_count,
                                              unsigned int flush_flags);
 int __must_check amd_iommu_flush_iotlb_all(struct domain *d);
 
xen/drivers/passthrough/amd/iommu_map.c +1 -1
index 54b991294a..ee6eb49bef 100644
@@ -362,7 +362,7 @@ static unsigned long flush_count(unsigned long dfn, unsigned int page_count,
 }
 
 int amd_iommu_flush_iotlb_pages(struct domain *d, dfn_t dfn,
-                                unsigned int page_count,
+                                unsigned long page_count,
                                 unsigned int flush_flags)
 {
     unsigned long dfn_l = dfn_x(dfn);
xen/drivers/passthrough/iommu.c +16 -19
index f27facd52d..e664f4ac79 100644
@@ -235,7 +235,7 @@ void iommu_domain_destroy(struct domain *d)
 }
 
 int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-              unsigned int page_order, unsigned int flags,
+              unsigned long page_count, unsigned int flags,
               unsigned int *flush_flags)
 {
     const struct domain_iommu *hd = dom_iommu(d);
@@ -245,10 +245,7 @@ int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
     if ( !is_iommu_enabled(d) )
         return 0;
 
-    ASSERT(IS_ALIGNED(dfn_x(dfn), (1ul << page_order)));
-    ASSERT(IS_ALIGNED(mfn_x(mfn), (1ul << page_order)));
-
-    for ( i = 0; i < (1ul << page_order); i++ )
+    for ( i = 0; i < page_count; i++ )
     {
         rc = iommu_call(hd->platform_ops, map_page, d, dfn_add(dfn, i),
                         mfn_add(mfn, i), flags, flush_flags);
@@ -278,25 +275,26 @@ int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
      * Something went wrong so, if we were dealing with more than a single
      * page, flush everything and clear flush flags.
      */
-    if ( page_order && unlikely(rc) && !iommu_iotlb_flush_all(d, *flush_flags) )
+    if ( page_count > 1 && unlikely(rc) &&
+         !iommu_iotlb_flush_all(d, *flush_flags) )
         *flush_flags = 0;
 
     return rc;
 }
 
 int iommu_legacy_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-                     unsigned int page_order, unsigned int flags)
+                     unsigned long page_count, unsigned int flags)
 {
     unsigned int flush_flags = 0;
-    int rc = iommu_map(d, dfn, mfn, page_order, flags, &flush_flags);
+    int rc = iommu_map(d, dfn, mfn, page_count, flags, &flush_flags);
 
     if ( !this_cpu(iommu_dont_flush_iotlb) && !rc )
-        rc = iommu_iotlb_flush(d, dfn, (1u << page_order), flush_flags);
+        rc = iommu_iotlb_flush(d, dfn, page_count, flush_flags);
 
     return rc;
 }
 
-int iommu_unmap(struct domain *d, dfn_t dfn, unsigned int page_order,
+int iommu_unmap(struct domain *d, dfn_t dfn, unsigned long page_count,
                 unsigned int *flush_flags)
 {
     const struct domain_iommu *hd = dom_iommu(d);
@@ -306,9 +304,7 @@ int iommu_unmap(struct domain *d, dfn_t dfn, unsigned int page_order,
     if ( !is_iommu_enabled(d) )
         return 0;
 
-    ASSERT(IS_ALIGNED(dfn_x(dfn), (1ul << page_order)));
-
-    for ( i = 0; i < (1ul << page_order); i++ )
+    for ( i = 0; i < page_count; i++ )
     {
         int err = iommu_call(hd->platform_ops, unmap_page, d, dfn_add(dfn, i),
                              flush_flags);
@@ -335,19 +331,20 @@ int iommu_unmap(struct domain *d, dfn_t dfn, unsigned int page_order,
      * Something went wrong so, if we were dealing with more than a single
      * page, flush everything and clear flush flags.
      */
-    if ( page_order && unlikely(rc) && !iommu_iotlb_flush_all(d, *flush_flags) )
+    if ( page_count > 1 && unlikely(rc) &&
+         !iommu_iotlb_flush_all(d, *flush_flags) )
         *flush_flags = 0;
 
     return rc;
 }
 
-int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned int page_order)
+int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned long page_count)
 {
     unsigned int flush_flags = 0;
-    int rc = iommu_unmap(d, dfn, page_order, &flush_flags);
+    int rc = iommu_unmap(d, dfn, page_count, &flush_flags);
 
     if ( !this_cpu(iommu_dont_flush_iotlb) && ! rc )
-        rc = iommu_iotlb_flush(d, dfn, (1u << page_order), flush_flags);
+        rc = iommu_iotlb_flush(d, dfn, page_count, flush_flags);
 
     return rc;
 }
@@ -363,7 +360,7 @@ int iommu_lookup_page(struct domain *d, dfn_t dfn, mfn_t *mfn,
     return iommu_call(hd->platform_ops, lookup_page, d, dfn, mfn, flags);
 }
 
-int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned int page_count,
+int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned long page_count,
                       unsigned int flush_flags)
 {
     const struct domain_iommu *hd = dom_iommu(d);
@@ -382,7 +379,7 @@ int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned int page_count,
     {
         if ( !d->is_shutting_down && printk_ratelimit() )
             printk(XENLOG_ERR
-                   "d%d: IOMMU IOTLB flush failed: %d, dfn %"PRI_dfn", page count %u flags %x\n",
+                   "d%d: IOMMU IOTLB flush failed: %d, dfn %"PRI_dfn", page count %lu flags %x\n",
                    d->domain_id, rc, dfn_x(dfn), page_count, flush_flags);
 
         if ( !is_hardware_domain(d) )
xen/drivers/passthrough/vtd/iommu.c +2 -2
index 607e8b5e65..68cf0e535a 100644
@@ -584,7 +584,7 @@ static int __must_check iommu_flush_all(void)
 
 static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn,
                                           bool_t dma_old_pte_present,
-                                          unsigned int page_count)
+                                          unsigned long page_count)
 {
     struct domain_iommu *hd = dom_iommu(d);
     struct acpi_drhd_unit *drhd;
@@ -632,7 +632,7 @@ static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn,
 
 static int __must_check iommu_flush_iotlb_pages(struct domain *d,
                                                 dfn_t dfn,
-                                                unsigned int page_count,
+                                                unsigned long page_count,
                                                 unsigned int flush_flags)
 {
     ASSERT(page_count && !dfn_eq(dfn, INVALID_DFN));
xen/drivers/passthrough/x86/iommu.c +1 -1
index aea07e47c4..f17b1820f4 100644
@@ -244,7 +244,7 @@ void __hwdom_init arch_iommu_hwdom_init(struct domain *d)
         else if ( paging_mode_translate(d) )
             rc = set_identity_p2m_entry(d, pfn, p2m_access_rw, 0);
         else
-            rc = iommu_map(d, _dfn(pfn), _mfn(pfn), PAGE_ORDER_4K,
+            rc = iommu_map(d, _dfn(pfn), _mfn(pfn), 1ul << PAGE_ORDER_4K,
                            IOMMUF_readable | IOMMUF_writable, &flush_flags);
 
         if ( rc )
xen/include/xen/iommu.h +6 -6
index 1831dc66b0..13f68dc93d 100644
@@ -146,23 +146,23 @@ enum
 #define IOMMU_FLUSHF_modified (1u << _IOMMU_FLUSHF_modified)
 
 int __must_check iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-                           unsigned int page_order, unsigned int flags,
+                           unsigned long page_count, unsigned int flags,
                            unsigned int *flush_flags);
 int __must_check iommu_unmap(struct domain *d, dfn_t dfn,
-                             unsigned int page_order,
+                             unsigned long page_count,
                              unsigned int *flush_flags);
 
 int __must_check iommu_legacy_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-                                  unsigned int page_order,
+                                  unsigned long page_count,
                                   unsigned int flags);
 int __must_check iommu_legacy_unmap(struct domain *d, dfn_t dfn,
-                                    unsigned int page_order);
+                                    unsigned long page_count);
 
 int __must_check iommu_lookup_page(struct domain *d, dfn_t dfn, mfn_t *mfn,
                                    unsigned int *flags);
 
 int __must_check iommu_iotlb_flush(struct domain *d, dfn_t dfn,
-                                   unsigned int page_count,
+                                   unsigned long page_count,
                                    unsigned int flush_flags);
 int __must_check iommu_iotlb_flush_all(struct domain *d,
                                        unsigned int flush_flags);
@@ -281,7 +281,7 @@ struct iommu_ops {
     void (*share_p2m)(struct domain *d);
     void (*crash_shutdown)(void);
     int __must_check (*iotlb_flush)(struct domain *d, dfn_t dfn,
-                                    unsigned int page_count,
+                                    unsigned long page_count,
                                     unsigned int flush_flags);
     int __must_check (*iotlb_flush_all)(struct domain *d);
     int (*get_reserved_device_memory)(iommu_grdm_t *, void *);
Signature · 1 line
2.20.1
#6 lore raw

Re: [PATCH v5 4/8] iommu: make map and unmap take a page count, similar to flush

Quoted 6 lines · Paul Durrant
xen/arch/x86/mm.c +0 -0
@@ -2966,9 +2966,11 @@ static int _get_page_type(struct page_info *page, unsigned long type,
             mfn_t mfn = page_to_mfn(page);
 
             if ( (x & PGT_type_mask) == PGT_writable_page )

xen/arch/x86/mm.c static int _get_page_type(struct page_info *page, unsigned long type, · line 2976

Quoted 10 lines
xen/arch/x86/mm.c +4 -2
-                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)), PAGE_ORDER_4K);
+                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)),
+                                        1ul << PAGE_ORDER_4K);
             else
-                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn, PAGE_ORDER_4K,
+                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn,
+                                      1ul << PAGE_ORDER_4K,
                                       IOMMUF_readable | IOMMUF_writable);
 
             if ( unlikely(rc) )

A few hundred lines up from here there is

            int rc2 = iommu_legacy_unmap(d, _dfn(mfn), PAGE_ORDER_4K);

in cleanup_page_mappings().

Quoted 12 lines
xen/common/grant_table.c +1 -1
@@ -1225,7 +1225,7 @@ map_grant_ref(
             kind = IOMMUF_readable;
         else
             kind = 0;
-        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 0, kind) )
+        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 1ul, kind) )
         {
             double_gt_unlock(lgt, rgt);
             rc = GNTST_general_error;

xen/common/grant_table.c unmap_common( · line 1484

Quoted 8 lines
xen/common/grant_table.c +2 -2
@@ -1479,9 +1479,9 @@ unmap_common(
 
         kind = mapkind(lgt, rd, op->mfn);
         if ( !kind )
-            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 0);
+            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 1ul);
         else if ( !(kind & MAPKIND_WRITE) )
-            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 0,
+            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 1ul,

For all three of these, I guess either 1ul << PAGE_ORDER_4K or simply 1? (Given that the code didn't use PAGE_ORDER_4K so far, I'd slightly prefer the latter. I'd be fine making the change while committing, but it looks like v6 is going to be needed anyway.)

xen/drivers/passthrough/amd/iommu_map.c static unsigned long flush_count(unsigned long dfn, unsigned int page_count, · line 365

Quoted 8 lines
xen/drivers/passthrough/amd/iommu_map.c +1 -1
@@ -362,7 +362,7 @@ static unsigned long flush_count(unsigned long dfn, unsigned int page_count,
 }
 
 int amd_iommu_flush_iotlb_pages(struct domain *d, dfn_t dfn,
-                                unsigned int page_count,
+                                unsigned long page_count,

This ought to be accompanied by a similar change to its flush_count() helper.

xen/drivers/passthrough/amd/iommu_map.c static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn, · line 638

Quoted 9 lines
xen/drivers/passthrough/amd/iommu_map.c +1 -1
@@ -632,7 +632,7 @@ static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn,
 
 static int __must_check iommu_flush_iotlb_pages(struct domain *d,
                                                 dfn_t dfn,
-                                                unsigned int page_count,
+                                                unsigned long page_count,
                                                 unsigned int flush_flags)
 {
     ASSERT(page_count && !dfn_eq(dfn, INVALID_DFN));

This similarly ought to be accompanied by a change to its iommu_flush_iotlb() helper.

