Re: [Xen-devel] [PATCH V5 15/15] Add ARM EFI boot support

[Roy Franz <roy.franz@xxxxxxxxxx>]

On Mon, Sep 22, 2014 at 4:20 AM, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:
> On Thu, 2014-09-18 at 15:50 -0700, Roy Franz wrote:
>
>> diff --git a/xen/common/efi/efi.h b/xen/common/efi/efi.h
>> index a80d5f1..09b0b0d 100644
>> --- a/xen/common/efi/efi.h
>> +++ b/xen/common/efi/efi.h
>> @@ -28,8 +28,6 @@ extern EFI_RUNTIME_SERVICES *efi_rs;
>>  extern UINTN efi_memmap_size, efi_mdesc_size;
>>  extern void *efi_memmap;
>>
>> -extern l4_pgentry_t *efi_l4_pgtable;
>
> FWIW l4_pgentry_t is an x86 specific type and to add to the confusion
> ARM counts the other way (so L4==root,L1==leaf on x86 and
> L0==root,L3==leaf on ARM...)
>
> But I presume handling that correctly is being put off until RTS
> support, which is fine by me.

Yes, this extern was moved to the x86 specific header file, and the
equivalent is not
needed for ARM until runtime services are implemented.
>
>> diff --git a/xen/include/asm-arm/efi-boot.h b/xen/include/asm-arm/efi-boot.h
>> new file mode 100644
>> index 0000000..4a1877a
>> --- /dev/null
>> +++ b/xen/include/asm-arm/efi-boot.h
>> @@ -0,0 +1,593 @@
>> +/*
>> + * Architecture specific implementation for EFI boot code.  This file
>> + * is intended to be included by XXX _only_, and therefore can define
>
> I guess you meant to come back to the XXX here (I didn't check if the
> x86 version had something similar).
>
> If XXX is another header then it wouldn't be unreasonable to have an
> #ifndef check for that header's guard variable and a #error.
>
> Ian.
>
oops :)  That xxx needs to be fixed up - it should be xen/common/efi/boot.c.

Roy

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