Re: [PATCH 1/1] EFI: Fix gdt load

[Edgar Hucek <hostmaster@ed-soft.at>]

Edgar Hucek wrote:
> This patch makes the kernel bootable again on ia32 EFI systems.
> 
> Signed-off-by: Edgar Hucek <hostmaster@ed-soft.at>
> 
> 
> ------------------------------------------------------------------------
> 
> diff -uNr linux-2.6.16-rc5/arch/i386/kernel/efi.c linux-2.6.16-rc5.efi/arch/i386/kernel/efi.c
> --- linux-2.6.16-rc5/arch/i386/kernel/efi.c	2006-03-02 14:08:06.000000000 +0100
> +++ linux-2.6.16-rc5.efi/arch/i386/kernel/efi.c	2006-03-02 14:04:44.000000000 +0100
> @@ -70,7 +70,8 @@
>  {
>  	unsigned long cr4;
>  	unsigned long temp;
> -
> +	struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, 0);
> +	
>  	spin_lock(&efi_rt_lock);
>  	local_irq_save(efi_rt_eflags);
>  
> @@ -103,18 +104,17 @@
>  	 */
>  	local_flush_tlb();
>  
> -	per_cpu(cpu_gdt_descr, 0).address =
> -				 __pa(per_cpu(cpu_gdt_descr, 0).address);
> -	load_gdt((struct Xgt_desc_struct *)__pa(&per_cpu(cpu_gdt_descr, 0)));
> +	cpu_gdt_descr->address = __pa(cpu_gdt_descr->address);
> +	load_gdt(cpu_gdt_descr);
>  }
>  
>  static void efi_call_phys_epilog(void)
>  {
>  	unsigned long cr4;
> +	struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, 0);
>  
> -	per_cpu(cpu_gdt_descr, 0).address =
> -			(unsigned long)__va(per_cpu(cpu_gdt_descr, 0).address);
> -	load_gdt((struct Xgt_desc_struct *)__va(&per_cpu(cpu_gdt_descr, 0)));
> +	cpu_gdt_descr->address = __va(cpu_gdt_descr->address);
> +	load_gdt(cpu_gdt_descr);
>  
>  	cr4 = read_cr4();
>  

I don't know if it is a race condition. At least it makes the kernel bootable agin on my box.
Any idea what i could try to make a better patch ?

cu

Edgar (gimli) Hucek
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/