Jan

#7 lore raw

RE: [PATCH v5 4/8] iommu: make map and unmap take a page count, similar to flush

Quoted 27 lines

-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 10 September 2020 13:48
To: Paul Durrant <[email protected]>
Cc: [email protected]; Paul Durrant <[email protected]>; Andrew Cooper <[email protected]>; Wei Liu <[email protected]>; Roger Pau Monné <[email protected]>; George Dunlap <[email protected]>; Ian Jackson <[email protected]>; Julien Grall <[email protected]>; Stefano Stabellini <[email protected]>; Jun Nakajima <[email protected]>; Kevin Tian <[email protected]>
Subject: Re: [PATCH v5 4/8] iommu: make map and unmap take a page count, similar to flush

Quoted 15 lines · Paul Durrant
xen/arch/x86/mm.c +4 -2
@@ -2966,9 +2966,11 @@ static int _get_page_type(struct page_info *page, unsigned long type,
             mfn_t mfn = page_to_mfn(page);
 
             if ( (x & PGT_type_mask) == PGT_writable_page )
-                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)), PAGE_ORDER_4K);
+                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)),
+                                        1ul << PAGE_ORDER_4K);
             else
-                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn, PAGE_ORDER_4K,
+                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn,
+                                      1ul << PAGE_ORDER_4K,
                                       IOMMUF_readable | IOMMUF_writable);
 
Quoted 7 lines

xen/arch/x86/mm.c static int _get_page_type(struct page_info *page, unsigned long type, · line 2976

Quoted 1 line
xen/arch/x86/mm.c +0 -0
             if ( unlikely(rc) )

A few hundred lines up from here there is

            int rc2 = iommu_legacy_unmap(d, _dfn(mfn), PAGE_ORDER_4K);

in cleanup_page_mappings().

Oh yes. Fixed.

Quoted 14 lines
Quoted 14 lines
xen/common/grant_table.c +1 -1
@@ -1225,7 +1225,7 @@ map_grant_ref(
             kind = IOMMUF_readable;
         else
             kind = 0;
-        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 0, kind) )
+        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 1ul, kind) )
         {
             double_gt_unlock(lgt, rgt);
             rc = GNTST_general_error;
@@ -1479,9 +1479,9 @@ unmap_common(
 
Quoted 12 lines

xen/common/grant_table.c unmap_common( · line 1484

Quoted 7 lines
xen/common/grant_table.c +2 -2
         kind = mapkind(lgt, rd, op->mfn);
         if ( !kind )
-            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 0);
+            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 1ul);
         else if ( !(kind & MAPKIND_WRITE) )
-            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 0,
+            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 1ul,

For all three of these, I guess either 1ul << PAGE_ORDER_4K or simply 1? (Given that the code didn't use PAGE_ORDER_4K so far, I'd slightly prefer the latter. I'd be fine making the change while committing, but it looks like v6 is going to be needed anyway.)

Ok. Changed.

Quoted 23 lines

xen/drivers/passthrough/amd/iommu_map.c static unsigned long flush_count(unsigned long dfn, unsigned int page_count, · line 365

Quoted 9 lines
xen/drivers/passthrough/amd/iommu_map.c +1 -1
@@ -362,7 +362,7 @@ static unsigned long flush_count(unsigned long dfn, unsigned int page_count,
 }
 
 int amd_iommu_flush_iotlb_pages(struct domain *d, dfn_t dfn,
-                                unsigned int page_count,
+                                unsigned long page_count,

This ought to be accompanied by a similar change to its flush_count() helper.

xen/drivers/passthrough/amd/iommu_map.c static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn, · line 638

Quoted 9 lines
xen/drivers/passthrough/amd/iommu_map.c +1 -1
@@ -632,7 +632,7 @@ static int __must_check iommu_flush_iotlb(struct domain *d, dfn_t dfn,
 
 static int __must_check iommu_flush_iotlb_pages(struct domain *d,
                                                 dfn_t dfn,
-                                                unsigned int page_count,
+                                                unsigned long page_count,
                                                 unsigned int flush_flags)
 {
     ASSERT(page_count && !dfn_eq(dfn, INVALID_DFN));
Quoted 3 lines

This similarly ought to be accompanied by a change to its iommu_flush_iotlb() helper.

It is... in the previous hunk.

Paul

Quoted 1 line

Jan

#8 lore raw

PATCH v5 5/8 remove remaining uses of iommu_legacy_map/unmap

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

The 'legacy' functions do implicit flushing so amend the callers to do the appropriate flushing.

Unfortunately, because of the structure of the P2M code, we cannot remove the per-CPU 'iommu_dont_flush_iotlb' global and the optimization it facilitates. It is now checked directly iommu_iotlb_flush(). This is safe because callers of iommu_iotlb_flush() on another CPU will not be affected, and hence a flush will be done. Also, 'iommu_dont_flush_iotlb' is now declared as bool (rather than bool_t) and setting/clearing it are no longer pointlessly gated on is_iommu_enabled() returning true. (Arguably it is also pointless to gate the call to iommu_iotlb_flush() on that condition - since it is a no-op in that case - but the if clause allows the scope of a stack variable to be restricted).

NOTE: The code in memory_add() now sets 'ret' if iommu_map() or
iommu_iotlb_flush() fails. This seems to be have been missed before, meaning the error path could actually return 0.

---
Cc: Jan Beulich <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Julien Grall <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Jun Nakajima <[email protected]>
Cc: Kevin Tian <[email protected]>

v5:
- Re-base
- Removed failure case on overflow of unsigned int as it is no longer necessary

v3:
- Same as v2; elected to implement batch flushing in the grant table code as a subsequent patch

v2:
- Shorten the diff (mainly because of a prior patch introducing automatic flush-on-fail into iommu_map() and iommu_unmap())

---
 xen/arch/x86/mm.c               | 24 +++++++++++++++++-------
 xen/arch/x86/mm/p2m-ept.c       | 21 +++++++++++++--------
 xen/arch/x86/mm/p2m-pt.c        | 16 ++++++++++++----
 xen/arch/x86/mm/p2m.c           | 25 +++++++++++++++++--------
 xen/arch/x86/x86_64/mm.c        | 20 +++++++-------------
 xen/common/grant_table.c        | 29 ++++++++++++++++++++++-------
 xen/common/memory.c             |  7 +++----
 xen/drivers/passthrough/iommu.c | 25 +------------------------
 xen/include/xen/iommu.h         | 21 +++++----------------
 9 files changed, 97 insertions(+), 91 deletions(-)
xen/arch/x86/mm.c +17 -7
index e133ccf0d8..f9a4ccf8d5 100644
@@ -2439,10 +2439,16 @@ static int cleanup_page_mappings(struct page_info *page)
 
         if ( d && unlikely(need_iommu_pt_sync(d)) && is_pv_domain(d) )
         {
-            int rc2 = iommu_legacy_unmap(d, _dfn(mfn), PAGE_ORDER_4K);
+            unsigned int flush_flags = 0;
+            int err;
+
+            err = iommu_unmap(d, _dfn(mfn), 1ul << PAGE_ORDER_4K, &flush_flags);
+            if ( !err )
+                err = iommu_iotlb_flush(d, _dfn(mfn), 1ul << PAGE_ORDER_4K,
+                                        flush_flags);
 
             if ( !rc )
-                rc = rc2;
+                rc = err;
         }
 
         if ( likely(!is_special_page(page)) )
@@ -2964,14 +2970,18 @@ static int _get_page_type(struct page_info *page, unsigned long type,
         if ( d && unlikely(need_iommu_pt_sync(d)) && is_pv_domain(d) )
         {
             mfn_t mfn = page_to_mfn(page);
+            dfn_t dfn = _dfn(mfn_x(mfn));
+            unsigned int flush_flags = 0;
 
             if ( (x & PGT_type_mask) == PGT_writable_page )
-                rc = iommu_legacy_unmap(d, _dfn(mfn_x(mfn)),
-                                        1ul << PAGE_ORDER_4K);
+                rc = iommu_unmap(d, dfn, 1ul << PAGE_ORDER_4K, &flush_flags);
             else
-                rc = iommu_legacy_map(d, _dfn(mfn_x(mfn)), mfn,
-                                      1ul << PAGE_ORDER_4K,
-                                      IOMMUF_readable | IOMMUF_writable);
+                rc = iommu_map(d, dfn, mfn, 1ul << PAGE_ORDER_4K,
+                               IOMMUF_readable | IOMMUF_writable, &flush_flags);
+
+            if ( !rc )
+                rc = iommu_iotlb_flush(d, dfn, 1ul << PAGE_ORDER_4K,
+                                       flush_flags);
 
             if ( unlikely(rc) )
             {
xen/arch/x86/mm/p2m-ept.c +13 -8
index 12cf38f6eb..3821063487 100644
@@ -842,15 +842,20 @@ out:
     if ( rc == 0 && p2m_is_hostp2m(p2m) &&
          need_modify_vtd_table )
     {
-        if ( iommu_use_hap_pt(d) )
-            rc = iommu_iotlb_flush(d, _dfn(gfn), 1ul << order,
-                                   (iommu_flags ? IOMMU_FLUSHF_added : 0) |
-                                   (vtd_pte_present ? IOMMU_FLUSHF_modified
-                                                    : 0));
-        else if ( need_iommu_pt_sync(d) )
+        unsigned int flush_flags = 0;
+
+        if ( need_iommu_pt_sync(d) )
             rc = iommu_flags ?
-                iommu_legacy_map(d, _dfn(gfn), mfn, 1ul << order, iommu_flags) :
-                iommu_legacy_unmap(d, _dfn(gfn), 1ul << order);
+                iommu_map(d, _dfn(gfn), mfn, 1ul << order, iommu_flags,
+                          &flush_flags) :
+                iommu_unmap(d, _dfn(gfn), 1ul << order, &flush_flags);
+        else if ( iommu_use_hap_pt(d) )
+            flush_flags =
+                (iommu_flags ? IOMMU_FLUSHF_added : 0) |
+                (vtd_pte_present ? IOMMU_FLUSHF_modified : 0);
+
+        if ( !rc )
+            rc = iommu_iotlb_flush(d, _dfn(gfn), 1ul << order, flush_flags);
     }
 
     unmap_domain_page(table);
xen/arch/x86/mm/p2m-pt.c +12 -4
index 3af51be78e..942d73d845 100644
@@ -678,10 +678,18 @@ p2m_pt_set_entry(struct p2m_domain *p2m, gfn_t gfn_, mfn_t mfn,
 
     if ( need_iommu_pt_sync(p2m->domain) &&
          (iommu_old_flags != iommu_pte_flags || old_mfn != mfn_x(mfn)) )
-        rc = iommu_pte_flags
-             ? iommu_legacy_map(d, _dfn(gfn), mfn, 1ul << page_order,
-                                iommu_pte_flags)
-             : iommu_legacy_unmap(d, _dfn(gfn), 1ul << page_order);
+    {
+        unsigned int flush_flags = 0;
+
+        rc = iommu_pte_flags ?
+            iommu_map(d, _dfn(gfn), mfn, 1ul << page_order, iommu_pte_flags,
+                      &flush_flags) :
+            iommu_unmap(d, _dfn(gfn), 1ul << page_order, &flush_flags);
+
+        if ( !rc )
+            rc = iommu_iotlb_flush(d, _dfn(gfn), 1ul << page_order,
+                                   flush_flags);
+    }
 
     /*
      * Free old intermediate tables if necessary.  This has to be the
xen/arch/x86/mm/p2m.c +17 -8
index 928344be30..01ff92862d 100644
@@ -1350,11 +1350,15 @@ int set_identity_p2m_entry(struct domain *d, unsigned long gfn_l,
 
     if ( !paging_mode_translate(p2m->domain) )
     {
-        if ( !is_iommu_enabled(d) )
-            return 0;
-        return iommu_legacy_map(d, _dfn(gfn_l), _mfn(gfn_l),
-                                1ul << PAGE_ORDER_4K,
-                                IOMMUF_readable | IOMMUF_writable);
+        unsigned int flush_flags = 0;
+
+        ret = iommu_map(d, _dfn(gfn_l), _mfn(gfn_l), 1ul << PAGE_ORDER_4K,
+                        IOMMUF_readable | IOMMUF_writable, &flush_flags);
+        if ( !ret )
+            ret = iommu_iotlb_flush(d, _dfn(gfn_l), 1ul << PAGE_ORDER_4K,
+                                    flush_flags);
+
+        return ret;
     }
 
     gfn_lock(p2m, gfn, 0);
@@ -1442,9 +1446,14 @@ int clear_identity_p2m_entry(struct domain *d, unsigned long gfn_l)
 
     if ( !paging_mode_translate(d) )
     {
-        if ( !is_iommu_enabled(d) )
-            return 0;
-        return iommu_legacy_unmap(d, _dfn(gfn_l), 1ul << PAGE_ORDER_4K);
+        unsigned int flush_flags = 0;
+
+        ret = iommu_unmap(d, _dfn(gfn_l), 1ul << PAGE_ORDER_4K, &flush_flags);
+        if ( !ret )
+            ret = iommu_iotlb_flush(d, _dfn(gfn_l), 1ul << PAGE_ORDER_4K,
+                                    flush_flags);
+
+        return ret;
     }
 
     gfn_lock(p2m, gfn, 0);
xen/arch/x86/x86_64/mm.c +7 -13
index 6f2c43181a..13b2f973fa 100644
@@ -1478,21 +1478,15 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
          !iommu_use_hap_pt(hardware_domain) &&
          !need_iommu_pt_sync(hardware_domain) )
     {
-        for ( i = spfn; i < epfn; i++ )
-            if ( iommu_legacy_map(hardware_domain, _dfn(i), _mfn(i),
-                                  1ul << PAGE_ORDER_4K,
-                                  IOMMUF_readable | IOMMUF_writable) )
-                break;
-        if ( i != epfn )
-        {
-            while (i-- > old_max)
-                /* If statement to satisfy __must_check. */
-                if ( iommu_legacy_unmap(hardware_domain, _dfn(i),
-                                        1ul << PAGE_ORDER_4K) )
-                    continue;
+        unsigned int flush_flags = 0;
+        unsigned long n = epfn - spfn;
 
+        ret = iommu_map(hardware_domain, _dfn(i), _mfn(i), n,
+                       IOMMUF_readable | IOMMUF_writable, &flush_flags);
+        if ( !ret )
+            ret = iommu_iotlb_flush(hardware_domain, _dfn(i), n, flush_flags);
+        if ( ret )
             goto destroy_m2p;
-        }
     }
 
     /* We can't revert any more */
xen/common/grant_table.c +22 -7
index c5a1ef5fc3..a844a94a5b 100644
@@ -1225,11 +1225,22 @@ map_grant_ref(
             kind = IOMMUF_readable;
         else
             kind = 0;
-        if ( kind && iommu_legacy_map(ld, _dfn(mfn_x(mfn)), mfn, 1ul, kind) )
+
+        if ( kind )
         {
-            double_gt_unlock(lgt, rgt);
-            rc = GNTST_general_error;
-            goto undo_out;
+            dfn_t dfn = _dfn(mfn_x(mfn));
+            unsigned int flush_flags = 0;
+            int err;
+
+            err = iommu_map(ld, dfn, mfn, 1ul, kind, &flush_flags);
+            if ( !err )
+                err = iommu_iotlb_flush(ld, dfn, 1ul, flush_flags);
+            if ( err )
+            {
+                double_gt_unlock(lgt, rgt);
+                rc = GNTST_general_error;
+                goto undo_out;
+            }
         }
     }
 
@@ -1473,19 +1484,23 @@ unmap_common(
     if ( rc == GNTST_okay && gnttab_need_iommu_mapping(ld) )
     {
         unsigned int kind;
+        dfn_t dfn = _dfn(mfn_x(op->mfn));
+        unsigned int flush_flags = 0;
         int err = 0;
 
         double_gt_lock(lgt, rgt);
 
         kind = mapkind(lgt, rd, op->mfn);
         if ( !kind )
-            err = iommu_legacy_unmap(ld, _dfn(mfn_x(op->mfn)), 1ul);
+            err = iommu_unmap(ld, dfn, 1ul, &flush_flags);
         else if ( !(kind & MAPKIND_WRITE) )
-            err = iommu_legacy_map(ld, _dfn(mfn_x(op->mfn)), op->mfn, 1ul,
-                                   IOMMUF_readable);
+            err = iommu_map(ld, dfn, op->mfn, 1ul, IOMMUF_readable,
+                            &flush_flags);
 
         double_gt_unlock(lgt, rgt);
 
+        if ( !err )
+            err = iommu_iotlb_flush(ld, dfn, 1ul, flush_flags);
         if ( err )
             rc = GNTST_general_error;
     }
xen/common/memory.c +3 -4
index 714077c1e5..fedbd9019e 100644
@@ -824,8 +824,7 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp,
     xatp->gpfn += start;
     xatp->size -= start;
 
-    if ( is_iommu_enabled(d) )
-       this_cpu(iommu_dont_flush_iotlb) = 1;
+    this_cpu(iommu_dont_flush_iotlb) = true;
 
     while ( xatp->size > done )
     {
@@ -845,12 +844,12 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp,
         }
     }
 
+    this_cpu(iommu_dont_flush_iotlb) = false;
+
     if ( is_iommu_enabled(d) )
     {
         int ret;
 
-        this_cpu(iommu_dont_flush_iotlb) = 0;
-
         ret = iommu_iotlb_flush(d, _dfn(xatp->idx - done), done,
                                 IOMMU_FLUSHF_added | IOMMU_FLUSHF_modified);
         if ( unlikely(ret) && rc >= 0 )
xen/drivers/passthrough/iommu.c +1 -24
index e664f4ac79..e0d36e6bef 100644
@@ -282,18 +282,6 @@ int iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
     return rc;
 }
 
-int iommu_legacy_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-                     unsigned long page_count, unsigned int flags)
-{
-    unsigned int flush_flags = 0;
-    int rc = iommu_map(d, dfn, mfn, page_count, flags, &flush_flags);
-
-    if ( !this_cpu(iommu_dont_flush_iotlb) && !rc )
-        rc = iommu_iotlb_flush(d, dfn, page_count, flush_flags);
-
-    return rc;
-}
-
 int iommu_unmap(struct domain *d, dfn_t dfn, unsigned long page_count,
                 unsigned int *flush_flags)
 {
@@ -338,17 +326,6 @@ int iommu_unmap(struct domain *d, dfn_t dfn, unsigned long page_count,
     return rc;
 }
 
-int iommu_legacy_unmap(struct domain *d, dfn_t dfn, unsigned long page_count)
-{
-    unsigned int flush_flags = 0;
-    int rc = iommu_unmap(d, dfn, page_count, &flush_flags);
-
-    if ( !this_cpu(iommu_dont_flush_iotlb) && ! rc )
-        rc = iommu_iotlb_flush(d, dfn, page_count, flush_flags);
-
-    return rc;
-}
-
 int iommu_lookup_page(struct domain *d, dfn_t dfn, mfn_t *mfn,
                       unsigned int *flags)
 {
@@ -367,7 +344,7 @@ int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned long page_count,
     int rc;
 
     if ( !is_iommu_enabled(d) || !hd->platform_ops->iotlb_flush ||
-         !page_count || !flush_flags )
+         !page_count || !flush_flags || this_cpu(iommu_dont_flush_iotlb) )
         return 0;
 
     if ( dfn_eq(dfn, INVALID_DFN) )
xen/include/xen/iommu.h +5 -16
index 13f68dc93d..a2eefe8582 100644
@@ -151,16 +151,8 @@ int __must_check iommu_map(struct domain *d, dfn_t dfn, mfn_t mfn,
 int __must_check iommu_unmap(struct domain *d, dfn_t dfn,
                              unsigned long page_count,
                              unsigned int *flush_flags);
-
-int __must_check iommu_legacy_map(struct domain *d, dfn_t dfn, mfn_t mfn,
-                                  unsigned long page_count,
-                                  unsigned int flags);
-int __must_check iommu_legacy_unmap(struct domain *d, dfn_t dfn,
-                                    unsigned long page_count);
-
 int __must_check iommu_lookup_page(struct domain *d, dfn_t dfn, mfn_t *mfn,
                                    unsigned int *flags);
-
 int __must_check iommu_iotlb_flush(struct domain *d, dfn_t dfn,
                                    unsigned long page_count,
                                    unsigned int flush_flags);
@@ -369,15 +361,12 @@ void iommu_dev_iotlb_flush_timeout(struct domain *d, struct pci_dev *pdev);
 
 /*
  * The purpose of the iommu_dont_flush_iotlb optional cpu flag is to
- * avoid unecessary iotlb_flush in the low level IOMMU code.
- *
- * iommu_map_page/iommu_unmap_page must flush the iotlb but somethimes
- * this operation can be really expensive. This flag will be set by the
- * caller to notify the low level IOMMU code to avoid the iotlb flushes.
- * iommu_iotlb_flush/iommu_iotlb_flush_all will be explicitly called by
- * the caller.
+ * avoid unnecessary IOMMU flushing while updating the P2M.
+ * Setting the value to true will cause iommu_iotlb_flush() to return without
+ * actually performing a flush. A batch flush must therefore be done by the
+ * calling code after setting the value back to false.
  */
-DECLARE_PER_CPU(bool_t, iommu_dont_flush_iotlb);
+DECLARE_PER_CPU(bool, iommu_dont_flush_iotlb);
 
 extern struct spinlock iommu_pt_cleanup_lock;
 extern struct page_list_head iommu_pt_cleanup_list;
Signature · 1 line
2.20.1
Reviewed-by on v5 5/8
#9 lore raw

Re: [PATCH v5 5/8] remove remaining uses of iommu_legacy_map/unmap

Quoted 3 lines · Paul Durrant

NOTE: The code in memory_add() now sets 'ret' if iommu_map() or
iommu_iotlb_flush() fails. This seems to be have been missed before, meaning the error path could actually return 0.

I agree this is the better way, but I'm not sure it really was
unintended: It could well have been considered a "best effort" approach to get IOMMU mappings in place, back at the time.

Quoted 1 line

with one nit and one further remark:

Quoted 13 lines
xen/arch/x86/x86_64/mm.c +0 -7
@@ -1478,21 +1478,15 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
          !iommu_use_hap_pt(hardware_domain) &&
          !need_iommu_pt_sync(hardware_domain) )
     {
-        for ( i = spfn; i < epfn; i++ )
-            if ( iommu_legacy_map(hardware_domain, _dfn(i), _mfn(i),
-                                  1ul << PAGE_ORDER_4K,
-                                  IOMMUF_readable | IOMMUF_writable) )
-                break;
-        if ( i != epfn )
-        {

xen/arch/x86/x86_64/mm.c int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm) · line 1485

Quoted 10 lines
xen/arch/x86/x86_64/mm.c +4 -5
-            while (i-- > old_max)
-                /* If statement to satisfy __must_check. */
-                if ( iommu_legacy_unmap(hardware_domain, _dfn(i),
-                                        1ul << PAGE_ORDER_4K) )
-                    continue;
+        unsigned int flush_flags = 0;
+        unsigned long n = epfn - spfn;
 
+        ret = iommu_map(hardware_domain, _dfn(i), _mfn(i), n,
+                       IOMMUF_readable | IOMMUF_writable, &flush_flags);

There's one blank too little here.

xen/arch/x86/x86_64/mm.c int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned long page_count, · line 348

Quoted 7 lines
xen/arch/x86/x86_64/mm.c +1 -1
@@ -367,7 +344,7 @@ int iommu_iotlb_flush(struct domain *d, dfn_t dfn, unsigned long page_count,
     int rc;
 
     if ( !is_iommu_enabled(d) || !hd->platform_ops->iotlb_flush ||
-         !page_count || !flush_flags )
+         !page_count || !flush_flags || this_cpu(iommu_dont_flush_iotlb) )
         return 0;

I'm still somewhat uneasy about this change, because there could in theory be fallout. I don't think there is in practice, so I'm willing to let it be this way. Long term we wan want to introduce a flag into flush_flags to allow overriding this behavior.

Jan

#10 lore raw

PATCH v5 6/8 common/grant_table: batch flush I/O TLB

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

This patch avoids calling iommu_iotlb_flush() for each individual GNTTABOP and instead calls iommu_iotlb_flush_all() at the end of a batch. This should mean non-singleton map/unmap operations perform better.

NOTE: A batch is the number of operations done before a pre-emption check and,
in the case of unmap, a TLB flush.

---
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Jan Beulich <[email protected]>
Cc: Julien Grall <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Wei Liu <[email protected]>

v5:
- Add batching to gnttab_map_grant_ref() to handle flushing before pre- emption check
- Maintain per-op flushing in the case of singletons

v3:
- New in v3

---
 xen/common/grant_table.c | 199 ++++++++++++++++++++++++++-------------
 1 file changed, 134 insertions(+), 65 deletions(-)
xen/common/grant_table.c +134 -65
index a844a94a5b..8c6d1dcea7 100644
@@ -241,7 +241,13 @@ struct gnttab_unmap_common {
     grant_ref_t ref;
 };
 
-/* Number of unmap operations that are done between each tlb flush */
+/* Number of map operations that are done between each pre-emption check */
+#define GNTTAB_MAP_BATCH_SIZE 32
+
+/*
+ * Number of unmap operations that are done between each tlb flush and
+ * pre-emption check.
+ */
 #define GNTTAB_UNMAP_BATCH_SIZE 32
 
 
@@ -979,7 +985,7 @@ static unsigned int mapkind(
 
 static void
 map_grant_ref(
-    struct gnttab_map_grant_ref *op)
+    struct gnttab_map_grant_ref *op, bool do_flush, unsigned int *flush_flags)
 {
     struct domain *ld, *rd, *owner = NULL;
     struct grant_table *lgt, *rgt;
@@ -1229,12 +1235,11 @@ map_grant_ref(
         if ( kind )
         {
             dfn_t dfn = _dfn(mfn_x(mfn));
-            unsigned int flush_flags = 0;
             int err;
 
-            err = iommu_map(ld, dfn, mfn, 1ul, kind, &flush_flags);
-            if ( !err )
-                err = iommu_iotlb_flush(ld, dfn, 1ul, flush_flags);
+            err = iommu_map(ld, dfn, mfn, 1ul, kind, flush_flags);
+            if ( do_flush && !err )
+                err = iommu_iotlb_flush(ld, dfn, 1ul, *flush_flags);
             if ( err )
             {
                 double_gt_unlock(lgt, rgt);
@@ -1319,29 +1324,60 @@ static long
 gnttab_map_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_map_grant_ref_t) uop, unsigned int count)
 {
-    int i;
-    struct gnttab_map_grant_ref op;
+    struct domain *currd = current->domain;
+    unsigned int done = 0;
+    int rc = 0;
 
-    for ( i = 0; i < count; i++ )
+    while ( count )
     {
-        if ( i && hypercall_preempt_check() )
-            return i;
+        unsigned int i, c = min_t(unsigned int, count, GNTTAB_MAP_BATCH_SIZE);
+        unsigned int flush_flags = 0;
 
-        if ( unlikely(__copy_from_guest_offset(&op, uop, i, 1)) )
-            return -EFAULT;
+        for ( i = 0; i < c; i++ )
+        {
+            struct gnttab_map_grant_ref op;
 
-        map_grant_ref(&op);
+            if ( unlikely(__copy_from_guest(&op, uop, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
 
-        if ( unlikely(__copy_to_guest_offset(uop, i, &op, 1)) )
-            return -EFAULT;
+            map_grant_ref(&op, c == 1, &flush_flags);
+
+            if ( unlikely(__copy_to_guest(uop, &op, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
+
+            guest_handle_add_offset(uop, 1);
+        }
+
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);
+
+            if ( !rc )
+                rc = err;
+        }
+
+        if ( rc )
+            break;
+
+        count -= c;
+        done += c;
+
+        if ( count && hypercall_preempt_check() )
+            return done;
     }
 
-    return 0;
+    return rc;
 }
 
 static void
 unmap_common(
-    struct gnttab_unmap_common *op)
+    struct gnttab_unmap_common *op, bool do_flush, unsigned int *flush_flags)
 {
     domid_t          dom;
     struct domain   *ld, *rd;
@@ -1485,22 +1521,21 @@ unmap_common(
     {
         unsigned int kind;
         dfn_t dfn = _dfn(mfn_x(op->mfn));
-        unsigned int flush_flags = 0;
         int err = 0;
 
         double_gt_lock(lgt, rgt);
 
         kind = mapkind(lgt, rd, op->mfn);
         if ( !kind )
-            err = iommu_unmap(ld, dfn, 1ul, &flush_flags);
+            err = iommu_unmap(ld, dfn, 1ul, flush_flags);
         else if ( !(kind & MAPKIND_WRITE) )
             err = iommu_map(ld, dfn, op->mfn, 1ul, IOMMUF_readable,
-                            &flush_flags);
+                            flush_flags);
 
         double_gt_unlock(lgt, rgt);
 
-        if ( !err )
-            err = iommu_iotlb_flush(ld, dfn, 1ul, flush_flags);
+        if ( do_flush && !err )
+            err = iommu_iotlb_flush(ld, dfn, 1ul, *flush_flags);
         if ( err )
             rc = GNTST_general_error;
     }
@@ -1599,8 +1634,8 @@ unmap_common_complete(struct gnttab_unmap_common *op)
 
 static void
 unmap_grant_ref(
-    struct gnttab_unmap_grant_ref *op,
-    struct gnttab_unmap_common *common)
+    struct gnttab_unmap_grant_ref *op, struct gnttab_unmap_common *common,
+    bool do_flush, unsigned int *flush_flags)
 {
     common->host_addr = op->host_addr;
     common->dev_bus_addr = op->dev_bus_addr;
@@ -1612,7 +1647,7 @@ unmap_grant_ref(
     common->rd = NULL;
     common->mfn = INVALID_MFN;
 
-    unmap_common(common);
+    unmap_common(common, do_flush, flush_flags);
     op->status = common->status;
 }
 
@@ -1621,31 +1656,55 @@ static long
 gnttab_unmap_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_unmap_grant_ref_t) uop, unsigned int count)
 {
-    int i, c, partial_done, done = 0;
-    struct gnttab_unmap_grant_ref op;
-    struct gnttab_unmap_common common[GNTTAB_UNMAP_BATCH_SIZE];
+    struct domain *currd = current->domain;
+    unsigned int done = 0;
+    int rc = 0;
 
-    while ( count != 0 )
+    while ( count )
     {
-        c = min(count, (unsigned int)GNTTAB_UNMAP_BATCH_SIZE);
-        partial_done = 0;
+        struct gnttab_unmap_common common[GNTTAB_UNMAP_BATCH_SIZE];
+        unsigned int i, c, partial_done = 0;
+        unsigned int flush_flags = 0;
+
+        c = min_t(unsigned int, count, GNTTAB_UNMAP_BATCH_SIZE);
 
         for ( i = 0; i < c; i++ )
         {
+            struct gnttab_unmap_grant_ref op;
+
             if ( unlikely(__copy_from_guest(&op, uop, 1)) )
-                goto fault;
-            unmap_grant_ref(&op, &common[i]);
+            {
+                rc = -EFAULT;
+                break;
+            }
+
+            unmap_grant_ref(&op, &common[i], c == 1, &flush_flags);
             ++partial_done;
+
             if ( unlikely(__copy_field_to_guest(uop, &op, status)) )
-                goto fault;
+            {
+                rc = -EFAULT;
+                break;
+            }
+
             guest_handle_add_offset(uop, 1);
         }
 
-        gnttab_flush_tlb(current->domain);
+        gnttab_flush_tlb(currd);
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);
+
+            if ( !rc )
+                rc = err;
+        }
 
         for ( i = 0; i < partial_done; i++ )
             unmap_common_complete(&common[i]);
 
+        if ( rc )
+            break;
+
         count -= c;
         done += c;
 
