* Re: Completely new idea to virtual memory
@ 2003-04-12 9:52 Chuck Ebbert
2003-04-12 10:29 ` Nigel Cunningham
2003-04-12 11:10 ` John Bradford
0 siblings, 2 replies; 4+ messages in thread
From: Chuck Ebbert @ 2003-04-12 9:52 UTC (permalink / raw)
To: John Bradford; +Cc: linux-kernel
>> Why not also write pages out to swap before it's really necessary?
>> If they were left mapped but marked as having up-to-date copies
>> on swap, they could be discarded immediately if the system needed
>> memory. (Of course if they got written to they would have to be
>> paged out again.)
>
>Hmmm, interesting...
The gains from this might be largely illusion but you could get some
theoretical benefits:
o could write out lots of pages to swap without having to flush TLBs
like you would if doing normal swap, since page remains present
o if swsusp does not need to write already-backed pages it could
reduce suspend time
Maybe something like this?
if about_to_spin_down_disks || swap_disk_idle
if lots_of_memory_free
page_in
if little_memory_free
copy_pages_to_swap
--
"Let's fight till six, and then have dinner," said Tweedledum.
--Lewis Carroll, _Through the Looking Glass_
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Completely new idea to virtual memory
2003-04-12 9:52 Completely new idea to virtual memory Chuck Ebbert
@ 2003-04-12 10:29 ` Nigel Cunningham
2003-04-12 11:10 ` John Bradford
1 sibling, 0 replies; 4+ messages in thread
From: Nigel Cunningham @ 2003-04-12 10:29 UTC (permalink / raw)
To: Chuck Ebbert; +Cc: John Bradford, Linux Kernel Mailing List
On Sat, 2003-04-12 at 21:52, Chuck Ebbert wrote:
> o if swsusp does not need to write already-backed pages it could
> reduce suspend time
Hmmm. Not sure if it's worth the time taken to code & run figuring out
if it's already backed, though :>. Would we ever be talking about a
significant proportion of pages?
Regards,
Nigel
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Completely new idea to virtual memory
2003-04-12 9:52 Completely new idea to virtual memory Chuck Ebbert
2003-04-12 10:29 ` Nigel Cunningham
@ 2003-04-12 11:10 ` John Bradford
1 sibling, 0 replies; 4+ messages in thread
From: John Bradford @ 2003-04-12 11:10 UTC (permalink / raw)
To: Chuck Ebbert; +Cc: John Bradford, linux-kernel
> Maybe something like this?
>
> if about_to_spin_down_disks || swap_disk_idle
> if lots_of_memory_free
> page_in
> if little_memory_free
> copy_pages_to_swap
Exactly - plus, if the disk is only, or mainly, used for swap, you
could purposely spin it down after doing the speculative swapout.
John.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Completely new idea to virtual memory
@ 2003-04-12 16:47 Chuck Ebbert
0 siblings, 0 replies; 4+ messages in thread
From: Chuck Ebbert @ 2003-04-12 16:47 UTC (permalink / raw)
To: Nigel Cunningham; +Cc: linux-kernel
Nigel Cunningham wrote:
>> o if swsusp does not need to write already-backed pages it could
>> reduce suspend time
>
> Hmmm. Not sure if it's worth the time taken to code & run figuring out
> if it's already backed, though :>. Would we ever be talking about a
> significant proportion of pages?
Depends on how aggressively you write them out, I guess.
Is there even a way to write pages to swap without unmapping them?
--
"Let's fight till six, and then have dinner," said Tweedledum.
--Lewis Carroll, _Through the Looking Glass_
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2003-04-12 16:41 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-04-12 9:52 Completely new idea to virtual memory Chuck Ebbert
2003-04-12 10:29 ` Nigel Cunningham
2003-04-12 11:10 ` John Bradford
-- strict thread matches above, loose matches on Subject: below --
2003-04-12 16:47 Chuck Ebbert
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.