Re: [Xen-devel] [PATCH v4 12/19] efi: introduce EFI_RS to ease control on runtime services usage

[Daniel Kiper <daniel.kiper@xxxxxxxxxx>]

On Thu, Aug 18, 2016 at 05:18:01AM -0600, Jan Beulich wrote:
> >>> On 18.08.16 at 12:30, <daniel.kiper@xxxxxxxxxx> wrote:
> > On Wed, Aug 17, 2016 at 10:12:32AM -0600, Jan Beulich wrote:
> >> >>> On 06.08.16 at 01:04, <daniel.kiper@xxxxxxxxxx> wrote:
> >>
> >> Apart from the question of this probably better getting merged with
> >> the previous patch ...
> >>
> >> > --- a/xen/common/efi/boot.c
> >> > +++ b/xen/common/efi/boot.c
> >> > @@ -936,6 +936,10 @@ efi_start(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE
> > *SystemTable)
> >> >
> >> >      __set_bit(EFI_BOOT, &efi.flags);
> >> >
> >> > +#ifndef CONFIG_ARM /* Disabled until runtime services implemented. */
> >> > +    __set_bit(EFI_RS, &efi.flags);
> >> > +#endif
> >>
> >> ... this needs to be paralleled by a __clear_bit() when "efi=no-rs"
> >> was given (and then efi_rs_enable) should go away. Oh, looks
> >> like that's the next patch, but I'd then again question the split.
> >
> > Do you suggest that patches 11-13 should be merged into one thing?
> > If it is OK for you I can do that.
> >
> >> > --- a/xen/include/xen/efi.h
> >> > +++ b/xen/include/xen/efi.h
> >> > @@ -12,6 +12,7 @@
> >> >  struct efi {
> >> >      unsigned long flags;        /* Bit fields representing available EFI
> > features/properties */
> >> >  #define EFI_BOOT        0       /* Were we booted from EFI? */
> >> > +#define EFI_RS          2       /* Can we use runtime services? */
> >>
> >> Why is this not the sequentially next number?
> >
> > I wish that EFI_LOADER (added in patch 16) has number 1 (next to EFI_BOOT).
>
> That'll then too end up more natural by merging the patches.

Potentially we can do that but patch 14 and 15 use efi_enabled().
So, we should merge patches 11-13 if you wish and leave 14-16 as is.

Daniel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel