Re: [PATCH] linux-2.6.18/x86_64: remove an inapplicable feature flag

[Jeremy Fitzhardinge <jeremy@goop.org>]

On 11/03/2010 05:13 AM, Jan Beulich wrote:
>>>> On 03.11.10 at 09:59, "Jan Beulich" <JBeulich@novell.com> wrote:
>> Signed-off-by: Jan Beulich <jbeulich@novell.com>
>>
>> --- a/arch/x86_64/kernel/head-xen.S
>> +++ b/arch/x86_64/kernel/head-xen.S
>> @@ -206,6 +206,6 @@
>> 	ELFNOTE(Xen, XEN_ELFNOTE_ENTRY,          .quad,  startup_64)
>>  	ELFNOTE(Xen, XEN_ELFNOTE_HYPERCALL_PAGE, .quad,  hypercall_page)
>>  	ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID,   .quad,  
>> _PAGE_PRESENT,_PAGE_PRESENT)
>> -	ELFNOTE(Xen, XEN_ELFNOTE_FEATURES,       .asciz, "writable_page_tables|writable_descriptor_tables|auto_translated_physmap|pae_pgdir_above_4gb|supervisor_mode_kernel")
>> +	ELFNOTE(Xen, XEN_ELFNOTE_FEATURES,       .asciz, "writable_page_tables|writable_descriptor_tables|auto_translated_physmap|supervisor_mode_kernel")
>>  	ELFNOTE(Xen, XEN_ELFNOTE_LOADER,         .asciz, "generic")
>>  	ELFNOTE(Xen, XEN_ELFNOTE_SUSPEND_CANCEL, .long,  1)
> If the flag was removed from the pv-ops 64-bit kernel too, we could
> even start warning about inapplicable flags in hypervisor/tools.
>
> What puzzles me more though is that the pv-ops kernel requires
> the writable_page_tables feature. Why is that?

It did in the very early days, but it hasn't used it since converting to
PV MMU (couldn't, since it doesn't exist any more).  It lists
"!writable_page_table", which I read as meaning "writable page tables
not supported"; does it actually mean "must"?

    J