libxc: add MMIO hole size parameter to xc_hvm_build()

Add a parameter for the MMIO hole size when building a HVM domain.

This is useful for specifying a larger than normal MMIO hole to ensure
that no PCIe device's MMIO region overlaps with RAM in a guest's
physical address space.  This is needed on certain systems with PCIe
switches with a broken ACS capability.  On these systems, if a device
downstream of the switch attempts a DMA to a guest physical address
that overlaps with the MMIO region of another downstream device, then
the switch routes the transfer directly to the device and the read or
write never hits RAM.

Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Acked-by: George Dunlap <george.dunlap@eu.citrix.com>
Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>

2 files changed, 19 insertions, 11 deletions

diff --git a/tools/libxc/xc_hvm_build.c b/tools/libxc/xc_hvm_build.c
index d08b54bed4..780b23f5b2 100644
--- a/tools/libxc/xc_hvm_build.c
+++ b/tools/libxc/xc_hvm_build.c
@@ -46,7 +46,8 @@
 #define NR_SPECIAL_PAGES     5
 #define special_pfn(x) (0xff000u - NR_SPECIAL_PAGES + (x))
 
-static void build_hvm_info(void *hvm_info_page, uint64_t mem_size)
+static void build_hvm_info(void *hvm_info_page, uint64_t mem_size,
+                           uint64_t mmio_start, uint64_t mmio_size)
 {
     struct hvm_info_table *hvm_info = (struct hvm_info_table *)
         (((unsigned char *)hvm_info_page) + HVM_INFO_OFFSET);
@@ -54,10 +55,10 @@ static void build_hvm_info(void *hvm_info_page, uint64_t mem_size)
     uint8_t sum;
     int i;
 
-    if ( lowmem_end > HVM_BELOW_4G_RAM_END )
+    if ( lowmem_end > mmio_start )
     {
-        highmem_end = lowmem_end + (1ull<<32) - HVM_BELOW_4G_RAM_END;
-        lowmem_end = HVM_BELOW_4G_RAM_END;
+        highmem_end = (1ull<<32) + (lowmem_end - mmio_start);
+        lowmem_end = mmio_start;
     }
 
     memset(hvm_info_page, 0, PAGE_SIZE);
@@ -126,10 +127,10 @@ static int loadelfimage(
  * Check whether there exists mmio hole in the specified memory range.
  * Returns 1 if exists, else returns 0.
  */
-static int check_mmio_hole(uint64_t start, uint64_t memsize)
+static int check_mmio_hole(uint64_t start, uint64_t memsize,
+                           uint64_t mmio_start, uint64_t mmio_size)
 {
-    if ( start + memsize <= HVM_BELOW_4G_MMIO_START ||
-         start >= HVM_BELOW_4G_MMIO_START + HVM_BELOW_4G_MMIO_LENGTH )
+    if ( start + memsize <= mmio_start || start >= mmio_start + mmio_size )
         return 0;
     else
         return 1;
@@ -142,6 +143,8 @@ static int setup_guest(xc_interface *xch,
     xen_pfn_t *page_array = NULL;
     unsigned long i, nr_pages = args->mem_size >> PAGE_SHIFT;
     unsigned long target_pages = args->mem_target >> PAGE_SHIFT;
+    uint64_t mmio_start = (1ull << 32) - args->mmio_size;
+    uint64_t mmio_size = args->mmio_size;
     unsigned long entry_eip, cur_pages, cur_pfn;
     void *hvm_info_page;
     uint32_t *ident_pt;
@@ -188,8 +191,8 @@ static int setup_guest(xc_interface *xch,
 
     for ( i = 0; i < nr_pages; i++ )
         page_array[i] = i;
-    for ( i = HVM_BELOW_4G_RAM_END >> PAGE_SHIFT; i < nr_pages; i++ )
-        page_array[i] += HVM_BELOW_4G_MMIO_LENGTH >> PAGE_SHIFT;
+    for ( i = mmio_start >> PAGE_SHIFT; i < nr_pages; i++ )
+        page_array[i] += mmio_size >> PAGE_SHIFT;
 
     /*
      * Allocate memory for HVM guest, skipping VGA hole 0xA0000-0xC0000.
@@ -230,7 +233,8 @@ static int setup_guest(xc_interface *xch,
         if ( ((count | cur_pfn) & (SUPERPAGE_1GB_NR_PFNS - 1)) == 0 &&
              /* Check if there exists MMIO hole in the 1GB memory range */
              !check_mmio_hole(cur_pfn << PAGE_SHIFT,
-                              SUPERPAGE_1GB_NR_PFNS << PAGE_SHIFT) )
+                              SUPERPAGE_1GB_NR_PFNS << PAGE_SHIFT,
+                              mmio_start, mmio_size) )
         {
             long done;
             unsigned long nr_extents = count >> SUPERPAGE_1GB_SHIFT;
@@ -327,7 +331,7 @@ static int setup_guest(xc_interface *xch,
               xch, dom, PAGE_SIZE, PROT_READ | PROT_WRITE,
               HVM_INFO_PFN)) == NULL )
         goto error_out;
-    build_hvm_info(hvm_info_page, v_end);
+    build_hvm_info(hvm_info_page, v_end, mmio_start, mmio_size);
     munmap(hvm_info_page, PAGE_SIZE);
 
     /* Allocate and clear special pages. */
@@ -408,6 +412,9 @@ int xc_hvm_build(xc_interface *xch, uint32_t domid,
     if ( args.mem_target == 0 )
         args.mem_target = args.mem_size;
 
+    if ( args.mmio_size == 0 )
+        args.mmio_size = HVM_BELOW_4G_MMIO_LENGTH;
+
     /* An HVM guest must be initialised with at least 2MB memory. */
     if ( args.mem_size < (2ull << 20) || args.mem_target < (2ull << 20) )
         return -1;
diff --git a/tools/libxc/xenguest.h b/tools/libxc/xenguest.h
index f52ca74231..8d885d3133 100644
--- a/tools/libxc/xenguest.h
+++ b/tools/libxc/xenguest.h
@@ -174,6 +174,7 @@ int xc_linux_build_mem(xc_interface *xch,
 struct xc_hvm_build_args {
     uint64_t mem_size;           /* Memory size in bytes. */
     uint64_t mem_target;         /* Memory target in bytes. */
+    uint64_t mmio_size;          /* Size of the MMIO hole in bytes. */
     const char *image_file_name; /* File name of the image to load. */
 };