On 16.01.20 10:26, Greg Kroah-Hartman wrote:
On Thu, Jan 16, 2020 at 09:59:44AM +0100, David Hildenbrand wrote:
On 16.01.20 09:54, Greg Kroah-Hartman wrote:
On Thu, Jan 16, 2020 at 09:42:51AM +0100, David Hildenbrand wrote:
On 16.01.20 09:34, Greg Kroah-Hartman wrote:
On Wed, Jan 15, 2020 at 04:54:59PM +0100, David Hildenbrand wrote:
> > And why would 4.9 and 4.4 care about them?
The crashes can be trigger under 4.9 and 4.4. If we decide that we do not care, then this series can be dropped.
Do we have users of memory hotplug that are somehow stuck at those old versions that can not upgrade? Obviously this didn't work previously for them, so moving to a modern kernel might be a good reason to get this new feature :)
That's a good point - but usually when you experience a crash it's too late for you to realize that you have to move to a newer release :) It used to work before 4.4 IIRC.
(one case I am concerned with is when memory onlining after memory hotplug failed (e.g., because the was an OOM event happening concurrently) - then memory hotunplug will crash your system.)
But yeah, I am not aware of a report where somebody actually hit any of these issues on a stable kernel.
Just to clarify: I can reproduce them of course :)
Ok, let's start with 4.19 and 4.14 for these for now. Should make things easier, right?
What do you mean with "start with"? Drop this series and not do the backport, meaning people should switch to a stable kernel > 4.19 if they don't want surprises on memory unplug?
No, I'm saying I want to take this for 4.19, and 4.14 if you have it.
Minor correction: I meant 4.19 and 4.14, not 4.4 :/ Sorry for the confusion. Will try to prepare the 4.14 backports as well.