[PATCH] Re: [Xen-devel] Re: [PATCH] to fix ACPI slit table access at runtime

["Jan Beulich" <JBeulich@xxxxxxxxxx>]

>>> Keir Fraser <keir.fraser@xxxxxxxxxxxxx> 25.02.10 14:16 >>>
>Given we agree the current situation is a bug, could you make up a patch
>along the lines of what you prefer to see? Because my default action will be
>simply to ifdef the code, as NUMA-on-i386 is not much a concern of mine.

Signed-off-by: Jan Beulich <jbeulich@xxxxxxxxxx>

--- 2010-02-19.orig/xen/arch/x86/numa.c 2010-01-13 18:58:11.000000000 +0100
+++ 2010-02-19/xen/arch/x86/numa.c      2010-02-25 14:27:14.000000000 +0100
@@ -30,7 +30,7 @@ struct node_data node_data[MAX_NUMNODES]
 
 /* Mapping from pdx to node id */
 int memnode_shift;
-static typeof(*memnodemap) _memnodemap[2];
+static typeof(*memnodemap) _memnodemap[64];
 unsigned long memnodemapsize;
 u8 *memnodemap;
 
@@ -90,6 +90,7 @@ static int __init populate_memnodemap(co
 
 static int __init allocate_cachealigned_memnodemap(void)
 {
+#ifndef __i386__
        unsigned long size = PFN_UP(memnodemapsize * sizeof(*memnodemap));
        unsigned long mfn = alloc_boot_pages(size, 1);
 
@@ -108,6 +109,13 @@ static int __init allocate_cachealigned_
        memnodemapsize = size / sizeof(*memnodemap);
 
        return 0;
+#else
+       printk(KERN_ERR
+              "Memory to Node hash needs %lu entries, got only %zu\n",
+              memnodemapsize, ARRAY_SIZE(_memnodemap));
+       memnodemapsize = 0;
+       return -1;
+#endif
 }
 
 /*

Attachment: i386-numa-not-boot-alloc.patch
Description: Text document

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel