Re: [PATCH v2 02/18] loader/linux: support passing rsdp address via boot params

[Juergen Gross]

On 18/10/2018 16:30, Daniel Kiper wrote:
> On Thu, Oct 18, 2018 at 04:18:26PM +0200, Juergen Gross wrote:
>> On 18/10/2018 16:13, Daniel Kiper wrote:
>>> On Tue, Oct 09, 2018 at 01:03:01PM +0200, Juergen Gross wrote:
>>>> Xen PVH guests will have the RSDP at an arbitrary address. Support that
>>>> by passing the RSDP address via the boot parameters to Linux.
>>>>
>>>> The new protocol version 2.14 requires to set version to 0x8000 ored
>>>> with the actually use protocol version (the minimum of the kernel
>>>> supplied protocol version and the grub2 supported protocol version)
>>>> if 2.14 or higher are in effect.
>>>>
>>>> Signed-off-by: Juergen Gross <address@hidden>
>>>> ---
>>>> V2: add oring 0x8000 to version field
>>>> ---
>>>>  grub-core/loader/i386/linux.c | 9 +++++++++
>>>>  include/grub/i386/linux.h     | 5 ++++-
>>>>  2 files changed, 13 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/grub-core/loader/i386/linux.c b/grub-core/loader/i386/linux.c
>>>> index 4eab55a2d..f96309476 100644
>>>> --- a/grub-core/loader/i386/linux.c
>>>> +++ b/grub-core/loader/i386/linux.c
>>>> @@ -35,6 +35,7 @@
>>>>  #include <grub/i18n.h>
>>>>  #include <grub/lib/cmdline.h>
>>>>  #include <grub/linux.h>
>>>> +#include <grub/machine/kernel.h>
>>>
>>> Probably this change belongs to another patch.
>>
>> I don't think so.
> 
> You do not add anything to this header here and out of the blue you
> include it in this file. So, why it is needed here?

Ah, now I see your problem.

machine/kernel.h will be the header which eventually defines
GRUB_KERNEL_USE_RSDP_ADDR.

> 
>>>>  GRUB_MOD_LICENSE ("GPLv3+");
>>>>
>>>> @@ -750,6 +751,14 @@ grub_cmd_linux (grub_command_t cmd __attribute__ 
>>>> ((unused)),
>>>>        prot_init_space = page_align (prot_size) * 3;
>>>>      }
>>>>
>>>> +#ifdef GRUB_KERNEL_USE_RSDP_ADDR
>>>> +  if (grub_le_to_cpu16 (lh.version) >= 0x020e)
>>>> +    lh.acpi_rsdp_addr = grub_le_to_cpu64 (grub_rsdp_addr);
>>>> +#endif
>>>
>>> It seems to me that this belongs to patch #8. Does not it?
>>
>> No, it is a generic new interface.
>>
>> It will be used by PVH first, but can be used in other environments,
>> too.
> 
> I do not like this much but if you want me to get it you have to put
> a blurb in the commit message why this is not enabled here. And when
> it will be enabled...

Okay, will do.


Juergen