[PATCH] efi: Disable interrupts around EFI calls, not in the epilog/prolog calls

[Ingo Molnar]

* Tapasweni Pathak <tapaswenipathak@xxxxxxxxx> wrote:

> Disabling interrupts around kmalloc() is less than ideal. Move it
> after the kmalloc().
> 
> Found using Coccinelle.
> 
> Signed-off-by: Tapasweni Pathak <tapaswenipathak@xxxxxxxxx>
> Suggested-by: Matt Fleming <matt.fleming@xxxxxxxxx>
> ---
>  arch/x86/platform/efi/efi_64.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c
> index 17e80d8..5d6af59 100644
> --- a/arch/x86/platform/efi/efi_64.c
> +++ b/arch/x86/platform/efi/efi_64.c
> @@ -88,10 +88,10 @@ void __init efi_call_phys_prolog(void)
>  		return;
> 
>  	early_code_mapping_set_exec(1);
> -	local_irq_save(efi_flags);
> 
>  	n_pgds = DIV_ROUND_UP((max_pfn << PAGE_SHIFT), PGDIR_SIZE);
>  	save_pgd = kmalloc(n_pgds * sizeof(pgd_t), GFP_KERNEL);
> +	local_irq_save(efi_flags);
> 
>  	for (pgd = 0; pgd < n_pgds; pgd++) {
>  		save_pgd[pgd] = *pgd_offset_k(pgd * PGDIR_SIZE);

So why are interrupts disabled around page table operations to begin 
with? It's not like any of this can execute on two CPUs at once, nor 
can this be executed from interrupt context.

So, shouldn't we only protect the EFI calls themselves? If we did that 
we could save the efi_flags global variable ugliness as well.

I.e. something like the attached patch. (not tested and all that)

Thanks,

	Ingo

===================>