On Thu, 13 Nov 2008 12:38:07 +0200
Izik Eidus [EMAIL PROTECTED] wrote:
If KSM pages are on radix-tree, it will be accounted automatically.
Now, we have Unevictable LRU and mlocked() pages are smartly isolated
into its
own LRU. So, just doing
- inode's radix-tree
- make all pages
ציטוט KAMEZAWA Hiroyuki:
Thank you for answers.
On Wed, 12 Nov 2008 13:11:12 +0200
Izik Eidus [EMAIL PROTECTED] wrote:
Avi Kivity wrote:
KAMEZAWA Hiroyuki wrote:
Can I make a question ? (I'm working for memory cgroup.)
Now, we do charge to anonymous page when
- charge(+1)
Avi Kivity wrote:
KAMEZAWA Hiroyuki wrote:
Can I make a question ? (I'm working for memory cgroup.)
Now, we do charge to anonymous page when
- charge(+1) when it's mapped firstly (mapcount 0-1)
- uncharge(-1) it's fully unmapped (mapcount 1-0) vir
page_remove_rmap().
My quesion is
-
KAMEZAWA Hiroyuki wrote:
Can I make a question ? (I'm working for memory cgroup.)
Now, we do charge to anonymous page when
- charge(+1) when it's mapped firstly (mapcount 0-1)
- uncharge(-1) it's fully unmapped (mapcount 1-0) vir page_remove_rmap().
My quesion is
- PageKSM pages are not
On Tue, Nov 11, 2008 at 09:10:45PM -0600, Christoph Lameter wrote:
get_user_pages() cannot get to it since the pagetables have already been
modified. If get_user_pages runs then the fault handling will occur
which will block the thread until migration is complete.
migrate.c does nothing for
On Wed, 2008-11-12 at 18:32 +0100, Andrea Arcangeli wrote:
On Tue, Nov 11, 2008 at 09:10:45PM -0600, Christoph Lameter wrote:
get_user_pages() cannot get to it since the pagetables have already been
modified. If get_user_pages runs then the fault handling will occur
which will block the
On Wed, 12 Nov 2008, Andrea Arcangeli wrote:
On Tue, Nov 11, 2008 at 09:10:45PM -0600, Christoph Lameter wrote:
get_user_pages() cannot get to it since the pagetables have already been
modified. If get_user_pages runs then the fault handling will occur
which will block the thread until
On Wed, 12 Nov 2008, Lee Schermerhorn wrote:
Might want/need to check for migration entry in do_swap_page() and loop
back to migration_entry_wait() call when the changed pte is detected
rather than returning an error to the caller.
Does that sound reasonable?
The reference count freezing
On Wed, Nov 12, 2008 at 05:09:03PM -0500, Lee Schermerhorn wrote:
Maybe not so wild, given the complexity of these interactions...
Perhaps Christoph's right it's just wild ideas, but see below.
You both seem to agree the first theory of the tree_lock is bogus
as it's lockless for
On Thu, Nov 13, 2008 at 03:00:59AM +0100, Andrea Arcangeli wrote:
CPU0 migrate.cCPU1 filemap.c
--- --
find_get_page
radix_tree_lookup_slot returns the oldpage
On Thursday 13 November 2008 13:31, Andrea Arcangeli wrote:
On Thu, Nov 13, 2008 at 03:00:59AM +0100, Andrea Arcangeli wrote:
CPU0 migrate.c CPU1 filemap.c
--- --
find_get_page
Thank you for answers.
On Wed, 12 Nov 2008 13:11:12 +0200
Izik Eidus [EMAIL PROTECTED] wrote:
Avi Kivity wrote:
KAMEZAWA Hiroyuki wrote:
Can I make a question ? (I'm working for memory cgroup.)
Now, we do charge to anonymous page when
- charge(+1) when it's mapped firstly (mapcount
From: Izik Eidus [EMAIL PROTECTED]
this function is needed in cases you want to change the userspace
virtual mapping into diffrent physical page,
KSM need this for merging the identical pages.
this function is working by removing the oldpage from the rmap and
calling put_page on it, and by
On Tue, 11 Nov 2008 15:21:39 +0200
Izik Eidus [EMAIL PROTECTED] wrote:
From: Izik Eidus [EMAIL PROTECTED]
this function is needed in cases you want to change the userspace
virtual mapping into diffrent physical page,
Not sure that I understand that description. We want to replace a live
Andrew Morton wrote:
On Tue, 11 Nov 2008 15:21:39 +0200
Izik Eidus [EMAIL PROTECTED] wrote:
From: Izik Eidus [EMAIL PROTECTED]
this function is needed in cases you want to change the userspace
virtual mapping into diffrent physical page,
Not sure that I understand that description.
