Re: [5.15-stable PATCH 0/2] Copy-on-write hwpoison recovery

[Greg KH <gregkh@linuxfoundation.org>]

On Tue, Jun 20, 2023 at 11:36:13AM -0700, Jane Chu wrote:
> Hi, Greg,
> 
> On 6/19/2023 1:28 AM, Greg KH wrote:
> > On Wed, Jun 14, 2023 at 07:52:53PM -0600, Jane Chu wrote:
> > > I was able to reproduce crash on 5.15.y kernel during COW, and
> > > when the grandchild process attempts a write to a private page
> > > inherited from the child process and the private page contains
> > > a memory uncorrectable error. The way to reproduce is described
> > > in Tony's patch, using his ras-tools/einj_mem_uc.
> > > And the patch series fixed the panic issue in 5.15.y.
> > 
> > But you are skipping 6.1.y, which is not ok as it would cause
> > regressions when you upgrade.
> > 
> > I'll drop this from my review queue now, please provide working
> > backports for this and newer releases, and I'll be glad to take them.
> > 
> 
> Thanks for the guidance, will do.
> To confirm, you're looking for backport to both 6.1.y and 5.15.y, and
> nothing else, correct?  Just curious, why 6.1.y in particular?


If you don't think it needs to go to any kernels older than 5.15.y, that would
be fine.

And as for 6.1.y, look at the front page of www.kernel.org, it shows the
active kernel versions.  We can't apply a change to an older kernel tree
only because if you upgrade to a newer one (i.e. from 5.15.y to 6.1.y),
you would have a regression which we don't ever want.

thanks,

greg k-h