Re: [PATCH v2 6.1.y] smb/dfs_cache: Fix NULL pointer dereference on session connection failure

On Tue, Mar 31, 2026 at 10:54:08AM +0300, Ghadi Rahme wrote:
> > Then why not backport the specific changes that caused newer kernels to
> not be affected?
> 
> Based on the documentation [1] for submitting patches to stable, the patch
> cannot be over 100 lines long with context.
> 
> The upstream patch exceeds this limit by a lot and cherry picking the
> specific changes from it that remove this function is not feasible without
> causing the driver to break. In other words the removal of
> "find_ipc_from_server_path" is dependent on this refactor.
> 
> > Backport the same changes?
> 
> I can go ahead with this solution, given I get the green light to ignore the
> 100 line rule.
> 
> [1] https://www.kernel.org/doc/html/v4.11/process/stable-kernel-rules.html#

We take "large" backports from mainline all the time, as long as the
maintainers for the subsystem are ok with it.  to take one-off patches
is usually much harder as the change itself is almost always "wrong",
and then future changes are even harder to backport.

So I recommend working with the maintainers here please and see what
they want to do, if anything.

thanks,

greg k-h