From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Jan Beulich <JBeulich@suse.com>
Cc: George Dunlap <George.Dunlap@eu.citrix.com>,
xen-devel <xen-devel@lists.xenproject.org>,
Keir Fraser <keir@xen.org>
Subject: Re: [PATCH 3/3] domctl: pause vCPU for context reads
Date: Wed, 5 Feb 2014 15:29:19 +0000 [thread overview]
Message-ID: <52F258CF.50706@citrix.com> (raw)
In-Reply-To: <52F25EB702000078001196EC@nat28.tlf.novell.com>
[-- Attachment #1.1: Type: text/plain, Size: 2626 bytes --]
On 05/02/14 14:54, Jan Beulich wrote:
> "Base" context reads already paused the subject vCPU when being the
> current one, but that special case isn't being properly dealt with
> anyway (at the very least when x86's fsgsbase feature is in use), so
> just disallow it.
>
> "Extended" context reads so far didn't do any pausing.
>
> While we can't avoid the reported data being stale by the time it
> arrives at the caller, this way we at least guarantee that it is
> consistent.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
Now I come to think about it, is this an ABI change, as we are now
disallowing a control domain to issue these hypercalls on itself?
~Andrew
>
> --- a/xen/arch/x86/domctl.c
> +++ b/xen/arch/x86/domctl.c
> @@ -819,7 +819,13 @@ long arch_do_domctl(
>
> if ( domctl->cmd == XEN_DOMCTL_get_ext_vcpucontext )
> {
> + if ( v == current ) /* no vcpu_pause() */
> + break;
> +
> evc->size = sizeof(*evc);
> +
> + vcpu_pause(v);
> +
> if ( is_pv_domain(d) )
> {
> evc->sysenter_callback_cs =
> @@ -849,6 +855,7 @@ long arch_do_domctl(
> evc->vmce.mci_ctl2_bank1 = v->arch.vmce.bank[1].mci_ctl2;
>
> ret = 0;
> + vcpu_unpause(v);
> copyback = 1;
> }
> else
> --- a/xen/common/domctl.c
> +++ b/xen/common/domctl.c
> @@ -675,11 +675,9 @@ long do_domctl(XEN_GUEST_HANDLE_PARAM(xe
> struct vcpu *v;
>
> ret = -EINVAL;
> - if ( op->u.vcpucontext.vcpu >= d->max_vcpus )
> - goto getvcpucontext_out;
> -
> - ret = -ESRCH;
> - if ( (v = d->vcpu[op->u.vcpucontext.vcpu]) == NULL )
> + if ( op->u.vcpucontext.vcpu >= d->max_vcpus ||
> + (v = d->vcpu[op->u.vcpucontext.vcpu]) == NULL ||
> + v == current ) /* no vcpu_pause() */
> goto getvcpucontext_out;
>
> ret = -ENODATA;
> @@ -694,14 +692,12 @@ long do_domctl(XEN_GUEST_HANDLE_PARAM(xe
> if ( (c.nat = xmalloc(struct vcpu_guest_context)) == NULL )
> goto getvcpucontext_out;
>
> - if ( v != current )
> - vcpu_pause(v);
> + vcpu_pause(v);
>
> arch_get_info_guest(v, c);
> ret = 0;
>
> - if ( v != current )
> - vcpu_unpause(v);
> + vcpu_unpause(v);
>
> #ifdef CONFIG_COMPAT
> if ( !is_pv_32on64_vcpu(v) )
>
>
>
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
[-- Attachment #1.2: Type: text/html, Size: 3392 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
next prev parent reply other threads:[~2014-02-05 15:29 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-02-05 14:48 [PATCH 0/3] guest context management adjustments Jan Beulich
2014-02-05 14:52 ` [PATCH 1/3] x86: fix FS/GS base handling when using the fsgsbase feature Jan Beulich
2014-02-05 15:50 ` Andrew Cooper
2014-02-05 14:54 ` [PATCH 2/3] domctl: also pause domain for extended context updates Jan Beulich
2014-02-05 15:05 ` Andrew Cooper
2014-02-05 14:54 ` [PATCH 3/3] domctl: pause vCPU for context reads Jan Beulich
2014-02-05 15:29 ` Andrew Cooper [this message]
2014-02-05 15:39 ` Jan Beulich
2014-02-05 15:43 ` Andrew Cooper
2014-02-05 14:55 ` [PATCH 0/3] guest context management adjustments Jan Beulich
2014-02-05 16:02 ` George Dunlap
2014-02-05 15:35 ` Keir Fraser
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=52F258CF.50706@citrix.com \
--to=andrew.cooper3@citrix.com \
--cc=George.Dunlap@eu.citrix.com \
--cc=JBeulich@suse.com \
--cc=keir@xen.org \
--cc=xen-devel@lists.xenproject.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).