@@ -1653,20 +1712,13 @@ gnttab_unmap_grant_ref(
             return done;
     }
 
-    return 0;
-
-fault:
-    gnttab_flush_tlb(current->domain);
-
-    for ( i = 0; i < partial_done; i++ )
-        unmap_common_complete(&common[i]);
-    return -EFAULT;
+    return rc;
 }
 
 static void
 unmap_and_replace(
-    struct gnttab_unmap_and_replace *op,
-    struct gnttab_unmap_common *common)
+    struct gnttab_unmap_and_replace *op, struct gnttab_unmap_common *common,
+    bool do_flush, unsigned int *flush_flags)
 {
     common->host_addr = op->host_addr;
     common->new_addr = op->new_addr;
@@ -1678,7 +1730,7 @@ unmap_and_replace(
     common->rd = NULL;
     common->mfn = INVALID_MFN;
 
-    unmap_common(common);
+    unmap_common(common, do_flush, flush_flags);
     op->status = common->status;
 }
 
@@ -1686,31 +1738,55 @@ static long
 gnttab_unmap_and_replace(
     XEN_GUEST_HANDLE_PARAM(gnttab_unmap_and_replace_t) uop, unsigned int count)
 {
-    int i, c, partial_done, done = 0;
-    struct gnttab_unmap_and_replace op;
-    struct gnttab_unmap_common common[GNTTAB_UNMAP_BATCH_SIZE];
+    struct domain *currd = current->domain;
+    unsigned int done = 0;
+    int rc = 0;
 
-    while ( count != 0 )
+    while ( count )
     {
-        c = min(count, (unsigned int)GNTTAB_UNMAP_BATCH_SIZE);
-        partial_done = 0;
+        struct gnttab_unmap_common common[GNTTAB_UNMAP_BATCH_SIZE];
+        unsigned int i, c, partial_done = 0;
+        unsigned int flush_flags = 0;
+
+        c = min_t(unsigned int, count, GNTTAB_UNMAP_BATCH_SIZE);
 
         for ( i = 0; i < c; i++ )
         {
+            struct gnttab_unmap_and_replace op;
+
             if ( unlikely(__copy_from_guest(&op, uop, 1)) )
-                goto fault;
-            unmap_and_replace(&op, &common[i]);
+            {
+                rc = -EFAULT;
+                break;
+            }
+
+            unmap_and_replace(&op, &common[i], c == 1, &flush_flags);
             ++partial_done;
+
             if ( unlikely(__copy_field_to_guest(uop, &op, status)) )
-                goto fault;
+            {
+                rc = -EFAULT;
+                break;
+            }
+
             guest_handle_add_offset(uop, 1);
         }
 
-        gnttab_flush_tlb(current->domain);
+        gnttab_flush_tlb(currd);
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);
+
+            if ( !rc )
+                rc = err;
+        }
 
         for ( i = 0; i < partial_done; i++ )
             unmap_common_complete(&common[i]);
 
+        if ( rc )
+            break;
+
         count -= c;
         done += c;
 
@@ -1718,14 +1794,7 @@ gnttab_unmap_and_replace(
             return done;
     }
 
-    return 0;
-
-fault:
-    gnttab_flush_tlb(current->domain);
-
-    for ( i = 0; i < partial_done; i++ )
-        unmap_common_complete(&common[i]);
-    return -EFAULT;
+    return rc;
 }
 
 static int
Signature · 1 line
2.20.1
#11 lore raw

Re: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 5 lines · Paul Durrant
xen/common/grant_table.c +0 -0
@@ -241,7 +241,13 @@ struct gnttab_unmap_common {
     grant_ref_t ref;
 };

xen/common/grant_table.c struct gnttab_unmap_common { · line 251

Quoted 10 lines
xen/common/grant_table.c +7 -1
 
-/* Number of unmap operations that are done between each tlb flush */
+/* Number of map operations that are done between each pre-emption check */
+#define GNTTAB_MAP_BATCH_SIZE 32
+
+/*
+ * Number of unmap operations that are done between each tlb flush and
+ * pre-emption check.
+ */
 #define GNTTAB_UNMAP_BATCH_SIZE 32

Interesting - I don't think I've ever seen preemption spelled like this (with a hyphen). In the interest of grep-ability, would you mind changing to the more "conventional" spelling? Albeit I now notice we have two such spellings in the tree already ...

xen/common/grant_table.c static unsigned int mapkind( · line 988

Quoted 6 lines
xen/common/grant_table.c +1 -1
@@ -979,7 +985,7 @@ static unsigned int mapkind(
 
 static void
 map_grant_ref(
-    struct gnttab_map_grant_ref *op)
+    struct gnttab_map_grant_ref *op, bool do_flush, unsigned int *flush_flags)

Why two parameters? Simply pass NULL for singletons instead? Although, you'd need another local variable then, which maybe isn't overly much nicer.

xen/common/grant_table.c static long · line 1327

Quoted 7 lines
xen/common/grant_table.c +1 -2
@@ -1319,29 +1324,60 @@ static long
 gnttab_map_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_map_grant_ref_t) uop, unsigned int count)
 {
-    int i;
-    struct gnttab_map_grant_ref op;
+    struct domain *currd = current->domain;

Is this a worthwhile variable to have in this case? It's used exactly once, granted in the loop body, but still not the inner one.

Quoted 30 lines
xen/common/grant_table.c +17 -8
+    unsigned int done = 0;
+    int rc = 0;
 
-    for ( i = 0; i < count; i++ )
+    while ( count )
     {
-        if ( i && hypercall_preempt_check() )
-            return i;
+        unsigned int i, c = min_t(unsigned int, count, GNTTAB_MAP_BATCH_SIZE);
+        unsigned int flush_flags = 0;
 
-        if ( unlikely(__copy_from_guest_offset(&op, uop, i, 1)) )
-            return -EFAULT;
+        for ( i = 0; i < c; i++ )
+        {
+            struct gnttab_map_grant_ref op;
 
-        map_grant_ref(&op);
+            if ( unlikely(__copy_from_guest(&op, uop, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
 
-        if ( unlikely(__copy_to_guest_offset(uop, i, &op, 1)) )
-            return -EFAULT;
+            map_grant_ref(&op, c == 1, &flush_flags);
+
+            if ( unlikely(__copy_to_guest(uop, &op, 1)) )
+            {

xen/common/grant_table.c static long · line 1359

Quoted 10 lines
xen/common/grant_table.c +10 -0
+                rc = -EFAULT;
+                break;
+            }
+
+            guest_handle_add_offset(uop, 1);
+        }
+
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);

There's still some double flushing involved in the error case here. Trying to eliminate this (by having map_grant_ref() write zero into *flush_flags) may not be overly important, but I thought I'd still mention it.

Jan

#12 lore raw

RE: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 17 lines

-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 10 September 2020 14:49
To: Paul Durrant <[email protected]>
Cc: [email protected]; Paul Durrant <[email protected]>; Andrew Cooper <[email protected]>; George Dunlap <[email protected]>; Ian Jackson <[email protected]>; Julien Grall <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>
Subject: Re: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 6 lines · Paul Durrant
xen/common/grant_table.c +0 -0
@@ -241,7 +241,13 @@ struct gnttab_unmap_common {
     grant_ref_t ref;
 };
 
Quoted 15 lines

xen/common/grant_table.c struct gnttab_unmap_common { · line 251

Quoted 9 lines
xen/common/grant_table.c +7 -1
-/* Number of unmap operations that are done between each tlb flush */
+/* Number of map operations that are done between each pre-emption check */
+#define GNTTAB_MAP_BATCH_SIZE 32
+
+/*
+ * Number of unmap operations that are done between each tlb flush and
+ * pre-emption check.
+ */
 #define GNTTAB_UNMAP_BATCH_SIZE 32

Interesting - I don't think I've ever seen preemption spelled like this (with a hyphen). In the interest of grep-ability, would you mind changing to the more "conventional" spelling? Albeit I now notice we have two such spellings in the tree already ...

Sure, I'll change it.

Quoted 11 lines

xen/common/grant_table.c static unsigned int mapkind( · line 988

Quoted 6 lines
xen/common/grant_table.c +1 -1
@@ -979,7 +985,7 @@ static unsigned int mapkind(
 
 static void
 map_grant_ref(
-    struct gnttab_map_grant_ref *op)
+    struct gnttab_map_grant_ref *op, bool do_flush, unsigned int *flush_flags)

Why two parameters? Simply pass NULL for singletons instead? Although, you'd need another local variable then, which maybe isn't overly much nicer.

No, I think the extra arg is clearer.

Quoted 12 lines

xen/common/grant_table.c static long · line 1327

Quoted 7 lines
xen/common/grant_table.c +1 -2
@@ -1319,29 +1324,60 @@ static long
 gnttab_map_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_map_grant_ref_t) uop, unsigned int count)
 {
-    int i;
-    struct gnttab_map_grant_ref op;
+    struct domain *currd = current->domain;

Is this a worthwhile variable to have in this case? It's used exactly once, granted in the loop body, but still not the inner one.

I thought it was nicer for consistency with the unmap function (where curd is used more than once) but I'll drop it.

Quoted 37 lines
Quoted 37 lines
xen/common/grant_table.c +24 -8
+    unsigned int done = 0;
+    int rc = 0;
 
-    for ( i = 0; i < count; i++ )
+    while ( count )
     {
-        if ( i && hypercall_preempt_check() )
-            return i;
+        unsigned int i, c = min_t(unsigned int, count, GNTTAB_MAP_BATCH_SIZE);
+        unsigned int flush_flags = 0;
 
-        if ( unlikely(__copy_from_guest_offset(&op, uop, i, 1)) )
-            return -EFAULT;
+        for ( i = 0; i < c; i++ )
+        {
+            struct gnttab_map_grant_ref op;
 
-        map_grant_ref(&op);
+            if ( unlikely(__copy_from_guest(&op, uop, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
 
-        if ( unlikely(__copy_to_guest_offset(uop, i, &op, 1)) )
-            return -EFAULT;
+            map_grant_ref(&op, c == 1, &flush_flags);
+
+            if ( unlikely(__copy_to_guest(uop, &op, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
+
+            guest_handle_add_offset(uop, 1);
+        }
+
Quoted 9 lines

xen/common/grant_table.c static long · line 1359

Quoted 3 lines
xen/common/grant_table.c +3 -0
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);

There's still some double flushing involved in the error case here. Trying to eliminate this (by having map_grant_ref() write zero into *flush_flags) may not be overly important, but I thought I'd still mention it.

This only kicks in if there's more than one operation and is it safe to squash the flush_flags if some ops succeed and others fail? If all ops fail then flush_flags should still be zero because the call to iommu_map() is the last failure point in map_grant_ref() anyway.

Paul

Quoted 1 line

Jan

Reviewed-by
#13 lore raw

Re: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 104 lines · Paul Durrant
Quoted 10 lines

From: Jan Beulich <[email protected]>
Sent: 10 September 2020 14:49

Quoted 6 lines · Paul Durrant
xen/common/grant_table.c +0 -0
@@ -241,7 +241,13 @@ struct gnttab_unmap_common {
     grant_ref_t ref;
 };
 
Quoted 15 lines

xen/common/grant_table.c struct gnttab_unmap_common { · line 251

Quoted 9 lines
xen/common/grant_table.c +7 -1
-/* Number of unmap operations that are done between each tlb flush */
+/* Number of map operations that are done between each pre-emption check */
+#define GNTTAB_MAP_BATCH_SIZE 32
+
+/*
+ * Number of unmap operations that are done between each tlb flush and
+ * pre-emption check.
+ */
 #define GNTTAB_UNMAP_BATCH_SIZE 32

Interesting - I don't think I've ever seen preemption spelled like this (with a hyphen). In the interest of grep-ability, would you mind changing to the more "conventional" spelling? Albeit I now notice we have two such spellings in the tree already ...

