| Message ID | 20170720113556.21490-1-wei.liu2@citrix.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > The code says it defaults to false. > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > --- > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > Cc: Jan Beulich <jbeulich@suse.com> > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > Cc: Stefano Stabellini <sstabellini@kernel.org> > Cc: Tim Deegan <tim@xen.org> > Cc: Wei Liu <wei.liu2@citrix.com> > --- > docs/misc/xen-command-line.markdown | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > index 3f90c3b7a8..f524294aa6 100644 > --- a/docs/misc/xen-command-line.markdown > +++ b/docs/misc/xen-command-line.markdown > @@ -136,7 +136,7 @@ mode during S3 resume. > ### allowsuperpage > > `= <boolean>` > > -> Default: `true` > +> Default: `false` > > Permit Xen to use superpages when performing memory management. I'm not an expert on Xen MM code, but isn't this intended for PV guests? The description above makes it look like this is for Xen itself, but AFAICT from skimming over the code this seems to be a PV feature, in which case the text above should be fixed to prevent confusion. Roger.
On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > The code says it defaults to false. > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > --- > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > Cc: Jan Beulich <jbeulich@suse.com> > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > Cc: Tim Deegan <tim@xen.org> > > Cc: Wei Liu <wei.liu2@citrix.com> > > --- > > docs/misc/xen-command-line.markdown | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > index 3f90c3b7a8..f524294aa6 100644 > > --- a/docs/misc/xen-command-line.markdown > > +++ b/docs/misc/xen-command-line.markdown > > @@ -136,7 +136,7 @@ mode during S3 resume. > > ### allowsuperpage > > > `= <boolean>` > > > > -> Default: `true` > > +> Default: `false` > > > > Permit Xen to use superpages when performing memory management. > > I'm not an expert on Xen MM code, but isn't this intended for PV > guests? The description above makes it look like this is for Xen > itself, but AFAICT from skimming over the code this seems to be a PV > feature, in which case the text above should be fixed to prevent > confusion. I believe it is PV only, but I'm not 100% sure. I would love to fix the text as well if possible.
On 20/07/17 12:47, Wei Liu wrote: > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: >> On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: >>> The code says it defaults to false. >>> >>> Signed-off-by: Wei Liu <wei.liu2@citrix.com> >>> --- >>> Cc: Andrew Cooper <andrew.cooper3@citrix.com> >>> Cc: George Dunlap <George.Dunlap@eu.citrix.com> >>> Cc: Ian Jackson <ian.jackson@eu.citrix.com> >>> Cc: Jan Beulich <jbeulich@suse.com> >>> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> >>> Cc: Stefano Stabellini <sstabellini@kernel.org> >>> Cc: Tim Deegan <tim@xen.org> >>> Cc: Wei Liu <wei.liu2@citrix.com> >>> --- >>> docs/misc/xen-command-line.markdown | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown >>> index 3f90c3b7a8..f524294aa6 100644 >>> --- a/docs/misc/xen-command-line.markdown >>> +++ b/docs/misc/xen-command-line.markdown >>> @@ -136,7 +136,7 @@ mode during S3 resume. >>> ### allowsuperpage >>> > `= <boolean>` >>> >>> -> Default: `true` >>> +> Default: `false` >>> >>> Permit Xen to use superpages when performing memory management. >> I'm not an expert on Xen MM code, but isn't this intended for PV >> guests? The description above makes it look like this is for Xen >> itself, but AFAICT from skimming over the code this seems to be a PV >> feature, in which case the text above should be fixed to prevent >> confusion. > I believe it is PV only, but I'm not 100% sure. > > I would love to fix the text as well if possible. I'm fairly sure this option applies exclusively to PV superpages. Double check the logic through the code, but I think (since dropping 32bit support), we have no configuration where Xen might not be able to use superpages. ~Andrew
On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: > On 20/07/17 12:47, Wei Liu wrote: > > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > > > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > > > The code says it defaults to false. > > > > > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > > > --- > > > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > > > Cc: Jan Beulich <jbeulich@suse.com> > > > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > > > Cc: Tim Deegan <tim@xen.org> > > > > Cc: Wei Liu <wei.liu2@citrix.com> > > > > --- > > > > docs/misc/xen-command-line.markdown | 2 +- > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > > > index 3f90c3b7a8..f524294aa6 100644 > > > > --- a/docs/misc/xen-command-line.markdown > > > > +++ b/docs/misc/xen-command-line.markdown > > > > @@ -136,7 +136,7 @@ mode during S3 resume. > > > > ### allowsuperpage > > > > > `= <boolean>` > > > > -> Default: `true` > > > > +> Default: `false` > > > > Permit Xen to use superpages when performing memory management. > > > I'm not an expert on Xen MM code, but isn't this intended for PV > > > guests? The description above makes it look like this is for Xen > > > itself, but AFAICT from skimming over the code this seems to be a PV > > > feature, in which case the text above should be fixed to prevent > > > confusion. > > I believe it is PV only, but I'm not 100% sure. > > > > I would love to fix the text as well if possible. > > I'm fairly sure this option applies exclusively to PV superpages. Double > check the logic through the code, but I think (since dropping 32bit > support), we have no configuration where Xen might not be able to use > superpages. > So we can just delete this option and make Xen always use superpage? That would be fine by me, too.
