On Thu, 2012-01-12 at 08:51 +0200, Michael S. Tsirkin wrote:
> One othe rinteresting thing is that
> in practice host often runs on the same CPU as guest.
That's surprising. On Power with our huge exit cost, I would think that
would hurt more than anything else.
> This is the fastest way to run
On Thu, Jan 12, 2012 at 05:28:34PM +1100, Benjamin Herrenschmidt wrote:
> On Thu, 2012-01-12 at 08:09 +0200, Michael S. Tsirkin wrote:
> > On Thu, Jan 12, 2012 at 03:31:59PM +1100, Benjamin Herrenschmidt wrote:
> > > However I can see at least one advantage of what you've done :-) You
> > > never h
On Thu, Jan 12, 2012 at 05:13:42PM +1100, Benjamin Herrenschmidt wrote:
> On Thu, 2012-01-12 at 07:29 +0200, Michael S. Tsirkin wrote:
> > On Thu, Jan 12, 2012 at 09:56:39AM +1100, Benjamin Herrenschmidt wrote:
> > > I'd suggest doing a simple user space app that creates such a ring,
> > > forks, a
On Thu, 2012-01-12 at 08:09 +0200, Michael S. Tsirkin wrote:
> On Thu, Jan 12, 2012 at 03:31:59PM +1100, Benjamin Herrenschmidt wrote:
> > However I can see at least one advantage of what you've done :-) You
> > never have to deal with holes in the ring.
>
> Another advantage is the design goal fo
On Thu, 2012-01-12 at 07:29 +0200, Michael S. Tsirkin wrote:
> On Thu, Jan 12, 2012 at 09:56:39AM +1100, Benjamin Herrenschmidt wrote:
> > I'd suggest doing a simple user space app that creates such a ring,
> > forks, and produce / consume. We can then run that through perf and
> > analyze the cach
On Thu, Jan 12, 2012 at 03:31:59PM +1100, Benjamin Herrenschmidt wrote:
> However I can see at least one advantage of what you've done :-) You
> never have to deal with holes in the ring.
Another advantage is the design goal for that ring:
host never needs to copy even if it completes
descriptors
On Thu, Jan 12, 2012 at 12:31:09PM +1030, Rusty Russell wrote:
> On Wed, 11 Jan 2012 12:21:30 +0200, "Michael S. Tsirkin"
> wrote:
> > On Wed, Jan 11, 2012 at 10:55:52AM +1030, Rusty Russell wrote:
> > > On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
> > > wrote:
> > > > On Wed, Dec 2
Subject: vhost-net: add module alias (v2.1)
By adding some module aliases, programs (or users) won't have to explicitly
call modprobe. Vhost-net will always be available if built into the kernel.
It does require assigning a permanent minor number for depmod to work.
Also:
- use C99 style initia
On Thu, Jan 12, 2012 at 09:56:39AM +1100, Benjamin Herrenschmidt wrote:
> I'd suggest doing a simple user space app that creates such a ring,
> forks, and produce / consume. We can then run that through perf and
> analyze the cache behaviour, maximum throughput, etc
>
> Cheers,
> Ben.
Sure. Y
(I called the tag to-linus. I think it worked).
To g...@github.com:rustyrussell/linux.git
+ 3ed0016...b6c96c0 master -> master (forced update)
+ 815645d...f8e8df5 to-linus -> to-linus (forced update)
+ git request-pull remotes/origin/master git://github.com/rustyrussell/linux.git
The following
On Thu, 2012-01-12 at 12:31 +1030, Rusty Russell wrote:
> > Are we going to keep guest endian for e.g. virtio net header?
> > If yes the benefit of switching config space is not that big.
> > And changes in devices would affect non-PCI transports.
>
> Yep. It would only make sense if we do it fo
On Wed, 11 Jan 2012 12:21:30 +0200, "Michael S. Tsirkin"
wrote:
> On Wed, Jan 11, 2012 at 10:55:52AM +1030, Rusty Russell wrote:
> > On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
> > wrote:
> > > On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
> > > > Yes. The idea th
On Wed, 11 Jan 2012 07:30:34 -0600, Anthony Liguori
wrote:
> I think the more important thing to do is require accesses to integers in the
> config space to always be aligned and to use the appropriate accessor.
> Non-integer fields should be restricted to byte access.
>
> That limits config s
On Thu, 12 Jan 2012 00:02:33 +0200, "Michael S. Tsirkin"
wrote:
> Look, we have a race currently. Let us not tie a bug fix to a huge
> rewrite with unclear performance benefits, please.