Sure, I'll change it.

Quoted 11 lines

xen/common/grant_table.c static unsigned int mapkind( · line 988

Quoted 6 lines
xen/common/grant_table.c +1 -1
@@ -979,7 +985,7 @@ static unsigned int mapkind(
 
 static void
 map_grant_ref(
-    struct gnttab_map_grant_ref *op)
+    struct gnttab_map_grant_ref *op, bool do_flush, unsigned int *flush_flags)

Why two parameters? Simply pass NULL for singletons instead? Although, you'd need another local variable then, which maybe isn't overly much nicer.

No, I think the extra arg is clearer.

Quoted 12 lines

xen/common/grant_table.c static long · line 1327

Quoted 7 lines
xen/common/grant_table.c +1 -2
@@ -1319,29 +1324,60 @@ static long
 gnttab_map_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_map_grant_ref_t) uop, unsigned int count)
 {
-    int i;
-    struct gnttab_map_grant_ref op;
+    struct domain *currd = current->domain;

Is this a worthwhile variable to have in this case? It's used exactly once, granted in the loop body, but still not the inner one.

I thought it was nicer for consistency with the unmap function (where curd is used more than once) but I'll drop it.

Quoted 37 lines
Quoted 37 lines
xen/common/grant_table.c +24 -8
+    unsigned int done = 0;
+    int rc = 0;
 
-    for ( i = 0; i < count; i++ )
+    while ( count )
     {
-        if ( i && hypercall_preempt_check() )
-            return i;
+        unsigned int i, c = min_t(unsigned int, count, GNTTAB_MAP_BATCH_SIZE);
+        unsigned int flush_flags = 0;
 
-        if ( unlikely(__copy_from_guest_offset(&op, uop, i, 1)) )
-            return -EFAULT;
+        for ( i = 0; i < c; i++ )
+        {
+            struct gnttab_map_grant_ref op;
 
-        map_grant_ref(&op);
+            if ( unlikely(__copy_from_guest(&op, uop, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
 
-        if ( unlikely(__copy_to_guest_offset(uop, i, &op, 1)) )
-            return -EFAULT;
+            map_grant_ref(&op, c == 1, &flush_flags);
+
+            if ( unlikely(__copy_to_guest(uop, &op, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
+
+            guest_handle_add_offset(uop, 1);
+        }
+
Quoted 9 lines

xen/common/grant_table.c static long · line 1359

Quoted 3 lines
xen/common/grant_table.c +3 -0
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);

There's still some double flushing involved in the error case here. Trying to eliminate this (by having map_grant_ref() write zero into *flush_flags) may not be overly important, but I thought I'd still mention it.

Quoted 1 line

This only kicks in if there's more than one operation and is it safe to squash the flush_flags if some ops succeed and others fail? If all ops fail then flush_flags should still be zero because the call to iommu_map() is the last failure point in map_grant_ref() anyway.

Well, yes, not a common thing to run into. With the small changes further up

Jan

#14 lore raw

RE: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 116 lines

-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 10 September 2020 15:39
To: [email protected]
Cc: [email protected]; 'Paul Durrant' <[email protected]>; 'Andrew Cooper' <[email protected]>; 'George Dunlap' <[email protected]>; 'Ian Jackson' <[email protected]>; 'Julien Grall' <[email protected]>; 'Stefano Stabellini' <[email protected]>; 'Wei Liu' <[email protected]>
Subject: Re: [PATCH v5 6/8] common/grant_table: batch flush I/O TLB

Quoted 55 lines · Paul Durrant
Quoted 10 lines

From: Jan Beulich <[email protected]>
Sent: 10 September 2020 14:49

Quoted 6 lines · Paul Durrant
xen/common/grant_table.c +0 -0
@@ -241,7 +241,13 @@ struct gnttab_unmap_common {
     grant_ref_t ref;
 };
 
Quoted 15 lines

xen/common/grant_table.c struct gnttab_unmap_common { · line 251

Quoted 9 lines
xen/common/grant_table.c +7 -1
-/* Number of unmap operations that are done between each tlb flush */
+/* Number of map operations that are done between each pre-emption check */
+#define GNTTAB_MAP_BATCH_SIZE 32
+
+/*
+ * Number of unmap operations that are done between each tlb flush and
+ * pre-emption check.
+ */
 #define GNTTAB_UNMAP_BATCH_SIZE 32

Interesting - I don't think I've ever seen preemption spelled like this (with a hyphen). In the interest of grep-ability, would you mind changing to the more "conventional" spelling? Albeit I now notice we have two such spellings in the tree already ...

Sure, I'll change it.

Quoted 11 lines

xen/common/grant_table.c static unsigned int mapkind( · line 988

Quoted 6 lines
xen/common/grant_table.c +1 -1
@@ -979,7 +985,7 @@ static unsigned int mapkind(
 
 static void
 map_grant_ref(
-    struct gnttab_map_grant_ref *op)
+    struct gnttab_map_grant_ref *op, bool do_flush, unsigned int *flush_flags)

Why two parameters? Simply pass NULL for singletons instead? Although, you'd need another local variable then, which maybe isn't overly much nicer.

No, I think the extra arg is clearer.

Quoted 12 lines

xen/common/grant_table.c static long · line 1327

Quoted 7 lines
xen/common/grant_table.c +1 -2
@@ -1319,29 +1324,60 @@ static long
 gnttab_map_grant_ref(
     XEN_GUEST_HANDLE_PARAM(gnttab_map_grant_ref_t) uop, unsigned int count)
 {
-    int i;
-    struct gnttab_map_grant_ref op;
+    struct domain *currd = current->domain;

Is this a worthwhile variable to have in this case? It's used exactly once, granted in the loop body, but still not the inner one.

Quoted 1 line

I thought it was nicer for consistency with the unmap function (where curd is used more than once)

but I'll drop it.

Quoted 48 lines
Quoted 47 lines
Quoted 31 lines
xen/common/grant_table.c +17 -8
+    unsigned int done = 0;
+    int rc = 0;
 
-    for ( i = 0; i < count; i++ )
+    while ( count )
     {
-        if ( i && hypercall_preempt_check() )
-            return i;
+        unsigned int i, c = min_t(unsigned int, count, GNTTAB_MAP_BATCH_SIZE);
+        unsigned int flush_flags = 0;
 
-        if ( unlikely(__copy_from_guest_offset(&op, uop, i, 1)) )
-            return -EFAULT;
+        for ( i = 0; i < c; i++ )
+        {
+            struct gnttab_map_grant_ref op;
 
-        map_grant_ref(&op);
+            if ( unlikely(__copy_from_guest(&op, uop, 1)) )
+            {
+                rc = -EFAULT;
+                break;
+            }
 
-        if ( unlikely(__copy_to_guest_offset(uop, i, &op, 1)) )
-            return -EFAULT;
+            map_grant_ref(&op, c == 1, &flush_flags);
+
+            if ( unlikely(__copy_to_guest(uop, &op, 1)) )
+            {

xen/common/grant_table.c static long · line 1359

Quoted 10 lines
xen/common/grant_table.c +10 -0
+                rc = -EFAULT;
+                break;
+            }
+
+            guest_handle_add_offset(uop, 1);
+        }
+
+        if ( c > 1 )
+        {
+            int err = iommu_iotlb_flush_all(currd, flush_flags);

There's still some double flushing involved in the error case here. Trying to eliminate this (by having map_grant_ref() write zero into *flush_flags) may not be overly important, but I thought I'd still mention it.

Quoted 8 lines
Quoted 1 line

This only kicks in if there's more than one operation and is it safe to squash the flush_flags if

some ops succeed and others fail? If all ops fail then flush_flags should still be zero because the call to iommu_map() is the last failure point in map_grant_ref() anyway.

Well, yes, not a common thing to run into. With the small changes further up

Thanks. Will send v6 shortly.

Paul

Quoted 1 line

Jan

#15 lore raw

PATCH v5 7/8 iommu: remove the share_p2m operation

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

Sharing of HAP tables is now VT-d specific so the operation is never defined for AMD IOMMU any more. There's also no need to pro-actively set vtd.pgd_maddr when using shared EPT as it is straightforward to simply define a helper function to return the appropriate value in the shared and non-shared cases.

NOTE: This patch also modifies unmap_vtd_domain_page() to take a const
pointer since the only thing it calls, unmap_domain_page(), also takes a const pointer.

---
Cc: Jan Beulich <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Kevin Tian <[email protected]>

v5:
- Pass 'nr_pt_levels' into domain_pgd_maddr() directly

v2:
- Put the PGD level adjust into the helper function too, since it is irrelevant in the shared EPT case

---
 xen/arch/x86/mm/p2m.c                 |  3 -
 xen/drivers/passthrough/iommu.c       |  8 ---
 xen/drivers/passthrough/vtd/extern.h  |  2 +-
 xen/drivers/passthrough/vtd/iommu.c   | 90 +++++++++++++++------------
 xen/drivers/passthrough/vtd/x86/vtd.c |  2 +-
 xen/include/xen/iommu.h               |  3 -
 6 files changed, 52 insertions(+), 56 deletions(-)
xen/arch/x86/mm/p2m.c +0 -3
index 01ff92862d..d382199c88 100644
@@ -726,9 +726,6 @@ int p2m_alloc_table(struct p2m_domain *p2m)
 
     p2m->phys_table = pagetable_from_mfn(top_mfn);
 
-    if ( hap_enabled(d) )
-        iommu_share_p2m_table(d);
-
     p2m_unlock(p2m);
     return 0;
 }
xen/drivers/passthrough/iommu.c +0 -8
index e0d36e6bef..e9b6c9a1ec 100644
@@ -500,14 +500,6 @@ int iommu_do_domctl(
     return ret;
 }
 
-void iommu_share_p2m_table(struct domain* d)
-{
-    ASSERT(hap_enabled(d));
-
-    if ( iommu_use_hap_pt(d) )
-        iommu_get_ops()->share_p2m(d);
-}
-
 void iommu_crash_shutdown(void)
 {
     if ( !iommu_crash_disable )
xen/drivers/passthrough/vtd/extern.h +1 -1
index ada3c3098c..9cf5b578c9 100644
@@ -72,7 +72,7 @@ void flush_all_cache(void);
 uint64_t alloc_pgtable_maddr(unsigned long npages, nodeid_t node);
 void free_pgtable_maddr(u64 maddr);
 void *map_vtd_domain_page(u64 maddr);
-void unmap_vtd_domain_page(void *va);
+void unmap_vtd_domain_page(const void *va);
 int domain_context_mapping_one(struct domain *domain, struct vtd_iommu *iommu,
                                u8 bus, u8 devfn, const struct pci_dev *);
 int domain_context_unmap_one(struct domain *domain, struct vtd_iommu *iommu,
xen/drivers/passthrough/vtd/iommu.c +50 -40

