All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: speck@linutronix.de
Subject: [MODERATED] Re: ***UNCHECKED*** [patch V5 05/10] KVM magic # 5
Date: Tue, 3 Jul 2018 10:23:11 +0200	[thread overview]
Message-ID: <46b8ee53-ad99-4c1c-f3c6-5aedd1edc168@redhat.com> (raw)
In-Reply-To: <dc14f2c1-f69c-e690-dad2-8b4f1b772ee3@suse.de>

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]

On 02/07/2018 23:19, speck for Alexander Graf wrote:
>> +		vcpu->arch.l1tf_flush_l1d = true;
> This means we flush when an external interrupt is the trap reason, but
> we may as well get the interrupt only after interrupts are enabled again
> from vcpu_enter_guest(), no? How is an interrupt executing "normally"
> any different from an interrupt that causes a #vmexit?
> 
> Or in other words: Is it worth it to have anything but always/never as
> an option? Has anyone done benchmarks?

Yes, and "always" is terrible. :((  See my reply to Boris a couple days
ago.  A good long term plan is to switch from blacklist to whitelist,
and run the whitelisted vmexits with interrupts disabled; another
possibility is to set a flag whenever softirqs run, and I'll send an RFC
on top of the basic pieces.

Paolo


  parent reply	other threads:[~2018-07-03  8:23 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-02 15:44 [patch V5 00/10] KVM magic # 0 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 01/10] KVM magic # 1 Thomas Gleixner
2018-07-02 16:22   ` [MODERATED] " Konrad Rzeszutek Wilk
2018-07-02 17:10     ` Thomas Gleixner
2018-07-02 16:25   ` [MODERATED] " Borislav Petkov
2018-07-02 15:44 ` [patch V5 02/10] KVM magic # 2 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 03/10] KVM magic # 3 Thomas Gleixner
2018-07-02 23:42   ` [MODERATED] " Jon Masters
2018-07-02 15:44 ` [patch V5 04/10] KVM magic # 4 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 05/10] KVM magic # 5 Thomas Gleixner
2018-07-02 16:35   ` [MODERATED] " Konrad Rzeszutek Wilk
2018-07-02 17:01     ` Thomas Gleixner
2018-07-02 17:24   ` [MODERATED] " Paolo Bonzini
2018-07-03 15:21     ` Thomas Gleixner
2018-07-02 21:19   ` [MODERATED] Re: ***UNCHECKED*** " Alexander Graf
2018-07-02 23:45     ` Andrew Cooper
2018-07-03  1:33       ` Linus Torvalds
2018-07-03  8:23     ` Paolo Bonzini [this message]
2018-07-03 14:29       ` [MODERATED] Re: ***UNCHECKED*** " Alexander Graf
2018-07-05 19:08         ` Jon Masters
2018-07-05 21:43         ` Paolo Bonzini
2018-07-05 21:50           ` Linus Torvalds
2018-07-02 15:44 ` [patch V5 06/10] KVM magic # 6 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 07/10] KVM magic # 7 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 08/10] KVM magic # 8 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 09/10] KVM magic # 9 Thomas Gleixner
2018-07-02 15:44 ` [patch V5 10/10] KVM magic # 10 Thomas Gleixner
2018-07-02 17:29   ` [MODERATED] " Paolo Bonzini
2018-07-02 17:41     ` Thomas Gleixner
2018-07-02 22:11       ` Thomas Gleixner
2018-07-03  9:05         ` [MODERATED] " Paolo Bonzini
2018-07-02 16:25 ` [patch V5 00/10] KVM magic # 0 Thomas Gleixner
2018-07-02 22:14 ` [MODERATED] [patch v5 11/10] Linux+KVM magic # 1 Jiri Kosina
2018-07-05 23:56   ` [MODERATED] " Josh Poimboeuf
2018-07-06 11:54     ` Jiri Kosina
2018-07-06 13:05       ` Konrad Rzeszutek Wilk
2018-07-06 15:11       ` Jon Masters

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=46b8ee53-ad99-4c1c-f3c6-5aedd1edc168@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=speck@linutronix.de \
    /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.