x86: 64 bit print out absent pages num too
authorYinghai Lu <yinghai@kernel.org>
Wed, 15 Oct 2008 01:59:18 +0000 (18:59 -0700)
committerIngo Molnar <mingo@elte.hu>
Tue, 28 Oct 2008 15:50:49 +0000 (16:50 +0100)
so users are not confused with memhole causing big total ram

we don't need to worry about 32 bit, because memhole is always
above max_low_pfn.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86/mm/init_64.c

index f79a02f64d108dde4de4a35055240f64fae2371b..ad38648bddbd3b66862c5418231c197c2dd3167e 100644 (file)
@@ -884,6 +884,7 @@ static struct kcore_list kcore_mem, kcore_vmalloc, kcore_kernel,
 void __init mem_init(void)
 {
        long codesize, reservedpages, datasize, initsize;
+       unsigned long absent_pages;
 
        start_periodic_check_for_corruption();
 
@@ -899,8 +900,9 @@ void __init mem_init(void)
 #else
        totalram_pages = free_all_bootmem();
 #endif
-       reservedpages = max_pfn - totalram_pages -
-                                       absent_pages_in_range(0, max_pfn);
+
+       absent_pages = absent_pages_in_range(0, max_pfn);
+       reservedpages = max_pfn - totalram_pages - absent_pages;
        after_bootmem = 1;
 
        codesize =  (unsigned long) &_etext - (unsigned long) &_text;
@@ -917,10 +919,11 @@ void __init mem_init(void)
                                 VSYSCALL_END - VSYSCALL_START);
 
        printk(KERN_INFO "Memory: %luk/%luk available (%ldk kernel code, "
-                               "%ldk reserved, %ldk data, %ldk init)\n",
+                        "%ldk absent, %ldk reserved, %ldk data, %ldk init)\n",
                (unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
                max_pfn << (PAGE_SHIFT-10),
                codesize >> 10,
+               absent_pages << (PAGE_SHIFT-10),
                reservedpages << (PAGE_SHIFT-10),
                datasize >> 10,
                initsize >> 10);