132 lines not on this page · raw message

xen/drivers/passthrough/vtd/x86/vtd.c +1 -1
index bbe358dc36..6681dccd69 100644
@@ -42,7 +42,7 @@ void *map_vtd_domain_page(u64 maddr)
     return map_domain_page(_mfn(paddr_to_pfn(maddr)));
 }
 
-void unmap_vtd_domain_page(void *va)
+void unmap_vtd_domain_page(const void *va)
 {
     unmap_domain_page(va);
 }
xen/include/xen/iommu.h +0 -3
index a2eefe8582..373145266f 100644
@@ -270,7 +270,6 @@ struct iommu_ops {
 
     int __must_check (*suspend)(void);
     void (*resume)(void);
-    void (*share_p2m)(struct domain *d);
     void (*crash_shutdown)(void);
     int __must_check (*iotlb_flush)(struct domain *d, dfn_t dfn,
                                     unsigned long page_count,
@@ -347,8 +346,6 @@ void iommu_resume(void);
 void iommu_crash_shutdown(void);
 int iommu_get_reserved_device_memory(iommu_grdm_t *, void *);
 
-void iommu_share_p2m_table(struct domain *d);
-
 #ifdef CONFIG_HAS_PCI
 int iommu_do_pci_domctl(struct xen_domctl *, struct domain *d,
                         XEN_GUEST_HANDLE_PARAM(xen_domctl_t));
Signature · 1 line
2.20.1
Reviewed-by on v5 7/8
#16 lore raw

Re: [PATCH v5 7/8] iommu: remove the share_p2m operation

Quoted 10 lines · Paul Durrant
xen/drivers/passthrough/vtd/iommu.c +4 -0
@@ -318,6 +318,48 @@ static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc)
     return pte_maddr;
 }
 
+static uint64_t domain_pgd_maddr(struct domain *d, unsigned int nr_pt_levels)
+{
+    struct domain_iommu *hd = dom_iommu(d);
+    uint64_t pgd_maddr;

xen/drivers/passthrough/vtd/iommu.c static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc) · line 334

Quoted 10 lines
xen/drivers/passthrough/vtd/iommu.c +10 -0
+    unsigned int agaw;
+
+    ASSERT(spin_is_locked(&hd->arch.mapping_lock));
+
+    if ( iommu_use_hap_pt(d) )
+    {
+        mfn_t pgd_mfn =
+            pagetable_get_mfn(p2m_get_pagetable(p2m_get_hostp2m(d)));
+
+        return pagetable_get_paddr(pagetable_from_mfn(pgd_mfn));

I thought I had already asked about this odd going through a pagetable_t a 2nd time without a clear need. Why not simply

    if ( iommu_use_hap_pt(d) )
    {
        pagetable_t pgt = p2m_get_pagetable(p2m_get_hostp2m(d));

        return pagetable_get_paddr(pgt);
    }

?

xen/drivers/passthrough/vtd/iommu.c static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc) · line 339

Quoted 5 lines
xen/drivers/passthrough/vtd/iommu.c +5 -0
+    }
+
+    if ( !hd->arch.vtd.pgd_maddr )
+    {
+        addr_to_dma_page_maddr(d, 0, 1);

Ahead of this, would you mind retaining ...

Quoted 6 lines
xen/drivers/passthrough/vtd/iommu.c +0 -1
@@ -1340,37 +1382,18 @@ int domain_context_mapping_one(
     if ( iommu_hwdom_passthrough && is_hardware_domain(domain) )
     {
         context_set_translation_type(*context, CONTEXT_TT_PASS_THRU);
-        agaw = level_to_agaw(iommu->nr_pt_levels);
     }

xen/drivers/passthrough/vtd/iommu.c int domain_context_mapping_one( · line 1352

Quoted 10 lines
xen/drivers/passthrough/vtd/iommu.c +2 -3
     else
     {
         spin_lock(&hd->arch.mapping_lock);
 
-        /* Ensure we have pagetables allocated down to leaf PTE. */
-        if ( hd->arch.vtd.pgd_maddr == 0 )
+        pgd_maddr = domain_pgd_maddr(domain, iommu->nr_pt_levels);
+        if ( !pgd_maddr )
         {
-            addr_to_dma_page_maddr(domain, 0, 1);

... the comment you remove here?

With these adjustments

Jan

#17 lore raw

RE: [PATCH v5 7/8] iommu: remove the share_p2m operation

Quoted 34 lines

-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 10 September 2020 15:20
To: Paul Durrant <[email protected]>
Cc: [email protected]; Paul Durrant <[email protected]>; Andrew Cooper <[email protected]>; George Dunlap <[email protected]>; Wei Liu <[email protected]>; Roger Pau Monné <[email protected]>; Kevin Tian <[email protected]>
Subject: Re: [PATCH v5 7/8] iommu: remove the share_p2m operation

Quoted 10 lines · Paul Durrant
xen/drivers/passthrough/vtd/iommu.c +4 -0
@@ -318,6 +318,48 @@ static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc)
     return pte_maddr;
 }
 
+static uint64_t domain_pgd_maddr(struct domain *d, unsigned int nr_pt_levels)
+{
+    struct domain_iommu *hd = dom_iommu(d);
+    uint64_t pgd_maddr;

xen/drivers/passthrough/vtd/iommu.c static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc) · line 334

Quoted 10 lines
xen/drivers/passthrough/vtd/iommu.c +10 -0
+    unsigned int agaw;
+
+    ASSERT(spin_is_locked(&hd->arch.mapping_lock));
+
+    if ( iommu_use_hap_pt(d) )
+    {
+        mfn_t pgd_mfn =
+            pagetable_get_mfn(p2m_get_pagetable(p2m_get_hostp2m(d)));
+
+        return pagetable_get_paddr(pagetable_from_mfn(pgd_mfn));

I thought I had already asked about this odd going through a pagetable_t a 2nd time without a clear need. Why not simply

Quoted 8 lines
    if ( iommu_use_hap_pt(d) )
    {
        pagetable_t pgt = p2m_get_pagetable(p2m_get_hostp2m(d));

        return pagetable_get_paddr(pgt);
    }

?

It's code movement (which IIRC you had acknowledged) but I can change it if you wish.

Quoted 19 lines

xen/drivers/passthrough/vtd/iommu.c static u64 addr_to_dma_page_maddr(struct domain *domain, u64 addr, int alloc) · line 339

Quoted 6 lines
xen/drivers/passthrough/vtd/iommu.c +5 -0
+    }
+
+    if ( !hd->arch.vtd.pgd_maddr )
+    {
+        addr_to_dma_page_maddr(d, 0, 1);

Ahead of this, would you mind retaining ...

Quoted 10 lines
xen/drivers/passthrough/vtd/iommu.c +0 -1
@@ -1340,37 +1382,18 @@ int domain_context_mapping_one(
     if ( iommu_hwdom_passthrough && is_hardware_domain(domain) )
     {
         context_set_translation_type(*context, CONTEXT_TT_PASS_THRU);
-        agaw = level_to_agaw(iommu->nr_pt_levels);
     }
     else
     {
         spin_lock(&hd->arch.mapping_lock);
 
Quoted 9 lines

xen/drivers/passthrough/vtd/iommu.c int domain_context_mapping_one( · line 1352

Quoted 6 lines
xen/drivers/passthrough/vtd/iommu.c +2 -3
-        /* Ensure we have pagetables allocated down to leaf PTE. */
-        if ( hd->arch.vtd.pgd_maddr == 0 )
+        pgd_maddr = domain_pgd_maddr(domain, iommu->nr_pt_levels);
+        if ( !pgd_maddr )
         {
-            addr_to_dma_page_maddr(domain, 0, 1);

... the comment you remove here?

Sure.

Quoted 3 lines

With these adjustments

Thanks.

Paul

Quoted 1 line

Jan

#18 lore raw

PATCH v5 8/8 iommu: stop calling IOMMU page tables 'p2m tables'

In reply to Paul Durrant

From: Paul Durrant <[email protected]>

It's confusing and not consistent with the terminology introduced with 'dfn_t'. Just call them IOMMU page tables.

Also remove a pointless check of the 'acpi_drhd_units' list in vtd_dump_page_table_level(). If the list is empty then IOMMU mappings would not have been enabled for the domain in the first place.

NOTE: All calls to printk() have also been removed from
iommu_dump_page_tables(); the implementation specific code is now responsible for all output.
The check for the global 'iommu_enabled' has also been replaced by an ASSERT since iommu_dump_page_tables() is not registered as a key handler unless IOMMU mappings are enabled.
Error messages are now prefixed with the name of the function.

---
Cc: Jan Beulich <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: Kevin Tian <[email protected]>

v5:
- Make sure domain id is in the output
- Use VTDPREFIX in output for consistency

v2:
- Moved all output into implementation specific code

---
 xen/drivers/passthrough/amd/pci_amd_iommu.c | 22 +++++++--------
 xen/drivers/passthrough/iommu.c             | 21 ++++-----------
 xen/drivers/passthrough/vtd/iommu.c         | 30 ++++++++++++---------
 xen/include/xen/iommu.h                     |  2 +-
 4 files changed, 34 insertions(+), 41 deletions(-)
xen/drivers/passthrough/amd/pci_amd_iommu.c +11 -11
index 3390c22cf3..80b0f58f88 100644
@@ -491,8 +491,8 @@ static int amd_iommu_group_id(u16 seg, u8 bus, u8 devfn)
 
 #include <asm/io_apic.h>
 
-static void amd_dump_p2m_table_level(struct page_info* pg, int level, 
-                                     paddr_t gpa, int indent)
+static void amd_dump_page_table_level(struct page_info* pg, int level,
+                                      paddr_t gpa, int indent)
 {
     paddr_t address;
     struct amd_iommu_pte *table_vaddr;
@@ -504,7 +504,7 @@ static void amd_dump_p2m_table_level(struct page_info* pg, int level,
     table_vaddr = __map_domain_page(pg);
     if ( table_vaddr == NULL )
     {
-        printk("Failed to map IOMMU domain page %"PRIpaddr"\n", 
+        printk("%s: failed to map domain page %"PRIpaddr"\n", __func__,
                 page_to_maddr(pg));
         return;
     }
@@ -521,15 +521,15 @@ static void amd_dump_p2m_table_level(struct page_info* pg, int level,
 
         if ( pde->next_level && (pde->next_level != (level - 1)) )
         {
-            printk("IOMMU p2m table error. next_level = %d, expected %d\n",
-                   pde->next_level, level - 1);
+            printk("%s: table error. next_level = %d, expected %d\n",
+                   __func__, pde->next_level, level - 1);
 
             continue;
         }
 
         address = gpa + amd_offset_level_address(index, level);
         if ( pde->next_level >= 1 )
-            amd_dump_p2m_table_level(
+            amd_dump_page_table_level(
                 mfn_to_page(_mfn(pde->mfn)), pde->next_level,
                 address, indent + 1);
         else
@@ -542,16 +542,16 @@ static void amd_dump_p2m_table_level(struct page_info* pg, int level,
     unmap_domain_page(table_vaddr);
 }
 
