ACPI: Harden acpi_table_parse_entries() against BIOS bug
Fenghua Yu [Tue, 25 Sep 2012 18:11:43 +0000 (11:11 -0700)]
Parsing acpi table entries may fall into an infinite loop on a buggy BIOS
which has entry length=0 in acpi table.

Instead of kernel hang with few failure clue which leads to heavy lifting debug
effort, this patch hardens kernel boot by booting into non NUMA mode. The debug
info left in log buffer helps people identify the issue.

Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>

drivers/acpi/tables.c

index f336bca..2572d97 100644 (file)
@@ -240,10 +240,17 @@ acpi_table_parse_entries(char *id,
               table_end) {
                if (entry->type == entry_id
                    && (!max_entries || count++ < max_entries))
-                       if (handler(entry, table_end)) {
-                               early_acpi_os_unmap_memory((char *)table_header, tbl_size);
-                               return -EINVAL;
-                       }
+                       if (handler(entry, table_end))
+                               goto err;
+
+               /*
+                * If entry->length is 0, break from this loop to avoid
+                * infinite loop.
+                */
+               if (entry->length == 0) {
+                       pr_err(PREFIX "[%4.4s:0x%02x] Invalid zero length\n", id, entry_id);
+                       goto err;
+               }
 
                entry = (struct acpi_subtable_header *)
                    ((unsigned long)entry + entry->length);
@@ -255,6 +262,9 @@ acpi_table_parse_entries(char *id,
 
        early_acpi_os_unmap_memory((char *)table_header, tbl_size);
        return count;
+err:
+       early_acpi_os_unmap_memory((char *)table_header, tbl_size);
+       return -EINVAL;
 }
 
 int __init