All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Eduardo Habkost <ehabkost@redhat.com>
Cc: Anthony Liguori <anthony@codemonkey.ws>,
	avi@redhat.com, mtosatti@redhat.com, gleb@redhat.com,
	qemu-devel@nongnu.org, kvm@vger.kernel.org,
	Jan Kiszka <jan.kiszka@web.de>, Blue Swirl <blauwirbel@gmail.com>
Subject: Re: [Qemu-devel] [PATCHv4 3/4] cpuid: disable pv eoi for 1.1 and older compat types
Date: Wed, 29 Aug 2012 00:35:28 +0300	[thread overview]
Message-ID: <20120828213527.GB5817@redhat.com> (raw)
In-Reply-To: <20120828191338.GE6223@otherpad.lan.raisama.net>

On Tue, Aug 28, 2012 at 04:13:38PM -0300, Eduardo Habkost wrote:
> On Tue, Aug 28, 2012 at 08:43:52PM +0300, Michael S. Tsirkin wrote:
> > In preparation for adding PV EOI support, disable PV EOI by default for
> > 1.1 and older machine types, to avoid CPUID changing during migration.
> > 
> > PV EOI can still be enabled/disabled by specifying it explicitly.
> >     Enable for 1.1
> >     -M pc-1.1 -cpu kvm64,+kvm_pv_eoi
> >     Disable for 1.2
> >     -M pc-1.2 -cpu kvm64,-kvm_pv_eoi
> > 
> 
> What about users that are already running "qemu-1.1 -M pc-1.1" on a host
> kernel that supports PV EOI already? They would get PV EOI disabled when
> migrating to a destination running "qemu-1.2 -M pc-1.1".
> 
> (On the other hand, people running "qemu-1.1 -M pc-1.1" on a host kernel
> supporting PV EOI already have migration broken, so there's not much we
> can do for them)

Exactly.

Talked to Gleb, long term I think we should rework code to make
it forward-compatible wrt adding new MSRs:
- source gets list of MSRs to be migrated from KVM and simply sends them all
- send all MSRS in key/value format
- destination gets list of MSRs to be migrated from KVM and
  only restores the supported ones
Too late for 1.2?

> While we don't make the KVM feature-bit handling sane (with defaults
> that are not blindly derived from the host kernel capabilities), maybe
> the safest bet is to expect users to not migrate between hosts running
> kernels with different KVM capabilities? (I am not sure which option is
> better)

Sorry not sure what you talk about here. What has KVM feature-bit
handling to do with this patchset?

> 
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> >  hw/Makefile.objs  |  2 +-
> >  hw/cpu_flags.c    | 32 ++++++++++++++++++++++++++++++++
> >  hw/cpu_flags.h    |  9 +++++++++
> >  hw/pc_piix.c      |  2 ++
> >  target-i386/cpu.c |  8 ++++++++
> >  5 files changed, 52 insertions(+), 1 deletion(-)
> >  create mode 100644 hw/cpu_flags.c
> >  create mode 100644 hw/cpu_flags.h
> > 
> > diff --git a/hw/Makefile.objs b/hw/Makefile.objs
> > index 850b87b..3f2532a 100644
> > --- a/hw/Makefile.objs
> > +++ b/hw/Makefile.objs
> > @@ -1,5 +1,5 @@
> >  hw-obj-y = usb/ ide/
> > -hw-obj-y += loader.o
> > +hw-obj-y += loader.o cpu_flags.o
> >  hw-obj-$(CONFIG_VIRTIO) += virtio-console.o
> >  hw-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o
> >  hw-obj-y += fw_cfg.o
> > diff --git a/hw/cpu_flags.c b/hw/cpu_flags.c
> > new file mode 100644
> > index 0000000..7a633c0
> > --- /dev/null
> > +++ b/hw/cpu_flags.c
> > @@ -0,0 +1,32 @@
> > +/*
> > + * CPU compatibility flags.
> > + *
> > + * Copyright (c) 2012 Red Hat Inc.
> > + * Author: Michael S. Tsirkin.
> > + *
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License as published by
> > + * the Free Software Foundation; either version 2 of the License, or
> > + * (at your option) any later version.
> > + *
> > + * This program is distributed in the hope that it will be useful,
> > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > + * GNU General Public License for more details.
> > + *
> > + * You should have received a copy of the GNU General Public License along
> > + * with this program; if not, see <http://www.gnu.org/licenses/>.
> > + */
> > +#include "hw/cpu_flags.h"
> > +
> > +static bool kvm_pv_eoi_disabled_state;
> > +
> > +void disable_kvm_pv_eoi(void)
> > +{
> > +       kvm_pv_eoi_disabled_state = true;
> > +}
> > +
> > +bool kvm_pv_eoi_disabled(void)
> > +{
> > +       return kvm_pv_eoi_disabled_state;
> > +}
> > diff --git a/hw/cpu_flags.h b/hw/cpu_flags.h
> > new file mode 100644
> > index 0000000..05777b6
> > --- /dev/null
> > +++ b/hw/cpu_flags.h
> > @@ -0,0 +1,9 @@
> > +#ifndef HW_CPU_FLAGS_H
> > +#define HW_CPU_FLAGS_H
> > +
> > +#include <stdbool.h>
> > +
> > +void disable_kvm_pv_eoi(void);
> > +bool kvm_pv_eoi_disabled(void);
> > +
> > +#endif
> > diff --git a/hw/pc_piix.c b/hw/pc_piix.c
> > index 008d42f..bdbceda 100644
> > --- a/hw/pc_piix.c
> > +++ b/hw/pc_piix.c
> > @@ -46,6 +46,7 @@
> >  #ifdef CONFIG_XEN
> >  #  include <xen/hvm/hvm_info_table.h>
> >  #endif
> > +#include "cpu_flags.h"
> >  
> >  #define MAX_IDE_BUS 2
> >  
> > @@ -371,6 +372,7 @@ static QEMUMachine pc_machine_v1_2 = {
> >  
> >  static void pc_machine_v1_1_compat(void)
> >  {
> > +    disable_kvm_pv_eoi();
> >  }
> >  
> >  static void pc_init_pci_v1_1(ram_addr_t ram_size,
> > diff --git a/target-i386/cpu.c b/target-i386/cpu.c
> > index 120a2e3..0d02fd1 100644
> > --- a/target-i386/cpu.c
> > +++ b/target-i386/cpu.c
> > @@ -23,6 +23,7 @@
> >  
> >  #include "cpu.h"
> >  #include "kvm.h"
> > +#include "asm/kvm_para.h"
> >  
> >  #include "qemu-option.h"
> >  #include "qemu-config.h"
> > @@ -33,6 +34,7 @@
> >  #include "hyperv.h"
> >  
> >  #include "hw/hw.h"
> > +#include "hw/cpu_flags.h"
> >  
> >  /* feature flags taken from "Intel Processor Identification and the CPUID
> >   * Instruction" and AMD's "CPUID Specification".  In cases of disagreement
> > @@ -889,6 +891,12 @@ static int cpu_x86_find_by_name(x86_def_t *x86_cpu_def, const char *cpu_model)
> >  
> >      plus_kvm_features = ~0; /* not supported bits will be filtered out later */
> >  
> > +    /* Disable PV EOI for old machine types.
> > +     * Feature flags can still override. */
> > +    if (kvm_pv_eoi_disabled()) {
> > +        plus_kvm_features &= ~(0x1 << KVM_FEATURE_PV_EOI);
> > +    }
> > +
> >      add_flagname_to_bitmaps("hypervisor", &plus_features,
> >          &plus_ext_features, &plus_ext2_features, &plus_ext3_features,
> >          &plus_kvm_features, &plus_svm_features);
> > -- 
> > MST
> > 
> > 
> 
> -- 
> Eduardo

WARNING: multiple messages have this Message-ID (diff)
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Eduardo Habkost <ehabkost@redhat.com>
Cc: gleb@redhat.com, kvm@vger.kernel.org, mtosatti@redhat.com,
	qemu-devel@nongnu.org, Blue Swirl <blauwirbel@gmail.com>,
	Jan Kiszka <jan.kiszka@web.de>,
	avi@redhat.com, Anthony Liguori <anthony@codemonkey.ws>
Subject: Re: [Qemu-devel] [PATCHv4 3/4] cpuid: disable pv eoi for 1.1 and older compat types
Date: Wed, 29 Aug 2012 00:35:28 +0300	[thread overview]
Message-ID: <20120828213527.GB5817@redhat.com> (raw)
In-Reply-To: <20120828191338.GE6223@otherpad.lan.raisama.net>

On Tue, Aug 28, 2012 at 04:13:38PM -0300, Eduardo Habkost wrote:
> On Tue, Aug 28, 2012 at 08:43:52PM +0300, Michael S. Tsirkin wrote:
> > In preparation for adding PV EOI support, disable PV EOI by default for
> > 1.1 and older machine types, to avoid CPUID changing during migration.
> > 
> > PV EOI can still be enabled/disabled by specifying it explicitly.
> >     Enable for 1.1
> >     -M pc-1.1 -cpu kvm64,+kvm_pv_eoi
> >     Disable for 1.2
> >     -M pc-1.2 -cpu kvm64,-kvm_pv_eoi
> > 
> 
> What about users that are already running "qemu-1.1 -M pc-1.1" on a host
> kernel that supports PV EOI already? They would get PV EOI disabled when
> migrating to a destination running "qemu-1.2 -M pc-1.1".
> 
> (On the other hand, people running "qemu-1.1 -M pc-1.1" on a host kernel
> supporting PV EOI already have migration broken, so there's not much we
> can do for them)

Exactly.

Talked to Gleb, long term I think we should rework code to make
it forward-compatible wrt adding new MSRs:
- source gets list of MSRs to be migrated from KVM and simply sends them all
- send all MSRS in key/value format
- destination gets list of MSRs to be migrated from KVM and
  only restores the supported ones
Too late for 1.2?

> While we don't make the KVM feature-bit handling sane (with defaults
> that are not blindly derived from the host kernel capabilities), maybe
> the safest bet is to expect users to not migrate between hosts running
> kernels with different KVM capabilities? (I am not sure which option is
> better)

Sorry not sure what you talk about here. What has KVM feature-bit
handling to do with this patchset?

