From: Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
To: Anthony Liguori <anthony-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
Cc: kvm-devel
<kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org>,
virtualization
<virtualization-qjLDD68F18O7TbgM5vRIOg@public.gmane.org>
Subject: Re: [PATCH 4/5] KVM: Add hypercall queue for paravirt_ops implementation
Date: Mon, 18 Jun 2007 18:08:46 +0300 [thread overview]
Message-ID: <46769FFE.6040502@qumranet.com> (raw)
In-Reply-To: <4676905B.6000805-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
Anthony Liguori wrote:
> Avi Kivity wrote:
>>>> These numbers are pretty bad. I'd like to improve them, even
>>>> without PV.
>>>
>>> I agree. Do you know what's missing at this point? There isn't a
>>> whole lot of state saving going on for the light weight exit paths
>>> for SVM.
>>
>> The SVM code doesn't even have a lightweight vmexit path.
>
> Sure it does. Quite a lot is deferred to vcpu_{load,put}.
Ah, I forgot. Yes, the syscall msrs are deferred.
>
>> For every vmexit, it does the entire thing, including vmload/vmsave
>
> I haven't had a lot of luck eliminating vmload/vmsave.
>
For x86_64, the only issue I see is with TR. Unfortunately, I don't see
a way around it.
>> , fpu switch (if needed)
>
> The FPU switch can really be avoided? Is it safe to assume that the
> KVM code isn't going to use any FPU operations?
Generally, kernel code does not use the fpu (when it does, it calls
kernel_fpu_begin() and kernel_fpu_end()). The vmx code avoids the switch.
Of course, if the guest doesn't use the fpu, the switch is avoided anyway.
>>
>> For kbuild vs. kernbench, I suspect that -j4 causes the shadow page
>> table cache to thrash. 1024 pages may be enough for a single
>> instance but not -j4. Hopefully replacing the eviction algorithm
>> (currently FIFO) will help. Otherwise we'll need to resize the cache
>> again.
>
> I naively tried to bump it to 2048 and hit a kmalloc limitation.
>
struct kvm is 22K on x86_64. Adding 1024 pointers makes it 30K. What
error did you get?
We should probably make the hashtable a pointer, and allocate vcpus
separately as well.
--
error compiling committee.c: too many arguments to function
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
next prev parent reply other threads:[~2007-06-18 15:08 UTC|newest]
Thread overview: 85+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-06-18 2:56 [PATCH 0/5] KVM paravirt_ops implementation Anthony Liguori
[not found] ` <4675F462.1010708-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 2:58 ` [PATCH 1/5] KVM paravirt_ops core infrastructure Anthony Liguori
[not found] ` <4675F4C3.6050700-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 8:03 ` Avi Kivity
[not found] ` <46763C6B.9050004-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:25 ` Anthony Liguori
[not found] ` <467679C5.6030201-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 12:28 ` Avi Kivity
2007-06-26 8:04 ` Dor Laor
[not found] ` <64F9B87B6B770947A9F8391472E032160C73025E-yEcIvxbTEBqsx+V+t5oei8rau4O3wl8o3fe8/T/H7NteoWH0uzbU5w@public.gmane.org>
2007-06-26 8:45 ` Jun Koi
[not found] ` <fdaac4d50706260145x1ebceadt432edd5b6a6ac1f2-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2007-06-26 11:57 ` Anthony Liguori
2007-06-26 11:56 ` Anthony Liguori
2007-06-18 2:58 ` [PATCH 2/5] KVM: Implement CR read caching for KVM paravirt_ops Anthony Liguori
[not found] ` <4675F4F1.5090207-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 8:05 ` Avi Kivity
[not found] ` <46763CD3.3060704-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:26 ` Anthony Liguori
2007-06-18 8:11 ` Avi Kivity
[not found] ` <46763E35.8020108-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:27 ` Anthony Liguori
2007-06-18 3:00 ` [PATCH 3/5] KVM: Add paravirt MMU write support Anthony Liguori
[not found] ` <4675F533.40809-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 8:20 ` Avi Kivity
[not found] ` <46764061.9080705-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:33 ` Anthony Liguori
[not found] ` <46767B8C.9050001-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 12:38 ` Avi Kivity
[not found] ` <46767CD1.7030208-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:48 ` Anthony Liguori
2007-06-19 21:57 ` Anthony Liguori
[not found] ` <46785132.3070505-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-19 22:19 ` Jeremy Fitzhardinge
[not found] ` <4678567C.6040400-TSDbQ3PG+2Y@public.gmane.org>
2007-06-19 22:28 ` Anthony Liguori
2007-06-18 3:00 ` [PATCH 4/5] KVM: Add hypercall queue for paravirt_ops implementation Anthony Liguori
[not found] ` <4675F568.90608-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 4:00 ` Jeremy Fitzhardinge
[not found] ` <46760343.5070401-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 4:09 ` Jeremy Fitzhardinge
2007-06-18 12:22 ` Anthony Liguori
2007-06-18 9:07 ` Avi Kivity
[not found] ` <46764B47.5060403-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 12:40 ` Anthony Liguori
[not found] ` <46767D47.1010104-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 12:50 ` Avi Kivity
[not found] ` <46767F98.70109-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 13:03 ` Gregory Haskins
[not found] ` <1182171781.4593.38.camel-5CR4LY5GPkvLDviKLk5550HKjMygAv58XqFh9Ls21Oc@public.gmane.org>
2007-06-18 13:19 ` Anthony Liguori
[not found] ` <4676867E.1090208-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 13:25 ` Gregory Haskins
2007-06-18 13:22 ` Anthony Liguori
[not found] ` <46768724.3000509-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 13:35 ` Avi Kivity
[not found] ` <46768A3F.2010202-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 14:02 ` Anthony Liguori
[not found] ` <4676905B.6000805-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 15:08 ` Avi Kivity [this message]
[not found] ` <46769FFE.6040502-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 15:20 ` Anthony Liguori
[not found] ` <4676A2D4.2040704-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 16:01 ` Avi Kivity
2007-06-18 16:00 ` Avi Kivity
[not found] ` <4676AC10.3090007-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-18 17:47 ` Anthony Liguori
2007-06-18 3:03 ` [PATCH 5/5] KVM: paravirt time source Anthony Liguori
[not found] ` <4675F601.3090706-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 9:24 ` Avi Kivity
2007-06-18 19:11 ` Jeremy Fitzhardinge
[not found] ` <4676D8E4.3020806-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 21:52 ` Anthony Liguori
[not found] ` <4676FEB9.6060308-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 22:04 ` Jeremy Fitzhardinge
[not found] ` <46770162.6030101-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 23:33 ` Anthony Liguori
[not found] ` <4677163F.2030308-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 23:56 ` Jeremy Fitzhardinge
[not found] ` <46771BA0.2000308-TSDbQ3PG+2Y@public.gmane.org>
2007-06-19 0:53 ` Anthony Liguori
2007-06-19 7:44 ` Avi Kivity
[not found] ` <4677894D.3050500-atKUWr5tajBWk0Htik3J/w@public.gmane.org>
2007-06-19 8:04 ` Rusty Russell
2007-06-19 20:38 ` Anthony Liguori
[not found] ` <46783EDB.5010808-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-19 21:26 ` Jeremy Fitzhardinge
[not found] ` <46784A0E.3080502-TSDbQ3PG+2Y@public.gmane.org>
2007-06-19 21:38 ` Anthony Liguori
2007-06-21 7:04 ` Dong, Eddie
2007-06-18 3:19 ` [PATCH 0/5] KVM paravirt_ops implementation Jeremy Fitzhardinge
[not found] ` <4675F9DE.6080806-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 3:36 ` Anthony Liguori
[not found] ` <4675FDCA.4040006-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 4:21 ` Jeremy Fitzhardinge
[not found] ` <4676084F.3090901-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 12:46 ` Anthony Liguori
[not found] ` <46767EB2.4090707-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-18 14:46 ` Jeremy Fitzhardinge
[not found] ` <46769AD2.9080105-TSDbQ3PG+2Y@public.gmane.org>
2007-06-18 15:07 ` Anthony Liguori
2007-06-20 0:21 ` Zachary Amsden
[not found] ` <46787325.30804-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 14:22 ` Anthony Liguori
[not found] ` <4679381E.9090404-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-20 15:32 ` Jeremy Fitzhardinge
[not found] ` <46794899.6070708-TSDbQ3PG+2Y@public.gmane.org>
2007-06-20 19:35 ` Zachary Amsden
[not found] ` <46798174.2060304-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 19:47 ` Jeremy Fitzhardinge
2007-06-20 19:52 ` Anthony Liguori
[not found] ` <46798597.4020903-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-20 20:03 ` Zachary Amsden
[not found] ` <4679882B.7070605-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 20:16 ` Jeremy Fitzhardinge
[not found] ` <46798B16.5090407-TSDbQ3PG+2Y@public.gmane.org>
2007-06-20 20:27 ` Anthony Liguori
[not found] ` <46798D9B.5000400-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-20 20:33 ` Jeremy Fitzhardinge
[not found] ` <46798F1A.4090901-TSDbQ3PG+2Y@public.gmane.org>
2007-06-20 20:46 ` Zachary Amsden
[not found] ` <4679921A.8090607-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 20:55 ` Jeremy Fitzhardinge
2007-06-20 22:08 ` Anthony Liguori
[not found] ` <4679A54F.5020908-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-20 22:33 ` Jeremy Fitzhardinge
2007-06-20 20:43 ` Zachary Amsden
[not found] ` <46799157.3070805-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 20:53 ` Jeremy Fitzhardinge
[not found] ` <467993B1.6000307-TSDbQ3PG+2Y@public.gmane.org>
2007-06-20 21:08 ` Zachary Amsden
[not found] ` <46799755.8060405-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 21:48 ` Jeremy Fitzhardinge
2007-06-20 22:39 ` Anthony Liguori
2007-06-20 20:22 ` Anthony Liguori
[not found] ` <46798C99.8010303-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org>
2007-06-20 20:37 ` Zachary Amsden
[not found] ` <46799001.5020807-pghWNbHTmq7QT0dZR+AlfA@public.gmane.org>
2007-06-20 21:07 ` Jeremy Fitzhardinge
[not found] ` <46799716.9040402-TSDbQ3PG+2Y@public.gmane.org>
2007-06-20 21:27 ` ron minnich
2007-06-20 21:39 ` H. Peter Anvin
2007-06-19 23:49 ` Zachary Amsden
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=46769FFE.6040502@qumranet.com \
--to=avi-atkuwr5tajbwk0htik3j/w@public.gmane.org \
--cc=anthony-rdkfGonbjUSkNkDKm+mE6A@public.gmane.org \
--cc=kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org \
--cc=virtualization-qjLDD68F18O7TbgM5vRIOg@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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.