In theory, yes. In practice, we bandaid it.
I think in the short term we change ->get to get the entire sequ
On Thu, Jan 12, 2012 at 1:16 AM, Stephen Hemminger
wrote:
> By adding the correct module alias, programs won't have to explicitly
> call modprobe. Vhost-net will always be available if built into the kernel.
> It does require assigning a permanent minor number for depmod to work.
> Choose one next
On Thu, 2012-01-12 at 00:13 +0200, Michael S. Tsirkin wrote:
> > Well, I would argue that the network driver world has proven countless
> > times that those are good ideas :-)
>
> Below you seem to suggest that separate rings like
> virtio has now is better than a single ring like Rusty
> suggest
On Thu, 2012-01-12 at 00:13 +0200, Michael S. Tsirkin wrote:
> > We typically pre-populate the data rings with skb's for 1500 and 9000
> > bytes packets. Small packets come in immediately in the completion ring,
> > and large packets via the data ring.
>
> Won't real workloads suffer from packet
On Thu, 2012-01-12 at 00:02 +0200, Michael S. Tsirkin wrote:
> > We could probably have a helper library for sending control messages
> > which could handle waiting for a ring slot to be free (practically
> > always the case on control queues), writing the message, sending it
> and
> > waiting for
On Thu, Jan 12, 2012 at 08:13:42AM +1100, Benjamin Herrenschmidt wrote:
> On Wed, 2012-01-11 at 12:21 +0200, Michael S. Tsirkin wrote:
> >
> > > BenH also convinced me we should finally make the config space LE if
> > > we're going to change things. Since PCI is the most common transport,
> > > g
On Thu, Jan 12, 2012 at 08:02:06AM +1100, Benjamin Herrenschmidt wrote:
> On Wed, 2012-01-11 at 14:26 -0600, Anthony Liguori wrote:
> >
> > I'd say that's a special case but I see what you're getting at here.
> >
> > So what about keeping the config space read-only and using control
> > queues fo
On Wed, Jan 11, 2012 at 02:26:55PM -0600, Anthony Liguori wrote:
> On 01/11/2012 02:14 PM, Michael S. Tsirkin wrote:
> >On Wed, Jan 11, 2012 at 01:42:39PM -0600, Anthony Liguori wrote:
> >>On 01/11/2012 11:08 AM, Michael S. Tsirkin wrote:
> >>>
> >>>Not sure what you mean. Using VQ is DMA which is
On Wed, 2012-01-11 at 12:21 +0200, Michael S. Tsirkin wrote:
>
> > BenH also convinced me we should finally make the config space LE if
> > we're going to change things. Since PCI is the most common transport,
> > guest-endian confuses people. And it sucks for really weird machines.
>
> Are we
On Wed, 2012-01-11 at 14:26 -0600, Anthony Liguori wrote:
>
> I'd say that's a special case but I see what you're getting at here.
>
> So what about keeping the config space read-only and using control
> queues for
> everything else?
Which is exactly what Rusty and I are proposing :-) I would g
On Wed, 2012-01-11 at 09:16 -0800, Stephen Hemminger wrote:
> By adding the correct module alias, programs won't have to explicitly
> call modprobe. Vhost-net will always be available if built into the kernel.
> It does require assigning a permanent minor number for depmod to work.
> Choose one nex
On Wed, 2012-01-11 at 13:42 -0600, Anthony Liguori wrote:
> On 01/11/2012 11:08 AM, Michael S. Tsirkin wrote:
> >
> > Not sure what you mean. Using VQ is DMA which is pretty common for PCI.
>
> Do you know of a network device that obtains it's mac address via a DMA
> transaction?
I wouldn't be s
On Wed, 2012-01-11 at 17:21 +, Stefan Hajnoczi wrote:
> > It won't be hard to show siginificant performance regression if
> > we do this. Hard to justify for something as niche as nested virt.
>
> For x86 this should be mostly a nop.
No it won't be. Or rather, it will be as long as you map yo
On Wed, 2012-01-11 at 17:21 +0200, Michael S. Tsirkin wrote:
>
> Possible but doesn't let us layer nicely to allow unchanged drivers
> that work with all transports (new pci, old pci, non pci).
> Something like a command VQ would be a generic transport
> that can be hidden behind config->set(...)
On Wed, 2012-01-11 at 17:12 +0200, Michael S. Tsirkin wrote:
> This is similar to what we have now. But it's still buggy: e.g. if guest
> updates MAC byte by byte, we have no way to know when it's done doing
> so.
Do like real HW, there's plenty of options:
- (better) Have a command "update MA
On Wed, 2012-01-11 at 14:28 +, Stefan Hajnoczi wrote:
> On Wed, Jan 11, 2012 at 9:10 AM, Benjamin Herrenschmidt
> wrote:
> > On Wed, 2012-01-11 at 08:47 +, Stefan Hajnoczi wrote:
> >>
> >> This is also an opportunity to stop using CPU physical addresses in
> >> the ring and instead perform
On 01/11/2012 02:14 PM, Michael S. Tsirkin wrote:
On Wed, Jan 11, 2012 at 01:42:39PM -0600, Anthony Liguori wrote:
On 01/11/2012 11:08 AM, Michael S. Tsirkin wrote:
Not sure what you mean. Using VQ is DMA which is pretty common for PCI.
Do you know of a network device that obtains it's mac a
On Wed, Jan 11, 2012 at 01:42:39PM -0600, Anthony Liguori wrote:
> On 01/11/2012 11:08 AM, Michael S. Tsirkin wrote:
> >
> >Not sure what you mean. Using VQ is DMA which is pretty common for PCI.
>
> Do you know of a network device that obtains it's mac address via a DMA
> transaction?
Sure.
See
On Wed, Jan 11, 2012 at 08:54:26AM -0800, Stephen Hemminger wrote:
> On Wed, 11 Jan 2012 15:43:42 +0800
> Amos Kong wrote:
>
> > On Wed, Jan 11, 2012 at 12:54 PM, Stephen Hemminger
> > wrote:
> >
> > > By adding the a module alias, programs (or users) won't have to explicitly
> > > call modprobe
On 01/11/2012 11:08 AM, Michael S. Tsirkin wrote:
Not sure what you mean. Using VQ is DMA which is pretty common for PCI.
Do you know of a network device that obtains it's mac address via a DMA
transaction?
Regards,
Anthony Liguori
___
Virt
On Wed, Jan 11, 2012 at 09:16:53AM -0800, Stephen Hemminger wrote:
> By adding the correct module alias, programs won't have to explicitly
> call modprobe. Vhost-net will always be available if built into the kernel.
> It does require assigning a permanent minor number for depmod to work.
> Choose
On Wed, Jan 11, 2012 at 05:21:53PM +, Stefan Hajnoczi wrote:
> On Wed, Jan 11, 2012 at 3:39 PM, Michael S. Tsirkin wrote:
> > On Wed, Jan 11, 2012 at 02:28:48PM +, Stefan Hajnoczi wrote:
> >> On Wed, Jan 11, 2012 at 9:10 AM, Benjamin Herrenschmidt
> >> wrote:
> >> > On Wed, 2012-01-11 at
On Wed, Jan 11, 2012 at 05:36:41PM +0100, Andrew Jones wrote:
> Add a description to the config menu for xen tmem.
We will have another config option asked during 'make oldconfig' right?
I thought part of this cleanup patch is to remove some of this
complexity - not make it _more_ complex.
>
>
On Wed, Jan 11, 2012 at 05:36:39PM +0100, Andrew Jones wrote:
> PV-on-HVM guests may want to use the xen keyboard/mouse frontend, but
> they don't use the xen frame buffer frontend. For this case it doesn't
> make much sense for INPUT_XEN_KBDDEV_FRONTEND to depend on
> XEN_FBDEV_FRONTEND. The oppos
On Wed, Jan 11, 2012 at 05:36:38PM +0100, Andrew Jones wrote:
> When XEN_XENBUS_FRONTEND gets selected as a module it can lead to
> unbootable configs. If we need it, then we should just build it in.
Hm, don't the frontends by themsevles load this module? So if you
do 'modprobe xen-pcifront' it wo
On Wed, Jan 11, 2012 at 12:19:11PM -0400, Konrad Rzeszutek Wilk wrote:
> > > If the root complaint is that "customers think that anything set in
> > > .config is a supported feature", then the solutions are to support
> > > all
> > > the features in .config, re-educate the customers that they're wr
On Wed, Jan 11, 2012 at 3:39 PM, Michael S. Tsirkin wrote:
> On Wed, Jan 11, 2012 at 02:28:48PM +, Stefan Hajnoczi wrote:
>> On Wed, Jan 11, 2012 at 9:10 AM, Benjamin Herrenschmidt
>> wrote:
>> > On Wed, 2012-01-11 at 08:47 +, Stefan Hajnoczi wrote:
>> >>
>> >> This is also an opportunity
By adding the correct module alias, programs won't have to explicitly
call modprobe. Vhost-net will always be available if built into the kernel.
It does require assigning a permanent minor number for depmod to work.
Choose one next to TUN since this driver is related to it.
Also, use C99 style in
On Wed, Jan 11, 2012 at 17:58, Stephen Hemminger wrote:
> On Wed, 11 Jan 2012 11:07:47 +0400
> Michael Tokarev wrote:
>
>> On 11.01.2012 08:54, Stephen Hemminger wrote:
>> > By adding the a module alias, programs (or users) won't have to explicitly
>> > call modprobe. Vhost-net will always be ava
On Wed, Jan 11, 2012 at 10:02:51AM -0600, Anthony Liguori wrote:
> On 01/11/2012 09:45 AM, Michael S. Tsirkin wrote:
> >On Wed, Jan 11, 2012 at 09:28:27AM -0600, Anthony Liguori wrote:
> >>On 01/11/2012 09:21 AM, Michael S. Tsirkin wrote:
> >>>On Wed, Jan 11, 2012 at 09:15:49AM -0600, Anthony Liguo
On Wed, 11 Jan 2012 11:07:47 +0400
Michael Tokarev wrote:
> On 11.01.2012 08:54, Stephen Hemminger wrote:
> > By adding the a module alias, programs (or users) won't have to explicitly
> > call modprobe. Vhost-net will always be available if built into the kernel.
> > It does require assigning a
On Wed, 11 Jan 2012 15:43:42 +0800
Amos Kong wrote:
> On Wed, Jan 11, 2012 at 12:54 PM, Stephen Hemminger
> wrote:
>
> > By adding the a module alias, programs (or users) won't have to explicitly
> > call modprobe. Vhost-net will always be available if built into the kernel.
> > It does require
Add a description to the config menu for xen tmem.
Signed-off-by: Andrew Jones
---
drivers/xen/Kconfig |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
index 1d24061..7e8d728 100644
--- a/drivers/xen/Kconfig
+++ b/drivers/xen/Kco
Describe dom0 support in the config menu and supply help text for it.
v2 adds 'if EXPERT' to keep it out of the "standard" menu.
Signed-off-by: Andrew Jones
---
arch/x86/xen/Kconfig |7 ++-
1 files changed, 6 insertions(+), 1 deletions(-)
diff --git a/arch/x86/xen/Kconfig b/arch/x86/xe
PV-on-HVM guests may want to use the xen keyboard/mouse frontend, but
they don't use the xen frame buffer frontend. For this case it doesn't
make much sense for INPUT_XEN_KBDDEV_FRONTEND to depend on
XEN_FBDEV_FRONTEND. The opposite direction always makes more sense, i.e.
if you're using xenfb, the
When XEN_XENBUS_FRONTEND gets selected as a module it can lead to
unbootable configs. If we need it, then we should just build it in.
Signed-off-by: Andrew Jones
---
drivers/xen/Kconfig |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kc
- Original Message -
> On Mon, Jan 09, 2012 at 06:51:41PM +0100, Andrew Jones wrote:
> > PV-on-HVM guests may want to use the xen keyboard/mouse frontend,
> > but
> > they don't use the xen frame buffer frontend. For this case it
> > doesn't
>
> Ok, but PV does?
> > make much sense for I
> > If the root complaint is that "customers think that anything set in
> > .config is a supported feature", then the solutions are to support
> > all
> > the features in .config, re-educate the customers that they're wrong,
> > or
> > maintain a local patch to do this stuff.
>
> If only re-educat
On Mon, Jan 09, 2012 at 06:51:41PM +0100, Andrew Jones wrote:
> PV-on-HVM guests may want to use the xen keyboard/mouse frontend, but
> they don't use the xen frame buffer frontend. For this case it doesn't
Ok, but PV does?
> make much sense for INPUT_XEN_KBDDEV_FRONTEND to depend on
> XEN_FBDEV_F
On 01/11/2012 09:45 AM, Michael S. Tsirkin wrote:
On Wed, Jan 11, 2012 at 09:28:27AM -0600, Anthony Liguori wrote:
On 01/11/2012 09:21 AM, Michael S. Tsirkin wrote:
On Wed, Jan 11, 2012 at 09:15:49AM -0600, Anthony Liguori wrote:
This is similar to what we have now. But it's still buggy: e.g.