> 
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> >  hw/Makefile.objs  |  2 +-
> >  hw/cpu_flags.c    | 32 ++++++++++++++++++++++++++++++++
> >  hw/cpu_flags.h    |  9 +++++++++
> >  hw/pc_piix.c      |  2 ++
> >  target-i386/cpu.c |  8 ++++++++
> >  5 files changed, 52 insertions(+), 1 deletion(-)
> >  create mode 100644 hw/cpu_flags.c
> >  create mode 100644 hw/cpu_flags.h
> > 
> > diff --git a/hw/Makefile.objs b/hw/Makefile.objs
> > index 850b87b..3f2532a 100644
> > --- a/hw/Makefile.objs
> > +++ b/hw/Makefile.objs
> > @@ -1,5 +1,5 @@
> >  hw-obj-y = usb/ ide/
> > -hw-obj-y += loader.o
> > +hw-obj-y += loader.o cpu_flags.o
> >  hw-obj-$(CONFIG_VIRTIO) += virtio-console.o
> >  hw-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o
> >  hw-obj-y += fw_cfg.o
> > diff --git a/hw/cpu_flags.c b/hw/cpu_flags.c
> > new file mode 100644
> > index 0000000..7a633c0
> > --- /dev/null
> > +++ b/hw/cpu_flags.c
> > @@ -0,0 +1,32 @@
> > +/*
> > + * CPU compatibility flags.
> > + *
> > + * Copyright (c) 2012 Red Hat Inc.
> > + * Author: Michael S. Tsirkin.
> > + *
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License as published by
> > + * the Free Software Foundation; either version 2 of the License, or
> > + * (at your option) any later version.
> > + *
> > + * This program is distributed in the hope that it will be useful,
> > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > + * GNU General Public License for more details.
> > + *
> > + * You should have received a copy of the GNU General Public License along
> > + * with this program; if not, see <http://www.gnu.org/licenses/>.
> > + */
> > +#include "hw/cpu_flags.h"
> > +
> > +static bool kvm_pv_eoi_disabled_state;
> > +
> > +void disable_kvm_pv_eoi(void)
> > +{
> > +       kvm_pv_eoi_disabled_state = true;
> > +}
> > +
> > +bool kvm_pv_eoi_disabled(void)
> > +{
> > +       return kvm_pv_eoi_disabled_state;
> > +}
> > diff --git a/hw/cpu_flags.h b/hw/cpu_flags.h
> > new file mode 100644
> > index 0000000..05777b6
> > --- /dev/null
> > +++ b/hw/cpu_flags.h
> > @@ -0,0 +1,9 @@
> > +#ifndef HW_CPU_FLAGS_H
> > +#define HW_CPU_FLAGS_H
> > +
> > +#include <stdbool.h>
> > +
> > +void disable_kvm_pv_eoi(void);
> > +bool kvm_pv_eoi_disabled(void);
> > +
> > +#endif
> > diff --git a/hw/pc_piix.c b/hw/pc_piix.c
> > index 008d42f..bdbceda 100644
> > --- a/hw/pc_piix.c
> > +++ b/hw/pc_piix.c
> > @@ -46,6 +46,7 @@
> >  #ifdef CONFIG_XEN
> >  #  include <xen/hvm/hvm_info_table.h>
> >  #endif
> > +#include "cpu_flags.h"
> >  
> >  #define MAX_IDE_BUS 2
> >  
> > @@ -371,6 +372,7 @@ static QEMUMachine pc_machine_v1_2 = {
> >  
> >  static void pc_machine_v1_1_compat(void)
> >  {
> > +    disable_kvm_pv_eoi();
> >  }
> >  
> >  static void pc_init_pci_v1_1(ram_addr_t ram_size,
> > diff --git a/target-i386/cpu.c b/target-i386/cpu.c
> > index 120a2e3..0d02fd1 100644
> > --- a/target-i386/cpu.c
> > +++ b/target-i386/cpu.c
> > @@ -23,6 +23,7 @@
> >  
> >  #include "cpu.h"
> >  #include "kvm.h"
> > +#include "asm/kvm_para.h"
> >  
> >  #include "qemu-option.h"
> >  #include "qemu-config.h"
> > @@ -33,6 +34,7 @@
> >  #include "hyperv.h"
> >  
> >  #include "hw/hw.h"
> > +#include "hw/cpu_flags.h"
> >  
> >  /* feature flags taken from "Intel Processor Identification and the CPUID
> >   * Instruction" and AMD's "CPUID Specification".  In cases of disagreement
> > @@ -889,6 +891,12 @@ static int cpu_x86_find_by_name(x86_def_t *x86_cpu_def, const char *cpu_model)
> >  
> >      plus_kvm_features = ~0; /* not supported bits will be filtered out later */
> >  
> > +    /* Disable PV EOI for old machine types.
> > +     * Feature flags can still override. */
> > +    if (kvm_pv_eoi_disabled()) {
> > +        plus_kvm_features &= ~(0x1 << KVM_FEATURE_PV_EOI);
> > +    }
> > +
> >      add_flagname_to_bitmaps("hypervisor", &plus_features,
> >          &plus_ext_features, &plus_ext2_features, &plus_ext3_features,
> >          &plus_kvm_features, &plus_svm_features);
> > -- 
> > MST
> > 
> > 
> 
> -- 
> Eduardo

