x86, hibernate: Initialize mmu_cr4_features during boot

Restore the initialization of mmu_cr4_features during boot, which was
removed without comment in checkin e5f15b45ddf3afa2bbbb10c7ea34fb32b6de0a0e

x86: Cleanup highmap after brk is concluded

thereby breaking resume from hibernate.  This restores previous
functionality in approximately the same place, and corrects the
reading of %cr4 on pre-CPUID hardware (%cr4 exists if and only if
CPUID is supported.)

However, part of the problem is that the hibernate suspend/resume
sequence should manage the save/restore of %cr4 explicitly.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cc: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <201104020154.57136.rjw@sisk.pl>

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 5a0484a95ad6b99ad1824058efdcb238170a9f93..4be9b398470ea7ed399fcd72d39c67681ae98e3f 100644 (file)
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -976,6 +976,11 @@ void __init setup_arch(char **cmdline_p)
        paging_init();
        x86_init.paging.pagetable_setup_done(swapper_pg_dir);
 
+       if (boot_cpu_data.cpuid_level >= 0) {
+               /* A CPU has %cr4 if and only if it has CPUID */
+               mmu_cr4_features = read_cr4();
+       }
+
 #ifdef CONFIG_X86_32
        /* sync back kernel address range */
        clone_pgd_range(initial_page_table + KERNEL_PGD_BOUNDARY,