- Original Message -
> Describe dom0 support in the config menu and supply help text for it.
>
> v2 adds 'if EXPERT' to keep it out of the "standard" menu.
>
> Signed-off-by: Andrew Jones
> ---
> arch/x86/xen/Kconfig |7 ++-
> 1 files changed, 6 insertions(+), 1 deletions(-)
>
On Wed, Jan 11, 2012 at 09:28:27AM -0600, Anthony Liguori wrote:
> On 01/11/2012 09:21 AM, Michael S. Tsirkin wrote:
> >On Wed, Jan 11, 2012 at 09:15:49AM -0600, Anthony Liguori wrote:
> >>>This is similar to what we have now. But it's still buggy: e.g. if guest
> >>>updates MAC byte by byte, we ha
On Wed, Jan 11, 2012 at 02:28:48PM +, Stefan Hajnoczi wrote:
> On Wed, Jan 11, 2012 at 9:10 AM, Benjamin Herrenschmidt
> wrote:
> > On Wed, 2012-01-11 at 08:47 +, Stefan Hajnoczi wrote:
> >>
> >> This is also an opportunity to stop using CPU physical addresses in
> >> the ring and instead
- Original Message -
> On 01/10/2012 05:44 AM, Stefano Stabellini wrote:
> > On Mon, 9 Jan 2012, Andrew Jones wrote:
> >> I guess if we did the s/XEN_DOM0/LOCAL_APIC && IO_APIC && ACPI/ in
> >> arch/x86/pci/xen.c it would be pretty easy to review for
> >> equivalence.
> >> Then keep CONFI
On 01/11/2012 09:21 AM, Michael S. Tsirkin wrote:
On Wed, Jan 11, 2012 at 09:15:49AM -0600, Anthony Liguori wrote:
This is similar to what we have now. But it's still buggy: e.g. if guest
updates MAC byte by byte, we have no way to know when it's done doing
so.
This is no different than a norm
On Wed, Jan 11, 2012 at 09:15:49AM -0600, Anthony Liguori wrote:
> On 01/11/2012 09:12 AM, Michael S. Tsirkin wrote:
> >On Wed, Jan 11, 2012 at 07:30:34AM -0600, Anthony Liguori wrote:
> >>On 01/10/2012 06:25 PM, Rusty Russell wrote:
> >>>On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
>
On 01/11/2012 09:12 AM, Michael S. Tsirkin wrote:
On Wed, Jan 11, 2012 at 07:30:34AM -0600, Anthony Liguori wrote:
On 01/10/2012 06:25 PM, Rusty Russell wrote:
On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
wrote:
On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
Yes
On Wed, Jan 11, 2012 at 07:30:34AM -0600, Anthony Liguori wrote:
> On 01/10/2012 06:25 PM, Rusty Russell wrote:
> >On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
> >wrote:
> >>On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
> >>>Yes. The idea that we can alter fields in
On Wed, Jan 11, 2012 at 9:10 AM, Benjamin Herrenschmidt
wrote:
> On Wed, 2012-01-11 at 08:47 +, Stefan Hajnoczi wrote:
>>
>> This is also an opportunity to stop using CPU physical addresses in
>> the ring and instead perform DMA like a normal PCI device (use bus
>> addresses).
>
> Euh why ?
B
On 01/10/2012 06:25 PM, Rusty Russell wrote:
On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
wrote:
On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
Yes. The idea that we can alter fields in the device-specific config
area is flawed. There may be cases where it doesn'
On Wed, Jan 11, 2012 at 10:55:52AM +1030, Rusty Russell wrote:
> On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
> wrote:
> > On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
> > > Yes. The idea that we can alter fields in the device-specific config
> > > area is flawed.
On Wed, 2012-01-11 at 08:47 +, Stefan Hajnoczi wrote:
>
> This is also an opportunity to stop using CPU physical addresses in
> the ring and instead perform DMA like a normal PCI device (use bus
> addresses).
Euh why ?
That would mean in many cases adding a layer of iommu, which will slow
th
On Wed, Jan 11, 2012 at 12:25 AM, Rusty Russell wrote:
> On Tue, 10 Jan 2012 19:03:36 +0200, "Michael S. Tsirkin"
> wrote:
>> On Wed, Dec 21, 2011 at 11:03:25AM +1030, Rusty Russell wrote:
>> > Yes. The idea that we can alter fields in the device-specific config
>> > area is flawed. There may
66 matches
Mail list logo