-static void amd_dump_p2m_table(struct domain *d)
+static void amd_dump_page_tables(struct domain *d)
 {
     const struct domain_iommu *hd = dom_iommu(d);
 
     if ( !hd->arch.amd.root_table )
         return;
 
-    printk("p2m table has %u levels\n", hd->arch.amd.paging_mode);
-    amd_dump_p2m_table_level(hd->arch.amd.root_table,
-                             hd->arch.amd.paging_mode, 0, 0);
+    printk("AMD IOMMU %pd table has %u levels\n", d, hd->arch.amd.paging_mode);
+    amd_dump_page_table_level(hd->arch.amd.root_table,
+                              hd->arch.amd.paging_mode, 0, 0);
 }
 
 static const struct iommu_ops __initconstrel _iommu_ops = {
@@ -578,7 +578,7 @@ static const struct iommu_ops __initconstrel _iommu_ops = {
     .suspend = amd_iommu_suspend,
     .resume = amd_iommu_resume,
     .crash_shutdown = amd_iommu_crash_shutdown,
-    .dump_p2m_table = amd_dump_p2m_table,
+    .dump_page_tables = amd_dump_page_tables,
 };
 
 static const struct iommu_init_ops __initconstrel _iommu_init_ops = {
xen/drivers/passthrough/iommu.c +5 -16
index e9b6c9a1ec..f5cd04fb3e 100644
@@ -22,7 +22,7 @@
 #include <xen/keyhandler.h>
 #include <xsm/xsm.h>
 
-static void iommu_dump_p2m_table(unsigned char key);
+static void iommu_dump_page_tables(unsigned char key);
 
 unsigned int __read_mostly iommu_dev_iotlb_timeout = 1000;
 integer_param("iommu_dev_iotlb_timeout", iommu_dev_iotlb_timeout);
@@ -212,7 +212,7 @@ void __hwdom_init iommu_hwdom_init(struct domain *d)
     if ( !is_iommu_enabled(d) )
         return;
 
-    register_keyhandler('o', &iommu_dump_p2m_table, "dump iommu p2m table", 0);
+    register_keyhandler('o', &iommu_dump_page_tables, "dump iommu page tables", 0);
 
     hd->platform_ops->hwdom_init(d);
 }
@@ -535,16 +535,12 @@ bool_t iommu_has_feature(struct domain *d, enum iommu_feature feature)
     return is_iommu_enabled(d) && test_bit(feature, dom_iommu(d)->features);
 }
 
-static void iommu_dump_p2m_table(unsigned char key)
+static void iommu_dump_page_tables(unsigned char key)
 {
     struct domain *d;
     const struct iommu_ops *ops;
 
-    if ( !iommu_enabled )
-    {
-        printk("IOMMU not enabled!\n");
-        return;
-    }
+    ASSERT(iommu_enabled);
 
     ops = iommu_get_ops();
 
@@ -555,14 +551,7 @@ static void iommu_dump_p2m_table(unsigned char key)
         if ( is_hardware_domain(d) || !is_iommu_enabled(d) )
             continue;
 
-        if ( iommu_use_hap_pt(d) )
-        {
-            printk("\ndomain%d IOMMU p2m table shared with MMU: \n", d->domain_id);
-            continue;
-        }
-
-        printk("\ndomain%d IOMMU p2m table: \n", d->domain_id);
-        ops->dump_p2m_table(d);
+        ops->dump_page_tables(d);
     }
 
     rcu_read_unlock(&domlist_read_lock);
xen/drivers/passthrough/vtd/iommu.c +17 -13
index 2ced3a5fa5..d84c000a8e 100644
@@ -2582,8 +2582,8 @@ static void vtd_resume(void)
     }
 }
 
-static void vtd_dump_p2m_table_level(paddr_t pt_maddr, int level, paddr_t gpa, 
-                                     int indent)
+static void vtd_dump_page_table_level(paddr_t pt_maddr, int level, paddr_t gpa,
+                                      int indent)
 {
     paddr_t address;
     int i;
@@ -2596,7 +2596,8 @@ static void vtd_dump_p2m_table_level(paddr_t pt_maddr, int level, paddr_t gpa,
     pt_vaddr = map_vtd_domain_page(pt_maddr);
     if ( pt_vaddr == NULL )
     {
-        printk("Failed to map VT-D domain page %"PRIpaddr"\n", pt_maddr);
+        printk("%s: failed to map domain page %"PRIpaddr"\n", __func__,
+               pt_maddr);
         return;
     }
 
@@ -2612,8 +2613,8 @@ static void vtd_dump_p2m_table_level(paddr_t pt_maddr, int level, paddr_t gpa,
 
         address = gpa + offset_level_address(i, level);
         if ( next_level >= 1 ) 
-            vtd_dump_p2m_table_level(dma_pte_addr(*pte), next_level, 
-                                     address, indent + 1);
+            vtd_dump_page_table_level(dma_pte_addr(*pte), next_level,
+                                      address, indent + 1);
         else
             printk("%*sdfn: %08lx mfn: %08lx\n",
                    indent, "",
@@ -2624,17 +2625,20 @@ static void vtd_dump_p2m_table_level(paddr_t pt_maddr, int level, paddr_t gpa,
     unmap_vtd_domain_page(pt_vaddr);
 }
 
-static void vtd_dump_p2m_table(struct domain *d)
+static void vtd_dump_page_tables(struct domain *d)
 {
-    const struct domain_iommu *hd;
+    const struct domain_iommu *hd = dom_iommu(d);
 
-    if ( list_empty(&acpi_drhd_units) )
+    if ( iommu_use_hap_pt(d) )
+    {
+        printk(VTDPREFIX " %pd sharing EPT table\n", d);
         return;
+    }
 
-    hd = dom_iommu(d);
-    printk("p2m table has %d levels\n", agaw_to_level(hd->arch.vtd.agaw));
-    vtd_dump_p2m_table_level(hd->arch.vtd.pgd_maddr,
-                             agaw_to_level(hd->arch.vtd.agaw), 0, 0);
+    printk(VTDPREFIX" %pd table has %d levels\n", d,
+           agaw_to_level(hd->arch.vtd.agaw));
+    vtd_dump_page_table_level(hd->arch.vtd.pgd_maddr,
+                              agaw_to_level(hd->arch.vtd.agaw), 0, 0);
 }
 
 static int __init intel_iommu_quarantine_init(struct domain *d)
@@ -2734,7 +2738,7 @@ static struct iommu_ops __initdata vtd_ops = {
     .iotlb_flush = iommu_flush_iotlb_pages,
     .iotlb_flush_all = iommu_flush_iotlb_all,
     .get_reserved_device_memory = intel_iommu_get_reserved_device_memory,
-    .dump_p2m_table = vtd_dump_p2m_table,
+    .dump_page_tables = vtd_dump_page_tables,
 };
 
 const struct iommu_init_ops __initconstrel intel_iommu_init_ops = {
xen/include/xen/iommu.h +1 -1
index 373145266f..7a539522b1 100644
@@ -276,7 +276,7 @@ struct iommu_ops {
                                     unsigned int flush_flags);
     int __must_check (*iotlb_flush_all)(struct domain *d);
     int (*get_reserved_device_memory)(iommu_grdm_t *, void *);
-    void (*dump_p2m_table)(struct domain *d);
+    void (*dump_page_tables)(struct domain *d);
 
 #ifdef CONFIG_HAS_DEVICE_TREE
     /*
Signature · 1 line
2.20.1
Reviewed-by on v5 8/8
#19 lore raw

Re: [PATCH v5 8/8] iommu: stop calling IOMMU page tables 'p2m tables'

xen/drivers/passthrough/amd/pci_amd_iommu.c static int amd_iommu_group_id(u16 seg, u8 bus, u8 devfn) · line 494

Quoted 9 lines · Paul Durrant
xen/drivers/passthrough/amd/pci_amd_iommu.c +1 -2
@@ -491,8 +491,8 @@ static int amd_iommu_group_id(u16 seg, u8 bus, u8 devfn)
 
 #include <asm/io_apic.h>
 
-static void amd_dump_p2m_table_level(struct page_info* pg, int level, 
-                                     paddr_t gpa, int indent)
+static void amd_dump_page_table_level(struct page_info* pg, int level,

Could you flip * and space here as you touch the line anyway?

xen/drivers/passthrough/amd/pci_amd_iommu.c static void amd_dump_p2m_table_level(struct page_info* pg, int level, · line 508

Quoted 7 lines
xen/drivers/passthrough/amd/pci_amd_iommu.c +1 -1
@@ -504,7 +504,7 @@ static void amd_dump_p2m_table_level(struct page_info* pg, int level,
     table_vaddr = __map_domain_page(pg);
     if ( table_vaddr == NULL )
     {
-        printk("Failed to map IOMMU domain page %"PRIpaddr"\n", 
+        printk("%s: failed to map domain page %"PRIpaddr"\n", __func__,
                 page_to_maddr(pg));

Why the addition of __func__? Personally I'd rather see most of its uses go away ... (There are two more further down.) Preferably with this dropped again

Jan

#20 lore raw

RE: [PATCH v5 8/8] iommu: stop calling IOMMU page tables 'p2m tables'

Quoted 13 lines

-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 10 September 2020 15:27
To: Paul Durrant <[email protected]>
Cc: [email protected]; Paul Durrant <[email protected]>; Andrew Cooper <[email protected]>; Kevin Tian <[email protected]>
Subject: Re: [PATCH v5 8/8] iommu: stop calling IOMMU page tables 'p2m tables'

Quoted 4 lines · Paul Durrant
xen/drivers/passthrough/amd/pci_amd_iommu.c +0 -0
@@ -491,8 +491,8 @@ static int amd_iommu_group_id(u16 seg, u8 bus, u8 devfn)
 
Quoted 8 lines

xen/drivers/passthrough/amd/pci_amd_iommu.c static int amd_iommu_group_id(u16 seg, u8 bus, u8 devfn) · line 494

Quoted 5 lines
xen/drivers/passthrough/amd/pci_amd_iommu.c +1 -2
 #include <asm/io_apic.h>
 
-static void amd_dump_p2m_table_level(struct page_info* pg, int level,
-                                     paddr_t gpa, int indent)
+static void amd_dump_page_table_level(struct page_info* pg, int level,

Could you flip * and space here as you touch the line anyway?

Sure.

Quoted 11 lines

xen/drivers/passthrough/amd/pci_amd_iommu.c static void amd_dump_p2m_table_level(struct page_info* pg, int level, · line 508

Quoted 7 lines
xen/drivers/passthrough/amd/pci_amd_iommu.c +1 -1
@@ -504,7 +504,7 @@ static void amd_dump_p2m_table_level(struct page_info* pg, int level,
     table_vaddr = __map_domain_page(pg);
     if ( table_vaddr == NULL )
     {
-        printk("Failed to map IOMMU domain page %"PRIpaddr"\n",
+        printk("%s: failed to map domain page %"PRIpaddr"\n", __func__,
                 page_to_maddr(pg));

Why the addition of __func__? Personally I'd rather see most of its uses go away ... (There are two more further down.) Preferably with this dropped again

Ok.

Quoted 2 lines

Thanks.

Paul

Quoted 1 line

Jan