On Fri, Oct 18, 2024 at 10:18:19AM +0200, Vlastimil Babka wrote:
> On 10/18/24 07:58, Greg KH wrote:
> > On Thu, Oct 17, 2024 at 02:58:01PM -0700, chrisl@xxxxxxxxxx wrote:
> >> A few commits from Yu Zhao have been merged into 6.12.
> >> They need to be backported to 6.11.
> >
> > Why?
> >
> >> - c2a967f6ab0ec ("mm/hugetlb_vmemmap: don't synchronize_rcu() without HVO")
> >> - 95599ef684d01 ("mm/codetag: fix pgalloc_tag_split()")
> >> - e0a955bf7f61c ("mm/codetag: add pgalloc_tag_copy()")
> >
> > For mm changes, we need an explicit ack from the mm maintainers to take
> > patches into the stable tree. Why were these not tagged with the normal
> > "cc: stable@" tag in the first place?
>
> The two codetags commit actually were tagged.
Ah, I missed that, I only looked at the first one and stopped there.
> c2a967f6ab0ec seems to me an
> unrelated performance optimization, could make sense in a LTS kernel but is
> there a lot of benefit expected for the lifetime of 6.11?
I can't take any of these as they aren't properly signed-off so I'll
wait for a resubmit to consider any of these (but patch 1 better have
some justification and ack from the mm developers that we can take
it...)
thanks,
greg k-h
