Re: [PATCH] x86-64: use EFI to deal with platform wall clock

[Jan Beulich]

>>> On 15.05.12 at 14:47, Matthew Garrett <mjg59@xxxxxxxxxxxxx> wrote:
> On Tue, May 15, 2012 at 01:18:19PM +0100, Jan Beulich wrote:
> 
>> Simply removing the #ifdef around the respective code isn't enough,
>> however: The runtime code must not only be forced to be executable, it
>> also must have a proper virtual address set (which on at least the
>> system I'm testing on isn't the case for all necessary regions, or at
>> least not as early as they're now being required).
> 
> I don't understand this. The get_time pointer won't be updated to the 
> virtual function until the end of efi_enter_virtual_mode, at which point 
> all runtime regions should have a virtual address mapped. We also call 
> runtime_code_page_mkexec() immediately after updating that pointer, 
> although maybe the order should be swapped. So I think the bug you're 
> fixing is not the bug you think you're fixing...

I would have expected that things work that way, but they
don't. In particular is the function in efi_64.c that's being
modified here called from efi_call_phys_{pro,epi}log(), and at
that point we can't expect virtual addresses to be uniformly
set yet. So it's a physical call that requires the fixup done
here, as efi_set_executable() simply expects ->virt_addr to
be valid. I suspect that no physical calls other than
phys_efi_set_virtual_address_map() were being done so far
at all on 64-bit, hiding the problem.

Jan

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/