kvm: fix unaligned slots

kvm_set_phys_mem() may be passed sections that are not aligned to a page
boundary.  The current code simply brute-forces the alignment which leads
to an inconsistency and an abort().

Fix by aligning the start and the end of the section correctly, discarding
and unaligned head or tail.

This was triggered by a guest sizing a 64-bit BAR that is smaller than a page
with PCI_COMMAND_MEMORY enabled and the upper dword clear.

Signed-off-by: Avi Kivity <avi@redhat.com>

1 files changed, 12 insertions, 3 deletions

diff --git a/kvm-all.c b/kvm-all.c
index c4babda..4b7a4ae 100644
--- a/kvm-all.c
+++ b/kvm-all.c
@@ -541,17 +541,26 @@ static void kvm_set_phys_mem(MemoryRegionSection *section, bool add)
     target_phys_addr_t start_addr = section->offset_within_address_space;
     ram_addr_t size = section->size;
     void *ram = NULL;
+    unsigned delta;
 
     /* kvm works in page size chunks, but the function may be called
        with sub-page size and unaligned start address. */
-    size = TARGET_PAGE_ALIGN(size);
-    start_addr = TARGET_PAGE_ALIGN(start_addr);
+    delta = TARGET_PAGE_ALIGN(size) - size;
+    if (delta > size) {
+        return;
+    }
+    start_addr += delta;
+    size -= delta;
+    size &= TARGET_PAGE_MASK;
+    if (!size || (start_addr & ~TARGET_PAGE_MASK)) {
+        return;
+    }
 
     if (!memory_region_is_ram(mr)) {
         return;
     }
 
-    ram = memory_region_get_ram_ptr(mr) + section->offset_within_region;
+    ram = memory_region_get_ram_ptr(mr) + section->offset_within_region + delta;
 
     while (1) {
         mem = kvm_lookup_overlapping_slot(s, start_addr, start_addr + size);