On Tue, Nov 11, 2008 at 11:45:55AM -0800, Andrew Morton wrote:
page migration already kinda does that. Is there common ground?
btw, page_migration likely is buggy w.r.t. o_direct too (and now
unfixable with gup_fast until the 2.4 brlock is added around it or
similar) if it does the same thing
On Tue, 11 Nov 2008, Izik Eidus wrote:
yes but it replace it with kernel allocated page.
page migration already kinda does that. Is there common ground?
page migration as far as i saw cant migrate anonymous page into kernel page.
if you want we can change page_migration to do that, but
Christoph Lameter wrote:
page migration as far as i saw cant migrate anonymous page into kernel page.
if you want we can change page_migration to do that, but i thought you will
rather have ksm changes separate.
What do you mean by kernel page? The kernel can allocate a page and then
On Tue, Nov 11, 2008 at 03:21:45PM -0600, Christoph Lameter wrote:
What do you mean by kernel page? The kernel can allocate a page and then
point a user space pte to it. That is how page migration works.
Just to make an example, remove_migration_pte adds the page back to
rmap layer. We can't do
On Tue, 11 Nov 2008, Andrea Arcangeli wrote:
btw, page_migration likely is buggy w.r.t. o_direct too (and now
unfixable with gup_fast until the 2.4 brlock is added around it or
similar) if it does the same thing but without any page_mapcount vs
page_count check.
Details please?
Christoph Lameter wrote:
Currently page migration assumes that the page will continue to be part
of the existing file or anon vma.
exactly, and ksm really need it to get out of the existing anon vma!
What you want sounds like assigning a swap pte to an anonymous page? That
way a anon
Christoph Lameter wrote:
page migration requires the page to be on the LRU. That could be changed
if you have a different means of isolating a page from its page tables.
Isn't rmap the means of isolating a page from its page tables? I guess
I'm misunderstanding something.
--
I have a
On Tue, 11 Nov 2008, Izik Eidus wrote:
What do you mean by kernel page? The kernel can allocate a page and then
point a user space pte to it. That is how page migration works.
i mean filebacked page (!AnonPage())
ok.
ksm need the pte inside the vma to point from anonymous page into
On Tue, 11 Nov 2008, Avi Kivity wrote:
Christoph Lameter wrote:
page migration requires the page to be on the LRU. That could be changed
if you have a different means of isolating a page from its page tables.
Isn't rmap the means of isolating a page from its page tables? I guess I'm
Christoph Lameter wrote:
On Tue, 11 Nov 2008, Avi Kivity wrote:
Christoph Lameter wrote:
page migration requires the page to be on the LRU. That could be changed
if you have a different means of isolating a page from its page tables.
Isn't rmap the means of isolating a page
On Tue, Nov 11, 2008 at 03:26:57PM -0600, Christoph Lameter wrote:
On Tue, 11 Nov 2008, Andrea Arcangeli wrote:
btw, page_migration likely is buggy w.r.t. o_direct too (and now
unfixable with gup_fast until the 2.4 brlock is added around it or
similar) if it does the same thing but
On Tue, Nov 11, 2008 at 03:31:18PM -0600, Christoph Lameter wrote:
ksm need the pte inside the vma to point from anonymous page into filebacked
page
can migrate.c do it without changes?
So change anonymous to filebacked page?
Currently page migration assumes that the page will continue
Christoph Lameter wrote:
On Tue, 11 Nov 2008, Avi Kivity wrote:
Christoph Lameter wrote:
page migration requires the page to be on the LRU. That could be changed
if you have a different means of isolating a page from its page tables.
Isn't rmap the means of isolating a page
On Tue, Nov 11, 2008 at 04:30:22PM -0600, Christoph Lameter wrote:
On Tue, 11 Nov 2008, Andrea Arcangeli wrote:
this page_count check done with only the tree_lock won't prevent a
task to start O_DIRECT after page_count has been read in the above line.
If a thread starts O_DIRECT on the
On Wed, Nov 12, 2008 at 12:17:22AM +0100, Andrea Arcangeli wrote:
We don't have to check the page_count vs mapcount later in
replace_page because we know if anybody started an O_DIRECT read from
disk, it would have triggered a cow, and the pte_same check that we
have to do for other reasons
On Wed, 12 Nov 2008, Andrea Arcangeli wrote:
O_DIRECT does not take a refcount on the page in order to prevent this?
It definitely does, it's also the only thing it does.
Then page migration will not occur because there is an unresolved
reference.
The whole point is that O_DIRECT can start
On Tue, 11 Nov 2008 23:24:21 +0100
Andrea Arcangeli [EMAIL PROTECTED] wrote:
On Tue, Nov 11, 2008 at 03:31:18PM -0600, Christoph Lameter wrote:
ksm need the pte inside the vma to point from anonymous page into
filebacked
page
can migrate.c do it without changes?
So change
On Tue, Nov 11, 2008 at 06:27:09PM -0600, Christoph Lameter wrote:
Then page migration will not occur because there is an unresolved
reference.
So are you checking if there's an unresolved reference only in the
very place I just quoted in the previous email? If answer is yes: what
should
On Wed, 12 Nov 2008, Andrea Arcangeli wrote:
So are you checking if there's an unresolved reference only in the
very place I just quoted in the previous email? If answer is yes: what
should prevent get_user_pages from running in parallel from another
thread? get_user_pages will trigger a
34 matches
Mail list logo
