lkml.org 
[lkml]   [2012]   [Oct]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH -v3 0/7] x86: Use BRK to pre mapping page table to make xen happy
From
On Wed, Oct 10, 2012 at 6:47 AM, Konrad Rzeszutek Wilk
<konrad@kernel.org> wrote:
> On Tue, Oct 09, 2012 at 04:58:28PM -0700, Yinghai Lu wrote:
>> on top of tip/x86/mm2, but please zap last patch in that branch.
>
> So while I appreciate you actively looking at this and iteratively
> sending snapshots of the progress - I think it would be easier if
> you posted a patchset that has rework done completly per what
> Peter described.

it should be done this time, so asked Stefano to test it again.

>
> That way folks who are reviewing will know when they can focus their
> time on reviewing the whole thing in one go instead of doing it
> step by step - as some of the patches still haven't addressed the
> review comments that were given the first time.

I addressed one that i can do.

>
>
>> 1. use brk to mapping first PMD_SIZE range.
>> 2. top down to initialize page table range by range.
>> 3. get rid of calculate page table, and find_early_page_table.
>> 4. remove early_ioremap in page table accessing.
>>
>> v2: changes, update xen interface about pagetable_reserve, so not
>> use pgt_buf_* in xen code directly.
>> v3: use range top-down to initialize page table, so will not use
>> calculating/find early table anymore.
>> also reorder the patches sequence.
>>
>> could be found at:
>> git://git.kernel.org/pub/scm/linux/kernel/git/yinghai/linux-yinghai.git for-x86-mm
>>
>> later we could get rid of workaround about xen_mapping_pagetable_reserve, that
>> could kill another 50 lines codes. --- will do that later because x86/mm2 is
>> not updated to linus/master yet. If we do that now, will have merge conflicts.
>
> I am confused. Why do you need x86/mm2? If you do need it, you need to
> describe in this writeup why you depend on it, and what is there.
>
> You can't base it on linus's tree?

no, some init_memory_mapping related cleanups are in x86/mm2.

the whole story: while reviewing x86/mm2,
Stefano say it would be better to get rid of ioremap to access page table area.
so according to hpa's concept, I use some pages in BRK to pre-map page
table at first.
on -v2, with xen Stefano found panic on system with more than 4g.
To address that panic, we come out -v3 that will map memory top-down
and range size for
each step is from PMD_SIZE to more bigger...

Thanks

Yinghai


\
 
 \ /
  Last update: 2012-10-10 17:21    [W:0.238 / U:0.528 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site