  reply	other threads:[~2012-08-28 21:34 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-28 17:43 [PATCHv4 0/4] migrate PV EOI MSR Michael S. Tsirkin
2012-08-28 17:43 ` [Qemu-devel] " Michael S. Tsirkin
2012-08-28 17:43 ` [PATCHv4 1/4] linux-headers: update to 3.6-rc3 Michael S. Tsirkin
2012-08-28 17:43   ` [Qemu-devel] " Michael S. Tsirkin
2012-08-28 17:43 ` [PATCHv4 2/4] pc: refactor compat code Michael S. Tsirkin
2012-08-28 17:43   ` [Qemu-devel] " Michael S. Tsirkin
2012-08-29 14:49   ` Anthony Liguori
2012-08-29 14:49     ` [Qemu-devel] " Anthony Liguori
2012-08-28 17:43 ` [PATCHv4 3/4] cpuid: disable pv eoi for 1.1 and older compat types Michael S. Tsirkin
2012-08-28 17:43   ` [Qemu-devel] " Michael S. Tsirkin
2012-08-28 19:13   ` Eduardo Habkost
2012-08-28 19:13     ` Eduardo Habkost
2012-08-28 21:35     ` Michael S. Tsirkin [this message]
2012-08-28 21:35       ` Michael S. Tsirkin
2012-08-28 22:02       ` Eduardo Habkost
2012-08-28 22:02         ` Eduardo Habkost
2012-08-28 22:21         ` Michael S. Tsirkin
2012-08-28 22:21           ` Michael S. Tsirkin
2012-08-28 22:25           ` Michael S. Tsirkin
2012-08-28 22:25             ` [Qemu-devel] " Michael S. Tsirkin
2012-08-28 23:50             ` Eduardo Habkost
2012-08-28 23:50               ` Eduardo Habkost
2012-08-29 10:06               ` Michael S. Tsirkin
2012-08-29 10:06                 ` Michael S. Tsirkin
2012-08-29 12:56                 ` Eduardo Habkost
2012-08-29 12:56                   ` Eduardo Habkost
2012-08-29 13:18                   ` Michael S. Tsirkin
2012-08-29 13:18                     ` [Qemu-devel] " Michael S. Tsirkin
2012-08-29 13:49                     ` Eduardo Habkost
2012-08-29 13:49                       ` Eduardo Habkost
2012-08-29 14:11                       ` Michael S. Tsirkin
2012-08-29 14:11                         ` Michael S. Tsirkin
2012-08-29 14:21                         ` Eduardo Habkost
2012-08-29 14:21                           ` Eduardo Habkost
2012-08-29  9:59           ` Marcelo Tosatti
2012-08-29  9:59             ` Marcelo Tosatti
2012-08-29 10:03             ` Marcelo Tosatti
2012-08-29 10:03               ` Marcelo Tosatti
2012-08-29 10:32               ` Michael S. Tsirkin
2012-08-29 10:32                 ` Michael S. Tsirkin
2012-08-29 10:23             ` Michael S. Tsirkin
2012-08-29 10:23               ` Michael S. Tsirkin
2012-08-29 10:31               ` Michael S. Tsirkin
2012-08-29 10:31                 ` Michael S. Tsirkin
2012-08-29 14:43       ` Juan Quintela
2012-08-29 14:43         ` [Qemu-devel] " Juan Quintela
2012-08-29 13:36   ` Anthony Liguori
2012-08-29 13:36     ` [Qemu-devel] " Anthony Liguori
2012-08-29 13:40     ` Gleb Natapov
2012-08-29 13:40       ` [Qemu-devel] " Gleb Natapov
2012-08-29 14:09       ` Anthony Liguori
2012-08-29 14:09         ` [Qemu-devel] " Anthony Liguori
2012-08-29 14:29         ` Michael S. Tsirkin
2012-08-29 14:29           ` [Qemu-devel] " Michael S. Tsirkin
2012-08-29 14:41         ` Eduardo Habkost
2012-08-29 15:04           ` [QEMU 1.2 PATCH] i386: kvm: have a predefined set of default KVM feature bits Eduardo Habkost
2012-08-29 15:04             ` [Qemu-devel] " Eduardo Habkost
2012-08-29 15:10             ` Michael S. Tsirkin
2012-08-29 15:10               ` [Qemu-devel] " Michael S. Tsirkin
2012-08-29 15:46               ` Marcelo Tosatti
2012-08-29 15:46                 ` [Qemu-devel] " Marcelo Tosatti
2012-08-29 14:03     ` [Qemu-devel] [PATCHv4 3/4] cpuid: disable pv eoi for 1.1 and older compat types Eduardo Habkost
2012-08-29 14:03       ` Eduardo Habkost
2012-08-28 17:43 ` [PATCHv4 4/4] kvm: get/set PV EOI MSR Michael S. Tsirkin
2012-08-28 17:43   ` [Qemu-devel] " Michael S. Tsirkin

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=20120828213527.GB5817@redhat.com \
    --to=mst@redhat.com \
    --cc=anthony@codemonkey.ws \
    --cc=avi@redhat.com \
    --cc=blauwirbel@gmail.com \
    --cc=ehabkost@redhat.com \
    --cc=gleb@redhat.com \
    --cc=jan.kiszka@web.de \
    --cc=kvm@vger.kernel.org \
    --cc=mtosatti@redhat.com \
    --cc=qemu-devel@nongnu.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.