* KSM without VT / KSM for all memory
@ 2010-03-26 18:22 Gordan Bobic
2010-03-26 22:46 ` Chris Wright
0 siblings, 1 reply; 4+ messages in thread
From: Gordan Bobic @ 2010-03-26 18:22 UTC (permalink / raw)
To: kvm
Hi,
Is it possible to use KSM:
1) Without hardware VT support
2) For all memory in a system, without patching all applications to
register with it
TIA.
Gordan
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: KSM without VT / KSM for all memory
2010-03-26 18:22 KSM without VT / KSM for all memory Gordan Bobic
@ 2010-03-26 22:46 ` Chris Wright
2010-03-26 23:47 ` Gordan Bobic
0 siblings, 1 reply; 4+ messages in thread
From: Chris Wright @ 2010-03-26 22:46 UTC (permalink / raw)
To: Gordan Bobic; +Cc: kvm
* Gordan Bobic (gordan@bobich.net) wrote:
> Is it possible to use KSM:
> 1) Without hardware VT support
Yes.
KSM and hardware virtualization support are technically not related.
> 2) For all memory in a system, without patching all applications to
> register with it
No.
Right now, an app must be modified to call madvise(MADV_MERGEABLE).
Further, the core scanning loop that ksmd performs is based on per-process
virtual memory regions rather than physical memory.
thanks,
-chris
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: KSM without VT / KSM for all memory
2010-03-26 22:46 ` Chris Wright
@ 2010-03-26 23:47 ` Gordan Bobic
2010-03-27 0:13 ` Chris Wright
0 siblings, 1 reply; 4+ messages in thread
From: Gordan Bobic @ 2010-03-26 23:47 UTC (permalink / raw)
To: KVM list
Chris Wright wrote:
>> 2) For all memory in a system, without patching all applications to
>> register with it
>
> No.
>
> Right now, an app must be modified to call madvise(MADV_MERGEABLE).
> Further, the core scanning loop that ksmd performs is based on per-process
> virtual memory regions rather than physical memory.
You mean only pages within the same process are de-duplicated?
Gordan
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: KSM without VT / KSM for all memory
2010-03-26 23:47 ` Gordan Bobic
@ 2010-03-27 0:13 ` Chris Wright
0 siblings, 0 replies; 4+ messages in thread
From: Chris Wright @ 2010-03-27 0:13 UTC (permalink / raw)
To: Gordan Bobic; +Cc: KVM list
* Gordan Bobic (gordan@bobich.net) wrote:
> Chris Wright wrote:
>>> 2) For all memory in a system, without patching all applications to
>>> register with it
>>
>> No.
>>
>> Right now, an app must be modified to call madvise(MADV_MERGEABLE).
>> Further, the core scanning loop that ksmd performs is based on per-process
>> virtual memory regions rather than physical memory.
>
> You mean only pages within the same process are de-duplicated?
No, de-duplication can span processes. Each process registers some range(s)
of its virtual address space with the scanning daemon. The scanning daemon
will then search all of the registered memory ranges for duplicates. This
is an opt-in thing for each process though.
thanks,
-chris
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2010-03-27 0:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-03-26 18:22 KSM without VT / KSM for all memory Gordan Bobic
2010-03-26 22:46 ` Chris Wright
2010-03-26 23:47 ` Gordan Bobic
2010-03-27 0:13 ` Chris Wright
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox