From: Andrew Morton <akpm@osdl.org>
arch/i386/kernel/efi.c: In function 'efi_call_phys_epilog':
arch/i386/kernel/efi.c:116: error: memory input 0 is not directly addressable
These patches now worry me.
I see that store_tr() grew an unchangelogged `=r' constraint too...
(It appears that we've accidnetally fixed a bug here. The efi code had its
lgdt argument in the output operand slot and not the input operand slot)
Anyway. Please triple-check these patches again, let me know?
Cc: Zachary Amsden <zach@vmware.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---
arch/i386/kernel/efi.c | 2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
diff -puN arch/i386/kernel/efi.c~i386-inline-assembler-cleanup-encapsulate-descriptor-and-task-register-management-fix arch/i386/kernel/efi.c
--- 25-alpha/arch/i386/kernel/efi.c~i386-inline-assembler-cleanup-encapsulate-descriptor-and-task-register-management-fix 2005-07-30 18:26:45.000000000 -0700
+++ 25-alpha-akpm/arch/i386/kernel/efi.c 2005-07-30 18:27:06.000000000 -0700
@@ -113,7 +113,7 @@ static void efi_call_phys_epilog(void)
cpu_gdt_descr[0].address =
(unsigned long) __va(cpu_gdt_descr[0].address);
- load_gdt(&cpu_gdt_descr);
+ load_gdt(&cpu_gdt_descr[0]);
cr4 = read_cr4();
if (cr4 & X86_CR4_PSE) {
_