Re: [Xen-devel] [PATCH] x86/efi: move the logic to detect PE build support

[Daniel Kiper <daniel.kiper@xxxxxxxxxx>]

On Mon, Jul 16, 2018 at 03:52:49AM -0600, Jan Beulich wrote:
> >>> On 16.07.18 at 11:25, <roger.pau@xxxxxxxxxx> wrote:
> > On Mon, Jul 16, 2018 at 03:18:13AM -0600, Jan Beulich wrote:
> >> >>> On 16.07.18 at 11:12, <roger.pau@xxxxxxxxxx> wrote:
> >> > On Mon, Jul 16, 2018 at 02:55:16AM -0600, Jan Beulich wrote:
> >> >> >>> On 16.07.18 at 10:26, <roger.pau@xxxxxxxxxx> wrote:
> >> >> > On Mon, Jul 16, 2018 at 01:59:15AM -0600, Jan Beulich wrote:
> >> >> >> >>> On 13.07.18 at 18:02, <roger.pau@xxxxxxxxxx> wrote:
> >> >> >> > --- a/xen/arch/x86/Makefile
> >> >> >> > +++ b/xen/arch/x86/Makefile
> >> >> >> > @@ -168,6 +168,16 @@ $(TARGET).efi: ALT_BASE = 0x$(shell $(NM)
> >> >> >> > efi/relocs-dummy.o | sed -n 's, A ALT_
> >> >> >> >  # Don't use $(wildcard ...) here - at least make 3.80 expands 
> >> >> >> > this too
> >> >> > early!
> >> >> >> >  $(TARGET).efi: guard = $(if $(shell echo efi/dis* | grep 
> >> >> >> > disabled),:)
> >> >> >> >
> >> >> >> > +# Check if the build system supports PE.
> >> >> >> > +efi := y$(shell rm -f efi/disabled)
> >> >> >> > +efi := $(if $(efi),$(shell $(CC) $(filter-out $(CFLAGS-y) 
> >> >> >> > .%.d,$(CFLAGS))
> > -c
> >> >> > efi/check.c -o efi/check.o 2>efi/disabled && echo y))
> >> >> >> > +efi := $(if $(efi),$(shell $(LD) -mi386pep --subsystem=10 -o 
> >> >> >> > efi/check.efi
> >> >> > efi/check.o 2>efi/disabled && echo y))
> >> >> >> > +efi := $(if $(efi),$(shell rm efi/disabled)y)
> >> >> >> > +export BUILD_PE := $(efi)
> >> >> >> > +ifeq ($(efi),y)
> >> >> >> > +CFLAGS += -DBUILD_PE
> >> >> >> > +endif
> >> >> >>
> >> >> >> For one I'm not really happy about this being moved here: I did 
> >> >> >> place it
> >> >> >> in efi/Makefile for the simple reason of having as much as possible 
> >> >> >> of
> >> >> >> the EFI specifics in that single file.
> >> >> >>
> >> >> >> Additionally I think it would be better if setting propagated 
> >> >> >> through the
> >> >> >> environment had XEN_ prefixes.
> >> >> >>
> >> >> >> > --- a/xen/arch/x86/xen.lds.S
> >> >> >> > +++ b/xen/arch/x86/xen.lds.S
> >> >> >> > @@ -304,7 +304,9 @@ SECTIONS
> >> >> >> >    } :text
> >> >> >> >  #endif
> >> >> >> >
> >> >> >> > -  efi = DEFINED(efi) ? efi : .;
> >> >> >> > +#ifndef BUILD_PE
> >> >> >> > +  efi = .;
> >> >> >> > +#endif
> >> >> >>
> >> >> >> And then I don't really understand how this is different from the
> >> >> >> original #ifndef EFI that Daniel had problems with.
> >> >> >
> >> >> > As I understand it EFI only signals whether a PE binary will be
> >> >> > created,
> >> >>
> >> >> But that is my point: BUILD_PE signals exactly that aiui.
> >> >
> >> > No, BUILD_PE signals whether the binary will use runtime.c instead of
> >> > stub.c, and thus have the efi symbol defined.
> >>
> >> But in that case - why did you chose this particular name?
> >
> > Because that seems to be what the tests actually do. AFAICT it tests
> > that the compiler supports the MS ABI and that the linker is able to
> > generate PE binaries.
> >
> > I think this is slightly wrong, because for multiboot2 support you
> > likely only need the compiler MS ABI support, but not the linker PE
> > support?
>
> Indeed. But too strict checking is less of a problem than too lax one.
> Otoh people are far more likely to have a suitable gcc but no suitably
> configured binutils, so relaxing this would likely be worthwhile. Daniel?

Yes, it is true and I think that it is worth relaxing the check.

> > How about naming this BUILD_EFI_SERVICES or some such?
>
> Perhaps just BUILD_EFI, but yes. EFI, as used in xen.lds.S, would

s/BUILD_EFI/XEN_BUILD_EFI/?

> then perhaps better be renamed into BUILD_PE (or some such).

XEN_BUILD_PE or XEN_LINK_PE?

Daniel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel