[SRU,J,2/2] UBUNTU: [Config] CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE

From: Michael Reed <Michael.Reed@canonical.com>

BugLink: https://bugs.launchpad.net/bugs/2028158
By default this is enabled, but it is causing an issue reconfiguring device dax memory, thus
it is being disabled.

Signed-off-by: Michael Reed <Michael.Reed@canonical.com>
---
 debian.master/config/annotations | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

On 22.09.23 19:09, Michael Reed wrote:
> From: Michael Reed <Michael.Reed@canonical.com>
> 
> BugLink: https://bugs.launchpad.net/bugs/2028158
> By default this is enabled, but it is causing an issue reconfiguring device dax memory, thus
> it is being disabled.
> 
> Signed-off-by: Michael Reed <Michael.Reed@canonical.com>
> ---
>   debian.master/config/annotations | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/debian.master/config/annotations b/debian.master/config/annotations
> index 5f854837a810..b9081681267c 100644
> --- a/debian.master/config/annotations
> +++ b/debian.master/config/annotations
> @@ -348,7 +348,7 @@ CONFIG_LOAD_PPC_KEYS                            note<'LP:1866909'>
>   CONFIG_LOCALVERSION_AUTO                        policy<{'amd64': 'n', 'arm64': 'n', 'armhf': 'n', 'ppc64el': 'n', 's390x': 'n'}>
>   CONFIG_LOCALVERSION_AUTO                        note<'triggers packaging failures'>
>   
> -CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64': 'y', 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
> +CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64': 'n', 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
>   CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            note<'LP:1709171 LP:1848492'>
>   
>   CONFIG_MFD_SM501                                policy<{'amd64': 'm', 'arm64': 'm', 'armhf': 'y', 'ppc64el': 'm', 's390x': 'n'}>

That would be a more fundamental behavior change which at least should 
be added as a reference to the annotation below. Seems for ppc64el and 
s390x this was changed in the past but I am not sure whether this might 
break assumptions or use cases on the x86 side.
Adding a few more cc's for discussion/awareness.

Hi Stefan,

Would you also recommend that I disable this for arm64?

-Michael


On Thu, Sep 28, 2023 at 11:03 AM Stefan Bader <stefan.bader@canonical.com>
wrote:

> On 22.09.23 19:09, Michael Reed wrote:
> > From: Michael Reed <Michael.Reed@canonical.com>
> >
> > BugLink: https://bugs.launchpad.net/bugs/2028158
> > By default this is enabled, but it is causing an issue reconfiguring
> device dax memory, thus
> > it is being disabled.
> >
> > Signed-off-by: Michael Reed <Michael.Reed@canonical.com>
> > ---
> >   debian.master/config/annotations | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/debian.master/config/annotations
> b/debian.master/config/annotations
> > index 5f854837a810..b9081681267c 100644
> > --- a/debian.master/config/annotations
> > +++ b/debian.master/config/annotations
> > @@ -348,7 +348,7 @@ CONFIG_LOAD_PPC_KEYS
> note<'LP:1866909'>
> >   CONFIG_LOCALVERSION_AUTO                        policy<{'amd64': 'n',
> 'arm64': 'n', 'armhf': 'n', 'ppc64el': 'n', 's390x': 'n'}>
> >   CONFIG_LOCALVERSION_AUTO                        note<'triggers
> packaging failures'>
> >
> > -CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64': 'y',
> 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
> > +CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64': 'n',
> 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
> >   CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            note<'LP:1709171
> LP:1848492'>
> >
> >   CONFIG_MFD_SM501                                policy<{'amd64': 'm',
> 'arm64': 'm', 'armhf': 'y', 'ppc64el': 'm', 's390x': 'n'}>
>
> That would be a more fundamental behavior change which at least should
> be added as a reference to the annotation below. Seems for ppc64el and
> s390x this was changed in the past but I am not sure whether this might
> break assumptions or use cases on the x86 side.
> Adding a few more cc's for discussion/awareness.
>
> --
> - Stefan
>
>

On 02.10.23 18:54, Michael Reed wrote:
> Hi Stefan,
> 
> Would you also recommend that I disable this for arm64?

Sorry this has been neglected a bit. We did discuss this again recently 
and generally the reaction is more hesitant. First, as mentioned before, 
this is an option which changes behavior in general. And none of us 
knows what the exact effect is for x86. Could this result in a server 
coming up with less memory than expected (than it was before)?
Is it really absolutely necessary to change general behavior to fix a 
single driver? And if it is necessary, then this should be changed in 
all kernels following (Lunar, Mantic, ...).

-Stefan

> 
> -Michael
> 
> 
> On Thu, Sep 28, 2023 at 11:03 AM Stefan Bader 
> <stefan.bader@canonical.com <mailto:stefan.bader@canonical.com>> wrote:
> 
>     On 22.09.23 19:09, Michael Reed wrote:
>      > From: Michael Reed <Michael.Reed@canonical.com
>     <mailto:Michael.Reed@canonical.com>>
>      >
>      > BugLink: https://bugs.launchpad.net/bugs/2028158
>     <https://bugs.launchpad.net/bugs/2028158>
>      > By default this is enabled, but it is causing an issue
>     reconfiguring device dax memory, thus
>      > it is being disabled.
>      >
>      > Signed-off-by: Michael Reed <Michael.Reed@canonical.com
>     <mailto:Michael.Reed@canonical.com>>
>      > ---
>      >   debian.master/config/annotations | 2 +-
>      >   1 file changed, 1 insertion(+), 1 deletion(-)
>      >
>      > diff --git a/debian.master/config/annotations
>     b/debian.master/config/annotations
>      > index 5f854837a810..b9081681267c 100644
>      > --- a/debian.master/config/annotations
>      > +++ b/debian.master/config/annotations
>      > @@ -348,7 +348,7 @@ CONFIG_LOAD_PPC_KEYS                         
>        note<'LP:1866909'>
>      >   CONFIG_LOCALVERSION_AUTO                       
>     policy<{'amd64': 'n', 'arm64': 'n', 'armhf': 'n', 'ppc64el': 'n',
>     's390x': 'n'}>
>      >   CONFIG_LOCALVERSION_AUTO                        note<'triggers
>     packaging failures'>
>      >
>      > -CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64':
>     'y', 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
>      > +CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE            policy<{'amd64':
>     'n', 'arm64': 'y', 'ppc64el': 'n', 's390x': 'n'}>
>      >   CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE           
>     note<'LP:1709171 LP:1848492'>
>      >
>      >   CONFIG_MFD_SM501                               
>     policy<{'amd64': 'm', 'arm64': 'm', 'armhf': 'y', 'ppc64el': 'm',
>     's390x': 'n'}>
> 
>     That would be a more fundamental behavior change which at least should
>     be added as a reference to the annotation below. Seems for ppc64el and
>     s390x this was changed in the past but I am not sure whether this might
>     break assumptions or use cases on the x86 side.
>     Adding a few more cc's for discussion/awareness.
> 
>     -- 
>     - Stefan
>