On 20/07/17 13:57, Wei Liu wrote: > On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: >> On 20/07/17 12:47, Wei Liu wrote: >>> On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: >>>> On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: >>>>> The code says it defaults to false. >>>>> >>>>> Signed-off-by: Wei Liu <wei.liu2@citrix.com> >>>>> --- >>>>> Cc: Andrew Cooper <andrew.cooper3@citrix.com> >>>>> Cc: George Dunlap <George.Dunlap@eu.citrix.com> >>>>> Cc: Ian Jackson <ian.jackson@eu.citrix.com> >>>>> Cc: Jan Beulich <jbeulich@suse.com> >>>>> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> >>>>> Cc: Stefano Stabellini <sstabellini@kernel.org> >>>>> Cc: Tim Deegan <tim@xen.org> >>>>> Cc: Wei Liu <wei.liu2@citrix.com> >>>>> --- >>>>> docs/misc/xen-command-line.markdown | 2 +- >>>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>>> >>>>> diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown >>>>> index 3f90c3b7a8..f524294aa6 100644 >>>>> --- a/docs/misc/xen-command-line.markdown >>>>> +++ b/docs/misc/xen-command-line.markdown >>>>> @@ -136,7 +136,7 @@ mode during S3 resume. >>>>> ### allowsuperpage >>>>> > `= <boolean>` >>>>> -> Default: `true` >>>>> +> Default: `false` >>>>> Permit Xen to use superpages when performing memory management. >>>> I'm not an expert on Xen MM code, but isn't this intended for PV >>>> guests? The description above makes it look like this is for Xen >>>> itself, but AFAICT from skimming over the code this seems to be a PV >>>> feature, in which case the text above should be fixed to prevent >>>> confusion. >>> I believe it is PV only, but I'm not 100% sure. >>> >>> I would love to fix the text as well if possible. >> I'm fairly sure this option applies exclusively to PV superpages. Double >> check the logic through the code, but I think (since dropping 32bit >> support), we have no configuration where Xen might not be able to use >> superpages. >> > So we can just delete this option and make Xen always use superpage? > That would be fine by me, too. No - my point was that this option now exclusively controls PV superpages, IIRC. ~Andrew
On Fri, Jul 21, 2017 at 05:21:26PM +0100, Andrew Cooper wrote: > On 20/07/17 13:57, Wei Liu wrote: > > On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: > > > On 20/07/17 12:47, Wei Liu wrote: > > > > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > > > > > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > > > > > The code says it defaults to false. > > > > > > > > > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > > > > > --- > > > > > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > > > > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > > > > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > > > > > Cc: Jan Beulich <jbeulich@suse.com> > > > > > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > > > > > Cc: Tim Deegan <tim@xen.org> > > > > > > Cc: Wei Liu <wei.liu2@citrix.com> > > > > > > --- > > > > > > docs/misc/xen-command-line.markdown | 2 +- > > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > > > > > index 3f90c3b7a8..f524294aa6 100644 > > > > > > --- a/docs/misc/xen-command-line.markdown > > > > > > +++ b/docs/misc/xen-command-line.markdown > > > > > > @@ -136,7 +136,7 @@ mode during S3 resume. > > > > > > ### allowsuperpage > > > > > > > `= <boolean>` > > > > > > -> Default: `true` > > > > > > +> Default: `false` > > > > > > Permit Xen to use superpages when performing memory management. > > > > > I'm not an expert on Xen MM code, but isn't this intended for PV > > > > > guests? The description above makes it look like this is for Xen > > > > > itself, but AFAICT from skimming over the code this seems to be a PV > > > > > feature, in which case the text above should be fixed to prevent > > > > > confusion. > > > > I believe it is PV only, but I'm not 100% sure. > > > > > > > > I would love to fix the text as well if possible. > > > I'm fairly sure this option applies exclusively to PV superpages. Double > > > check the logic through the code, but I think (since dropping 32bit > > > support), we have no configuration where Xen might not be able to use > > > superpages. > > > > > So we can just delete this option and make Xen always use superpage? > > That would be fine by me, too. > > No - my point was that this option now exclusively controls PV superpages, > IIRC. > OK. I misunderstood. In that case. We can change the text to: Permit PV guests to use suerpages.
On Thu, Jul 20, 2017 at 01:57:17PM +0100, Wei Liu wrote: > On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: > > On 20/07/17 12:47, Wei Liu wrote: > > > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > > > > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > > > > The code says it defaults to false. > > > > > > > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > > > > --- > > > > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > > > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > > > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > > > > Cc: Jan Beulich <jbeulich@suse.com> > > > > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > > > > Cc: Tim Deegan <tim@xen.org> > > > > > Cc: Wei Liu <wei.liu2@citrix.com> > > > > > --- > > > > > docs/misc/xen-command-line.markdown | 2 +- > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > > > > index 3f90c3b7a8..f524294aa6 100644 > > > > > --- a/docs/misc/xen-command-line.markdown > > > > > +++ b/docs/misc/xen-command-line.markdown > > > > > @@ -136,7 +136,7 @@ mode during S3 resume. > > > > > ### allowsuperpage > > > > > > `= <boolean>` > > > > > -> Default: `true` > > > > > +> Default: `false` > > > > > Permit Xen to use superpages when performing memory management. > > > > I'm not an expert on Xen MM code, but isn't this intended for PV > > > > guests? The description above makes it look like this is for Xen > > > > itself, but AFAICT from skimming over the code this seems to be a PV > > > > feature, in which case the text above should be fixed to prevent > > > > confusion. > > > I believe it is PV only, but I'm not 100% sure. > > > > > > I would love to fix the text as well if possible. > > > > I'm fairly sure this option applies exclusively to PV superpages. Double > > check the logic through the code, but I think (since dropping 32bit > > support), we have no configuration where Xen might not be able to use > > superpages. > > > > So we can just delete this option and make Xen always use superpage? > That would be fine by me, too. Can we just nuke the code altogther? Oracle is not using it anymore. > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xen.org > https://lists.xen.org/xen-devel
On Fri, Jul 21, 2017 at 12:44:18PM -0400, Konrad Rzeszutek Wilk wrote: > On Thu, Jul 20, 2017 at 01:57:17PM +0100, Wei Liu wrote: > > On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: > > > On 20/07/17 12:47, Wei Liu wrote: > > > > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > > > > > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > > > > > The code says it defaults to false. > > > > > > > > > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > > > > > --- > > > > > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > > > > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > > > > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > > > > > Cc: Jan Beulich <jbeulich@suse.com> > > > > > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > > > > > Cc: Tim Deegan <tim@xen.org> > > > > > > Cc: Wei Liu <wei.liu2@citrix.com> > > > > > > --- > > > > > > docs/misc/xen-command-line.markdown | 2 +- > > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > > > > > index 3f90c3b7a8..f524294aa6 100644 > > > > > > --- a/docs/misc/xen-command-line.markdown > > > > > > +++ b/docs/misc/xen-command-line.markdown > > > > > > @@ -136,7 +136,7 @@ mode during S3 resume. > > > > > > ### allowsuperpage > > > > > > > `= <boolean>` > > > > > > -> Default: `true` > > > > > > +> Default: `false` > > > > > > Permit Xen to use superpages when performing memory management. > > > > > I'm not an expert on Xen MM code, but isn't this intended for PV > > > > > guests? The description above makes it look like this is for Xen > > > > > itself, but AFAICT from skimming over the code this seems to be a PV > > > > > feature, in which case the text above should be fixed to prevent > > > > > confusion. > > > > I believe it is PV only, but I'm not 100% sure. > > > > > > > > I would love to fix the text as well if possible. > > > > > > I'm fairly sure this option applies exclusively to PV superpages. Double > > > check the logic through the code, but I think (since dropping 32bit > > > support), we have no configuration where Xen might not be able to use > > > superpages. > > > > > > > So we can just delete this option and make Xen always use superpage? > > That would be fine by me, too. > > Can we just nuke the code altogther? > > Oracle is not using it anymore. Sure! I was about to ask you about that. I'm happy to submit patches to nuke it from both the hypervisor and toolstack.
On Fri, Jul 21, 2017 at 05:51:02PM +0100, Wei Liu wrote: > On Fri, Jul 21, 2017 at 12:44:18PM -0400, Konrad Rzeszutek Wilk wrote: > > On Thu, Jul 20, 2017 at 01:57:17PM +0100, Wei Liu wrote: > > > On Thu, Jul 20, 2017 at 12:49:37PM +0100, Andrew Cooper wrote: > > > > On 20/07/17 12:47, Wei Liu wrote: > > > > > On Thu, Jul 20, 2017 at 12:45:38PM +0100, Roger Pau Monné wrote: > > > > > > On Thu, Jul 20, 2017 at 12:35:56PM +0100, Wei Liu wrote: > > > > > > > The code says it defaults to false. > > > > > > > > > > > > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com> > > > > > > > --- > > > > > > > Cc: Andrew Cooper <andrew.cooper3@citrix.com> > > > > > > > Cc: George Dunlap <George.Dunlap@eu.citrix.com> > > > > > > > Cc: Ian Jackson <ian.jackson@eu.citrix.com> > > > > > > > Cc: Jan Beulich <jbeulich@suse.com> > > > > > > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > > > > Cc: Stefano Stabellini <sstabellini@kernel.org> > > > > > > > Cc: Tim Deegan <tim@xen.org> > > > > > > > Cc: Wei Liu <wei.liu2@citrix.com> > > > > > > > --- > > > > > > > docs/misc/xen-command-line.markdown | 2 +- > > > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > > > > > diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown > > > > > > > index 3f90c3b7a8..f524294aa6 100644 > > > > > > > --- a/docs/misc/xen-command-line.markdown > > > > > > > +++ b/docs/misc/xen-command-line.markdown > > > > > > > @@ -136,7 +136,7 @@ mode during S3 resume. > > > > > > > ### allowsuperpage > > > > > > > > `= <boolean>` > > > > > > > -> Default: `true` > > > > > > > +> Default: `false` > > > > > > > Permit Xen to use superpages when performing memory management. > > > > > > I'm not an expert on Xen MM code, but isn't this intended for PV > > > > > > guests? The description above makes it look like this is for Xen > > > > > > itself, but AFAICT from skimming over the code this seems to be a PV > > > > > > feature, in which case the text above should be fixed to prevent > > > > > > confusion. > > > > > I believe it is PV only, but I'm not 100% sure. > > > > > > > > > > I would love to fix the text as well if possible. > > > > > > > > I'm fairly sure this option applies exclusively to PV superpages. Double > > > > check the logic through the code, but I think (since dropping 32bit > > > > support), we have no configuration where Xen might not be able to use > > > > superpages. > > > > > > > > > > So we can just delete this option and make Xen always use superpage? > > > That would be fine by me, too. > > > > Can we just nuke the code altogther? > > > > Oracle is not using it anymore. > > Sure! I was about to ask you about that. > > I'm happy to submit patches to nuke it from both the hypervisor and toolstack. Feel free to add Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> on them :-)
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown index 3f90c3b7a8..f524294aa6 100644 --- a/docs/misc/xen-command-line.markdown +++ b/docs/misc/xen-command-line.markdown @@ -136,7 +136,7 @@ mode during S3 resume. ### allowsuperpage > `= <boolean>` -> Default: `true` +> Default: `false` Permit Xen to use superpages when performing memory management.
The code says it defaults to false. Signed-off-by: Wei Liu <wei.liu2@citrix.com> --- Cc: Andrew Cooper <andrew.cooper3@citrix.com> Cc: George Dunlap <George.Dunlap@eu.citrix.com> Cc: Ian Jackson <ian.jackson@eu.citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Stefano Stabellini <sstabellini@kernel.org> Cc: Tim Deegan <tim@xen.org> Cc: Wei Liu <wei.liu2@citrix.com> --- docs/misc/xen-command-line.markdown | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)