[PATCH v2 22/29] maple_tree: Use maple copy node for mas_wr_rebalance() operation

[Andrew Morton]

On Fri, 30 Jan 2026 11:42:49 -0500 "Liam R. Howlett" <Liam.Howlett at oracle.com> wrote:

> * Andrew Morton <akpm at linux-foundation.org> [260127 18:16]:
> > On Tue, 27 Jan 2026 23:05:48 +0000 Mark Brown <broonie at kernel.org> wrote:
> > 
> > > On Wed, Jan 21, 2026 at 11:45:19AM -0500, Liam R. Howlett wrote:
> > > > Stop using the maple big node for rebalance operations by changing to
> > > > more align with spanning store.  The rebalance operation needs its own
> > > > data calculation in rebalance_data().
> > > > 
> > > > In the event of too much data, the rebalance tries to push the data
> > > > using push_data_sib().  If there is insufficient data, the rebalance
> > > > operation will rebalance against a sibling (found with rebalance_sib()).
> > > 
> > > I'm seeing a test failure in the LTP linkat02 test on arm64 which we're
> > > also seeing on a range of platforms in the Arm lab.  A NULL pointer
> > > deference is generated handling the syscall in the updated code (log for
> > > the actual next-20260126 commit):
> > 
> > Cool, thanks.
> > 
> > Liam, I'll move this series back into mm-new to avoid disrupting
> > linux-next testing.
> > 
> > 
> 
> I know you like fixup patches on larger sets, but I've done three things
> to the patch set now:
> 
> 1. A small fix for this issue
> 2. Added a new test in its own patch for this issue
> 3. Cleaned up the old compiler pointer warning in a more readable way
> 
> Would you like me to send the delta or a new patch set?

A new patchset works.  I'll send my heres-what-changed-since v2 email to
aid people who reviewed v2 and so you can scan it and think "yup, I
meant to do that".