Re: [PATCH 1/2] cxl/region: fix region leak when attach_target fails in cxl_add_to_region

[Gregory Price <gourry@gourry.net>]

On Tue, Feb 24, 2026 at 04:15:33PM +0000, Alejandro Lucero Palau wrote:
> 
> On 2/21/26 05:17, Gregory Price wrote:
> > On Fri, Feb 20, 2026 at 11:30:12PM -0500, Gregory Price wrote:
> > > cxl_add_to_region() ignores the return value of attach_target().  When
> > > attach_target() fails (e.g. cxl_port_setup_targets() returns -ENXIO),
> > > the auto-discovered region remains registered with its HPA resource
> > > consumed but never reaches COMMIT state.  Subsequent region creation
> > > attempts fail with -ENOSPC because the HPA range is already reserved.
> > > 
> > > Track whether this call to cxl_add_to_region() created the region, and
> > > call drop_region() on attach_target() failure to unregister it and
> > > release the HPA resource.  Pre-existing regions are left alone since
> > > other endpoints may already be attached.
> > > 
> > > Signed-off-by: Gregory Price <gourry@gourry.net>
> > BAH - disregard this patch, it uses drop_region which is introduced by
> > Alejandro here:
> > 
> > https://lore.kernel.org/linux-cxl/20260201155438.2664640-20-alejandro.lucero-palau@amd.com/
> > 
> Feel free to add it to this series. I have started to send individual series
> as you know but the part changing the region creation will require more work
> than the already sent.
> 
> About this fix, it looks good to me, although I have to admit I'm a bit lost
> after following the discussion Allison points to. If we want to keep the
> state of failure for forensics, not sure if the debugging/tracing or default
> error info in this case will be enough.
> 
> In any case:
> 
> Reviewed-by: Alejandro Lucero <alucerop@amd.com>
> 

Yeah i don't quite follow the want to keep the objects around, it seems
to cause more issues than it solves - but then i also don't think this
is going to be a particularly common problem

~Gregory